2011 International Melanoma Congress Advancement Through Collaboration joins the following four annual meetings in 2011:

  1. Top of page
  2. 2011 International Melanoma Congress Advancement Through Collaboration joins the following four annual meetings in 2011:
  3. Congress Agenda
  4. Congress Proceedings Abstracts of the 8th International Congress of the Society for Melanoma Research November 9–11, 2011 Invited Faculty Abstracts
  5. Abstracts of the 5th Meeting of Interdisciplinary Melanoma/Skin Cancer Centres November 11–12, 2011 Invited Faculty Abstracts
  6. Abstracts of the 4th Melanoma Pathology Symposium of the International Melanoma Pathology Working Group November 13, 2011 Invited Faculty Abstracts
  7. Abstracts of the 8th International Congress of The Society for Melanoma Research November 9–11, 2011 Proffered Abstracts Selected for Oral Presentation
  8. Abstracts of the 8th International Congress of The Society for Melanoma Research November 9–11, 2011 Proffered Abstracts Selected for Poster Presentation
  9. Abstracts of the 5th International Meeting of Interdisciplinary Melanoma/Skin Cancer Centres November 11–12, 2011 Proffered Abstracts Selected for Poster Presentation

8th International Congress of The Society for Melanoma Research

5th Meeting of Interdisciplinary Melanoma / Skin Cancer Centres

2nd Melanoma Update for Primary Care Physicians, Surgeons, Oncologists, and Dermatologists

4th Melanoma Pathology Symposium of the International Melanoma Pathology Working Group

Congress Organizers: Vernon Sondak, Keiran Smalley, Boris Bastian, Raymond Barnhill, Jane Messina, Jeffrey Weber

November 9–13, 2011

Marriott Waterside Hotel & Marina

700 South Florida Avenue, Tampa, Florida, USA 33602-5404

Congress Agenda

  1. Top of page
  2. 2011 International Melanoma Congress Advancement Through Collaboration joins the following four annual meetings in 2011:
  3. Congress Agenda
  4. Congress Proceedings Abstracts of the 8th International Congress of the Society for Melanoma Research November 9–11, 2011 Invited Faculty Abstracts
  5. Abstracts of the 5th Meeting of Interdisciplinary Melanoma/Skin Cancer Centres November 11–12, 2011 Invited Faculty Abstracts
  6. Abstracts of the 4th Melanoma Pathology Symposium of the International Melanoma Pathology Working Group November 13, 2011 Invited Faculty Abstracts
  7. Abstracts of the 8th International Congress of The Society for Melanoma Research November 9–11, 2011 Proffered Abstracts Selected for Oral Presentation
  8. Abstracts of the 8th International Congress of The Society for Melanoma Research November 9–11, 2011 Proffered Abstracts Selected for Poster Presentation
  9. Abstracts of the 5th International Meeting of Interdisciplinary Melanoma/Skin Cancer Centres November 11–12, 2011 Proffered Abstracts Selected for Poster Presentation

8th International Congress of The Society for Melanoma Research

November 9–11, 2011

Wednesday November 9

6:00–7:00 pm
Grand Ballroom
Society for Melanoma Research Annual Congress Welcome and Keynote Address
Session Chairs: Boris Bastian, David Fisher
Welcome to SMR meeting participantsBoris Bastian
Keynote Address: ‘The biology of cancer metastasis’Joan Massagué
Welcome Reception to follow
Grand Ballroom Gallery and Foyer

Thursday November 10

6:50–7:50 am
Florida Ballroom Salons I-IV and Foyer
Melanoma clinical trials of greatest relevance to the scientist
Axel Hauschild, Rick Kefford, John Kirkwood
Florida Ballroom Salons V-VI and Foyer
So you think you have an invention worth patenting
Jarrett Rieger, Pedram Gerami, Josep Malvehy
8:00–10:15 am
Grand Ballroom
SMR Plenary Session I
New insights into melanoma biology
Session Chairs: Keiran Smalley, Glenn Merlino
8:00–8:30Roles for MITF in melanomagenesisDavid Fisher
8:30–9:00Mice offer new wrinkles on UV damage and melanoma progressionGlenn Merlino
9:00–9:30 Marisol Soengas
9:30–10:00Senescence and melanoma suppressionDaniel Peeper
10:15–10:30 am Refreshment Break, Networking and Exhibit Viewing
Grand Ballroom Foyer
10:30 am–12:45 pm
Grand Ballroom
SMR Plenary Session II
Novel drivers of melanoma initiation, progression and resistance
Session Chairs: Richard Marais, Meenhard Herlyn
10:30–11:00BAP1 mutations in melanomaJW Harbour
11:00–11:30 Levi Garraway
11:30–12:00Vemurafenib (PLX4032) promotes epigenetic changes in melanoma cells leading to development of more invasive metastatic diseaseGavin Robertson
12:00–12:30 Richard Marais
12:45–1:50 pm Networking Lunch and Poster and Exhibit Viewing
Grand Ballroom Gallery and Foyer
1:50–2:30 pm
Grand Ballroom Salons A-E
Special SMR Session
Late-breaking clinical trial results (previously unreported data), Part 1’
Session Chairs: Axel Hauschild, Dirk Schadendorf
1:50–2:00(LBA1 1) BREAK-2: A Phase IIA trial of the selective BRAF kinase inhibitor GSK2118436 in patients with BRAF (V600E / K) positive metastatic melanomaUwe Trefzer
2:00–2:10(LBA1 2) Vemurafenib improves overall survival compared to dacarbazine in advanced BRAFV600E-mutated melanoma: An update from the Phase III randomized, open-label, multicenter BRIM3 trialAxel Hauschild
2:10–2:20(LBA1 3) Phase II study of the MEK1 / MEK2 inhibitor GSK1120212 in metastatic BRAF V600E / K mutant cutaneous melanoma patients previously treated with or without a BRAF inhibitorKevin Kim
2:20–2:30(LBA1 4) Phase I / II expansion cohort of BRAF inhibitor GSK2118436 + MEK inhibitor GSK1120212 in patients with BRAF mutant metastatic melanoma who progressed on a prior BRAF inhibitorKeith Flaherty
2:30–4:15 pm
Grand Ballroom Salons A-E
Breakout Session IA – SMR
Understanding melanoma biology through model systems
Session Chairs: Sandy Anderson, Marcus Bosenberg
2:30–2:50Murine melanoma modelsLionel Larue
2:50–3:10Elucidating the role of PI3-kinase pathway activation in melanoma progression in vivoMartin McMahon
3:10–3:30Getting old and misbehaving: Can stromal aging drive melanoma initiation?Sandy Anderson
3:30–3:40(IA 1) New animal models for the analysis and visualization of lymphangiogenesis and metastasis of malignant melanomaDavid Olmeda
3:40–3:50(IA 2) Why red-heads are at increased risk of melanoma: A novel BRAF mutant mouse modelDevarti Mitra
3:50–4:00(IA 3) Dynamic intra-tumoral evolution of human melanomaMark Shackleton
Grand Ballroom Salon F
Breakout Session IB – SMR
Transcriptional control/epigenetic regulation of melanoma
Session Chairs: Silvya Stuchi Maria-Engler, Anja Bosserhoff
2:30–2:50Signaling and transcription in melanoma subpopulationsColin Goding
2:50–3:10c-Jun, a main regulator in melanomaAnja Bosserhoff
3:10–3:30Developmental chemical screens in zebrafish melanoma modelsRichard White
3:30–3:40(IB 1) MDMX is a key therapeutic target in melanomaJean-Christophe Marine
3:40–3:50(IB 2) SOX10 is an essential regulator of melanoma cell proliferationJulie Cronin
3:50–4:00(IB 3) Transcriptional plasticity and cycling between proliferative and invasive gene expression signatures: Clinical implicationsReinhard Dummer
4:15–4:30 pm Refreshment Break, Networking and Exhibit Viewing
4:30–6:15 pm
Breakout Session IIA – SMR
New therapeutic targets and strategies in melanoma
Session Chairs: Andrew Aplin, Nikolas Haass
4:30–4:50RAF265 inhibits the growth of advanced human melanoma tumorsAnn Richmond
4:50–5:10Mutant BRAF signaling and resistance to RAF inhibitorsAndrew Aplin
5:10–5:30A unified strategy for overcoming multiple mechanisms of vemurafenib resistanceKeiran Smalley
5:30–5:40(IIA 1) WIPI-1 links oncogenic BRAF to deregulated autophagy in melanomaPenny Lovat
5:40–5:50(IIA 2) ROCK and JAK1 signaling cooperate to control actomyosin contractility in melanomaVictoria Sanz-Moreno
5:50–6:00(IIA 3) Investigation of the melanoma genetic landscape identifies the glutamate pathway as a major player in the diseaseYardena Samuels
Breakout Session IIB – SMR-Pan-American Society for Pigment Cell Research Joint Session
Implications of melanocyte development for understanding and treating melanoma
Session Chairs: Frank Meyskens, Andrzej Slominski
4:30–4:50Melanogenesis in melanomas: therapeutic and metabolic implicationsAndrzej Slominski
4:50–5:10Sox2 and MITF cross regulatory interactions establish progenitor and melanocyte lineages in the neural crestPatrik Ernfors
5:10–5:30Implications of redox metabolism for cutaneous melanomaFrank Meyskens
5:30–5:40(IIB 1) Induction of primary brain melanoma by congenital expression of oncogenic NRAS in mouse melanocytesMalin Pedersen
5:40–5:50(IIB 2) Abrogation of oncogene-induced senescence by PI3K activation contributes to human melanomagenesisPatricia A. Possik
5:50–6:00(IIB 3) Dual role of apoptosis-associated speck-like protein containing a CARD (ASC) in tumorigenesis of human melanomaWeimin Liu
6:00–6:15Panel discussion/Q&A 

Joint session of The Society for Melanoma Research and The Interdisciplinary Melanoma / Skin Cancer Centres Friday November 11

6:50–7:45 am
Meet The Professor Breakfast 2A
Pathways 101: Molecular biology the clinician needs to know
Michael Davies, Caroline Robert, Grant McArthur, Richard Marais
Meet The Professor Breakfast 2B
Tissue is the Issue: Acquiring, banking and using primary and metastatic melanoma specimens
Ed Seijo, Shane Huntsman, Amod Sarnaik, Richard Scolyer, Jeff Gershenwald
7:50–10:15 am
7:50–8:00Welcome to Melanoma Centres participantsVernon Sondak, Claus Garbe
JOINT SYMPOSIUM I: Society for Melanoma Research Annual Congress and Interdisciplinary Melanoma/Skin Cancer Centres Meeting
Melanoma susceptibility, epidemiology and prevention
Session Chairs: Marianne Berwick, Sancy Leachman
8:00–8:15Genetic epidemiology of melanoma: What we know nowJulia Newton-Bishop
8:15–8:30Whole-genome and exome sequencing in melanoma familiesNicholas Hayward
8:30–8:45UV vs Vitamin D: How important is it?Marianne Berwick
8:45–9:00Overdiagnosis of melanoma: is it real?Martin Weinstock
9:00–9:15Using genetic test resultsSancy Leachman
9:15–9:30Personalized melanoma chemopreventionDoug Grossman
9:30–9:45National prevention efforts: EUJean-Jacques Grob
9:45–10:00Who is at high risk of getting lethal melanoma?Alan Geller
10:00–10:15Panel discussion/Q&A 
10:15–10:30 am Refreshment Break, Networking and Exhibit Viewing
10:30 am–12:30 pm
Joint SMR/Centres Symposium II
Immunology: From bench to bedside
Session Chairs: Antoni Ribas, Jeffrey Weber
10:30–10:45Immunobiology of immune checkpointsJedd Wolchok
10:45–11:00The melanoma tumor microenvironment as a predictive biomarker for response to immunotherapiesThomas Gajewski
11:00–11:15PD1 blockade in the treatment of metastatic melanomaMario Sznol
11:15–11:30Current and future status of adoptive cellular therapyJeffrey Weber
11:30–11:45Combining BRAF inhibitors with immunotherapyAntoni Ribas
11:45–12:00Combining ipilimumab with other drugsSteven Hodi
12:00–12:30Panel discussion/Q&A 
12:30–1:20 pm Networking Lunch and Poster and Exhibit Viewing
1:20–2:00 pm
Special Joint SMR/Centres Session
Late-breaking clinical trial results (previously unreported data) Part 2’
Session Chairs: Mark Middleton, Dirk Schadendorf
1:20–1:30(LBA2 1) Benchmarks for evaluating phase II clinical trials in stage IV melanoma: The recent SWOG experienceVernon Sondak
1:30–1:40(LBA2 2) Veliparib (ABT-888) plus temozolomide versus temozolomide alone: efficacy and safety in patients with metastatic melanoma in a randomized, double-blind, placebo-controlled trialMark Middleton
1:40–1:50(LBA2 3) Molecular testing for BRAF V600 mutations in the Phase III trial of the selective BRAF inhibitor vemurafenib in metastatic melanoma: A comparison of the cobas® 4800 BRAF V600E Mutation Test and Sanger sequencingJeffrey Lawrence
1:50–2:00(LBA2 4) Tasisulam sodium versus paclitaxel as second-line treatment in patients with metastatic melanoma: A randomized phase III studyOmid Hamid
2:00–4:15 pm
2:00–2:30 SMR Presidential Address
Session Chair: David Fisher
SMR Presidential Address: ‘The road ahead’Boris Bastian
Joint SMR/Centres Symposium III
Clinically important breakthroughs in melanoma genetics and genomics
Session Chairs: David Solit, Rick Kefford
2:30–2:50 Kevin Brown
2:50–3:10 Ze’ev Ronai
3:10–3:30Translating human melanoma geneticsMarcus Bosenberg
3:30–3:50Genetic predictors of RAF-dependenceDavid Solit
3:50–4:00(JIII 1) Squamous cell tumors from RAF inhibitor-treated patients have a distinct mutational profile supporting a mechanism of therapy-induced tumorigenesis in RAS-primed cellsDamien Kee
4:15–4:30 pm Refreshment Break, Networking and Exhibit Viewing
4:30–6:30 pm
Joint SMR/Centres Symposium IV
Working together to overcome drug resistance in melanoma
Session Chairs: Keith Flaherty, Michael Davies
4:30–4:45Overview of resistance pathways in melanomaMichael Davies
4:45–5:00Melanoma heterogeneity and therapyMeenhard Herlyn
5:00–5:15Acquired resistance mechanisms in melanoma patients treated with BRAF inhibitorsRoger Lo
5:15–5:30Update on BRAF resistanceKeith Flaherty
5:30–5:45Joint BRAF/MEKRick Kefford
5:45–6:00Joint MEK/AKTDirk Schadendorf
6:00–6:10(JIV 1) Therapeutic modulation of autophagy overcomes resistance to BRAF inhibition in BRAF mutant melanomaRavi Amaravadi
6:10–6:20(JIV 2) BRAF drugs accelerate the rate of growth of keratoacanthoma and squamous cell carcinomaAmaya Viros
6:20–6:30Panel discussion/Q&A 
7:00–9:30 pm
Society for Melanoma Research Awards Banquet and Dinner Cruise
Aboard the Yacht StarShip on Tampa Bay

Joint session of The Interdisciplinary Melanoma/Skin Cancer Centres and the Melanoma Update for Primary Care Physicians, Surgeons, Oncologists, and Dermatologists Saturday November 12

6:50–7:45 am
Meet The Professor Breakfast 3A
Immunology 101: Immunobiology the clinician needs to know
Antoni Ribas, Jedd Wolchok, James Mulé
Meet The Professor Breakfast 3B
Designing, implementing and using clinical pathways in the Melanoma Centre
Mark Schippits, Jonathan Zager, Claus Garbe
7:50–10:15 am
7:50–8:00Welcome to Melanoma Update participantsVernon Sondak
Joint Symposium I: Interdisciplinary Melanoma/Skin Cancer Centres Meeting and Melanoma Update for Primary Care Physicians, Surgeons, Oncologists & Dermatologists
30,000 foot view: What’s new in melanoma 2011
Session Chairs: John Thompson, Claus Garbe
8:00–8:20Histologic and molecular classification of melanomaRichard Scolyer
8:20–8:40Sentinel node biopsy/MSLT-1 updateJohn Thompson
8:40–9:00Adjuvant therapy overviewSanjiv Agarwala
9:00–9:20Immunotherapy for advanced diseaseAntoni Ribas
9:20–9:40Targeted therapy for advanced diseaseGrant McArthur
9:40–10:15Melanoma Centres Keynote Address: Melanoma in 2020: How we’ll treat it thenLex Eggermont
10:15–10:30 am Refreshment Break, Networking and Exhibit Viewing
10:30 am–12:30 pm
Breakout Session IIIA – Centres/Melanoma Update Surgery/Oncology Tracks
A closer look: Unanswered questions about adjuvant therapy for melanoma in 2011
Session Chairs: Paul Lorigan, Helen Gogas
10:30–10:45Are pegylated and standard interferon still relevant?Vernon Sondak
10:45–11:00Are European interferon regimens better?Claus Garbe
11:00–11:15Can we predict who will respond to interferon?Helen Gogas
11:15–11:30Adjuvant therapy with MAGE.A3 vaccines and bevacizumabPaul Lorigan
11:30–11:45Is checkpoint blockade the future of adjuvant therapy?Jeffrey Weber
11:45–12:00How do we move forward in the future?Paul Chapman
12:00–12:30Case-based panel discussion/Q&A 
Breakout Session IIIB – Melanoma Update Primary Care/Dermatology Tracks
Evaluation of pigmented skin lesions and the patient at high risk for melanoma
Session Chairs: Glen Bowen, Josep Malvehy
10:30–10:45Biopsy techniques: To shave or notMary Lien
10:45–11:00Mole mapping and digital photographyJosep Malvehy
11:00–11:30Dermoscopy & other diagnostic approachesAllan Halpern
11:30–11:45Topical imiquimod 5% cream for lentigo maligna followed by conservative staged excisions as a means of reducing surgical morbidityGlen Bowen
11:45–12:30Case-based panel discussion/Q&A 
12:30–1:30 pm Networking Lunch and Poster and Exhibit Viewing
1:30–2:00 pm
‘Hot Lunch’ Topics
2:00–4:15 pm
Breakout Session IVA – Centres and Melanoma Update Surgery Track
Advances in surgery and regional therapy for melanoma
Session Chairs: Merrick Ross, Jonathan Zager
2:00–2:15Indications for SLN biopsy in thin melanomaJeff Gershenwald
2:15–2:30New lymphatic mapping agentsVernon Sondak
2:30–2:45Limiting morbidity from inguinal lymphadenectomy: Videoscopic lymphadenectomy and other approachesKeith Delman
2:45–3:00Technical advances in melanoma surgeryJames Lewis
3:00–3:15Isolated limb infusion/perfusionJohn Thompson
3:15–3:30Intralesional therapy: Local/regional control and implications for systemic responseMerrick Ross
3:30–3:45Percutaneous hepatic perfusion for patients with melanoma liver metastasesJonathan Zager
Breakout Session IVB – Centres and Melanoma Update Medical Oncology Track
Case-based discussion: How do you treat metastatic melanoma today?
Tumor Board cases with audience response system
Session Chairs: Georgina Long, Christopher Lao
2:00–3:30Case-based panel discussion: How will we use the drugs we have? 
3:30–4:00Case-based panel discussion: Systemic approaches to brain metastasis 
Breakout Session IVC – Centres and Melanoma Update Primary Care/Dermatology Track
Cutaneous lymphoma and non-melanoma skin cancers
Session Chairs: Reinhard Dummer, Frank Glass
2:00–2:30Diagnosis and staging of mycosis fungoidesFrank Glass
2:30–3:00Peripheral T cell lymphoma treatmentReinhard Dummer
3:00–3:15The biology and treatment Merkel cell carcinomaRyan Fields
3:15–3:30Clinical, pathological and molecular study of the spectrum of skin tumors induced by RAF inhibitorsCaroline Robert
3:30–3:45Targeted therapy for non-melanoma skin cancerDirk Schadendorf
3:45–4:00Cutaneous sarcoma updateGrant McArthur
4:15–4:30 pm Refreshment Break, Networking and Exhibit Viewing
4:30–6:30 pm
Breakout Session VA – Centres and Melanoma Update All Tracks
Case-based discussion: Problem areas in melanoma treatment today
Tumor Board cases with audience response system
Session Chairs: Vernon Sondak, Katharina Kaehler
4:30–4:50Lentigo maligna melanoma 
4:50–5:10Desmoplastic melanoma 
5:10–6:00Primary and metastatic mucosal melanoma 
6:00–6:30Metastatic uveal melanoma 
Breakout Session VB – Melanoma Centres
Data management and clinical research challenges facing the Melanoma Centre
Session Chairs: Lauren Haydu, Mohammed Kashani-Sabet
4:30–4:45Defining the issuesLauren Haydu, Mohammed Kashani-Sabet
4:45–5:00Multinstitutional clinical and research datasetsJeffrey Gershenwald
5:00–5:15National melanoma data registriesClaus Garbe
5:15–5:30Challenges to clinical trial development and implementationRagini Kudchadkar
5:30–6:30Panel discussion/Q&A 

4th Melanoma Pathology Symposium of the International Melanoma Pathology Working Group Sunday November 13

8:20–10:00 am
Perspectives on classification and staging of melanocytic neoplasms
Session Chairs: Raymond Barnhill, Jane Messina
8:20–8:30Welcome and introductory remarksRaymond Barnhill
8:30–8:50A new perspective for the classification of melanocytic lesionsArtur Zembowicz
8:50–9:10The pathology report and prognostic factorsRichard Scolyer
9:10–9:30The AJCC and melanoma staging: A critical re-examinationAlain Spatz
9:30–9:50Sentinel lymph node biopsy: State of the artAlistair Cochran
9:50–10:10 am Refreshment Break, Networking and Exhibit Viewing
10:10 am–12:10 pm
Difficult case presentations with audience response system
Session Chairs: Raymond Barnhill, Martin Mihm
10:10–10:30Melanocytic nevi of special sitesPhilip Leboit
10:30–10:50Is nodular melanoma a distinct entity or a final common pathway?Martin Cook
10:50–11:10Spitz nevi & atypical Spitz tumorsLori Lowe
11:10–11:30Acral and subungual melanocytic lesionsKlaus Busam
11:30–11:50Melanocytic proliferations in sun-damaged skinJane Messina
11:50–12:10Desmoplastic melanocytic lesionsChristopher Shea
12:10–1:30 pm Lunch, Networking and Exhibit Viewing
1:30 pm–3:40 pm
Special techniques/molecular advances – Part I
Session Chairs: Raymond Barnhill, Klaus Busam
1:30–2:00Debate - Dysplastic nevi: Grading and guidelines for treatmentLyn Duncan/Frank Glass
2:00–2:20How to resolve the problems of dysplastic neviMichael Piepkorn
2:20–2:40Margins for melanocytic lesions: A re-appraisalScott Binder
2:40–3:00Mucosal melanocytic lesionsVictor Prieto
3:00–3:20Unusual presentations of metastatic melanomaGilles Landman
3:20–3:40Angiotropism and extravascular migratory metastasis: An updateRaymond Barnhill
3:40–4:00 pm Refreshment Break, Networking and Exhibit Viewing
4:00 pm–6:00 pm
Special techniques/molecular advances – Part II
Session Chairs: Raymond Barnhill, Klaus Busam
4:00–4:20Melanoma stem/initiating cells from the pathologist’s perspectiveJoost van den Oord
4:20–4:40Melanoma stem cells: Not rare but well-doneGeorge Murphy
4:40–5:00Update on FISH as a diagnostic and prognostic tool for melanocytic neoplasmsPedram Gerami
5:00–5:20Application of microRNAs to melanocytic lesionsVictor Tron
5:20–5:40Prognosis in thin melanoma: An updateDavid Elder
5:40–6:00Molecular advances in melanomaBoris Bastian

Congress Proceedings Abstracts of the 8th International Congress of the Society for Melanoma Research November 9–11, 2011 Invited Faculty Abstracts

  1. Top of page
  2. 2011 International Melanoma Congress Advancement Through Collaboration joins the following four annual meetings in 2011:
  3. Congress Agenda
  4. Congress Proceedings Abstracts of the 8th International Congress of the Society for Melanoma Research November 9–11, 2011 Invited Faculty Abstracts
  5. Abstracts of the 5th Meeting of Interdisciplinary Melanoma/Skin Cancer Centres November 11–12, 2011 Invited Faculty Abstracts
  6. Abstracts of the 4th Melanoma Pathology Symposium of the International Melanoma Pathology Working Group November 13, 2011 Invited Faculty Abstracts
  7. Abstracts of the 8th International Congress of The Society for Melanoma Research November 9–11, 2011 Proffered Abstracts Selected for Oral Presentation
  8. Abstracts of the 8th International Congress of The Society for Melanoma Research November 9–11, 2011 Proffered Abstracts Selected for Poster Presentation
  9. Abstracts of the 5th International Meeting of Interdisciplinary Melanoma/Skin Cancer Centres November 11–12, 2011 Proffered Abstracts Selected for Poster Presentation

Roles for MITF in melanomagenesis

D. E. Fisher

Department of Dermatology, Massachusetts General Hospital, Harvard Medical School, Boston, MA, USA

The MITF transcription factor plays a central role in regulation of melanocyte lineage development and survival. It is regulated by key differentiation pathways, such as the MSH/MC1R signaling pathway that controls pheomelanin-eumelanin switching, as well as melanoma risk. Through its transcriptional control of specific melanin producing factors, MITF is thought to contribute importantly to pigment production in melanocytes, which is likely to play a key role in determining melanoma risk, via skin lightness or darkness. It has also been shown that genomic amplification of MITF appears to be oncogenic in a fraction (~15%) of cutaneous or advanced melanomas. The mechanism(s) through which MITF contributes to tumorigenesis are unclear, particularly if they are distinct from MITF's roles in pigmentation. This presentation will present recent data which implicate a point mutant variant of MITF in conferring significantly elevated melanoma risk in human populations. The variant appears likely to exhibit increased biological activity of MITF, because it is associated with non-blue eye color phenotype in affected individuals. Additional studies will be presented which focus upon novel signaling regulatory circuits controlling MITF expression, and evidence for a previously unrecognized link between pheomelanin and BRAF-associated melanoma risk.

Mice offer new wrinkles on UV damage and melanoma progression

G. Merlino1, R. Zaidi1, C.-P. Day1, P. Mishra1,2, T. Guo1, P. Mishra1, T. Hornyak3, H. Arnheiter4, F. Noonan2, G. Trinchieri5, H. Young5, S. Davis6, P. Meltzer6, E. De Fabo2

1 Laboratory of Cancer Biology and Genetics, NCI, Bethesda, MD, USA; 2George Washington University Medical Center, Washington DC, USA; 3Dermatology Branch, NCI, Bethesda, MD, USA; 4NINDS, Bethesda, MD, USA; 5Cancer and Inflammation Program, NCI, Frederick, MD, USA; 6Genetics Branch, NCI, Bethesda, MD, USA

Ultraviolet radiation (UV) is a long-appreciated major risk factor for melanoma, but the underlying mechanisms that incite melanomagenesis remain poorly understood. We have generated a genetically engineered mouse model (iDCT-GFP) that expresses Green Fluorescent Protein specifically in all melanocytic cells in a doxycycline-regulated manner. This mouse model was used to study the effects of melanomagenic neonatal UVB on the transcriptome of melanocytes in their natural in vivo microenvironment. Expression profiling demonstrated a delayed upregulation of an Interferon (IFN) response gene signature in melanocytes activated by UVB and isolated by FACS. Antibody-mediated systemic blockade showed that IFN-γ was responsible for UVB-mediated melanocyte activation, associated with infiltrating macrophages recruited to the neonatal skin by UVB-induced CCR2 ligands. Neonatal skin macrophages enhanced tumor growth in an IFN-γ-dependent manner when admixed with mouse melanoma cells and transplanted subcutaneously into syngeneic mice. These data suggest that melanocytes actively participate in UVB-induced pro-tumorigenic skin inflammation through macrophage crosstalk, which features IFN-γ as a critical signaling component promoting melanocytic survival and immunoevasion.

Melanomas are among the most aggressive cancers, especially in their ability to metastasize. We hypothesized that advanced metastatic melanomas can exploit hard-wired pathways employed by migratory embryonic melanocytes to achieve a more aggressive malignant phenotype. The iDCT-GFP mouse was therefore used in concert with FACS to isolate melanoblasts from various developmental stages, sequence their transcriptomes using RNA-Seq, and identify genes/pathways common to metastatic melanoma. This study has already provided a full library of the genes expressed throughout embryonic development of the melanocyte. Currently, functional assays for a dozen promising gene candidates are being performed in human metastatic melanoma cell lines. We hope to identify novel therapeutic targets and biomarkers for late stage melanoma.

Senescence and melanoma suppression

D. S. Peeper

Division of Molecular Genetics, Netherlands Cancer Institute, Amsterdam, The Netherlands

Although nevi are considered to be precursors of melanoma, little is known about the mechanism underlying progression from nevus to melanoma. We and others have previously shown that, although initially, activated BRAF and NRAS oncogenes act mitogenically, eventually oncogene-induced senescence (OIS) ensues, suggesting that abrogation of OIS of nevus cells acts as a rate-limiting event in melanomagenesis. Favoring this model, while nevi and melanomas are commonly histologically associated, malignant melanoma can emerge within a nevus.

The involvement of p16INK4a in melanomagenesis is undisputed, but there is little evidence to support a non-redundant role for p16INK4a in BRAF(E600)- or NRAS(K61)-induced senescence, neither in vitro nor in vivo. Another common genetic event in melanoma is the activation of the PI3K pathway, which is seen in ~60% of cases. This is achieved by loss of PTEN expression, increased AKT3 activity or mutations in PIK3CA. Interestingly, some 20% of melanomas show concurrent mutation in BRAF and diminished expression of PTEN.

We have found that in cultured human melanocytes, BRAF(E600)-induced senescence is accompanied by suppression of AKT3. Activation of the PI3K pathway by ectopic expression of AKT or PIK3CA, or depletion of PTEN, abrogates senescence. Correspondingly, in a series of contiguous human nevus-melanoma specimens, we observed a decrease in PTEN and/or an increase in AKT3 in the melanoma relative to the adjacent nevus in >50% of the cases. In several of these, laser microdissection-guided genetic analysis revealed identical mutations in BRAF or NRAS in the nevus and contiguous melanoma, including a rare double mutation, supporting support a nevus-to-melanoma progression model. These findings indicate that PI3K pathway activation serves as a rate-limiting event in human melanoma progression, acting at least in part by abrogating OIS. This provides an explanation for the frequent co-occurrence of mutations in BRAF and the PI3K pathway in melanoma.

BAP1 mutations in melanoma

J. W. Harbour

Washington University School of Medicine, St. Louis, MO, USA

Recently, we reported that inactivating mutations in BAP1 (BRCA1-associated protein 1) are strongly associated with metastasis in uveal melanoma and can occasionally. Most of these mutations are somatic in origin, but they can also arise in the germline and lead to inherited melanoma syndromes. Subsequently, BAP1 mutations increasingly are being found in cutaneous melanoma, mesothelioma and other cancers. Despite the importance of BAP1 in melanoma, the normal function of BAP1 and the consequences of its loss in the melanocytic lineage remain unclear. This presentation will update the current state of knowledge regarding BAP1 in melanoma.

Vemurafenib (PLX4032) promotes epigenetic changes in melanoma cells leading to development of more invasive metastatic disease

C.-Y. Chung1,2, P. Khanna3, R. Gowda1,2,3, A. Sharma1,2,3, R. Neves1,2,4,5,6, C. Dong2,3,6, G. P. Robertson1,2,4,5,6,7,8

1 Department of Pharmacology, The Pennsylvania State University College of Medicine, Hershey, PA, USA; 2Penn State Melanoma Therapeutics Program, The Pennsylvania State University College of Medicine, Hershey, PA, USA; 3Department of Bioengineering, The Pennsylvania State University College of Medicine, Hershey, PA, USA; 4Department of Dermatology, The Pennsylvania State University College of Medicine, Hershey, PA, USA; 5Department of Surgery, The Pennsylvania State University College of Medicine, Hershey, PA, USA;6Penn State Melanoma Center, The Pennsylvania State University College of Medicine, Hershey, PA, USA; 7The Foreman Foundation for Melanoma Research, The Pennsylvania State University College of Medicine, Hershey, PA, USA; 8Department of Pathology, The Pennsylvania State University College of Medicine, Hershey, PA, USA

PLX4032, clinically known as vemurafenib, is a V600EBRAF selective inhibitor. It is effective in patients containing V600EBRAF protein, leading to an ~80% partial or complete anti-tumor response rate during the first 2 month treatment cycle. An average regression period of 2–18 and 6.2 months progression-free survival is observed but all patients eventually relapse developing drug resistant invasive disease. Recurrence can be caused by secondary BRAF mutations, alternate pathways of MAPK reactivation, or activation of compensating alternative survival pathways. The mechanisms promoting disease recurrence to BRAF targeting agents are an extremely important area of research for the clinical management of melanoma, which remains to be completely unraveled and the epigenetic contribution to this process in unknown. Once the mechanisms are completely elucidated, this information would be useful for designing better approaches to prevent resistance and disease recurrence. This study demonstrates that an acquired more invasive resistant phenotype can occur following treatment with BRAF inhibitors by increasing methyl transferases activity to promote epigenetic silencing of genes regulating this process. Vemurafenib treatment led to promoter methylation and silencing of the invasion suppressor CD82 in melanoma cells. Lack of CD82 in turn increased the invasive potential of the cells, promoting migration through vessel and capillary walls, thereby aiding development of metastases. Invasive metastatic disease mediated by silencing of CD82 could be reversed using 5-aza-2′-deoxycytidine (5AzaC), clinically known as decitabine, which then decreased the invasive phenotype mediated by these agents. These observations suggest that combining BRAF targeting with DNA demethylating agents might be one clinically effective approach to overcome the development of resistant invasive melanoma following treatment with agents such as vemurafenib.

Murine melanoma models

L. Larue

Institut Curie – Developmental Genetics of Melanocytes, UMR3347 CNRS, U1021 INSERM, Orsay, France

Melanocytes are derived from neural crest cells. Genetic defaults may lead to coat color abnormality, early greying and melanoma. Melanoma is a very aggressive tumor and is responsible of the death of 80% of the patients having a cancer of the skin. Incidence in western countries continue to raise. In order to better understand initiation and progression of melanoma, our research based on normal and pathological development of melanocytes combines (i) a molecular approach to better understand cellular signaling associated with β-catenin involving Wnt, PI3K, Map-kinase and cadherins, (ii) a cellular approach to reveal cellular specifity of melanocytes and (iii) a physiopathological approach, based on murine models to understand the establishment, maintenance and alteration (transformation and hypo/hyperpigmentation) of this lineage. We will present original murine models presenting coat color abnormalities and affecting BRAF, NRAS, INK4A, CTNNB1, PTEN and CDH1, which allowed a better understanding of proliferation, immortalization and invasion.

Elucidating the role of PI3’-kinase pathway activation in melanoma progression in vivo

V. Marsh1, J. Silva1, M. Bosenberg2, W. Phillips3, M. McMahon1

1 U. C. San Francisco/Helen Diller Family Comprehensive Cancer Center, CA, USA; 2Yale School of Medicine, New Haven, CT, USA; 3Peter MacCallum Cancer Centre, Melbourne, Vic., Australia

The majority of melanomas express mutationally activated BRAF leading to sustained activation of BRAF[RIGHTWARDS ARROW]MEK[RIGHTWARDS ARROW]ERK MAP kinase signaling. Progression of BRAF mutant melanomas is frequently associated with silencing of the PTEN tumor suppressor, a negative regulator of PI3’-kinase signaling. Whilst PTEN silencing is common, mutational activation of PIK3CA (encoding the catalytic subunit of PI3’-kinase-α) is extremely rare, even though PIK3CA mutations are common in other human cancers. Since PTEN encodes a direct antagonist of PI3’-kinase, this observation is surprising given that either PTEN silencing or PIK3CA mutational activation is predicted to promote PI3’-kinase signaling and might therefore be interchangeable. However, since PTEN is reported to have tumor suppressor functions independent of its lipid phosphatase activity, PTEN silencing may have more profound tumor promoting activity than mutational activation of PIK3CA. To test this, we generated Pik3calat−1047R mice carrying a conditional knock-in allele of Pik3ca that permits Cre-mediated conversion of normal PIK3CA to PIK3CAH1047R. To achieve temporal and spatial control of oncogene expression, we used Tyr::CreER mice carrying the melanocyte-specific Tyrosinase::CreERT2 transgene. Tamoxifen treatment of Tyr::CreER; Pik3caHR mice did not result in an obvious melanocytic phenotype even up to 6 months after administration. This is consistent with the lack of a melanocytic phenotype in tamoxifen treated Tyr::CreER; Ptenlox/lox mice. We have previously shown that melanocyte specific expression of BRAFV600E expression combined with PTEN silencing leads to metastatic melanoma in Tyr::CreER; BRafCA/+; Ptenlox/lox mice. To test whether PIK3CAH1047R would substitute for PTEN silencing in this model, we treated Tyr::CreER; BRafCA/+; Pik3calat−1047R/+ mice topically with tamoxifen. Preliminary data indicates that co-expression of BRAFV600E plus PIK3CAH1047R in melanocytes leads to primary cutaneous melanoma, in a manner that is similar to that arising in Tyr::CreER; BRafCA/+; Ptenlox/lox mice. These data suggest mutational activation of PI3’-kinase signaling cooperates with mutationally activated BRAFV600E in melanoma progression. This observation is intriguing given that PIK3CA mutation is rarely detected in human melanomas. Using genetic and pharmacological inhibitors of BRAFV600E[RIGHTWARDS ARROW]MEK[RIGHTWARDS ARROW]ERK or PI3’KH1047R[RIGHTWARDS ARROW]PDK[RIGHTWARDS ARROW]AKT signaling we are currently seeking to define mechanism(s) of oncogene cooperation between BRAFV600E and PIK3CAH1047R in melanoma progression.

Getting old and misbehaving: can stromal aging drive melanoma initiation?

E. Kim1, D. Basanta1, V. Rebecca2, J. Messina3, R. Mathew3, K. S. Smalley1,2,4, A. R. A. Anderson1

1 Integrated Mathematical Oncology Department, H. Lee Moffitt Cancer Center and Research Institute, Tampa, FL, USA; 2Department of Cutaneous Oncology, H. Lee Moffitt Cancer Center and Research Institute, Tampa, FL, USA; 3College of Medicine Pathology and Cell Biology, University of South Florida, Tampa, FL, USA; 4Department of Molecular Oncology, H. Lee Moffitt Cancer Center and Research Institute, Tampa, FL, USA

Melanoma initiation is known to involve genetic changes as well as the disruption of both cell-cell and cell-microenvironment interactions. However, the mechanisms by which the deregulated interactions lead to melanoma development have not yet been fully understood. It is our view, that we must first model normal skin form and the regulatory mechanisms that maintain skin homeostasis before we can understand cancer initiation. To this end, we have developed a hybrid multiscale mathematical model of normal skin (virtual skin). The model focuses on key cellular (melanocytes, keratinocytes, fibroblasts) and microenvironmental variables (growth factors, ECM) that regulate normal skin homeostasis.

Using our virtual skin model, we examined how microenvironmental perurbations affect skin structure and interactions between keratinocytes, melanocytes, and fibroblasts. The simulations show that the virtual skin recovers from small changes but fails to compensate for large perturbations due to altered feedback between keratinocytes, melanocytes and fibroblasts.

Based on our experiments, we found that as fibroblasts age and become senescent their behavior changes significantly, leading to the expression of multiple growth factors, matrix proteins and proteases. These age-associated phenotypic changes in fibroblasts can be viewed as a permanent perturbation of the skin microenvironment for both keratinocytes and melanocytes. We incorporated senescent fibroblasts into our virtual skin model to examine how these permanent microenvironmental changes affect skin structure. Interestingly model simulations suggest that aged fibroblasts bias the skin microenvironment to favor the development of nevi.

Collectively, microenvironmental changes affect melanocyte homeostasis greatly. However, successful melanoma initiation may require preexisting melanocyte mutations that allow them to exploit this stromal activation. To test this hypothesis we incorporated mutant melanocytes into the model and were able to recapitulate a spectrum of true aberrant clinical pathologies. We speculate that senescent fibroblasts may create a pro-oncogenic environment that synergizes with mutations to drive melanoma initiation and progression.

Signalling and transcription in melanoma subpopulations

C. R. Goding1,2

1 Ludwig Institute for Cancer Research; 2University of Oxford, Oxford, UK

Increasing evidence suggests that one of the major driving forces behind melanoma metastasis is microenvironment-mediated changes to the transcription program that underlies melanoma cell identity. The phenotype-switching model proposes that melanoma cell identity is plastic, and can reversibly switch from a proliferative to a differentiated or stem cell-like phenotype, and that activating mutations in oncogenes such as BRAF may render melanoma cells phenotypically unstable. The different biological characteristics of melanoma sub-populations are dependent on the activity and expression of the Microphthalmia-associated transcription factor MITF, with MITF–low cells adopting a stem-like invasive phenotype. However, siRNA mediated depletion of MITF in melanoma cell lines can lead to DNA-damage and senescence, raising the possibility that in vivo low levels of MITF may be counterbalanced by signals that suppress senescence. Moreover, it is also likely that melanoma stem-like cells will share characteristics with their physiological stem cell counterparts. Our results suggest that distinct melanoma subpopulations are defined by a complex interplay between signaling and transcription that regulates cell identity, suppresses senescence and presents remarkable parallels to that operating to regulate physiological melanocyte stem cells.

c-Jun, a main regulator in melanoma

A. K. Bosserhoff

Institute of Pathology, Molecular Pathology, University of Regensburg, Regensburg, Germany

A deregulation of transcriptional activity, which is often found during tumor development, alters several biological processes. The constitutive activity of the transcription factor superfamily AP-1 (hetero- or homodimers of c-Jun, c-Fos, JunB or further family members) influences the expression of a multitude of regulators of cell proliferation, migration and survival which are significantly involved in tumor development and metastasis. Although the relevance of AP-1 for cancerogenesis is known, molecular mechanisms for the regulation of constitutive activity in diverse tumors are largely unknown. Our experiments revealed that cell-cell contacts via E-cadherin have a negative impact on c-Jun expression in melanocytes. Concomitant with the loss of E-cadherin expression during melanoma development induction of c-Jun expression is acquired. To analyze the role of c-Jun in regulation of gene expression in melanoma, we transfected melanoma cells showing high levels of c-Jun using either a dominant negative mutant form of c-Jun or shRNA. Both attempts resulted in a strong reduction of AP-1 activity in the transiently transfected cells suggesting that c-Jun is the main regulator of strong AP-1 activity in melanoma. The importance of c-Jun for proliferation and survival of melanoma cells is underlined by the fact that stable transfection of shRNA against c-Jun led to mortality of all analyzed melanoma cell lines. Interestingly, we could further demonstrate that c-Jun is regulated by E-cadherin on post-transcriptional level and it seems that this adhesion-dependent regulation is mainly influenced by cytoskeletal organization. In summary, we defined the molecular signaling mechanisms of E-cadherin dependent gene regulation and determined the impact of c-Jun for melanoma development and progression.

Developmental chemical screens in zebrafish melanoma models

R. M. White1,2,3, L. I. Zon1,2,3

1 Dana Farber Cancer Institute, Boston, MA, USA; 2Children's Hospital Boston, Boston, MA, USA; 3Harvard Medical School, Boston, MA, USA

The developmental programs regulating neural crest stem/progenitor cells are recapitulated during melanoma genesis. To functionally characterize these programs, we developed a zebrafish melanoma model in which BRAFV600E is expressed under the MITF promoter. When crossed with p53−/− mutants, 100% of animals develop melanoma between 4 and 12 months. These animals exhibit abnormal overexpression of key neural crest factors such as SOX10, MITF and EDNRB. Using this model, we undertook an in vivo chemical genetic screen to identify suppressors of neural crest development that may have utility in melanoma. Of 2000 small molecules, 0.8% caused a diminished expression of crestin, a pan-neural crest marker. One class of compounds, inhibitors of DHODH such as leflunomide, caused a nearly complete loss of neural crest stem/progenitor cells in both zebrafish and mammalian systems. Leflunomide exerts this effect by modulating the transcriptional elongation of genes required for neural crest development, including MYC targets, MITF, SOX10 and EDNRB, but had no effect on control genes. DHODH genetically interacts with the SPT5 transcription elongation factor, leading to transcriptional stalling of these genes. ChIP-seq reveals that transcriptional stalling is widespread in human melanoma cells, affecting >30% of loci, and that leflunomide potentiates this effect. Treatment of human melanoma cells with leflunomide led to a marked abrogation of cell proliferation, an effect that was additive or synergistic with the BRAFV600E inhibitor PLX4720. In mouse xenografts, the combination of leflunomide + PLX led to a near complete abrogation of tumor growth. The results indicate that transcriptional elongation is a key regulatory mechanism in both neural crest and melanoma development, which can be therapeutically targeted by molecules such as leflunomide. Based on our preclinical data, we plan to initiate a phase II trial of leflunomide + PLX in the near future.

RAF265 inhibits the growth of advanced human melanoma tumors

Y. Su1, M. C. Kelley2, R. C. Splittgerber1, S. Kantrow3, S. P. Short1, T. Sobolik-Delmaire1, S. Zaja-Milatovic1, Y. Liu1, A. E. Vilgelm1, O. E. Hawkins1, K. B. Dahlman1, K. I. Amiri1, A. Jiang4, P. Lu4, Y. Shyr4, D. D. Stuart5, S. Levy6, J. A. Sosman7, A. Richmond1,8

1 Department of Cancer Biology, Vanderbilt-Ingram Cancer Center, Vanderbilt University Medical Center, Nashville, TN, USA; 2Division of Surgical Oncology, Department of Surgery, Vanderbilt University Medical Center, Nashville, TN, USA; 3Pathology Consultants, St. Thomas Hospital, Nashville, TN, USA; 4Division of Cancer Biostatistics, Department of Biostatistics, Vanderbilt University Medical Center, Nashville, TN, USA; 5Novartis Institutes for Biomedical Research, Emeryville, CA, USA; 6Department of Biochemistry, Vanderbilt University School of Medicine, Nashville, TN, USA; 7Division of Hematology/Oncology, Department of Internal Medicine, Vanderbilt University Medical Center, Nashville, TN, USA; 8Department of Veterans Affairs, Vanderbilt University Medical Center, Nashville, TN, USA

RAS-RAF-MEK-ERK signaling is a major factor for growth and progression of human melanoma tumors. RAF265 is a RAF inhibitor that targets all ARAF, BRAF and CRAF as well as mutant V600E BRAF, VEGFR2, PDGFR, CSF1R, RET and cKIT. The purpose of this study was to determine how mutation of BRAF affected the response to RAF265. We utilized a tumor orthotopic implant model of early passage melanoma tumors in nude mice from a series of 17 patients with advanced metastatic melanoma. Tumor growth was compared between RAF265 treatment (40 mg/kg, QD) and diluent control groups. The melanoma associated gene mutation profile and global gene expression profile were determined on these human melanoma samples by SNaPshot and Affymetrix Human Gene ST 1.0 Array, respectively. Nine of the 17 tumors (53%) evaluated for growth response to RAF265 exhibited the V600E mutation of BRAF, while 8 (47%) were wild type for BRAF. Tumor implants from 7 of 17 patients (41%) responded to RAF265 treatment, where response was defined as more than 50% tumor volume reduction in RAF265 treated groups relative to controls. Five of the 7 (71%) responders were BRAF wild type, while only two responders (29%) harbored V600E BRAF mutation. Of the seven responding tumors WT for BRAF, one exhibited N-RASQ61R mutation and another exhibited c-KITL576P mutation. For 5/7 tumors responding to RAF265 treatment, there was a significant inhibition of Ki67 staining with RAF265 treatment. Analysis of the microarray data revealed responders exhibited enriched expression of genes involved in cell cycle, apoptosis, cell-cell adhesion and initiation of epithelial/mesenchymal transition. It is concluded that RAF265 significantly inhibits the growth of a sub-population of V600E mutant and wild type BRAF human melanoma tumors in vivo and the gene expression profile of this subset of tumors that may predict response to RAF265. RAF265 may be a useful therapy for patients that do not harbor a BRAF mutation. Additional work will be needed to determine whether patients developing resistance to vemurafinib or other agents that specifically target the mutant BRAF will respond to RAF265.

Mutant BRAF signaling and resistance to RAF inhibitors

A. E. Aplin, E. V. Abel, K. Basile, F. Kaplan, Y. Shao

Department of Cancer Biology, Kimmel Cancer Center, Thomas Jefferson University, Philadelphia, PA, USA

Mutant BRAF signaling drives malignant properties in melanoma cells. Targeting mutant BRAF with PLX4032 (vemurafenib, Zelboraf®) frequently elicits clinical benefit that is short term due to acquired resistance. Recently, we demonstrated that the stemness factor, FOXD3, is up-regulated following inhibition of mutant BRAF signaling in melanoma cells. Preventing upregulation of FOXD3 rendered melanoma cells more susceptible to cell death induced by PLX4032/4720. Based on these findings, we sought to determine targets of FOXD3. Chromatin-immunoprecipitation followed by next generation sequencing (ChIP-seq) and expression array profiling highlighted ERBB3 as a direct target of FOXD3. Furthermore, ligand-induced activation of ErbB3-ErbB2 signaling is enhanced in PLX4032-treated and FOXD3-expressing cells. These data highlight a signaling pathway that is sensitized as an adaptive response to RAF inhibitors. While short-term responses may provide tolerance to RAF inhibitors, permanent changes such as acquired mutations in NRAS (Nazarian et al., Nature 2010) provide long-term resistance. Expression of mutant NRAS in mutant BRAF cells is sufficient to prevent PLX4720 inhibition of the ERK1/2 pathway and promote cell cycle progression and resistance to apoptosis in the presence of RAF inhibitor. We have isolated cells that are resistant to PLX4720 and have gained an acquired mutation in NRAS. These acquired resistant cells are dependent on NRAS and the RAS-RAF scaffold molecule, SHOC-2, to by-pass PLX4720 inhibition of the MEK-ERK1/2 pathway. Together, these data underscore the likelihood that multiple mechanisms confer resistance to RAF inhibitors.

A unified strategy for overcoming multiple mechanisms of vemurafenib resistance

K. S. M. Smalley1,2, K. H. T. Paraiso1, A. Sarnaik2, V. K. Sondak2, H. E. Haarberg1, J. Koomen1

1 Department of Molecular Oncology, The Moffitt Cancer Center & Research Institute, Tampa, FL, USA; 2Department of Cutaneous Oncology, The Moffitt Cancer Center & Research Institute, Tampa, FL, USA

Acquired and intrinsic vemurafenib resistance is a major factor that limits the long-term disease management of BRAF V600E mutant melanoma. To date, a diverse array of resistance mechanisms have been reported that are likely to complicate future clinical trial design. Here we demonstrate that most or all of the signaling proteins implicated thus far in the escape from vemurafenib therapy (ARAF, CRAF, COT, IGF1R, MEK1, mTOR, AKT, cyclin D1) are clients of heat shock protein (HSP)-90, and were rapidly degraded following treatment with an inhibitor of HSP90 (XL888). XL888 was found to inhibit the growth and survival of melanoma cell lines with intrinsic and acquired resistance to vemurafenib, whether mediated through COT overexpression, cyclin D1 overexpression, increased PDGFR expression, PTEN loss or the presence of mutant NRAS. HSP90 inhibition was also highly effective at overcoming vemurafenib resistance in melanoma cell lines grown under 3D cell culture conditions and as mouse xenografts. In long-term colony formation assays, vemurafenib alone led to the emergence of resistant colonies in every case, whereas XL888 treatment completely abrogated the emergence of resistant colonies. Mechanistically, XL888 was found to induce expression of BIM at the mRNA and protein levels and to reduce BIM phosphorylation at Ser69 as well as downregulating the expression of Mcl-1. These studies provide the preclinical rationale for the dual targeting of mutant BRAF and HSP90 and phase I clinical trials are being initiated to validate these data in melanoma patients.

Melanogenesis in melanomas: therapeutic and metabolic implications

A. Slominski

Department of Pathology and Laboratory Medicine, University of Tennessee Health Science Center, Memphis, TN, USA

Melanoma is a neoplasm of melanocytic origin. The main function of epidermal melanocytes is to produce melanin pigment, which protects cells against the harmful actions of UV radiation, and attenuates melanomagenesis. Paradoxically melanin can contribute to the resistance of melanoma to different forms of therapy including traditional (chemo- or radiotherapy) or more specific (photodynamic- or immunotherapy) types. This negative outcome is likely due to intrinsic properties of melanogenic pathway, e.g., it consists of series of tightly coupled oxidoreduction reactions that generate several intermediates displaying cytotoxic, genotoxic or mutagenic activities, it consume oxygen leading to relative intracellular hypoxia and melanin acts as a scavenger of free radicals, metal cations, cellular toxins including chemotherapeutics (Physiol. Rev. 84: 1155–1228, 2004). Based on the above, it has been proposed that uncontrolled melanogenesis may have a role, perhaps critical, in the progression of melanotic variants of melanoma (Anticancer Res. 18: 3709–3715, 1998; Physiol. Rev. 84: 1155–1228, 2004). Moreover, melanogenesis has immunosuppressive effects that could abrogate immunotherapy directed at melanoma cells and can affect expression of important intracellular regulatory pathways (NF-κB, vitamin D receptors, HIF-1) changing the behavior of the cells (Int. J. Cancer 124: 1470–1477 and 123: 1448–1456, 2009). Recently, using cultured in vitro human melanoma cells, we have obtained additional data supporting the concept that the melanotic melanoma phenotype attenuates the effects of radio-, chemotherapy and immunotherapy. Specifically, inhibition of melanogenesis increased the tumorocidal effects of radio-/chemotherapy in vitro, and enhanced lymphocyte mediated cytotoxicity towards melanoma cells. Therefore, a hypothesis was proposed that inhibition of melanogenesis will amplify the effectiveness of existing therapeutic strategies such as chemotherapy, radiotherapy, phototherapy and immunotherapy and, will even directly inhibit melanoma growth. In this direct or combined treatment inhibitors of the melanogenic pathway would be used to increase the efficacy of the main therapy.

Sox2 and MITF cross regulatory interactions establish progenitor and melanocyte lineages in the neural crest

I. Adameyko1, F. Lallemend1, A. Furlan1, N. Zinin2, S. Aranda1, S. S. Kitambi1, A. Blanchart1, R. Favaro3, S. Nicolis3, M. Lübke1, T. Müller4, C. Birchmeier4, U. Suter5, Y. Takahashi6, P. Ernfors1

1 Unit of Molecular Neurobiology, Department of Medical Biochemistry and Biophysics, Karolinska Institute, Stockholm, Sweden; 2Department of Microbiology, Tumor and Cell Biology, Karolinska Institute, Stockholm, Sweden; 3Department of Biotechnology and Biosciences, University of Milano-Bicocca, Milano, Italy; 4Max-Delbrück-Centrum for Molecular Medicine, Berlin, Germany; 5Institute of Cell Biology, Department of Biology, ETH Zürich, Zürich, Switzerland; 6Graduate School of Biological Sciences, Nara Institute of Science and Technology, Takayama, Ikoma, Nara, Japan

The cellular origin and molecular mechanisms regulating pigmentation of head and neck are largely unknown. Melanocyte specification is controlled by the transcriptional activity of MITF, but no general logic has emerged to explain how MITF and progenitor transcriptional activities determine the fate of melanocyte and progenitor cells. We find that cranial melanocytes arise from at least two different cellular sources, early from nerve-associated Schwann cell precursors (SCPs) and later from a cellular source independent of nerves. Unlike the midbrain-hindbrain cluster where melanoblasts arise independently of nerves, a large center of melanocytes in and around cranial nerves IX and X is derived from SCPs, evidenced by genetic cell lineage tracing and analysis of ErbB3 null mutant mice. Conditional gain and loss of function experiments show genetically that the establishment of cell fates in neural crest cells and SCPs involve the SRY transcription factor Sox2 and MITF whereby Sox2 maintains a SCP progenitor state and its cross-regulatory interactions with MITF underlies a commitment to melanocytes. Hence, a gradual downregulation of Sox2 in progenitors during development fates the emergence of both neural crest- and SCP-derived progenitors into melanocytes. By this mechanism, an initial small pool of nerve associated melanoblasts expands in number and disperse under the control by Endothelin receptor B (EDNRB) and Wnt5a signaling.

Implication of redox metabolism for prevention and treatment of cutaneous melanoma (CM)

F. L. Meyskens1–5, S. Yang1–5, F. Liu1,2, R. Dellinger1,2, T. Poulos6, H. Ji7, R. Silverman7

1 Chao Family Comprehensive Cancer Center, University of California, CA, USA; 2Department of Medicine, University of California, CA, USA; 3Department of Biological Chemistry, University of California, CA, USA; 4Department of Public Health, University of California, CA, USA; 5Department of Pharmaceutical Sciences, University of California, CA, USA; 6Department of Molecular Biology and Biochemistry, University of California, CA, USA; 7Department of Chemistry and Molecular Biosciences, Chemistry of Life Processes Institute, Center for Molecular Innovation and Drug Discover, Northwestern University, Evanston, IL, USA

The human melanocyte is constantly exposed to potential intrinsic (melanin synthesis, melanin) and extrinsic (UV-activation of many redox-sensitive pathways) redox active molecules. We will summarize updates on our work with apurinic endonuclease/Redox effector factor-1 (APE/Ref-1), including the design and synthesis of new inhibitors, and new directions involving the lipo-oxygenase (LOX) and NADPHoxidase (NOX) pathways. We will present also in some detail on our studies of neuronal nitric oxide synthase/nitric oxide (nNos)/NO), including its regulation by the APE/Ref-1 regulatory pathway and our progress in developing specific nNos inhibitors. We have shown: (i) progressive expression of nNos during melanomagenesis as measured in biopsies of CM; (ii) induction of nNOS and increase in NO by UV-A and UB-radiation of neonatal melanoctyes and CM cells; (iii) siRNA NOS knockdown inhibited melanoma cell invasion measured by Matrigel assay with marked downstream decrease of c-Jun, Bc1—2 and MMP; (iv) decrease of melanoma overgrowth by nNOS inhibitors in an arginine stimulated 3-D reconstruct CM model. We have also synthesized new highly specific nNOS inhibitors with high nNOS/iNOS and nNOS/eNOS ratios which should result in low toxicity in vivo. This approach opens up whole new strategies for the (chemo) prevention and treatment of CM.

Genetic epidemiology of melanoma susceptibility: what we know now

J. A Newton-Bishop

University of Leeds, UK, on behalf of the Section of Epidemiology and Biostatistics at the University of Leeds and GenoMEL (

Genetic epidemiology is the science of understanding genes determining health and how those genes interact with the environment and each other. Considerable progress has been in melanoma: in identifying the genes which increase susceptibility, and the correlations between those genes and well-known phenotypic risk factors. We will explore the biological implications of these genes and how they interact with the environment in the future.

It is traditional to divide susceptibility genes into highly penetrant genes associated with familial clustering and lower penetrance genes which increase risk in populations. This is a rather arbitary but useful division and using this division, the most recent developments have been in identifying the lower penetrance susceptibility genes using a series of genome wide association studies (GWAS). These large GWAS data sets have produced the first fruit: at least 14 melanoma risk loci, five associated with pigmentation variation (ASIP, MC1R, TYR, SLC45A2, and TYRP1), four associated with nevus count variation (MTAP/CDKN2A, PLA2G6, IRF4] and TERT/CLMPT1L), three not associated with pigmentation or nevus count variation (ATM, CASP8 and MX2) and two new loci on chromosome 1.

In the Leeds cohort data set we have explored the proportion of variance ‘explained’ by the identified nevus genes and shown that as yet the identified genes explain a small proportion so that additional nevus genes remain to be found.

In families with multiple cases of melanoma, CDKN2A and CDK4 mutations underlie susceptibility to melanoma in a proportion of families which increases with number of cases: these genes therefore ‘explain’ the majority of families with large number of cases but a smaller proportion where there are small numbers. Mutations are not found however in a significant number of families with large numbers of cases and there are therefore additional high penetrance genes which are as yet to be identified.

Whole-genome and exome sequencing in melanoma families

N. K. Hayward1, L. G. Aoude1, M. S. Stark1, M. Gartside1, K. Holohan2, S. Woods1, K. Dutton-Regester1, V. Bonazzi1, J. M. Palmer1, H. Schmid3, E. Holland3, Z. Z. Zhao1, V. Zismann2, M. Harland4, R. Kefford3, G. W. Montgomery1, N. G. Martin1, J. Trent2, J. Newton-Bishop4, D. T. Bishop4, A. Goldstein5, M. Tucker5, N. Gruis6, F. Demenais7, B. B. de Paillerets7, G. Mann3, A. Borg8, G. Jonsson8, K. M. Brown9

1 Queensland Institute of Medical Research, Brisbane, Queensland, Australia; 2Translational Genomics Research Institute, Phoenix, AZ, USA; 3Westmead Institute for Cancer Research, Sydney, NSW, Australia; 4Leeds Institute of Molecular Medicine, Leeds, UK; 5National Cancer Institute, Bethesda, MD, USA; 6Leiden University Medical Center, Leiden, The Netherlands; 7INSERM, Paris, France; 8University Hospital Lund, Lund, Sweden; 9National Cancer Institute, National Institutes of Health, Gaithersburg, MD, USA

To search for novel susceptibility genes responsible for predisposition in families with a combination of cutaneous and uveal melanoma we used whole-genome or exome sequencing of between 1 and 7 individuals (48 in total) in each of 23 families. Sequence reads were aligned to the human reference genome, hg19, and then variants called using SAMtools and annotated using ANNOVAR. Any SNP listed in dbSNP or the 1000 Genomes Project was not considered further in the initial phase of analysis. From the more refined list, select variants that were found in rational melanoma candidate genes were assessed for (i) segregation in the relevant family to determine whether mutations were common to all affected members and hence might be high penetrance mutations responsible for the increased melanoma risk in these families; and (ii) association analysis in a population-based case-control sample to determine if the variants may be low penetrance melanoma risk alleles. To date, none of the mutations tested have segregated in all affected members of the family in which they were discovered. A number of variants were found to be unique to the proband or a limited number of cases in certain families. The potential involvement of these variants in melanoma predisposition requires further investigation. None of the more common variants showed significant differences in allele frequency between cases and controls, and hence do not appear to contribute to melanoma susceptibility in the general population. Additional families and other unique variants are currently being analyzed.

UV versus vitamin D: how important is it?

M. Berwick

University of New Mexico, Albuquerque, NM, USA

Controversy swirls around the optimal requirement for Vitamin D and how to obtain it. The 2010 Institute of Medicine report on Vitamin D states that there is not adequate evidence to determine the optimal requirement for health conditions other than bone health. Numerous studies have been published that purport associations both for the benefits and risks of UV exposure and separately for serum vitamin. Few of these studies are compelling due to methodological issues. For example, multiple ecological analyses have associated latitude with the etiology and/or survival in melanoma. However, it is quite clear that individual behavior and phenotype at different latitudes are far more critical for the development of melanoma.

A number of studies with individual-level data do demonstrate positive associations between intake of supplements containing vitamin D and the development of melanoma, but none of these give dosage or type of supplement. Some laboratory data show a direct association between UVB dose and serum vitamin D. However, there is a very wide variation among individuals. Additionally, few studies have evaluated genetic factors well known to be associated with absorption of vitamin D. Additionally, inter-laboratory variability leads to the inability to draw conclusions ‘over’ studies. Therefore, it is critical to have robust data in free-living human populations to address this question.

Thomas et al. (2010) demonstrated that chronic sun exposure as opposed to intermittent sun exposure can be measured by solar elastosis. We have demonstrated an inverse association between solar elastosis and melanoma in large populations. Unfortunately, however, before we can answer whether UV or vitamin D are important for the development and progression of melanoma, it is critical to have more compelling data.

Using genetic test results

S. A. Leachman1, P. Cassidy1, D. Grossman1, Zalfa Abdel-Malek2, L. G. Aspinwall1

1 University of Utah; 2University of Cincinnati

Genetic testing for melanoma is transitioning from research use into clinical utility, largely because of a dynamic interaction between the fields of genetic epidemiology, social psychology, cellular and molecular biology, and clinical medicine. There is up to a 67% lifetime risk of melanoma development in individuals who carry a cyclin-dependent kinase inhibitor 2A (CDKN2A/p16) gene mutation; and there is growing international consensus that melanoma patients with a hereditary pattern of melanoma should be offered genetic counselling and appropriate testing. Members of CDKN2A/p16 melanoma families that receive genetic test results are statistically more likely to practice prevention behaviors, in some cases as long as 2 yr following receipt of the results. Improving prevention and early detection behaviors has the potential to reduce the incidence and mortality of melanoma in this high-risk population and may help to optimize prevention messages to melanoma patients in general. Since its discovery as a major melanoma predisposition gene in the early 1990s, the cellular and molecular role of CDKN2A/p16 as a key cell cycle regulatory protein has become increasingly well-defined. More recently, oxidative stress functions have been demonstrated for both CDKN2A/p16 and another, less penetrant, but more common melanoma predisposition gene, the melanocortin 1 receptor. The central role of oxidative stress pathways in melanoma predisposition creates the opportunity to develop and test targeted chemoprevention agents in these high-risk populations. Thus, the dynamic exchange of epidemiologic, psychologic, biologic, and medical information is leading to objective clinical benefit for high-risk melanoma patients through preventive behaviour modification and through development of targeted chemoprevention strategies.

Personalized melanoma chemoprevention

P. Cassidy1,2,3, S. Leachman1,3, D. Grossman1,3,4

1 Department of Dermatology, University of Utah, Utah; 2Department of Medicinal Chemistry, University of Utah, Utah; 3Huntsman Cancer Institute, Salt Lake City, Utah; 4Department of Oncological Sciences, University of Utah, Utah

Multiple oral agents have been proposed for melanoma chemoprevention. We propose a novel paradigm for melanoma chemoprevention using N-acetylcysteine (NAC), a potent antioxidant that acts as a glutathione (GSH) prodrug. We believe oxidative stress in nevi is a viable surrogate for melanoma risk, and propose that reduced melanoma risk in humans can be inferred by protection of nevi from acute UV-induced oxidative changes. We recently completed a pilot study demonstrating that 1200 mg NAC can be safely given orally (72 patients), its metabolites cysteine and GSH detected in nevi, and nevi (in nine of 19 patients treated) were protected from UV-induced oxidative stress (GSH depletion in post-drug versus pre-drug nevus). We also found the capacity of NAC to protect nevi correlated with magnitude of UV-induced GSH depletion in nevi not exposed to NAC, and was greater in patients without loss-of-function mutations in melanocortin-1 receptor (MC1R). We hypothesize that episodic administration of NAC around the time of UV exposure will reduce melanoma risk in high-risk patient populations with genetic susceptibility to UV-induced oxidative stress, and examination of key genetic variants will identify which individuals are most likely to benefit from chemoprotection. We are initiating a clinical trial designed to validate potential genetic (MC1R, p53, others) and functional markers (antioxidant response, NAC metabolism, GSH synthesis) of susceptibility to UV-induced oxidative stress and protection by NAC. Patients’ nevi will be UV-irradiated in vivo, using biologically equivalent (based on MED) doses, and NAC or placebo will be given orally both prior to and after UV treatment. If successfully implemented, such a strategy would obviate many of the pitfalls inherent in most conventional chemoprevention approaches, and may be useful for other cancers triggered by known environmental stimuli.

Who is at high risk of getting lethal melanoma?

A. Geller

Harvard School of Public Health

Incidence and mortality rates of melanoma throughout most of the developed world have risen in the past 25 yr although many parts of the world have experienced stabilization of these rates over the past 10–15 yr. Mortality rates are very disproportionate-men experience higher rates than men and rates are far greater for persons ages 60 and above compared with younger individuals. In the past years, sharper social class disparities have emerged such that persons of lower SES have been diagnosed with a greater proportion of advanced stage melanoma. Worldwide, the nodular melanoma (NM) subtype accounts for a disproportionate number of newly diagnosed thick melanoma (≥2 mm). In an analysis of more than 35 000 invasive melanomas from the National Cancer Institute's SEER data (1988–1999), nodular melanoma comprised only 9% of all lesions but accounted for nearly 50% of melanomas 2 mm or deeper when melanomas not otherwise specified (NOS) were excluded. We propose that reduction of deaths from melanoma can be best enhanced by strong collaborations between experts in dermatology, primary care, oncology, cancer education and health systems research, epidemiologists and behavioral scientists. A careful examination of advanced melanoma requires a critique of the epidemiology, social structures, and tumor biology compared with less invasive melanoma.

Immunobiology of immune checkpoints

J. D. Wolchok

Memorial Sloan-Kettering Cancer Center and Ludwig Institute for Cancer Research

Advances in basic immunology during the past 10–15 yr have revealed that the induction of a successful immune response is the result of a dynamic interplay of positive and negative signals between T cells and antigen presenting cells. Two critical immunologic checkpoints which normally function to restrain the immune response from self-reactivity and activation induced cell death are CTLA-4 and PD-1. Human antibodies blocking these negative regulatory proteins have been developed based on pre-clinical studies demonstrating more robust tumor immunity when the relevant checkpoints. In March, 2011, ipilimumab, a fully human IgG1 antibody which blocks CTLA-4 became the first drug approved for the treatment of metastatic melanoma in 13 yr based on a phase three trial results revealing significantly improved overall survival in treatment-refractory patients receiving ipilimumab, compared with a group randomized to receive a gp100 peptide vaccine. Subsequently, a second phase three trial showed that in treatment-naive patients with metastatic melanoma, randomization to treatment with ipilimumab + dacarbazine yielded a significant improvement in overall survival compared with patients receiving dacarbazine + placebo. In both studies, responses were durable, with some patients living 3 yr or longer and a mechanism-based series of immune related adverse events was observed. These instances of tissue-specific inflammation were concentrated in the skin and GI tract and were generally manageable with corticosteroids or other immune suppressant medications given according to algorithm-driven guidelines. Ipilimumab is now being investigated in combination with vemurafenib in an attempt to achieve an optimal balance of high response rate with durable disease control. MDX-1106 is a fully human antibody to PD-1. Data from a large phase 1 program has revealed the ability of treatment with this antibody to mediate durable responses, as well an immune related adverse events. An ongoing phase 1 trial is exploring the combination of ipilimumab and MDX-1106 in patients with advanced melanoma.

The melanoma tumor microenvironment as a predictive biomarker for response to immunotherapies

T. F. Gajewski

University of Chicago, Chicago, IL, USA

Immunotherapeutic approaches for the treatment of melanoma can boost immune responses but clinical responses as measured by tumor shrinkage occur in a minority of patients. This has prompted analysis of the tumor microenvironment to probe for biologic correlates to clinical response and also to identify mechanisms of resistance. Gene expression profiling and confirmatory assays have revealed two major categories of melanoma metastases. One subgroup of patients has an inflamed phenotype that includes expression of chemokines, T cell markers, and other immunoregulatory factors. Clinical responders to melanoma vaccines appear to fall within this subset. This group also contains the highest expression of negative regulatory factors, including PD-L1, IDO, and FoxP3, suggesting that these immune suppressive mechanisms may dominantly inhibit anti-tumor T cell function in those patients. In addition, absence of B7 expression suggests that classical T cell anergy is operational. Preclinical experiments have confirmed a critical role for all four of these mechanisms in limiting anti-tumor T cell efficacy in vivo, giving candidate treatment strategies for translation back into the clinic. A second subset of patients is represented by tumors which are non-inflamed and lack chemokines. Therefore, a major barrier in these cases appears to be failed T cell migration into tumor sites. Experimental strategies to augment T cell migration can have important anti-tumor effects in preclinical models. Finally, the presence of the ‘inflamed’ gene signature was associated with a type I IFN transcriptional profile, and murine models have confirmed a critical role for type I IFN signaling for innate immune ‘sensing’ of a tumor and promotion of adaptive immunity. Our results confirm that molecular profiling of melanoma metastases may be useful as a predictive biomarker for response to melanoma vaccines and other immunotherapies. Specific features identified in defined subsets of patients offer new therapeutic strategies based on overcoming immune resistance.

PD1 blockade in the treatment of metastatic melanoma

M. Sznol

Yale Cancer Center, Smilow Cancer Hospital, New Haven, CT, USA

PD1 is expressed on a subset of activated lymphocytes and serves as a co-inhibitory signal for lymphocyte function upon binding to its ligands. Tumor cell expression of one PD1 ligand, B7-H1 (PD-L1), has been found in the majority of melanomas, often co-localized to infiltrating T-cells. In vitro, tumor expression of B7-H1 can be up-regulated by exposure to interferons, and antibody blockade of PD1 increases T-cell cytokine production in mixed lymphocyte reactions or following antigen-specific recognition. Antibodies blocking PD1 or PD-L1, and a PD-L2 Fc fusion protein, have been advanced into clinical trials. One of these agents, MDX-1106 (BMS-936558/ONO-4538) was evaluated in a single and a multi-dose phase 1 trial including patients with metastatic melanoma. In the multi-dose phase 1 trial, MDX-1106 was given at 1, 3, and 10 mg/kg IV every 2 weeks. Each cycle consisted of four doses, and up to 12 cycles could be administered. The 10 mg/kg dose was well-tolerated, an MTD was not established, and adverse events were similar at the three dose levels. Up to 16 patients with melanoma were treated in expansion cohorts at each of the three dose levels. All patients had received at least one prior treatment for metastatic disease (but not immunomodulatory antibodies). For the total of 46 metastatic melanoma patients across the three dose levels for which response data are available (presented at ASCO, 2010), 15 (32.6%) objective responses were observed. Responses were observed in patients with bulky metastatic disease, and most responses have been durable, with the longest ongoing for up to 34+ months. These results indicate that PD1 blockade is a promising approach for treatment of metastatic melanoma, and additional single agent and combination development is warranted.

Current and future status of adoptive cell therapy

J. Weber

Department of Cutaneous Oncology, Moffitt Cancer Center, Tampa, FL, USA

Since the initial promising trials of adoptive cell therapy for melanoma in the 1980s at NCI's Surgery Branch, the use of tumor infiltrating lymphocytes after lymphoid depletion has gained momentum and is now being evaluated at several centers around the world. Trials of unselected, so-called ‘young’ TIL at Sheba Medical Center in Israel have shown 40% ORR with excellent OS; trials of selected TIL after non-myeloablative (NMA) chemotherapy at Moffitt and MD Anderson have demonstrated excellent responses and suggested that CD8 percentage and total numbers of cells infused were predictive of response. At the NCI, extensive analyses of an experience of over 200 patients have shown that OS and response were superior if 1200 R of TBI was added to NMA; that CD8 cells that had longer telomeres and were CD28+/CD27+ were associated with superior clinical efficacy, as was faster growth of the TIL cultures. Several groups, particularly at NCI, have attempted to use TcR gene-engineered PBMC to redirect the T cell receptor with some success. Current trials will evaluate the use of artificial aAPC for improved antigen specific stimulation; the use of anti-CD137 and anti-PD-1 antibodies in vitro for improved cell growth and in vivo to attempt to substitute for high dose IL-2; leaving out the high-dose IL-2 entirely to reduce toxicity and generation of T regulatory cells, and combination of other active agents like ipilimumab and vemurafenib with TIL. A major obstacle is that high numbers of TIL require the labor intensive use of large bags to propagate them. This might be overcome by the use of compact bioreactors like G-REX. The other obstacle is the prolonged time to grow the cells, often as long as 5–6 weeks. This is being overcome by early selection of rapidly growing TIL from fragments, not tumor digests. These advances have awakened commercial interest in this technology, and several commercial enterprises have proposed registration strategies for TIL that might allow this important treatment to be widely available in the near future.

Translating human melanoma genetics

M. Bosenberg

Departments of Dermatology and Pathology, Yale School of Medicine

The number of genetic changes in any given melanoma may range between a few to several thousand. The large number of changes makes the distinction between driver and passenger mutations difficult. Furthermore, the functional outcome of certain genetic changes depends on the presence of absence of additional specific changes. The large number of possible combinations of driver mutations is only beginning to be identified and understood. Ongoing efforts to identify novel genetic and epigenetic changes in human melanoma will be presented. Functional evaluation of putative human melanoma driver mutations in mouse melanoma models will be discussed as well.

Genetic predictors of RAF-dependence

D. B. Solit

Memorial Sloan-Kettering Cancer Center

There is a desperate need for more effective treatments for patients with advanced cancer. Approximately half of melanomas have an activating mutation of the BRAF gene. PLX4032 (also called vemurafenib) is a potent and selective inhibitor of the most common mutant form of BRAF. In a Phase I trial of PLX4032, approximately 80% of patients with melanoma whose tumors express the BRAF mutation experienced significant tumor shrinkage with minimal side effects. In contrast, none of the patients with melanomas without a BRAF mutation responded to the drug. These promising results demonstrate what can be achieved with personalized, molecularly targeted therapy for melanoma and other solid tumors. However, the degree of tumor shrinkage varied greatly among patients and many of the patients who initially responded to PLX4032 subsequently developed resistance. The experience with targeted therapies in other diseases suggests that understanding the causes of resistance can lead to improved patient selection for treatments like PLX4032 and the development of more effective drug combinations. My laboratory is focused on identifying the mechanisms of resistance to PLX4032 and other targeted kinase inhibitors. In this lecture, I will review the genetic basis for RAF inhibitor resistance and highlight ways in which real time genetic evaluation of human tumors is altering the development of novel targeted cancer therapies.

Overview of resistance pathways in melanoma

M. A. Davies

Department of Melanoma Medical Oncology, Department of Systems Biology, University of Texas MD Anderson Cancer Center, Houston, TX, USA

Research over the last 10 yr has greatly advanced the understanding of the molecular heterogeneity of melanoma. This information has recently led to personalized targeted therapy approaches which have demonstrated marked benefit in melanoma patients. However, the efficacy of these approaches has been limited by both primary and secondary resistance. Clinical experiences in other cancers have implicated alterations in the direct targets of therapeutic agents; compensatory signalling in targeted pathways; and activation of parallel signalling pathways as possible mechanisms of resistance. In this presentation we will review the development of targeted therapies for melanoma, and present an overview of pathways and mechanisms that have been implicated in melanoma to date. The presentation will also review critical information that has been learned from the development of these agents, and highlight both the opportunities and challenges that are high priorities to address in this field.

Melanoma heterogeneity and therapy

M. Herlyn, A. Roesch, R. Somasundaram, J. Villanueva, M. Fukunaga, C. Krepler, J. T. Lee, G. Zhang

Program of Molecular and Cellular Oncogenesis, Melanoma Research Center, The Wistar Institute, Philadelphia, PA, USA

Similar to observations in normal tissues, particularly in keratinocyte stem cells, each melanoma contains a slow-cycling, i.e. label retaining sub-population, which exhibits a proliferation cycle of >4 weeks asynchronous to the majority of the malignant cells, which have doublings of ~2 days. Our preliminary data indicate that the label-retaining, slow-cycling property of melanoma cells is linked to the expression of JARID1B. This slow-cycling population has self-renewal capacity and gives rise to a rapidly proliferating heterogeneous progeny. Slow-cycling cells defined by JARID1B are essential for tumor maintenance. However, these cells do not follow a unidirectional cancer stem cell model because both JARID1B+ and JARID1B-cells are tumorigenic. We have induced tumors with single cells from either group. The same results, tumorigenicity after injecting one melanoma cell not selected for any marker expression into NOD/LtSscidIL2Rgcnull (NGS) mice, were found in 8/9 additional cases confirming the data from the Morrison group. Moreover, JARID1B-cells can become over time JARID1B+ underscoring the phenotypic plasticity in this malignancy. We are determining how cells from the small JARID1B+ sub-population are resistant to state-of-the-art therapies and what the potential mechanisms are for resistance. Preliminary data suggest that intrinsic resistance to therapies due to the emergence of one or more sub-populations. Besides JARID1B we have similar results with CD20+ and CD133+ tumor cell subpopulations. Next to traditional therapy regimens with cytotoxic drugs, the melanoma field has focused on drugs that target proteins of the MAPK proliferation pathway. However, the BRAF or MEK inhibitors designed to block this pathway show limited effects on cells that do not proliferate. Thus the JARID1B+ cells are expected to remain intact after such an approach and repopulate the tumor. Indeed, our preliminary studies strongly support this concern because we see an accumulation of JARID1B+ tumor cells after therapy with classic chemotherapeutic drugs like temozolomide but also after treatment with signaling inhibitors targeting BRAFV600E (PLX4720), MEK (AZD6244), PI3K (LY294002), and proteasomes (bortezomib). Similarly, CD20+ tumor cells show also selected resistance to a variety of therapies, including T cell-mediated cytotoxicity. Thus, the clinical failures of melanoma therapy including even the newest drugs confirm that the mainstream concepts for cancer therapy need to be adjusted by taking into account tumor heterogeneity for the development of new strategies for therapy. In the future we expect the development of two therapies, one for the minor population with stem cell-like properties, the other for the major population representing the bulk of a tumor.

Acquired resistance mechanisms in melanoma patients treated with BRAF inhibitors

R. Lo

Dermatology/Medicine, Molecular & Medical Pharmacology, Jonsson Comprehensive Cancer Center and David Geffen School of Medicine at the University of California, Los Angeles, CA, USA

Activating BRAF V600 kinase mutations occur in ~60% of melanomas, and the ATP-competitive RAF inhibitor, PLX4032/vemurafenib, with its remarkable anti-tumor selectivity toward V600EBRAF melanomas, has now been approved for clinical use. Acquisition of drug resistance leading to clinical relapse, however, develops in the majority of patients treated with BRAF inhibitors (BRAFi). Cell lines adapted to growth in vemurafenib as well as objective genetic screens in vitro have provided valuable insights into potential BRAFi escape pathways. Evaluation and validation of the likely mechanisms in patient-derived tissues become critically important diagnostically to guide next generation clinical trials designed at preventing or overcoming BRAFi acquired resistance. Heterogeneous mechanisms of acquired BRAFi resistance hitherto uncovered fall into general MAPK-redundant, AKT-dependent or MAPK-reactivating pathways, suggesting specific translatable therapeutic strategies. Mechanisms leading to MAPK-redundant survival include PDGFRβ and IGF-1R RTK overexpression. Mechanisms leading to MAPK reactivation include acquisition of NRAS activating mutations and V600EBRAF alternative splicing, leading to an activated, N-terminally truncated V600EBRAF. Secondary mutations continue to remain undetected in clinical specimens. The number of disease progression specimens unaccounted for by known mechanisms may be due to an under-appreciation of the true RTK diversity, limitations imposed by current detection tools or sample heterogeneity, and/or the existence of novel mechanisms hitherto uncovered.

Abstracts of the 5th Meeting of Interdisciplinary Melanoma/Skin Cancer Centres November 11–12, 2011 Invited Faculty Abstracts

  1. Top of page
  2. 2011 International Melanoma Congress Advancement Through Collaboration joins the following four annual meetings in 2011:
  3. Congress Agenda
  4. Congress Proceedings Abstracts of the 8th International Congress of the Society for Melanoma Research November 9–11, 2011 Invited Faculty Abstracts
  5. Abstracts of the 5th Meeting of Interdisciplinary Melanoma/Skin Cancer Centres November 11–12, 2011 Invited Faculty Abstracts
  6. Abstracts of the 4th Melanoma Pathology Symposium of the International Melanoma Pathology Working Group November 13, 2011 Invited Faculty Abstracts
  7. Abstracts of the 8th International Congress of The Society for Melanoma Research November 9–11, 2011 Proffered Abstracts Selected for Oral Presentation
  8. Abstracts of the 8th International Congress of The Society for Melanoma Research November 9–11, 2011 Proffered Abstracts Selected for Poster Presentation
  9. Abstracts of the 5th International Meeting of Interdisciplinary Melanoma/Skin Cancer Centres November 11–12, 2011 Proffered Abstracts Selected for Poster Presentation

Histologic and molecular classification of melanoma

R. A. Scolyer1,2,3

1 Melanoma Institute Australia, Sydney, NSW, Australia; 2Tissue Pathology and Diagnostic Oncology, Royal Prince Alfred Hospital, Camperdown, NSW, Australia; 3Discipline of Pathology, Sydney Medical School, The University of Sydney, NSW, Australia

In the early 1800's melanoma was recognised as a disease entity. Initially it was subclassified on the basis of its association with, and the type of, any precursor lesion. Dr Wallace Clark proposed a histogenetic classification for melanoma in 1967 which subdivided into superficial spreading melanoma (SSM), lentigo maligna melanoma (LMM) and nodular melanoma (NM) mostly on the basis of histopathology of the insitu component of the tumor adjacent to any invasive component. These entities remain in the latest version of the WHO melanoma classification, although additional entities have been added. Because clinical features and non-tumorous histopathological features are included as criteria, defining feature overlap, poor predictability of patient outcome and lack of relevance to patient care, this classification scheme has been criticised. However, the subtype of melanoma highlights the various patterns of melanoma, both clinical and pathological, that clinicians and pathologists must recognize to avoid misdiagnosis. Recently advances in molecular biology have identified oncogenic mutations in NRAS (15–35%), BRAF (50%), CKIT (2%), and GNAQ/GNA11 (usually in uveal melanomas) in melanomas. Targeted therapies have been developed for patients with mutations in BRAF and CKIT with impressive clinically efficacy. In the future, it is likely that a melanoma classification that integrates clinical, pathologic and molecular features will be developed that defines biologically distinct melanoma subsets that share a oncogenic mechanism, share clinical behavioural features and require management that is similar. Already as a consequence of these findings, more rational and effective therapies for metastatic melanoma patients are now available and are already changing the care of metastatic melanoma patients.

Melanoma: targeted therapy for advanced disease

G. A. McArthur1, K. Waldeck1, J. Kelly2, W. Liu1,2, R. Young1

1 Peter MacCallum Cancer Centre; 2Victorian Melanoma Service, Alfred Hospital

The growing effort to identify oncogenic events in melanoma by gene sequencing is uncovering a number of important targets in melanoma. Over 60% of melanoma contain mutations in BRAF or NRAS that activate the MEK/ERK pathway to drive the proliferation and survival of melanoma cells. A number of receptor tyrosine kinases such as KIT are also mutated in melanoma. The accumulated data show that the majority of melanomas contain mutations in protein kinases that are readily amenable to inhibition by small molecule drugs. BRAF inhibitors such as vemurafenib and GSK2118436 show high response rates in patients with advanced disease that contain BRAF mutations. In the case of vemurafenib this has now been shown to improve overall survival. One of the well characterized oncogenic pathways in preclinical models of melanoma is the CCND1 (Cyclin-D1), CDKN2A (p16), CDK4 complex. We have examined amplification of CCND1 and CDK4 and loss of CDKN2A in 121 primary melanomas by fluorescence in-situ hybridisation. One or more of these genomic events were identified in over 70% of primary melanomas with CCND1 amplification identified in 18%, CDK4 amplification in 37% and hemizygous and homozygous loss of CDKN2A being identified in 50 and 5% respectively. Together these data indicate that the CDK4 kinase maybe activated in the majority of melanomas. Given a significant proportion of these genetic changes occur in melanomas with BRAF-mutations the data also raise the possibility of combination studies with BRAF inhibitors. The future of targeted therapies is likely to be in such combination approaches where cooperating oncogenic events are simultaneously targeted to increase cell death and improve survival. Early results with the combination of BRAF and MEK inhibitors are encouraging with tolerable toxicity and significant response rates. The combination of genomic efforts and development of new therapeutic approaches offers great promise in the targeted therapy of advanced melanoma.

Update on the adjuvant (PEG) interferon trials of the EORTC

A. M. M. Eggermont1, S. Suciu2, A. Testori3, W. Kruit4, C. J. A. Punt5, M. Santinami6, J. Marsden7, R. Dummer8, D. Schadendorf9, P. Patel10, C. Robert11, U. Keilholz11, A. Spatz12

1 Cancer Institute Gustave Roussy; 2EORTC HQ; 3European Institute of Oncology; 4Erasmus Univ MC; 5Nijmegen Univ MC; 6Istituto Tumori Milano; 7Birmingham Univ; 8Zuerich Univ; 9Essen Univ; 10Nottingham Univ; 11Charite Berlin; 12McGill University

EORTC 18991 is the largest adjuvant trial ever conducted in stage III melanoma. It assessed the efficacy and toxicity of long term PEG-IFN versus Observation (Obs.). PEG-IFN was recently FDA approved as an adjuvant therapy under the name Sylatron for Stage III melanoma patients post lymph node dissection. PEG-IFN (Induction at 6 μg/kg/week, sc, 8 weeks; followed by Maintenance at 3 μg/kg/week, sc) for a total treatment duration of up to 5 yr was compared to Obs. in 1256 patients (pts) with stage III melanoma (anyTN1-2M0 without in-transit metastases). Randomization was stratified for nodal involvement N1 (microscopic) versus N2 (palpable nodes), # of nodes, Breslow and ulceration of primary, sex and center. Relapse-Free Survival (RFS) was the pre-specified regulatory primary endpoint.

Median follow-up was 7.6 yr: ITT analysis: RFS [HR: 0.87 (0.76–1.00); P = 0.05]; DMFS [HR: 0.93 (0.81–1.07); P = 0.33]; OS [HR 0.96 (0.82–1.11); P = 0.57]; 7-yr rates: RFS: Obs 35%; PEG-IFN 39%; DMFS 40 versus 42%; OS 46 versus 48%. Median RFS was almost doubled in the PEG-IFN treated patients. Further analyses revealed that independent predictive factors for outcome were Stage III-N1 and Ulceration of the Primary. Results in this patient population N1 + Ulcerated Primary were: RFS [HR (99%CI) 0.72 (0.46–1.13); P = 0.06]; DMFS (99%CI) [0.96 (0.76–1.22) P = 0.02]; and for OS (99%CI) [0.59 (0.35–0.97) P = 0.006]; Median OS was 3.7 yr for Obs versus >8 yr for PEG-IFN.

Long term PEG-IFN therapy in stage III melanoma had a significant and sustained impact on RFS, but not on DMFS and OS. In the SN+/Ulc pts, the benefit was greatest and consistent regarding all endpoints, and was maintained at long term follow up. These effects have been observed in two consecutive EORTC trials (18952 and 18991) involving 2644 pts.

Are pegylated and standard interferon still relevant?

V. K. Sondak

Department of Cutaneous Oncology, Moffitt Cancer Center, Tampa, FL, USA

It is an extraordinary time for melanoma patients, with CTLA4 blockade and BRAF-targeted therapy improving survival in Stage IV disease. The adjuvant setting, however, remains problematic, with interferon-alfa still the only available therapy in many countries. The recent approval of pegylated interferon-alfa (peginterferon-alfa) in the USA and some other countries has only added to the confusion and made the decision-making process regarding patient selection more complex. Even after thousands of patients have been treated on adjuvant interferon trials evaluating dosing, schedule, toxicity and quality of life, controversies persist regarding the overall survival (OS) benefit associated with therapy and the importance of relapse-free survival (RFS) as an endpoint justifying substantial toxicity. In light of advances in treating metastatic disease, are interferon-alfa and peginterferon-alfa still relevant?

The available evidence supporting the RFS benefit of adjuvant interferon and peginterferon is convincing, and current and future adjuvant trials employ RFS as a primary endpoint because of concerns about the confounding effects on survival of postrelapse therapy. From the patient's perspective, delaying relapse is potentially more important than ever: we have more effective drugs for treating metastatic melanoma than ever, but their use has not been optimized. Two to three years from now, treatments for stage IV melanoma will likely be substantively more effective and less toxic than they are now. A few extra months could afford patients access to options that would otherwise not be available in time.

In contrast, evidence supporting a specific interferon dose, duration or formulation – or identifying subsets of patients most likely to benefit from therapy – is currently unconvincing. While the preferred management of patients with node-positive melanoma remains enrollment on clinical trials, the impact of interferon therapy on RFS cannot and should not be ignored. Until and unless something clearly better comes along, adjuvant interferon-alfa and peginterferon-alfa remain relevant.

Can we predict who will respond to interferon?

H. Gogas

1st Department of Medicine, University of Athens, Athens, Greece

Interferon alfa (IFNα) has been widely tested as adjuvant therapy of melanoma in patients with intermediate and higher risk for recurrence melanoma and currently the only approved agent for the adjuvant treatment in Europe and in the United States. However, the tolerability of this regimen has been an issue due to the frequent occurrence of flu-like symptoms, fatigue, anorexia, and occasional depression. Attempts to identify the subset of patients who benefit from adjuvant treatment with IFNα-2b have been undertaken but the results of these efforts have largely been disappointing to date. Subanalyses of data from a number of recent studies suggest that response varies according to disease burden. Exploratory analyses of the data from two EORTC trials suggested that ulcerated primary melanomas are more sensitive to PegIFNα2b versus non-ulcerated melanomas. Methylthioadenosine phosphorylase (MTAP) expression, STAT1 expression, S100B levels, and HLA class I and II type, have been correlated with prognosis but need further study to confirm their predictive value. A number of other potential predictive markers for IFN efficacy have been assessed in clinical trials. Those that have shown potential as predictors of treatment outcomes include methylthioadenosine phosphorylase (MTAP) expression, STAT1 expression, S100B levels, and HLA class I and II type, but need further study to confirm their predictive value. Others that did not show predictive value include pSTAT1/pSTAT3 ratio, ferritin and CRP, and some have shown mixed results, such as white blood cell count. Analysis of peripheral blood lymphocytes from 14 patients receiving neoadjuvant HDI induction indicated that differences in STAT1 activation pre- and post-HDI distinguished between patients who were likely to have a favorable versus unfavorable outcome, and may therefore be useful as a predictive indicator of response to therapy, although further verification is required.

Adjuvant therapy for malignant melanoma – MAGE.A3 vaccines and bevacizumab

P. Lorigan

Christie NHS Foundation Trust/University of Manchester, Manchester, UK

DERMA (ADjuvant ImmunothERapy with MAGE3 in MelanomA) is a double-blind, randomized, placebo controlled phase III study of recombinant MAGE-A3+ AS15 antigen specific cancer immunotherapeutic (ASCI) in patients with MAGE.A3 positive resected stage III melanoma. MAGE-A3 is expressed on approximately 60% of melanoma specimens. This study has completed accrual of 1300 patients. The primary endpoint is disease-free survival. Additionally, the study examines the role of a molecular predictor in identifying patients likely to benefit from treatment with the MAGE.A3 vaccine +AS15. Patients received an induction phase with five vaccinations given 3-weekly followed by eight vaccinations given every 3 months. DERMA is based on the EORTC phase II study in patients with stage IV M1A disease. This identified a superior survival benefit for patients receiving the MAGE–A3 vaccine treated with the AS15 rather than the ASO2B adjuvant [HR 0.55, (95% CI, 0.28; 1.06)]. Furthermore, a genetic predictor identified a group of patients receiving MAGE-A3+ AS15 ASCI with a better overall survival [HR 0.27 (95% CI, 0.08; 0.89)].

AVAST-M (Adjuvant AVAStin Trial in High-Risk Melanoma) is a randomized multi-centred phase III study evaluating bevacizumab as an adjuvant therapy for patients with high-risk [AJCC stage IIB (T3bN0M0 & T4aN0M0), IIC (T4bN0M0) and III (TxN1-3M0)] cutaneous melanoma. Patients are randomised to received bevacizumab 7.5 mg/kg IVQ 3 weeks for 1 yr, or routine follow-up. The primary endpoint is overall survival. Thirteen hundred and twenty patients are required to show an 8% survival benefit at 5 yr. The trial will complete accrual in early 2012. The concept of the study was based on the importance of angiogenic shift in development of the metastatic process, the correlation of increased expression of vascular markers with stage and prognosis in melanoma and the positive two outcomes seen in the combination of bevacizumab with chemotherapy in a number of solid tumours.

Is checkpoint blockade the future of adjuvant therapy?

J. S. Weber

Moffitt Cancer Center, Tampa, FL, USA

The recent enthusiasm for the use of immune checkpoint blockade as an anti-tumor strategy and the approval this year of ipilimumab for metastatic melanoma based on prolonged survival in two randomized trials has promoted its evaluation in the adjuvant setting. Newer agents that are checkpoint blockade inhibitors also look promising and are being tested in small pilot studies as adjuvant treatment. Ipilimumab was tested in a small phase I study in resected stages IIIC/IV melanoma looking at increasing doses given monthly, and over a 12 month duration, 1 mg/kg was tolerated well. Follow-up studies of the drug at doses from 3 to 10 mg/kg given every 8 weeks in resected stages IIIC and mostly stage IV melanoma suggested that excellent RFS (beyond 30 months) and OS (not reached at 30 months) were observed, and that dose limiting immune related adverse events (irAEs) were observed in up to 30% of patients at a dose of either 3 or 10 mg/kg. Some patients have continued to tolerate ipilimumab for up to 5 yr, after a one year course of the drug every 8 weeks, followed by ‘maintenance’ ipilimumab every 12 weeks. Patients who were treated beyond a year did not suffer any subsequent dose limiting irAEs. In that trial, there was no evidence in the subset of patients that received a multi-peptide vaccine that T cell reactivity was augmented; significant increases in CD8+DR+ and CD4DR+ T cells were observed in all patients, and pharmacodynamic markers were defined including ICOS, Ki-67 and granzyme. Dose limiting irAEs were associated with RFS. Novel predictive markers for outcome on T cells were also found. These data supported the carrying out of a large randomized phase III study through EORTC of ipilimumab at 10 mg/kg versus placebo for stage III A-B-C disease. That trial has accrued over 950 patients, and is in its follow-up phase, with results expected in 2–3 yr. The current ECOG adjuvant trial with ipilimumab randomly assigns patients with resected stage III B and C and IV disease to ipilimumab at 10 mg/kg or the standard one year high-dose interferon alpha-2b regimen. That trial just began accrual, which will run for at least 2 yr. Although the rate of dose limiting irAEs with ipilimumab at 10 mg/kg may be as high as 25%, the acute toxicity of interferon is also significant in most patients. The survival prolongation seen with ipilimumab in phase III studies in stage IV melanoma may give it the edge as an adjuvant therapy if that translates to prolonged survival in the current adjuvant trials.

Biopsy techniques: to shave or not?

M. Lien

Department of Dermatology, University of South Florida College of Medicine and Cutaneous Oncology Program, Moffitt Cancer Center, Tampa, FL, USA

Shave biopsy is a procedure commonly utilized by dermatologists and primary care providers. The main advantages of the shave biopsy are that it is simple and efficient to perform in the office. However, this technique has been criticized for possibly preventing the accurate diagnosis of melanoma and thus, hindering the appropriate staging and treatment decision-making. Currently, the excisional biopsy is the recommended diagnostic procedure to sample a skin lesion suspicious for melanoma. This rationale stems from concerns that other biopsy methods, including superficial shave biopsies, deep scallop shave biopsies, and punch biopsies can potentially cause inaccurate staging of the melanoma due to inadequate sampling of the tissue. Recently, a retrospective analysis of 600 patients concluded that deep shave biopsies are accurate in determining diagnosis and microstaging melanoma when compared with the final diagnosis at wide excision. Importantly, in the great majority of cases where shave biopsy was performed for melanoma, melanoma was not in the stated differential diagnosis. Shave biopsy is an acceptable method for the diagnosis of melanoma, and clinicians should not be afraid to perform shave biopsies as the risk of not biopsying a lesion that could be melanoma outweighs the consequences of inadvertently diagnosing melanoma by shave biopsy.

Mole mapping and digital photography

J. Malvehy, S. Puig

Department of Dermatology, Melanoma Unit, IDIBAPS, Hospital Clinic, Barcelona, Spain

Malignant melanoma (MM) may be clinically and dermatoscopically indistinguishable from melanocytic nevi making early recognition a diagnostic challenge, especially in incipient lesions. The use of baseline regional photographs, namely total body photography (TBP), might facilitate the detection of new lesions, and visual changes in pre-existing lesions, by providing a comparative reference point of areas of skin for subsequent examinations. Dermatoscopic documentation of melanocytic lesions for the comparison of current and previous images in search of subtle changes over time, known as digital follow-up (DFU), has been shown to be helpful in the diagnosis of early melanomas for which specific criteria for MM may not yet be present.

The combined use of TBP and digital dermatoscopy, called the ‘‘two-step method’’ of DFU, has been proposed by our group as an approach for the assessment of individuals at high risk, being potentially more accurate than the two strategies separately. It is well known that the DFU procedure is not only time-consuming but also a technique that requires training, experience, and specific equipment. Chances of success in DFU depend basically on the proper selection of patients.

In a recent study of 10 yr experience of our Group with total body photography and dermatoscopy in a selected population at high risk, it was demonstrated that early detection of melanomas was possible with a low rate of excisions. Long-term follow-up was required to allow the detection of slow-growing melanomas. Based on our 10-yr experience, melanomas can be diagnosed at any time, suggesting that in a population at high risk for melanoma, DFU should be maintained over time.

Dermoscopy and other diagnostic approaches

A. C. Halpern

Memorial Sloan-Kettering Cancer Center

Over the past decade, dermoscopy has been convincingly demonstrated to improve diagnostic accuracy for early melanoma. We will review some of the pooled data that supports this conclusion. We will review some of the melanoma specific structures/features that allow the dermoscopists to recognize early melanoma including atypical network, streaks, negative network, chrysalis, atypical dots and globules, atypical blotches, blue/white veil, regression structures, atypical vessels, and peripheral brown structureless areas. While terminology for dermoscopic features continues to evolve, there is growing appreciation of the desirability of developing a consensus approach to existing terms as well as the desirability of unified algorithms for teaching and applying dermoscopy. In recent years, dermoscopic equipment has evolved. It is now broadly appreciated that contact and polarized dermoscopy highlight different and complementary features. Contact dermoscopy is best for seeing very superficial structures (e.g., comedo-like openings) whereas polarized dermoscopy is better for discerning deeper and vascular structures. Not surprisingly, combination instruments which allow the user the flexibility of both polarized and contact dermoscopy have become increasingly popular. Several recent dermatoscopes also take advantage of smartphone capabilities to offer inexpensive digital dermoscopy solutions.

Multiple studies have indicated that in vivo laser scanning confocal microscopy (CLSM) can improve diagnostic accuracy above and beyond dermoscopy. Commercially available high quality CLSM microscopes have been developed and now provide a very intuitive graphical user interface. The major challenges to clinical implementation are the amount of time and effort required to attain the CLSM images relative to the ease of skin biopsy as well as the expertise required to interpret the CLSM images.

Given recent advances in information technology, there is little question that computer assisted melanoma diagnosis is on the horizon. However, there are still significant hurdles for the implementation of computerized diagnosis in the near future. In this session, we will review optical, molecular, and other approaches, and discuss some significant cultural and regulatory barriers to the diffusion of these technologies.

Topical imiquimod 5% cream for lentigo maligna followed by conservative staged excisions as a means of reducing surgical morbidity

M. A. Hyde, G. M. Bowen

Huntsman Cancer Institute, University of Utah School of Medicine

The widely accepted standard of care for treating lentigo maligna (LM) is staged excisions with confirmation of histologically negative margins before surgical repairs are performed. This approach has greatly reduced the rate of local recurrences but is associated with significant morbidity due to the location of most tumors in cosmetically sensitive areas on the head and neck as well as the requirement for relatively large resections to achieve negative margins (average 7.1 mm margins required). Imiquimod is an immune response modifier that binds to Toll-like receptors on effecter cells in the skin leading to an innate immune response. In an attempt to reduce surgical morbidity we have treated 123 LM sites off-label with topical imiquimod 5% cream for a period of 3 months followed by conservative staged excisions beginning with two millimeter margins. Eighty-four patients were male (68%) while 39 were female (32%) [male:female = 2.2:1] and the mean age was 67.7 yr. Ninety-five patients had negative excisional margins beginning with two millimeters (77.2%), 24 (19.5%) required two stages, while 4 (3.3%) required three or more stages. Two cases were found to have invasive melanoma at the time of surgery. There have been two local recurrences (2.7%): one in situ and one invasive (0.34 mm) recurrence and no disease-related deaths. The mean follow-up time is 60 months with 51 patients lost to follow-up. Seventeen patients have died due to causes unrelated to melanoma with a mean time of death at 27.9 months following surgery. Because approximately 33% of patients retain histologic evidence of LM after topical imiquimod therapy, using imiquimod as a replacement for surgery may put the patient at increased risk for local recurrence. However, pretreatment with imiquimod allows for much more conservative surgical margins with 5-yr recurrence rates comparable to more aggressive staged excisions without imiquimod.

Sentinel node biopsy

J. F. Thompson1,2

1 Melanoma Institute Australia, Sydney, NSW, Australia; 2Discipline of Surgery, The University of Sydney, Sydney, NSW, Australia

It is now well established that sentinel node biopsy (SNB) provides the most accurate staging currently available for patients who present with primary cutaneous melanoma. There is also compelling evidence from interim analyses of the first Multicenter Selective Lymphadenectomy Trial (MSLT-I) and several single institution studies that for those with intermediate thickness (1.2–3.5 mm) melanomas and metastatic disease in regional nodes, survival is improved if early completion lymph node dissection is performed when a positive SN is identified, rather than therapeutic lymph node dissection when disease in regional nodes becomes clinically palpable. Furthermore, the morbidity of regional node dissection is substantially reduced if it is performed early rather than later. Technical issues relating to SNB have been addressed, and experience with SNB is now substantial. However, false negative results have not been eliminated and methods of reducing false negative rates are being explored. It has been suggested that very low volume metastatic disease in sentinel nodes may have no prognostic significance, but longer follow-up is required; the results of another large multicenter study (MSLT-II) and the European MINITUB study should ultimately resolve the question. SNB for staging is likely to become even more important in the near future, by identifying melanoma patients at greatest risk of recurrence whose outcome can be improved by effective adjuvant therapies.

New lymphatic mapping agents

V. K. Sondak

Department of Cutaneous Oncology, Moffitt Cancer Center, Tampa, FL, USA

Vital blue dyes and radiolabeled colloids are widely used worldwide for intraoperative lymphatic mapping and sentinel lymph node biopsy. Until this year, no radiolabeled agents were specifically approved for lymphoscintigraphy in the USA, and none had been tested in prospective clinical trials. Recently, 99mTc-sulfur colloid was approved by the FDA for ‘the localization of lymph nodes drainage a primary tumor in patients with breast cancer’ based on retrospective data. 99mTc-Tilmanocept was designed for use in lymphoscintigraphy and sentinel node biopsy, and has been evaluated in phase II and phase III clinical trials. In prospective but non-randomized comparisons to blue dye, 99mTc-tilmanocept was reproducibly able to localize in regional lymph nodes after injection at the site of primary breast carcinomas and melanomas undergoing sentinel node biopsy, detecting nearly all visibly blue nodes and identifying more tumor-containing nodes overall without missing any blue tumor-containing nodes. Available evidence does not permit a definitive comparison to the other colloids that are commonly used for lymphoscintigraphy, but does suggest that 99mTc-tilmanocept is an effective option with some potential but as yet unproven advantages. Various nanoparticulate and fluorescent systems are being developed for lymphatic imaging, including quantum dots (tiny nanometer-sized light-emitting particles. Ultimately, these may eliminate the reliance on radionuclides for lymphatic mapping. Agents that traffic preferentially to tumor-containing nodes are under study and would represent a major advance over current options. For now, the single biggest innovation in lymphatic mapping is not a new agent but rather a new technique: SPECT-CT lymphoscintigraphy. By providing three-dimensional anatomic detail, this approach greatly facilitates sentinel node localization, particularly in the head and neck and for in transit nodes. We believe SPECT-CT lymphoscintigraphy is likely to decrease false-negative rates and shorten procedure times, and it is possible that new agents may be most advantageous when combined with SPECT-CT imaging.

Limiting morbidity from inguinal lymphadenectomy: videoscopic lymphadenectomy and other approaches

K. A. Delman1,2, M. Rizzo1,2, D. Kooby1,2, K. Ogan3, V. Master2,3

1 Division of Surgical Oncology, Department of Surgery, Emory University School of Medicine; 2Winship Cancer Institute, Emory University School of Medicine; 3Department of Urology, Emory University School of Medicine

Inguinal lymphadenectomy for metastatic cutaneous, genitourinary and occasional other rare tumors is the standard of care and offers excellent control of regional disease. Unfortunately, the procedure has been noted to have a significant risk of complications. A videoscopic approach to inguinal lymphadenectomy for metastatic disease to the groin has recently been shown to be a feasible option for potentially reducing the complications of groin surgery. Outcomes from 50 videoscopic lymphadenectomies have demonstrated similar nodal yield (median 11, range 4–24) to open lymphadenectomy in conjunction with a significantly lower rate of wound complications. Direct visual inspection of the groin through the sentinel node site was initially used to confirm a complete ‘anatomic’ dissection. During the 3 yr that this procedure has been performed no patient with disease detected by sentinel lymph node biopsy has recurred. Nodes as large as 5.6 cm have been removed. Mean operative time is 165 min, length of stay: 1 day, Oncologic outcomes such as recurrence and survival are still being measured in this patient population. While videoscopic inguinal lymphadenectomy appears to be an excellent option to reduce the morbidity of inguinal lymphadenectomy with equivalent short-term pathologic endpoints, other methods for minimizing morbidity from this operation exist and will also be discussed.

Technical advances in melanoma surgery

J. M. Lewis

Department of Surgical Oncology, University of Tennessee Graduate School of Medicine, Knoxville, TN, USA

In an age where adjuvant therapies are stealing the spotlight in the treatment of melanoma, it may serve us to look back to past surgical advances and forward to new surgical methods and modalities. In 1907 William Sampson Handley gave two Hunterian lectures on melanoma forming the basis of surgical treatment for over 50 yr. Handley's recommendations centered largely on surgical margins, a matter that continues to be pushed with the advent of Moh's microsurgical techniques. Moh's surgery is a relatively new technique that with current advances pushes the margin concept often causing anxiety to cutaneous surgeons. The advent of sentinel lymph node biopsy has quelled the fires that centered on elective lymph node dissection. New technologies may improve the sentinel lymph node biopsy technique, and for those with a positive sentinel node, the morbidity of a completion nodal basin dissection may be reduced with the advent of ultrasonic sheers. This dissection technique may lessen the time needed for nodal basin drainage and ultimately lessen the risk of infection and discomfort to patients.

In the setting of resectable metastatic disease PET probe guided technologies may improve success or serve to focus dissection efforts. Furthermore, the use of new radiolabled molecules such as thymidine may increase sensitivity of staging studies and serve to identify resecatable metastatic disease at an earlier stage allowing for initiation of surgical approach or earlier adjuvant therapy. Lastly in situations of skin-graft closure we will visit nodal basin graft procurement techniques, delayed closure approaches, and old stand-bys such as absorbable purse-string suturing to reduce the morbidity of grafting that was so prevalent at the recommendations of Sir Handley in 1907.

Intralesional therapy: local/regional control and implications for systemic response

M. I. Ross

MD Anderson Cancer Center, Houston, TX, USA

While systemic therapies have been the mainstay for recurrent melanoma, responses are infrequent and rarely durable. Such is particularly true for the patients who develop unresectable local/regional in-transit recurrences and have therefore been the focus for studying intralesional ablative therapies. Early anecdotal studies using intralesional BCG, led to the hypothesis that the products of ablation elicited an immune response resulting in the regression of distant sites of disease. Use fell out of favor because of serious allergic reactions, systemic BCG infections, and inconsistent effectiveness in subsequent series of patients. Several other locally ablative approaches have been studied, predominately in small single institutional cohorts of patients with mixed results; many however demonstrating relatively good local disease control. Agents and or strategies used include mechanical ablation such as cryotherapy and carbon dioxide laxer, cytokines (Interferon and IL-2), and electro-chemotherapy with Bleomycin. Large multi-institutional phase II and randomized phase III trials have been completed or are planned with novel agents (Alovectin-7, PV-10, and Oncovex) that not only have locally ablative qualities but also can elicit bystander responses in other sites of uninjected regional and distant metastases. Candidate trial patients have included those with stage III in-transit disease with or without stage IV (M1a, b, or c) disease or injectedable stage IV (M1a) with or without additional non-injectable distant disease. A combination of intralesional injection of IL-2 and topical Imiquimod and retinoids was recently published demonstrating a high rate of complete response in treated lesions in a small number of patients. Long-term results of the above phase II and III trials will determine the future use of such therapies and may provide the foundation for combination with such agents as CTLA-4 blockade and PD-1 or the BRAF inhibitors.

Percutaneous hepatic perfusion (PHP) versus best alternative care (BAC) for patients (pts) with melanoma liver metastases: updated data from the phase III trial

J. S. Zager1, J. F. Pingpank2

1 Department of Cutaneous Oncology, Moffitt Cancer Center, Tampa, FL, USA; 2Department of Surgical Oncology, University of Pittsburgh, Pittsburgh, PA, USA

There is no standard of care for metastatic melanoma to the liver. Systemic therapy has historically been very ineffective. PHP was designed to saturate the liver with high doses of chemotherapy, via a minimally invasive, catheter based approach. Reported are the data from the phase 3 multi-center randomized trial for pts with unresectable liver melanoma metastases. Patients were prospectively randomized 1:1 to PHP or BAC. PHP patients received melphalan via the hepatic artery over 30 min. Hepatic venous return was captured a specially-designed double-balloon catheter in the IVC, the melphalan was filtered extra-corporeally, and blood returned through the internal jugular vein. The procedure was repeated every 4–8 weeks on recovery from hematological toxicities. The primary endpoint was hepatic progression-free survival (hPFS). Secondary endpoints included safety, ORR, PFS, OS. Cross-over to PHP upon hepatic progression was permitted.

From February 2006 to July 2009, 93 patients were randomized to PHP (n = 44) or BAC (n = 49). Mean age was 54.8 yr with no significant differences in baseline demographics, pathology (uveal versus cutaneous), or previous treatments between the groups. AEs were primarily hematological (grade 3/4). As of 4/2011, hPFS was significantly improved in the PHP group, median 8.1 versus 1.6 months, HR 0.34, P < 0.0001. Overall PFS was improved in PHP (HR 0.41, P < 0.0001, median 6.1 versus 1.6 months). OS was not significantly different with1-yr OS of 29% (PHP) versus 26% (BAC), median OS PHP 11.4 versus BAC 9.9 months, due to 51% crossover from BAC to PHP. Crossover pts had a median hPFS from crossover date of 9.2 months and overall PFS 6.5 months.

This phase 3 study in patients with metastatic melanoma to the liver met its primary endpoint. hPFS, ORR and overall PFS were significantly improved with PHP versus BAC. PHP with melphalan should provide a new treatment option for unresectable metastatic melanoma in the liver.

Diagnosis and staging of mycosis fungoides (MF)

L. F. Glass

H. Lee Moffitt Cancer Center and Research Institute

Cutaneous T-cell lymphomas (CTCL) are a type of non-Hodgkin's lymphoma that develops primarily in the skin but ultimately may progress to involve lymph nodes, blood, and visceral organs. T cell neoplasms account for 65% of all primary cutaneous lymphomas. Mycosis fungoides is the most common type of CTCL, accounting for nearly 50% of all cases. It generally has an indolent course, and may show little or no progression over years to decade. In contrast, the leukemic variant of CTCL, Sézary Syndrome (SS), is associated with rapid progression and a poor prognosis.

In regards to staging, patients with limited patch-plaque mycosis fungoides generally have an excellent prognosis, with no apparent decrease in life expectancy compared to age, sex and race matched controls. In general, patients with stage IA, IB and IIA have a median survival of >12 yr, which drops to approximately 4 yr when tumors or erythroderma are present (stages IIB to III). The worst prognoses are associated with stage IV disease (histologically involved lymph nodes or viscera) and Sézary syndrome, with a median survival of <3 yr in each. Large cell transformation, a process in which lymphocytes typical of MF are replaced by large cells similar to those seen in ALCL, is also associated with an extremely poor prognosis. Opportunistic infections due to impaired immunity are the most common cause of death in CTCL.

The treatment of cutaneous T-cell lymphoma (CTCL) is largely dependent upon the stage of disease. While patients with localized disease can typically be successfully managed using skin-directed therapies, later CTCL stages often require systemic treatment with either biologic agents or chemotherapy. According to the NCCN, CTCL is such an infrequently encountered malignancy that is best managed with an individualized approach, and referral to a multidisciplinary academic specialty center for treatment may be preferable.

The biology and treatment of Merkel Cell carcinoma

R. C. Fields

Barnes-Jewish Hospital, Washington University School of Medicine, St. Louis, MO, USA

Merkel cell carcinoma (MCC) is a rare, cutaneous, neuroendocrine neoplasm with a propensity for lymphatic spread. The rarity of MCC has prevented any level-one evidence to guide treatment recommendations. Oncologists have relied on institutional case series, large registry reports (i.e. SEER, NCDB), and expert consensus to recommend treatment pathways. Generally, treatment for MCC is based on stage at presentation. Wide excision of the primary tumor with regional nodal sampling [i.e. sentinel lymph node biopsy (SLNB)] is recommended for clinically localized disease. Local and nodal adjuvant radiation therapy (RT) is controversial. Adjuvant chemotherapy is generally not recommended for clinically localized MCC. Outcomes after treatment for clinically localized disease are excellent. In contrast, patients presenting with clinically involved lymph nodes (stage 3b) have significantly decreased survival. Treatment consists of wide excision of the primary tumor with regional lymph node dissection. The use of RT and chemotherapy for stage 3 MCC is reasonable, given its high likelihood for both regional and distant recurrence. Adjuvant treatment for clinically localized MCC with a positive SLNB (Stage 3a) is also not standardized. Combinations of completion lymph node dissection, therapeutic or adjuvant RT, and chemotherapy have all been described. Recently, a polyoma virus integrated into the genome of patients with MCC has been characterized. The prognostic and therapeutic implications of this landmark finding are beginning to be understood and may dramatically change the way MCC is staged and treated.

Clinical, pathological and molecular study of the spectrum of skin tumors induced by RAF inhibitors

C. Robert1,2, L. Boussemart1,2, C. Mateus1, J. P. Arnault1, B. Escudier1, G. Tomasic,1, J. Wechsler1, J.-C. Soria1,2, A. Sarasin3, M. Larcher4, J. Andrée5,6, N. Kamsu-Kom2, L. Lacroix2, A. Spatz7, A. Eychène4, S. Vagner2, N. Dumaz5,6

1 Department of Medical Oncology, Institute Gustave Roussy; 2INSERM U981, Institute Gustave Roussy; 3CNRS, Institute Gustave Roussy, Villejuif, France; 4INSERM U1021 – CNRS UMR3347, Institute Curie, Orsay, France; 5INSERM, U976, Paris, France; 6Univ Paris Diderot, Paris, France; 7Department of Pathology, McGill University, Montreal, Canada

Clinical, pathological and molecular study of skin lesions occurring in patients receiving sorafenib, vemurafenib or show that these drugs can induce keratinocyte proliferation in vivo and a time and dose-dependent activation of the MAP-kinase pathway with ERK phosphorylation. They are associated with a spectrum of lesions ranging from benign follicular cysts, papillomatous or verrucous lesions, to keratoacanthomas (KA) and KA-like squamous cell carcinomas (SCC). Study of DNA extracted from the lesions and screened for mutation hotspots of several oncogenes like Ha-RAS, N-RAS, Ki-RAS, TP53, EGFR, BRAF, AKT, PI3K and PTEN shows that additional and potentially preexisting somatic genetic events, like UV-induced mutations of oncogenic proteins like ras or p53, might influence the evolution of benign lesions to more proliferative and malignant tumors. The fact that these skin lesions occur more frequently with more potent and specific BRAF inhibitors than with the pan-RAF inhibitor sorafenib could result from the differential effect of these drugs on CRAF or from various bioavailabilies of the drugs in the skin. These findings have other important clinical implications. Indeed, they also suggest that the use of RAF inhibitors might be pointless and even deleterious in normal tissues especially in cells harboring a RAS mutation, as it can be found in lung and colon tissue for example. Indeed, we cannot exclude the theoretical hypothesis that RAF inhibitors could promote cell transformation and/or proliferation in these wt BRAF cells with RAS activation and lead to internal tumors. Fortunately, these events have not been reported so far.

Altogether, the study of the effect of RAF inhibitors in the skin illustrates the complexity and the variety of the consequences of the systemic blockade of an universal pathway like the MAPK pathway and reminds us that efficient targeted therapies always have off-target effects that need to be carefully evaluated and monitored.

Cutaneous sarcoma update

G. A. McArthur

Peter MacCallum Cancer Centre

Cutaneous sarcomas are rare malignancies requiring a multidisciplinary approach to patient care. The most common sarcomas presenting to cutaneous oncology units are dermatofibrosarcoma protuberans (DFSP), leiomyosarcomas, atypical fibroxanthomas (AFX), angiosarcomas and Kaposi's sarcoma. Primary management is aggressive surgery with adjuvant radiotherapy or systemic therapy being used in some patients. Recent developments in managing these sarcomas have included the use of liposomal doxorubicin and taxanes in angiosarcomas and Kaposi's sarcomas, and platelet derived growth factor receptor (PDGFR) inhibitors for DFSP. The finding that over 90% of cases of DFSP contain translocations involving chromosomes 17 and 22 that result in the platelet derived growth factor-B gene coming under the control of the Collagen 1A1 locus led to the evaluation of PDGFR-inhibitors in DFSP. Results were striking with response rates of over 50% and neoadjuvant therapy being useful in some cases to allow complete disease resection. Cutaneous sarcomas represent a rare but important subset of cutaneous malignancies that require early multidisciplinary intervention to achieve the best patient outcomes.

Challenges to clinical trial development and implementation

R. R. Kudchadkar

H. Lee Moffitt Cancer Center, Department Cutaneous Oncology, Tampa, FL, USA

With clinical trials agents offering new hope and increased survival for patients with metastatic melanoma, the inadequacies of the current clinical research system have become more apparent. Rapid advances in our understanding of cancer biology have lead to a new era of oncologic drug development. There is currently more rigidity and complexity of the regulations than were ever used in past decades, thus a new era in clinical trial management is also present. This combination creates new challenges for institutions and individual investigators involved in oncologic drug development. From the preparation of the protocol, approval by multiple review boards, to budget contract negotiations; the initial implementation of a clinical trial in oncology has multiple barriers. Data management and continuing regulations consumes clinical trials offices throughout the United States far after the last patient is enrolled onto a given clinical trial. An increased focus on ensuring patient safety in clinical research via new regulations and monitoring has made the process cumbersome, though the toxicity of experimental therapy has not changed significantly since the 1970s. Hundreds of small hurdles in research have created this problem; therefore the solution is not to change just one given impediment. In order to resolve this problem, a new paradigm must be implemented that focuses on clinical progress. A look to our colleagues in pediatric oncology can provide an example how we can improve our system to become more streamlined throughout various cooperative-group, investigator-initiated, and pharmaceutical trials.

Abstracts of the 4th Melanoma Pathology Symposium of the International Melanoma Pathology Working Group November 13, 2011 Invited Faculty Abstracts

  1. Top of page
  2. 2011 International Melanoma Congress Advancement Through Collaboration joins the following four annual meetings in 2011:
  3. Congress Agenda
  4. Congress Proceedings Abstracts of the 8th International Congress of the Society for Melanoma Research November 9–11, 2011 Invited Faculty Abstracts
  5. Abstracts of the 5th Meeting of Interdisciplinary Melanoma/Skin Cancer Centres November 11–12, 2011 Invited Faculty Abstracts
  6. Abstracts of the 4th Melanoma Pathology Symposium of the International Melanoma Pathology Working Group November 13, 2011 Invited Faculty Abstracts
  7. Abstracts of the 8th International Congress of The Society for Melanoma Research November 9–11, 2011 Proffered Abstracts Selected for Oral Presentation
  8. Abstracts of the 8th International Congress of The Society for Melanoma Research November 9–11, 2011 Proffered Abstracts Selected for Poster Presentation
  9. Abstracts of the 5th International Meeting of Interdisciplinary Melanoma/Skin Cancer Centres November 11–12, 2011 Proffered Abstracts Selected for Poster Presentation

A new perspective for the classification of melanocytic lesions

A. Zembowicz

Associate Professor of Pathology, Tufts Medical School, Medical Director,, Boston, MA

Until recently, the prevailing paradigm in classification and clinical management of melanocytic proliferations mandated dichotomous classification of all melanocytic lesions as either entirely benign (nevus) or entirely malignant (melanoma). However, some diagnostically challenging lesions cannot be unequivocally classified as nevus or melanoma based on histologic evaluation of the primary tumor. Such lesions have been referred to as borderline or melanocytic tumors of uncertain malignant potential (MELTUMP). Recent evidence indicates that it may be appropriate to expand the classification scheme of melanocytic neoplasms to include a third diagnostic category of melanocytic lesions of intermediate malignant potential which are capable of metastasis to regional lymph nodes but have limited potential for distant spread. Dr. Richard Scolyer (Sydney Melanoma Institute of Australia)(1) and I recently proposed the term ‘melanocytoma’ for this group of lesions. We believe that a nevus/melanocytoma/ melanoma paradigm may provide a useful intellectual framework to understand, research and clinically manage borderline melanocytic tumors.


1. Zembowicz A. and Scolyer R. Arch. Pathol. Lab. Med. 2011; 135: 300–306.

The melanoma pathology report and prognostic factors

R. A. Scolyer1,2,3

1 Melanoma Institute Australia, Sydney, Australia; 2Tissue Pathology and Diagnostic Oncology, Royal Prince Alfred Hospital, Camperdown; 3Discipline of Pathology, Sydney Medical School, The University of Sydney, Sydney, NSW, Australia

Pathologic assessment of a tissue biopsy is a critical aspect for melanoma patient multidisciplinary management. Whilst melanoma may be clinically suspected, a definite diagnosis is established in most cases by pathologic assessment. The key criteria upon which the diagnosis was based should be documented in the pathology as well as important prognostic features and those that guide initial definitive management (e.g. the recommended width of excision margins and whether or not a sentinel node biopsy should be performed). Suboptimal care and poorer outcome may result from inadequate documentation. The components of the primary melanoma histopathology report that are of greatest importance in determining initial management and prognosis include its tumor (Breslow) thickness and whether or not the tumor has been completely excised. The tumor mitotic rate, ulceration, satellites and Clark level of invasion are other features important for staging and prognosis. The presence of lymphovascular invasion, neurotropism, a desmoplastic component, tumor-infiltrating lymphocytes, late regression and angiotropism may also influence prognosis and management. It has been shown that the use of a structured or synoptic pathology report format for cutaneous melanoma improves both the completeness of pathology reports and the validity of the reported data elements. Consistency and speed of reporting are also improved. Recently, there has been an international effort by pathology colleges in the United States, Canada, the United Kingdom and Australia to produce uniform guidelines for the reporting of various cancers including melanoma that can be utilized by pathologists throughout the world.

Sentinel lymph node biopsy (SLNB): state of the art

A. J. Cochran

Pathology, Laboratory Medicine and Surgery, David Geffen School of Medicine at UCLA, Jonsson Comprehensive Cancer Center and John Wayne Cancer Institute

Mature analysis of MSLT1 patients (median follow-up 9.2 yr) indicates a favorable outcome for melanoma patients managed by SLNB with immediate completion lymphadenectomy for tumor-positive SN. These patients achieve significantly longer regional and distant disease-free intervals and significantly better disease-specific survival. Successful lymphatic mapping and SLNB require specifically trained surgeons, nuclear medicine physicians and pathologists to precisely locate and remove the true sentinel node(s) (SN) and accurately determine the tumor-status of the SN. From MSLT1 and MSLT2 data and the literature it is essential to evaluate multiple levels of the SN, by HE histology and immunohistochemistry using antibodies to S-100, MART-1 and HMB-45. Debate persists on the optimal number and depth of tissue sections to be evaluated. ‘False negative’ SN are problematic; though not all indicate pathologist error. Some reflect errors in identification of true SN and others vagaries in the timing of metastatic extension through afferent lymphatics to regional nodes. Evaluation of nodes by RT-PCR has not been widely adopted clinically. Prediction of likelihood of SN metastases from primary tumor characteristics is possible but actual determination of SN status requires nodal biopsy. Likelihood of metastases in non-sentinel nodes (and need for completion lymphadenectomy), extra-regional melanoma metastases and death from melanoma can be predicted from the amount and anatomic location of SN tumor. This assigns patients to ‘risk-categories’ but is insufficiently accurate to tailor management decisions for individual patients. Individualized care may require combination of ‘standard’ laboratory evaluation with newer approaches, such as FISH, CGH and gene microarray assessment. Different gene signatures of primary melanomas and SN metastases have the potential to correlate with differing metastatic pathways. Even more precise prognostic data that can optimise surgical and medical management will likely derive from increasingly available sophisticated techniques, such as whole exome sequencing.

Is nodular melanoma a distinct entity or a final common pathway

M. Cook1, A. Viros2, R. Marais3, M. Mihm4

1 Dept of Hist, RSCH & Div of Clinical Medicine, Uni of Surrey, Guildford, UK; 2Inst of Cancer Research, London, UK; 3Inst of Cancer Research, London, UK; 4Dana Farber Inst, Boston, USA

Traditionally, melanoma is classified by clinical and histopathological features, summarised in the ‘histogenetic’ types formally proposed in the Sydney Classification 1969, and currently represented in the WHO classification of melanoma. A prominent criterion in the WHO classification of melanoma is the distribution of intraepidermal melanocytes during the early phase of melanoma growth when the tumours are expanding superficially, in the so called radial growth phase (RGP). Whereas the three types of superficial spreading melanoma (SSM), lentigo maligna melanoma (LMM), and acral lentiginous melanoma (ALM) are distinguished by their RGP pattern, the fourth type, nodular melanoma (NM), was proposed as a separate category based on the virtual absence of a RGP.

Whilst this classification corresponds well with clinicopathological observations it was said not to have prognostic implications and therefore its value was questionable (Koh et al 1984). Nodular melanoma came to be regarded as the final common pathway of other subtypes (Heenan et al. 1999 & 2003). On the other hand a large study (84 000 cases) suggested nodular and ALM did have significantly worse prognosis even allowing for thickness and other prognostic features (Chang et al. 1998).

Advances in molecular techniques have revealed that melanoma can also be classified according to its underlying genetic aberrations and these reflect both anatomical site and relationship to UV exposure and to a certain extent on the features of the RGP. NM comprises ~10% of all melanoma cases and in contrast to the other subtypes, does not show a particular association with body site, patterns of UV exposure, or particular age distribution and initial lesions present a subtle clinical appearance and often lack the ABCD criteria used to diagnose MM. This less aggressive clinical appearance in the early stage allows NM to escape early detection and to complicate matters further, NM patients generally lack classical high-risk melanoma phenotypes and present with few atypical naevi and low total body naevus counts. As a consequence of these clinical characteristics and a capacity to undergo rapid growth, at diagnosis NMs represent a large fraction of thick melanomas with poor prognoses. The question remains therefore is this a predetermined aggressive melanoma or is it the aggressive end stage of other types.

In common with other cutaneous melanomas, NRAS and BRAF are mutated in ~31% and 43% of cases respectively of NM, suggesting that the apparently distinctive clinical and pathological features of NM are driven by genetic aberrations other than these common driver oncogenes. Our colleagues at the ICR led by Professor Richard Marais have proposed that we should assess these questions by providing a comprehensive catalogue of the driving and cooperating mutations that occur in NM and to identify the genetic events that distinguish NM from other forms of cutaneous melanoma. We will present preliminary data from whole exome sequencing and array comparative genome hybridization (aCGH). We will overlay the aCGH and whole exome sequencing data to determine if specific genes are both mutated and amplified or lost to refine our ability to identify the driver oncogenes and tumour suppressors for further analysis.

Melanocytic proliferations in sun-damaged skin

J. L. Messina

Department of Cell Biology, University of South Florida College of Medicine, and Cutaneous Oncology Program, Moffitt Cancer Center, Tampa, FL

Accurate diagnosis of melanocytic proliferations in chronically sun-damaged skin is challenging, in both biopsy and therapeutic excision settings. The various surgical and pathologic methods utilized to effect the initial diagnosis of melanoma in situ in sun-damaged skin (lentigo maligna-LM), as well as to ensure its complete eradication, are presented herein. Ultimately, the optimal method(s) should be chosen based on the technical capabilities and resources of the institution and its faculty. Careful attention must be paid to regional variations in terminology, including the use and intended meaning of terms such as actinic melanocytic hyperplasia and atypical junctional melanocytic proliferation. Methods used to analyze pigmented lesions suspected to represent LM include immunohistochemical staining, image analysis of nuclear morphometric features, and in vivo reflectance confocal microscopy. We have demonstrated accurate distinction of LM from solar lentigo with immunohistochemical staining for micropthalmia transcription factor (MITF) in combination with nuclear morphometry measuring cellular density and nuclear diameter. Cutoff values that favored LM over solar lentigo included melanocyte density ≥10 nuclei per 200 mm, nuclear diameter ≥9 mm, and a product of density and diameter of 80 or more; each of these values demonstrated 100% specificity of LM diagnosis. Excision choices for LM are varied, including standard elliptical excision, marginal (“picture frame”) excision, Mohs surgery (and its “slow-Mohs variant); topical therapies such as imiquimod and radiation treatment have also been advocated. Optimal methods of pathologic evaluation of surgical specimens are also debated. Our group has demonstrated that effective removal of large LM with clinically indistinct borders can be achieved using staged contoured marginal excision, wherein thin strips of clinically uninvolved skin 5 mm beyond the clinical border are removed, analyzed en face in pathology, and final excision/repair accomplished in a separate procedure after clear margins are obtained.

Desmoplastic melanocytic lesions

C. R. Shea

Department of Medicine, Section of Dermatology, University of Chicago, Chicago, IL, USA

Desmoplasia (i.e., formation of fibrous connective tissue by proliferation of fibroblasts) is observed in various cutaneous and non-cutaneous neoplasms. It is a particular source of diagnostic concern in the pathology of melanocytic tumors. Desmoplastic malignant melanoma (DMM) is notoriously easy to miss, for several reasons. First, DMM cells are frequently obscured by the stromal response, simulating a scar; this pitfall is especially hazardous in re-excision specimens in which both true scar tissue and DMM are present. Second, in contrast to most types of primary cutaneous melanoma, an in-situ component is frequently absent from DMM. Third, DMM usually exhibits minimal if any immunohistochemical expression of specific markers of melanocytic differentiation such as gp100 (HMB-45), Melan A/MART-1, and microphthalmia transcription factor; however, expression of S-100 protein is uniformly strong. Clues to the diagnosis of DMM upon low-magnification inspection include the presence of deep infiltrates of lymphocytes deep in the dermis; at higher power, cytologically atypical spindle cells are evident. Of note, the presence of desmoplasia in a melanocytic neoplasm does not necessarily indicate DMM. Various types of melanocytic nevi, including Spitz nevi and blue nevi, commonly induce a desmoplastic tissue response and assume the appearance of desmoplastic nevi (DN). Such lesions may in some cases also simulate non-melanocytic lesions including dermatofibroma/benign cutaneous fibrohistiocytoma. Distinction of DMM from DN is crucial for appropriate clinical management; while some studies suggest that DMM when occurring in a pure form may have a better prognosis than conventional melanoma of equivalent staging characteristics, nonetheless DMM must be treated as a fully malignant tumor. Accurate pathologic diagnosis of DMM and its distinction from DN are facilitated by consideration of demographic data (e.g., tendency to occur on the head and neck of elderly patients), gross features (e.g., size), higher degree of cytological atypia, and immunohistochemical profile.

Grading and treatment of dysplastic melanocytic nevi

L. M. Duncan

Dermatopathology Unit, Massachusetts General Hospital and Harvard Medical School, Boston, MA, USA

Dysplastic nevi are important mimics, risk markers and precursors of melanoma. Treatment and risk assessment is determined by the histopathological grade of atypia. Criteria for diagnosis include architectural and cytological features; in a simple algorithm developed by the World Health Organization 20 yr ago, two major and four minor criteria were described. The major criteria include: (i) a basilar melanocytic proliferation at the dermal epidermal junction (extending at least three rete ridges lateral to any dermal melanocytic proliferation, a.k.a. a shoulder), (ii) cytological atypia of the intraepidermal nevomelanocytes. Both of the major criteria must be observed. The minor criteria include: (i) prominence of the superficial venular plexus, (ii) sub-epidermal fibrosis, in a lamellar or eosinophilic concentric pattern, (iii) inflammatory infiltrate, (iv) apparent fusion of the rete by confluent nests of nevomelanocytes (a.k.a. bridging). At least two of four minor criteria must be present. If the diagnostic criteria of dysplastic nevus are identified, the tumors are graded based upon the degree of cytological atypia.

Mildly atypical nevomelanocytes have round, oval or sickle-shaped nuclei 1.5–2× the size of a keratinocyte nucleus, chromatin is uniformly dispersed without identified nucleoli, cytoplasm is sparse or forms a perinuclear halo. Moderately atypical nevomelanocytes have features intermediate between mildly atypical nevomelanocytes and the melanoma-like cytology of severely atypical nevomelanocytes. Moderate atypia is often characterized by finely pigmented cytoplasm and enlarged polygonal nuclei. Severe nevomelanocytic atypia is manifested by nuclear size >2× a keratinocytic nucleus, chromatin clumping or nuclear membrane notching and prominent or multiple nucleoli.

Controversy exists regarding the treatment of dysplastic nevi. Our practice is to recommend complete excision of dysplastic nevi with moderate or severe cytological atypia. Complete excision is not recommended for slightly atypical dysplastic nevi unless there is significant clinical residual that may be associated with sampling error.

Grading of dysplastic nevi does not predict transformation to melanoma

L. F. Glass

H. Lee Moffitt Cancer Center and Research Institute

There has been considerable debate over the issue of whether ‘dysplastic nevi’ are potential precursors of melanoma, and debate over histologic grading of these lesions. Two recent studies demonstrate that the grade of histological dysplasia correlates with melanoma risk, one showing a 19.7% personal history of melanoma for nevi associated with severe dysplasia, and another, a 3.99 relative risk of melanoma with severe dysplasia. However, there is no evidence whatsoever supporting the notion that grade of dysplasia is associated with the capacity of a nevus to develop melanoma.

Dysplastic nevi may simulate melanomas both clinically and histologically. As opposed to melanomas, however, which are always malignant, dysplastic nevi are benign. They have the potential to be precursors of melanoma at some time in the future, but the overwhelming majority of these lesions never progress to melanoma, and there are no histologic criteria or grading system that can predict a transformation. The principal reason for histologic examination of these lesions is to distinguish them from melanoma, not to predict which ones will eventually transform into melanoma.

In fact, the likelihood that a given dysplastic nevus would actually produce a melanoma is remote. A more likely indicator of tumor progression would be a change in the appearance of a nevus, rather than grading of its histopathologic criteria. Although reproducibility for grading of dysplastic nevi may be satisfactory when using previously agreed upon histologic criteria, population-based studies show a variation in recognition between 7% and 32% among observers, likely because of different criteria for melanocytic dysplasia. In summary, determination of the grade of histologic dysplasia in a nevus may be relatively straightforward, but grading does not predict its progression to melanoma.

Unusual presentations of metastatic melanoma

G. Landman

Department of Pathology, EPM-UNIFESP, São Paulo, Brazil

Melanoma metastases may pose difficulties in sorting out the differential diagnosis with a broad spectrum of tumors. Lesions devoid of pigment can present as spindle cell, myxoid, chondroid, osteoid, small cell, signet-ring, and even be admixed to other tumors, such as salivary glands, leyomiomas, colon adenomas, ovarian tumors and so on. We present two unusual metastatic melanomas. Case 1: A 78-yr-old female renal transplant recipient with a melanoma on her back, non-ulcerated, Clark's IV, 2.2 mm. Left axillary sentinel lymph node biopsy revealed a metastatic lesion with large sheets of non-pigmented keratinocytes, and isolated or nested atypical melanocytes filled with melanin pigment. Keratinocytes were AE1/AE3 positive and pigmented atypical melanocytes stained for HMB-45, melan A and S-100 protein. The diagnosis of malignant melanoma in association with a SCC lymph node metastasis was made. Reviewing one of her left arm multiple SCCs, acquired over the years there were admixed atypical melanocytes. Primary melanomas have been described within of basal cell carcinoma and/or squamous cell proliferations, that conceivably could metastasize as a biphasic tumor. Alternatively, SCCs and melanoma isolated lesions could metastasize to the same lymph node. Both possibilities could explain the present case. Considering the close admixture of atypical melanocytes and keratinocytes, we favor a biphasic metastasis. Case 2: A white male, 51 yr old, presented in 2000 an acral-lentigious melanoma on the 1st left hand finger, Clark V, 19.5 mm in thickness. Three years later, a left axillary soft tissue tumor showed a malignant metastatic tumor with epithelioid atypical cells, containing bone trabecules. S-100 protein and Vimentin were positive. Both in the 4th year and 5th year, multiple pulmonary metastases had a chondroid background with lacunae filled with atypical neoplastic cells, Melan-A and HMB-45 posite. The final diagnosis was acral-lentiginous melanoma with osteoid and chondroid metastases, uncommon differentiations in primary and metastatic melanoma. The majority of primary melanomas with these differentiations are acral-lentiginous, but in general, they occur in metastases, a potential diagnostic pitfall.

Angiotropism and extravascular migratory metastasis in melanoma: an update

R. L. Barnhill1, J. S. Wilmott2, L. E. Haydu3, E. Zhang3, V. Jakrot3, J. F. Thompson2, R. A. Scolyer2, M. Bagot4, C. Lugassy1

1 University of California, Los Angeles, CA, USA; 2University of Sydney, Sydney, Australia; 3Melanoma Institute Australia, Sydney, Australia; 4Hopital Saint-Louis, Paris, France

Recent progress in angiotropism and extravascular migratory metastasis (EVMM) in melanoma have included the following investigations. A microarray analysis of angiotropic versus non-angiotropic melanomas has facilitated the identification of genes potentially directly involved in extravascular migratory metastasis (EVMM). Immunohistochemistry has shown the over-expression of TCOF1 in angiotropic melanoma, a gene involved in neural crest cell migration. Microscopic satellites (MS) are local micrometastases that correlate with melanoma recurrence and diminished survival. Analysis of the relationship between angiotropism and MS has confirmed that angiotropism is more prevalent in melanomas with MS (23/44, 52%) than in those without (12/44, 27%) (P = 0.017). Further angiotropism is the only factor among several other factors that independently predicts MS formation when analyzed by binary logistic regression (P < 0.02). Of particular interest, a subset of melanomas with diffuse (at least four MS) angiotropic MS seem to predict a fatal outcome. Inter-observer agreement among three pathologists for the recognition of angiotropism in human melanomas has recently been demonstrated. The Kappa statistic revealed agreement rates for pairs of observers as follows: 0.51, 0.65, and 0.75, indicating overall ‘moderate to good’ to ‘excellent’ (P-values from 0.01 to 0.0003) concordance for angiotropism. The simultaneous comparison of the three observers by intra-class correlation yielded a highly significant agreement of 0.775 (P = 0.001).

Melanoma stem/initiating cells from the pathologist's perspective

J. J. van den Oord

Lab Morphology & Molecular Pathology, University Hospitals Leuven, Katholieke Universiteit Leuven, Leuven, Belgium

The concept of cancer stem or initiating cells (CSC) implies the existence of (a subset of) tumor cells, capable to self-renewal, and to the establishment and maintenance of tumor growth in a hierarchically organized manner. Studies on cell lines strongly suggest the existence of such CSC in melanoma, based on cell surface markers (e.g. ABCB5, CD271, ALDH, JARID1B, CD20), xenotransplantation into immune-deficient mice, and in vitro spheroid formation. However, use of various mouse models and different matrix constituents has yielded discrepancies as to their relative number and phenotype; thus, the question remains whether in-vitro data from cell lines can be applied on CSC in melanoma tissue samples. Moreover, stemness in melanoma is likely to be a dynamic, temporary process in which CSCs exhibit a highly plastic phenotype, continuously changing to adapt to the environmental pressure. Studies on the control of this ‘phenotype switching’ suggests a role for environmental factors (hypoxia) and Mitf/Brn-2.

We have searched for melanoma CSC starting from fresh frozen patient melanoma samples using three approaches: (i) in fresh primary and metastatic melanomas, the side population (comprising the CSC) was studied by FACS and yielded an SP in all melanomas but not in nevi; gene expression analysis of this SP revealed several up-regulated CSC and epithelial-to-mesenchymal transition (EMT)-related genes; (ii) in patients undergoing isolated limb perfusion with Melphalan for multiple in-transit metastases, we observed up-regulation of CSC and EMT-related genes after 1 h of chemotherapy; relapsed lesions, emerging after initial total regression of in transit metastases were micro-arrayed for aberrant gene expression; and (iii) in a large series of primary and metastatic melanomas, immunohistochemical staining for Mitf revealed a subset of Mitf-negative cells that were particularly prominent in lesions showing tumor necrosis. Currently, the detailed phenotype of these cells is being studied using micro-dissection of fresh frozen samples.

Melanoma stem cells: not rare, but well-done

G. F. Murphy

Brigham and Women's Hospital and the Harvard Medical School

Melanoma is one of the most deadly tumors to which flesh is heir. Its virulence has recently been recognized to reside in a subpopulation of cells that confer aggressive behavior. Biomarkers now exist to identify and target these cells. Their characteristics include tumorigenic growth as a result of asymmetrical division and self-renewal, ability to invade and metastasize, and mechanisims to evade therapeutic agents and immune responses. Moreover, they exhibit plasticity that may further enhance their growth and well-being. New treatments designed to interrupt molecular pathways in melanoma as well as check-point blockade approaches have been preliminarily promising. However, combinatorial targeting to eradicate minority populations of the most virulent melanoma cells likely will be required to maximize therapeutic strategies to eliminate all tumor cells and attain long-term remissions and even cures.

Update on FISH as a diagnostic and prognostic tool for melanocytic neoplasms

P. Gerami

Northwestern University, Department of Dermatology and the Feinberg School of Medicine

Initial studies utilizing a heterogenous group of melanoma and nevi as a training set, identified 6p25, 6q23, Cep6 and 11q13 as the highest yielding four probe FISH set for differentiating melanoma from benign nevi. Since that time our group has investigated a variety of other probe sets for their diagnostic and prognostic potential. In particular we show that the addition of 9p21 as an adjunctive probe can significantly enhance the ability to identify melanomas with Spitzoid morphology. Also in a case controlled retrospective study of melanomas with and without metastasis, we identified 8q24 and 11q13 as highly indicative of aggressive behaviour in cases of unequivocal melanoma. Further we found that melanomas with copy number gains in 8q24 are an aggressive subgroup of melanomas with highly characteristic clinical and histopathologic features. Hence by recognizing these characteristic features this may help allow for the development of tailored probe sets based on the presenting clinical and histologic features. Hence using tailored probe sets including 9p21 for melanomas with spitzoid morphology and 8q24 for amelanotic nodular or primary dermal melanocytic neoplasms from non-chronically sun damaged skin we can increase our efficiency when using fluorescence in situ hybridization. Finally we present our data from a multicenter collaborative study evaluating the potential of these adjunctive probes in addition to the original melanoma FISH assay to predict outcome in a large set of ambiguous melanocytic tumors with known outcome.

Prognosis in thin melanoma: an update

D. Elder

University of Pennsylvania

In 1967, Breslow published a seminal study based on 98 melanoma cases, demonstrating that melanomas <0.76 mm in thickness had an excellent prognosis, with no metastases observed. The American Joint Committee on Cancer (AJCC) more recently selected a cutoff of 1.0 mm which, with ulceration and mitogenicity as stage modifiers, achieves a similar result, although some metastases and deaths do occur, especially in the thickest quartiles. A few years earlier, Clark had demonstrated an almost equally good prognosis for level II melanomas. The majority of these lesions did not form tumors or undergo mitosis in the dermis, and were therefore ‘nontumorigenic’ and ‘nonmitogenic’, and lacked competence for metastasis. Studies of these low risk melanomas led to the development of criteria for earlier diagnosis and improved prediction of prognosis. A multivariable model developed in the Penn Pigmented Lesion Group (PPLG) identified patients with thin melanomas whose prognosis varies from 99% to about 70% disease free survival. Prognosis declines with increasing thickness, modified by ulceration, mitotic rate, and other attributes. Mitogenicity appears to be of particular importance in thin melanomas, and has recently been added to the AJCC staging system. Patients with mitogenic melanomas, even when these are thin, are usually considered for sentinel node staging. The complete absence of mitoses defines a nonmitogenic melanoma and should be expressed as a rate of zero. The mitotic rate in mitogenic melanomas should be expressed as a whole number. Melanomas should not be exhaustively sectioned to determine this attribute, because this was not the method used in the original studies, and also to preserve tissue for molecular studies if needed for patient care. In the PPLG database, the frequency of mitogenicity declines with thickness, being <10% in melanomas thinner than 0.5 mm, consistent with the excellent prognosis of this subgroup.

Abstracts of the 8th International Congress of The Society for Melanoma Research November 9–11, 2011 Proffered Abstracts Selected for Oral Presentation

  1. Top of page
  2. 2011 International Melanoma Congress Advancement Through Collaboration joins the following four annual meetings in 2011:
  3. Congress Agenda
  4. Congress Proceedings Abstracts of the 8th International Congress of the Society for Melanoma Research November 9–11, 2011 Invited Faculty Abstracts
  5. Abstracts of the 5th Meeting of Interdisciplinary Melanoma/Skin Cancer Centres November 11–12, 2011 Invited Faculty Abstracts
  6. Abstracts of the 4th Melanoma Pathology Symposium of the International Melanoma Pathology Working Group November 13, 2011 Invited Faculty Abstracts
  7. Abstracts of the 8th International Congress of The Society for Melanoma Research November 9–11, 2011 Proffered Abstracts Selected for Oral Presentation
  8. Abstracts of the 8th International Congress of The Society for Melanoma Research November 9–11, 2011 Proffered Abstracts Selected for Poster Presentation
  9. Abstracts of the 5th International Meeting of Interdisciplinary Melanoma/Skin Cancer Centres November 11–12, 2011 Proffered Abstracts Selected for Poster Presentation


New animal models for the analysis and visualization of lymphangiogenesis and metastasis of malignant melanoma

D. Olmeda1, E. Riveiro-Falkenbach1, D. Alonso1, E. Pérez-Guijarro1, T. G. Calvo1, A. Angulo1, I. Martínez-Corral2, S. Ortega2, M. S. Soengas1

1 Melanoma Laboratory, Molecular Pathology Programme, Spanish National Cancer Research Centre (CNIO), Madrid, Spain; 2Transgenic Mice Unit, Biotechnology Programme, Spanish National Cancer Research Centre (CNIO), Madrid, Spain

Cutaneous melanoma is the 5th most frequent tumor, and the deadliest form of skin cancer. Although the presence of tumor cells in lymph nodes is a poor prognosis indicator, the specific contribution of lymphangiogenesis to the mortality associated with melanoma metastasis remains unknown. This is mainly due to the lack of tumor markers and experimental models for non-invasive imaging of lymphoangiogenesis in vivo. To address molecular mechanism(s) regulating metastatic spread of melanoma, we are using cell lines, xenografts and genetically engineered mice, in combination with a new ‘lymphoreporter model’ in which the luciferase protein is expressed under the control of the endogenous Flt4 (VEGFR3) promoter. Using these reporter animals we have found that the metastatic potential of melanoma cells is in fact tightly associated to their ability to induce neo-lymphangiogenesis both, at proximal sentinel lymph nodes and at distant sites. In fact, the induction of VEGFR3-dependent luciferase was detected before the onset of metastasis. Therefore, these results support the ‘seed and soil’ hypothesis of metastatic dissemination driving aggressive melanomas. Moreover, the Lymphoreporter mice have been very useful in the pharmacological characterization of anticancer agents. Altogether, we expect that the combination of tumor tissues and animal models will clarify one of the least understood aspects of metastatic melanoma, that is, the role of lymphangiogenesis and inflammation in tumor progression and chemoresistance.


Why red-heads are at increased risk of melanoma: a novel BRAF mutant mouse model

D. Mitra1, X. Luo2, J. Wargo3, M. McMahon4, M. Bosenberg5, K. Haigis2, D. Haber2, D. Fisher6

1 Harvard Medical School, Boston, MA, USA; 2Massachusetts General Hospital Cancer Center, Boston, MA, USA; 3Massachusetts General Hospital Surgical Oncology, Boston, MA, USA; 4University of California San Francisco Helen Diller Family Comprehensive Cancer Center, San Francisco, CA, USA; 5Yale Medical School Dermatopathology, New Haven, CT, USA; 6Massachusetts General Hospital Cutaneous Biology Research Center, Boston, MA, USA

The goal of our studies was to develop a novel mouse model to investigate the long-standing epidemiological observation that red-heads are at relatively high risk of developing melanoma. The most frequent genetic causes of the red-hair/pale skin phenotype are polymorphisms in the Melanocortin 1 Receptor (Mc1R) gene, which is normally responsible for stimulating production of dark eumelanin pigment.

Previously, it was hypothesized that individuals with Mc1R red-hair/fair-skin polymorphisms are at increased risk of melanoma because they lacked eumelanin, a known protective agent against genotoxic UV radiation from the sun. To investigate if red-heads have a UV independent risk of melanoma, three strains of C57B/6 mice were acquired that express (i) the ‘wild type’ Mc1R (black) phenotype, (ii) an Mc1R polymorphic ‘red-head’ phenotype, or (iii) a tyrosinase mutant ‘albino’ phenotype. Into each pigmentation-variant we introduced an inducible, melanocyte-specific, BRAF(V600E) oncogene, previously shown in murine models to induce melanoma when combined with inactivation of tumor suppressor genes. We utilized this murine BRAF (V600E) conditional knock-in model to ask whether the red-head (mc1re/e) background influences melanomagenesis.

We observed a striking enhancement in melanoma incidence as well as shortened latency to tumor formation following BRAF (V600E) induction in red-head mice, versus genetically matched black or albino animals. The tumors were diffusely positive for S100 and displayed some evidence of microscopic pigmentation, though they were grossly amelanotic.

It is notable that genetically matched black BRAF (V600E) conditional knock-in mice develop benign melanocytic proliferations (presumably akin to nevi) and only develop invasive melanomas after much longer latency than red-head mc1re/e matched animals. Importantly, the elevated melanoma predisposition of the red-head allele is entirely independent of UV exposure in this study. We are currently utilizing a set of genetically defined gene disruption models to gain additional insights into the precise mechanism(s) by which Mc1R polymorphisms amplify UV-independent melanoma risk.


Dynamic intra-tumoral evolution of human melanoma

M. Shackleton1,2, E. Quintana2, J. Li3, S. J. Morrison2

1 Melanoma Research Laboratory, Peter MacCallum Cancer Centre, East Melbourne, Vic., Australia; 2Howard Hughes Medical Institute, Life Sciences Institute, Center for Stem Cell Biology, Department of Internal Medicine, University of Michigan, Ann Arbor, MI, USA; 3Department of Human Genetics, University of Michigan, Ann Arbor, MI, USA

Malignant tumors are frequently incurable because of their evolutionary nature. However, revealing the extent of heritable molecular change in cancer has been difficult, as it has not been possible to study genetic and epigenetic variation among individual cells from the same tumor. In overcoming this, we have found that genetic and epigenetic changes are remarkably dynamic within human melanoma. Using as founder tumors melanomas from four separate patients, we performed serial in vivo xenotransplantation of single cells isolated from each melanoma, creating four multi-generational clonal melanoma tumor families. SNP genotyping and DNA methylation arrays were performed on gDNA extracted from each tumor, allowing inter-clonal comparison of CN versus and DNA methylation status. All founder tumors contained CNVs and DNA methylomes that were present in clonal progeny. However, new CNVs appeared frequently during clonal serial transplantation of three of the four melanomas studied, involving 8–15% of the genomes of these tumors. These new changes affected genes known to be functionally relevant in melanoma, such as N-RAS and BRAF. Furthermore, new DNA methylation changes were observed in all melanomas, including the genetically stable melanoma, and involved 0.007–8.6% of CpG sites. Sites of altered DNA methylation were randomly distributed across the genome and did not correlate with CN versus Principal component analysis showed expected inter-melanoma clustering of genetic/epigenetic changes, but also clustering according to intra-tumoral clonal subfamily, indicating heritability of new changes. These data reveal pervasive instability of the melanoma genome and epigenome.


MDMX is a key therapeutic target in melanoma

A. Gembarska1,2, F. Luciani1, J. de Lange3, A. Zwolinska1,2, D. Yip4, J. Goydos4, J. Haigh5, L. Larue6, A. Jochemsen2, G. Ghanem7, F. Bernal8, J.-C. Marine1,2

1 Laboratory for Molecular Cancer Biology, VIB-KULeuven, CME, Belgium; 2Laboratory for Molecular Cancer Biology, VIB-UGent, DMBR, Leuven, Belgium; 3Leiden University Medical Centre, Dep. of Molecular Cell Biology, Leiden, The Netherlands; 4UMDNJ-Robert Wood Johnson Medical School, CINJ, NJ, USA; 5Vascular Cell Biology, VIB-UGent, DMBR, Ghent, Belgium; 6Developmental Genetics of Melanocytes, Curie Institute, Orsay, France; 7Institute Jules Bordet, Brussels, Belgium; 8Metabolism Branch CCR/NCI/NIH, Bethesda, MD, USA

Despite its universal role in tumor suppression the relevance of p53 in melanomagenesis remains highly controversial, partly because the mechanisms that disrupt p53 function in melanoma remain elusive. In fact, <5% of melanomas carry mutations that inactivate p53 itself. Here, we show that MDMX, a negative regulator of p53, is overexpressed in 65% of human skin melanomas. We establish a causative link between MDMX overexpression and melanoma formation in vivo. Melanocyte-specific overexpression of Mdmx cooperates with the NrasQ61K mutation in promoting malignant melanoma formation in transgenic mice. Importantly, we show that most human malignant melanoma cells are addicted to high MDMX levels. MDMX promotes melanoma cell survival by antagonizing p53 pro-apoptotic function. Unexpectedly, MDMX is also required for the high proliferation capacity of melanoma cells independently of its ability to regulate p53. Finally, all melanoma cell lines expressing high MDMX are sensitive to a specific inhibitor of the MDMX-p53 interaction, alone or in combination with an MDM2 antagonist. In addition, the MDMX antagonist dramatically sensitizes malignant melanoma cells to chemotherapy. Together, these data identify a key mechanism that disrupt p53 function in melanoma and designate MDMX as an important therapeutic target for melanoma treatment.


SOX10 is an essential regulator of melanoma cell proliferation

J. C. Cronin1, S. K. Loftus1, B. Sharma2, N. Schönewolf3, N. Bouvier4, N. Hayward5, B. Bastian4, R. Dummer3, W. J. Pavan1

1 Genetic Disease Research Branch, National Human Genome Research Institute, NIH, Bethesda, MD, USA; 2Harry and Jeanette Weinberg Cancer Institute, Franklin Square Hospital Center, Baltimore, MD, USA; 3University Hospital of Zurich, Department of Dermatology, Zurich, Switzerland; 4Memorial Sloan-Kettering Cancer Center, Sloan-Kettering Institute, New York, NY, USA; 5Queensland Institute of Medical Research, Oncogenomics Department, Queeensland, Australia

SOX10 (Sry-like HMG box) is a member of the high-mobility-group domain SOX family of transcription factors, playing an essential role in the specification, survival, proliferation and differentiation of neural crest cell lineages through regulation of downstream target genes MITF and melanocyte differentiation genes. SOX10 is expressed in melanocytes, melanocytic neoplasms, and melanoma, and has recently been shown to be a useful diagnostic marker of melanocytic tumors. Understanding how SOX10 expression influences melanoma proliferation could provide diagnostic, prognostic and therapeutic value. Previously we reported that SOX10 is mutated in six primary melanoma tumors exhibiting non-cutaneous histology. To further explore the mutation spectrum of SOX10 in melanoma, we are analyzing SOX10 coding region sequence from additional metastatic and primary melanoma tumor samples from multiple histological subtypes. Consistent with the observed heterogeneity of melanoma lesions and cell lines derived from these tumors, we find a dynamic range of SOX10 expression across established melanoma cell lines, with an absence of SOX10 expression patterned with reduced cellular proliferation. We are currently addressing how SOX10 expression is being regulated and determining the functional consequences of stable reduction of SOX10 protein in melanoma cells. Short hairpin RNA technology was employed to generate melanoma lines with stable knockdown of SOX10 expression. Downregulation of SOX10 resulted in cellular senescence in four out of five cells lines, showing significant reduction in proliferation rate and increased SAHF (senescence associated heterochromatin foci). The mechanisms regulating this senescent phenotype are being investigated.


Transcriptional plasticity and cycling between proliferative and invasive gene expression signatures: clinical implications

R. Dummer, O. Eichhoff, D. Widmer, Ph. Cheng, B. Belloni, K. Hoek

Department of Dermatology, University Hospital Zurich

Human melanoma metastases as in other tumours include heterogeneous tumour cell populations and benign host cells including blood vessels and inflammatory cells. In vitro culturing of human melanoma biopsies results in tumour cells with different morphology. This morphology is associated with differences in proliferative and migratory capacity. A careful analysis of the gene expression profile of the proliferative and invasive phenotypes has been published and has identified MITF and pigmentation genes to be over expressed in the proliferative phenotype, in comparison to the invasive phenotype where many TGF-beta pathway genes and inflammatory cytokines are overexpressed. Phenotype switching is associated to changes in the WNT signalling pathway. We have recently shown that LEF1 is preferentially expressed by proliferative phenotype cells. In contrast, TCF4 expression is associated with the invasive phenotype cells. We have investigated RAF and MEK inhibitors in the context of the phenotype switching model. Interestingly, these kinase inhibitors targeting the MAPK pathway are efficient against proliferative cells but hardly efficient against invasive cells and inhibition seems to be independent of the BRAF mutation status.

The phenotype switching model and the effect of kinase inhibitors on the two phenotypes might explain the limited response rate and response duration of kinase inhibitors such as vemurafenib or MEK kinase inhibitors. The invasive phenotype cells display many features of slow cycling stem cell-like cells which might also explain other phenomena in melanoma such as very late recurrences and the failure of immunotherapy targeting melanocytic differentiation antigens.


WIPI-1 links oncogenic BRAF to deregulated autophagy in melanoma

J. L. Armstrong1, R. Ellis1, H. Polson2, N. Kirkham3, S. Horswell2, S Tooze2, P. E. Lovat1

1 Dermatological Sciences, Institute of Cellular Medicine, Newcastle University, Newcastle upon Tyne, UK; 2London Research Institute, Cancer Research UK, London, UK; 3Department of Pathology, Royal Victoria Infirmary, Newcastle upon Tyne, UK

Cutaneous melanoma is an increasing public health problem and novel treatment strategies remain in acute demand. Recent evidence suggests autophagy, the principle mechanism for the lysosomal degradation and recycling of surplus proteins and damaged organelles is important in tumourigenesis and a relevant target for cancer therapy. In the context of melanoma we have recently demonstrated oncogenic B-RAF is associated with deregulated autophagy. The aim of the present study was to decipher the molecular regulation of autophagy inhibition in this context.

Data demonstrate increased expression of the autophagy regulatory protein WIPI-1 in B-RAF mutated compared to B-RAF wild-type melanoma cell lines, while over-expression of WIPI-1 in B-RAF wild-type melanoma cells resulted in blockade of autophagy activation in response to nutrient deprivation, identifying WIPI-1 as a novel autophagy inhibitory protein in melanoma. In addition, WIPI-1 expression was increased by over-expression of B-RAF V600E in B-RAF wild-type cells, whereas inhibition of MEK signalling resulted in down-regulation of WIPI-1 expression in B-RAF mutated melanoma cells. Data also demonstrate primary early stage tumours derived from B-RAF mutated melanomas exhibit significantly increased WIPI-1 expression compared to B-RAF wild-type tumours (P < 0.001). As B-RAF signalling is activated in the majority of melanomas, these data suggest increased WIPI-1 expression and deregulated autophagy may play a role in the development and progression of melanoma.

The identification of novel proteins coupled with defective autophagy may therefore facilitate the discovery of novel therapeutic targets and biomarkers of melanoma progression.


ROCK and JAK1 signalling cooperate to control actomyosin contractility in melanoma

V. Sanz-Moreno1,2, M. Yeo1, F. Wallberg3, A. Viros4, S. Hooper5, R. Mitter6, M. Cook7, J. Larkin8, R. Marais4, E. Sahai5, C. J. Marshall1

1 Oncogene Team, Institute of Cancer Research, Section of Cell and Molecular Biology, Cancer Research UK Tumour Cell Signalling Unit, London, UK; 2Randall Division of Cell and Molecular Biophysics, New Hunt's House, London, UK; 3Institute of Cancer Research, Section of Cell and Molecular Biology, London, UK; 4Signal Transduction Team, Institute of Cancer Research, Section of Cell and Molecular Biology, Cancer Research UK Tumour Cell Signalling Unit, London, UK; 5Tumour Cell Biology Laboratory Cancer Research UK London Research Institute, Lincoln's Inn Fields, London, UK; 6Bioinformatics and Biostatistics Service Cancer Research UK London Research Institute, London, UK; 7Royal Surrey County Hospital, Guildford, UK; 8Skin Unit, Royal Marsden NHS Trust, London, UK

Cutaneous melanoma is the deadliest form of skin cancer due in part to its high potential for invasion and metastatic spread. In order to invade and metastasize, tumour cells can move as collective groups or as individual cells. Individual movement is necessary for the cells to cross basement membranes and enter blood vessels. Individual tumour cells have two different modes of movement:

1 Mesenchymal-type mode characterised by an elongated morphology, that requires extracellular proteolysis localized at cellular protrusions

2 Amoeboid mode where cells have a rounded morphology, movement is less dependent of proteases and relies on high acto-myosin contractility.

We and others have previously shown that these two modes of tumour cell movement are inter-convertible, depending on environmental conditions. Importantly in this study, we have found that the dermal invasive front of the majority of melanoma primary tumors consists of cells with a rounded morphology. Round cells predominated in the invasive front even when the cells in the body of the tumor had an elongated morphology.

Pro-inflammatory cytokines are frequently observed in the tumor microenvironment, and chronic inflammation is involved in cancer initiation and progression. We have now been able to show that cytokine signaling through the receptor subunit GP130-IL6ST and the kinase JAK1 generates actomyosin contractility through Rho-kinase dependent signaling. This pathway provides the high levels of actomyosin contractility required for migration of individual melanoma cells in the rounded, ‘‘amoeboid’’ mode.

Strikingly, actomyosin contractility itself positively modulates activity of the transcription factor STAT3 downstream of JAK1, demonstrating positive feedback within the signaling network. Therapeutic agents such as blocking antibodies against cytokines or small molecule inhibitors of JAK kinase or STAT activity may be useful agents to block invasion and metastasis.


Investigation of the melanoma genetic landscape identifies the glutamate pathway as a major player in the disease

X. Wei1, J. C. Lin2, S. C. J. Parker1, T. D. Prickett1, V. Walia1, J. K. Teer1, S. A. Rosenberg3, E. H. Margulies1, Y. Samuels1

1 National Institutes of Health, National Human Genome Research Institute, Bethesda, MD, USA; 2Ludwig Center for Cancer Genetics and Therapeutics and Howard Hughes Medical Institute at the Johns Hopkins Kimmel Cancer Center, Baltimore, MD, USA; 3National Institutes of Health, National Cancer Institute, Bethesda, MD, USA

Melanoma develops through acquired mutations in cancer genes. To comprehensively identify genetic alterations in melanoma we are using whole exome capture and whole genome sequencing of melanoma samples. We performed whole-genome sequencing and analysis of a low-passage cell culture of a metastatic melanoma tumor sample, the metastatic tumor from which it was derived, and the patient-matched normal genomes. Using the Illumina GAIIx and HiSeq 2000 platforms, we generated genome builds with at least 95% of the bases covered 10× or greater in each sample. When comparing somatic mutations identified in the cell culture and tissue genomes, we observe concordance at the majority of single nucleotide variants (Nature, under review).

To systematically survey mutations in melanoma coding regions in an unbiased fashion, we performed whole-exome sequencing of 14 matched normal and metastatic tumor DNAs. This allowed us to discover that (i) TRRAP, a histone acetyltransferase subunit, harbored a recurrent mutation and that (ii) GRIN2A, which encodes a glutamate receptor subunit, is mutated in 25% of melanoma samples (Nat. Genet. 2011, 43:442) and (iii) Glutamate is known to activate two different types of receptors: ionotropic glutamate receptors (iGluRs) and metabotropic glutamate receptors (mGluRs). iGluRs, to which GRIN2A belongs, are ligand-gated ion channels that allow cations such as calcium and potassium to pass through the plasma membrane of the cell after the binding of glutamate to the receptors. Intriguingly, a separate study in which we genetically analyzed the G-protein-coupled receptor family in melanoma identified that a metabotropic glutamate receptor (mGluR), GRM3, which is also activated by glutamate, is highly mutated in melanoma (Nat. Genet. 2011, in revision). The functional interplay between GRIN2A and GRM3 in melanoma will be discussed and preliminary data presented. Further investigation of the glutamate pathway in melanoma as well as development of glutamate pathway inhibitors is therefore warranted.


Induction of primary brain melanoma by congenital expression of oncogenic NRAS in mouse melanocytes

M. Pedersen, A. Viros, B. Sanchez-Laorden, R. Marais

Signal Transduction Team, The Institute of Cancer Research, London, UK

We previously reported that oncogenic BRAF (BRAFV600E) drives melanomagenesis with a median latency of ~12.5 months when expressed in mature mouse melanocytes. We therefore investigated whether oncogenic NRAS (NRASG12D) could also induce melanoma in mice. Surprisingly, when NRASG12D was expressed in mature mouse melanocytes, we did not observe the induction of cutaneous melanoma. However, we note that a high proportion of truly congenital nevi in humans harbour mutations in NRAS, so we investigated how NRASG12D expression in developing mouse melanocytes affected melanomagenesis. Mice were engineered to express NRASG12D from ~embryonic day 11.5 (E11.5). We did not observe any developmental abnormalities in these mice, although their skin darkened visibly within 4 weeks of birth. Despite this, the mice did not develop cutaneous melanoma, but with a median age of 4.5 months the mice rapidly developed neurological symptoms (motor dysfunction, hyperreactivity to normal stimuli, akathisia) that required their sacrifice. Macroscopic examination of the brains revealed heavily pigmented and thickened leptomeninges as well as pigmented tumour masses invading the brain parenchyma, which suggested leptomeningeal melanomatosis. The tumours stained positively for S100 and HMB45/MelanA (melanocytic markers), but were negative for GFAP (a glial cell marker). The tumour cells displayed high proliferative indices and stained positive for ERK phosphorylation. Cell lines derived from the tumours were sensitive to both MEK and AKT inhibitors. These data show that congenital expression of oncogenic NRAS in melanocytes does not drive the development of cutaneous melanoma, but predisposes the mice to primary brain melanoma.


Abrogation of oncogene-induced senescence by PI3K activation contributes to human melanomagenesis

L. C. W. Vredeveld1*, P. A. Possik1*, C. Michaloglou1, M. A. Smit1, H. M. Horlings2, A. Ajouaou2, P. C. Kortman3, W. J. Mooi3, D. S. Peeper1

1 Division of Molecular Genetics, The Netherlands Cancer Institute, Amsterdam, The Netherlands; 2Department of Pathology, The Netherlands Cancer Institute, Amsterdam, The Netherlands; 3Department of Pathology, Vrije University Medical Center, Amsterdam, The Netherlands

Human melanocytic nevi commonly harbor activating BRAF or NRAS mutations. Although considered to be precursors of melanoma, relatively little is known about the mechanism underlying progression from nevus to melanoma. Activated BRAF and NRAS initially act mitogenically, but eventually oncogene-induced senescence (OIS) ensues, suggesting that abrogation of OIS acts as a rate-limiting event in melanomagenesis. Favoring this model, nevi and melanomas are commonly histologically associated and melanoma can emerge within a nevus. In addition to the activation of the ERK pathway, another common genetic event in melanoma is PI3K activation. This can be achieved by loss of PTEN expression, AKT3 overexpression or mutation, or, to a lesser extent, by PIK3CA mutations. Interestingly, 20% of melanomas show concurrent mutation in BRAF and diminished expression of PTEN.

We have found that in human fibroblasts and melanocytes, activation of the PI3K pathway by ectopic expression of PIK3CA or AKT, or depletion of PTEN, abrogates BRAFE600-induced senescence. Correspondingly, in a series of contiguous human nevus-melanoma specimens, we observed a decrease in PTEN and/or an increase in AKT3 in the melanoma relative to the adjacent nevus in >50% of the cases. In several of these, laser microdissection-guided genetic analysis revealed identical mutations in BRAF or NRAS in the both the nevus and the contiguous melanoma, which is in support of a nevus-to-melanoma progression model. Finally, inhibition of the PI3K pathway in melanoma cell lines induced cell cycle arrest and enhanced the effect of PLX 4072, a BRAFE600 inhibitor, on cell survival.

These findings indicate that PI3K pathway activation serves as a rate-limiting event in melanoma progression, acting at least in part by abrogating OIS. This provides an explanation for the frequent co-occurrence of mutations in BRAF and in the PI3K pathway in melanoma. Furthermore, our data suggests that these genetic alterations can be exploited therapeutically.


Dual role of apoptosis-associated speck-like protein containing a CARD (ASC) in tumorigenesis of human melanoma

W. Liu1, J. H. Dunn1, D. A. Norris1, C. A. Dinarello2, M. Fujita1

1 Department of Dermatology, University of Colorado School of Medicine, Aurora, CO, USA; 2Department of Medicine, University of Colorado School of Medicine, Aurora, CO, USA

Apoptosis-associated speck-like protein containing a CARD (ASC) was originally named due to its ability to induce apoptosis in certain tumor cell lines. However, more recently ASC has been shown to mediate inflammatory signaling by directly activating caspase-1. While ASC has been examined in a variety of immune cells and autoinflammatory diseases, its role in cancer biology has not been well explored. Here we examined the role of ASC in human melanoma cells derived from primary stage and metastatic stage. Consistent with the previous reports, ASC protein expression was high in primary melanoma cells and was decreased in metastatic melanoma cells. However, overexpressing ASC by transfection in metastatic melanoma induced little change in cell viability whereas silencing ASC with short hairpin RNA decreased in vitro cell viability and in vivo tumorigenesis in metastatic melanoma. In primary melanoma, on the contrary, silencing ASC enhanced cell viability in vitro and tumorigenesis in vivo. Immunoprecipitation revealed interaction of ASC with NLRP3 and IKKα/β in both primary and metastatic melanoma cells. Consistent with this, ASC knockdown inhibited inflammasome-mediated caspase-1 dependent IL-1β secretion in primary and metastatic melanoma cells. However, ASC knockdown reduced phosphorylated IKKα/β expression and NF-κB activity in metastatic melanoma cells whereas it enhanced phosphorylated IKKα/β expression and NF-κB activity in early stage melanoma cells. These findings implicate the dual role of ASC in tumorigenesis. Despite relatively low levels of expression in metastatic melanoma, ASC induces tumorigenic pathways, most notably by activating caspase-1-dependent IL-1β secretion and enhancing autoinflammatory NF-κB activity in metastatic melanoma. On the other hand, high levels of ASC expression inhibit NF-κB activity through phosphorylation of IKKα/β in primary melanoma. To our knowledge, this is the first report to examine expression-dependent and/or stage-dependent roles of ASC in cancer cells.


Squamous cell tumors from RAF inhibitor-treated patients have a distinct mutational profile supporting a mechanism of therapy-induced tumorigenesis in RAS-primed cells

D. Kee1,2*, P. A. Oberholzer3,4*, P. Dziunycz5, A. Sucker6, N. Kamsukom7, R. Jones4, C. Roden4, C. Chalk3, K. Ardlie3, E. Palescandolo4, A. Piris8, L. E. MacConaill3,4, C. Robert7, G. F. L. Hofbauer5, G. A. McArthur1,2, D. Schadendorf6, L. A. Garraway3,4

1 Peter MacCallum Cancer Center, East Melbourne, Australia; 2University of Melbourne, Parkville, Australia; 3Broad Institute of M. I. T. and Harvard, Cambridge, MA, USA; 4Dana-Faber Cancer Institute, Boston, MA, USA; 5University Hospital Zürich, Zürich, Switzerland; 6University Hospital Essen, Essen, Germany; 7Gustave Roussy Institute, Villejuif Cadex, France; 8Massacheusetts General Hospital, Boston, MA, USA

Selective RAF-inhibitors are highly active in BRAFV600E mutant melanoma but are associated with keratoacanthoma (KA) and cutaneous squamous cell carcinoma (cSCC) formation. No extra-cutaneous tumors have yet been reported, but their potential to promote secondary malignancies is of major concern. Preclinical models suggest the potential for a paradoxical proliferative interaction between RAF-inhibitors and RAS-activated cells. We performed selective genotyping on these tumors in search of predisposing mutations. Four international centers contributed 237 KA/SCC FFPE tumor samples from RAF-inhibitor treated (vemurafanib or sorafenib, n = 19), immunosuppressed (n = 53), or untreated patients (n = 165). Samples were profiled for 396 known somatic mutations across 33 cancer genes using the mass-spectrometric genotyping platform (OncoMap 3).

Forty-four mutations were detected from 38 samples. Previously described in cutaneous squamous cell tumors were mutations in TP53, CDKN2A, HRAS and KRAS. Mutations in PIK3CA, MYC, FGFR3 and VHL were identified for the first time. Tumors from the RAF-inhibitor cohort were enriched for RAS mutations (21.1% versus 3.2%, P < 0.01) although overall mutation rates were similar between groups (RAF-inhibitor 21.1%, immunosuppressed 18.9%, and spontaneous 17.6%, P = ns). Tumor histology (KA versus cSCC), site (head and neck versus other), patient age (≤70, >70) and gender had no significant impact on mutation rate or type.

KA/SCCs from RAF-inhibitor treated patients are selectively enriched for RAS mutations supporting a clinically significant proliferative interaction between RAF-inhibitors and RAS-activated cells. RAS mutations are frequent early events in many extra-cutaneous premalignant lesions raising concerns of non-skin secondary malignancies. However, the proposed underlying mechanism also suggests that co-targeting MEK may prevent formation of these secondary tumors.


Therapeutic modulation of autophagy overcomes resistance to BRAF inhibition in BRAF mutant melanoma

X.-H. Ma1, S. Piao1, J. Villanueva2, G. Karakousis3, L. M. Shuchter1,4, J. Lynch1, M. Herlyn2, M. A. Davies4, R. K. Amaravadi1,4

1 Department of Medicine, University of Pennsylvania; 2The Wistar Institute; 3MD Anderson Cancer Center; 4Abramson Cancer Center, University of Pennsylvania

The role of autophagy in BRAF inhibitor (BRAFi) resistance is unclear. In serial tumor biopsies from BRAFV600E melanoma patients treated with vemurafenib, we found a 2–6 fold therapy-associated accumulation of autophagic vesicles (AV), and changes in autophagy markers ATG5 and p62 that supported robust autophagy induction. Elevated autophagy was also present in biopsies obtained at progression. The BRAFi PLX4720 (PLX) induced autophagy in PLX-sensitive (PLX-S) and PLX-resistant (PLX-R) BRAFV600E melanoma cell lines. Autophagy inhibition with hydroxychloroquine (HCQ) significantly enhanced PLX-induced cell death in PLX-R cells in 2D and 3D culture. Basal autophagy was higher and PLX-induced autophagy was blunted in two cell lines with acquired resistance (PLX-AR) to BRAFi compared to their BRAFi-sensitive parental lines. To maximize autophagy induction and sensitize PLX-R cells to HCQ, concurrent targeting of BRAF and mTOR was investigated. Separately, PLX or rapamycin treatment both induced early dephosphorylation and activation of the key autophagy regulator ULK1. While PLX treatment did not increase phospho-AMPK levels at either early or late timepoints, PLX treatment resulted in increased levels of ATF3 implicating PLX-induced ER stress as the mechanism of autophagy induction. Treatment of PLX-R cell lines with PLX, rapamycin, and HCQ lead to maximal activation of ULK1, accumulation of ineffective AV, and apoptosis compared to any single or doublet combination tested. The enhanced cytotoxicity of this triple drug combination to PLX-R and PLX-AR cell lines was confirmed by MTT and clonogenic assays. In conclusion, maximizing autophagy induction by combining BRAFi and mTORi may sensitize resistant cells to autophagy inhibition and prevent or delay recurrence. This study demonstrates for the first time that autophagy is a targetable resistance mechanism to BRAFi therapy, and provides the rationale for launching a phase I trial of a triple drug combination targeting BRAF, mTOR, and autophagy in BRAF mutant melanoma.


BRAF drugs accelerate the rate of growth of keratoacanthoma and squamous cell carcinoma

A. Viros1 C. Milagre1, R. Hayward1, F. Geyer2, N. Dhomen1, I. Niculescu-Duvaz3, A. Zambon3, D. Niculescu-Duvaz3, M. Martin1, C. J. Springer3, J. S. Reis-Filho2, R. Marais1

1 Signal Transduction Team, The Institute of Cancer Research, London, UK; 2Breakthrough Breast Cancer Research, The Institute of Cancer Research Centre, London, UK; 3Cancer Research UK Centre for Cancer Therapeutics The Institute of Cancer Research, Belmont, Sutton, Surrey, UK

The protein kinase BRAF, a component of the RAS/RAF/MEK/ERK signalling pathway, is mutated in about half of human melanomas and is a validated therapeutic target. BRAF-targeting drugs such as PLX4032 have been shown to lead to dramatic responses in melanoma patients with tumours expressing mutant BRAF, but not in patients with tumours expressing wild-type BRAF. Unexpectedly however, PLX4032 induces the development of keratoacanthomas and squamous cell carcinomas in 15–31% of patients. It has also recently been shown that in the presence of active or oncogenic RAS, BRAF drugs drive paradoxical MEK/ERK pathway activation through wild-type BRAF or the closely related protein CRAF. To investigate keratoacanthoma and squamous cell carcinoma induction by BRAF drugs, we used a two-stage skin carcinogenesis model driven by 7,12-dimethylbenz[α]anthracene (DMBA) and 12-O-tetradecanoylphorbol-13-acetate (TPA). We show that PLX4720, a close analogue of PLX4032, accelerated the development of keratoacanthomas and squamous cell carcinomas induced by DMBA/TPA. Importantly, PLX4720 did not cooperate with DMBA to induce tumour development and did not alter the histopathological features of the lesions induced by DMBA/TPA. The tumours that arose in the presence of PLX4720 carried oncogenic mutations in Hras, their initiation was blocked when MEK was inhibited, but MEK inhibitors did not block the growth of the established tumours. Notably, topical administration of the anti-proliferative agent 5-Fluorouracil (5-FU) elicited complete remissions of these tumours. We conclude that PLX4720 is not a tumour promoter per se, but rather that within the microenvironment induced by TPA it provides additional selective pressure that accelerates the growth of pre-existing mutant HRAS non-melanoma skin lesions through paradoxical activation of the MEK/ERK pathway.


BREAK-2: a phase IIA trial of the selective BRAF kinase inhibitor GSK2118436 in patients with BRAF mutation-positive (V600E/K) metastatic melanoma

U. Trefzer1, D. Minor2, A. Ribas3, C. Lebbe4, A. Siegfried5, N. Arya5, M. Guckert5, D. Schadendorf6, R. Kefford7, J-J. Grob8, O. Hamid9, R. Amaravadi10, E. Simeone11, T. Wilhelm1, K. Kim12, V. Goodman5, P. A. Ascierto11

1 Department of Dermatology, Charité– Universitätsmedizin Berlin, Berlin, Germany; 2California Pacific Center for Melanoma Research and Treatment, San Francisco, CA, USA; 3The Jonsson Comprehensive Cancer Center, University of California, Los Angeles, CA, USA; 4Dermatology Department, Policlinique de Dermatologie, Hôpital Saint Louis, Paris, France; 5GlaxoSmithKline Oncology, Collegeville, PA, USA; 6Department of Dermatology, University Hospital Essen, Essen, Germany; 7Melanoma Institute Australia, University of Sydney, Westmead Institute of Cancer Research, Westmead Millennium Institute, and Department of Medical Oncology, Sydney, Australia; 8Service de Dermatologie, Centre Hospitalo-Universitaire Sainte-Marguerite, Marseille, France; 9Experimental Therapeutics/Immunotherapy, The Angeles Clinic and Research Institute, Los Angeles, CA, USA; 10Department of Medicine, Abramson Cancer Center University of Pennsylvania, Philadelphia, PA, USA; 11Unit of Medical Oncology and Innovative Therapy Istituto Nazionale Tumori Fondazione ‘G. Pascale’, Napoli, Italy; 12Department of Melanoma Medical Oncology, The U. T. MD Anderson Cancer Center, Houston, TX, USA

BRAF mutations, identified in around 50% of cutaneous melanoma cases lead to constitutive activation of the MAP kinase pathway, making BRAF an attractive therapeutic target. Preclinical analyses of the BRAF inhibitor GSK2118436 in tumor cell lines and BRAF-mutant xenograft models have demonstrated biomarker suppression, antiproliferative activity and >100-fold selectivity for mutant over wild-type BRAF cell lines. In a first-time-in-human trial (NCT00880321), GSK2118436 was clinically active in patients with BRAF mutation-positive solid tumors that were unresponsive to standard therapy (77% unconfirmed response rate in patients with V600E mutant melanoma). BREAK-2 (NCT01153763) is an open-label, single-arm, Phase II study assessing the efficacy, safety and tolerability of GSK2118436 in patients with BRAF mutation-positive (V600E/K), cutaneous metastatic melanoma (Stage IV). Eligibility criteria included central laboratory confirmation of BRAF status, an ECOG performance status of 0–1, the absence of brain metastases and no prior treatment with BRAF or MEK inhibitors. GSK2118436 (150 mg) was administered orally, twice daily until disease progression, death or the occurrence of an unacceptable adverse event (AE). The primary endpoint was overall response rate (ORR), defined as the proportion of patients with a confirmed complete or partial response by investigator assessment. Main secondary endpoints included progression-free survival, duration of response, safety/tolerability and overall survival. In total, 92 patients (median age 55.5 [range 22–83] yr; 53% male; 83% V600E, 63% M1c-stage disease; 15% treatment-naïve for metastatic disease) were enrolled. The total incidence of AEs was 93%; the majority (57%) were Grade 1–2. Common AEs were arthralgia (33%), hyperkeratosis (27%), pyrexia (24%), fatigue (22%), headache (21%) and nausea (20%). There were eight cases of cutaneous squamous cell carcinoma, none of which required treatment withdrawal. The final ORR and PFS analysis will be presented.


Vemurafenib improves overall survival compared to dacarbazine in advanced BRAFV600E-mutated melanoma: an update from the phase III randomized, open-label, multicenter BRIM3 trial

A. Hauschild1, G. A. McArthur2, C. Robert3, J. B. Haanen4, P. Ascierto5, J. Larkin6, R. Dummer7, C. Garbe8, A. Testori9, M. Maio10, D. Hogg11, P. Lorigan12, C. Lebbe13, T. Jouary14, D. Schadendorf15, A. Ribas16, S. J. O’Day17, J. A. Sosman18, J. Kirkwood19, A. Eggermont3, B. Dreno20, K. Nolop21, J. Li22, B. Nelson23, A. Joe22, R. J. Lee22, K. T. Flaherty24, P. B. Chapman25

1 Universitaetsklinikum Schleswig-Holstein, Kiel Schleswig-Holstein, Germany; 2Department of Medical Oncology, Peter MacCallum Cancer Centre, Melbourne, Australia; 3Cancer Institute Gustave Roussy, Villejuif, France; 4The Netherlands Cancer Institute, Antoni van Leeuwenhoek Hospital, Amsterdam, the Netherlands; 5Instituto Nazionale Tumori Pascale, Naples, Italy; 6Department of Medical Oncology, Royal Marsden Hospital, London, UK; 7Department of Dermatology, University of Zurich, Zurich, Switzerland; 8University of Tübingen, Tübingen, Germany; 9Istituto Europeo di Oncologia, Milan, Italy; 10Medical Oncology and Immunotherapy, University Hospital, Siena, Italy; 11Princess Margaret Hospital and University Health Network, Toronto, ON, Canada; 12University of Manchester, Manchester, UK; 13Service de Dermatologie, Hôpital Saint Louis, Paris, France; 14Sainte Andre Hospital, Bordeaux, France; 15University Hospital Essen, Essen, Germany; 16Department of Medicine, David Geffen School of Medicine at UCLA, CA, USA; 17Angeles Clinic and Research Institute, CA, USA; 18Department of Medicine, Vanderbilt University School of Medicine, Nashville, TN, USA; 19Department of Medicine, University of Pittsburgh School of Medicine, Pittsburgh, USA; 20Department of Dermatology, Nantes University Hospital, Nantes, France; 21Plexxikon Inc., Berkeley, CA, USA; 22Hoffman La Roche, Nutley, NJ, USA; 23Genentech, South San Francisco, CA, USA; 24Massachusetts General Hospital, Boston, MA, USA; 25Memorial Sloan-Kettering Cancer Center, New York, NY, USA

Vemurafenib (Zelboraf®; PLX4032/RG7204/RO5185426) is an orally administered inhibitor of oncogenic BRAF kinase. The aim of the Phase III BRIM3 trial was to determine if vemurafenib improves overall survival (OS) in melanoma patients with the BRAFV600E mutation. Patients with previously untreated, unresectable stage IIIC or IV melanoma testing positive for the BRAFV600E mutation by the cobas® 4800 BRAF V600 Mutation Test (Roche Molecular Systems, Inc.) were randomized to vemurafenib (960 mg po bid) or dacarbazine (1000 mg/m2 IV q3w), stratified by performance status, stage, LDH, and region. At the planned OS interim analysis (50% of the 196 deaths needed for final analysis) the independent monitoring board recommended release of data because of a compelling OS hazard ratio (HR) of 0.37 (95% CI 0.26–0.55; P < 0.0001) and to permit crossover from dacarbazine to vemurafenib. An updated OS analysis with an additional 3 months’ follow-up was performed. Survival data were censored at crossover for the 50 dacarbazine patients who received vemurafenib after release of results.

Six hundred and seventy-five patients were enrolled at 104 centers worldwide (January to December 2010). In this update (March 2011), median follow-up was 6.21 months for vemurafenib (range <1–13.9) and 4.46 months for dacarbazine (range <1–11.7). The HR for OS was 0.44 (95% CI 0.33–0.59) favoring vemurafenib. The Kaplan–Meier (KM) estimate of median OS has not been reached in the vemurafenib group (95% CI 9.59–NR) and was 7.89 months (95% CI 7.26–9.63) with dacarbazine. The KM estimate of 6-month survival was 83% for vemurafenib and 63% for dacarbazine. The safety profile for vemurafenib was consistent with that reported in previous studies. In this update of the BRIM3 trial in patients with treatment-naive BRAFV600E mutation-positive melanoma, OS with vemurafenib remains superior to dacarbazine; after 6.2 months’ follow-up, median OS has not been reached.


A phase II study of the MEK1/MEK2 inhibitor GSK1120212 in metastatic BRAF-V600E or K mutant cutaneous melanoma patients previously treated with or without a BRAF inhibitor

K. B. Kim1, K. Lewis2, A. Pavlick3, J. R. Infante4, A. Ribas5, J. A. Sosman6, L. A. Fecher7, M. Millward8, G. A. McArthur9, P. Hwu1, R. Gonzalez2, P. A. Ott3, G. Long10, O. S. Gardner11, D. Ouellet11, Y. Xu11, D. J. DeMarini11, N. Le11, K. Patel11, R. Kefford10

1 The University of Texas MD Anderson Cancer Center, Houston, TX, USA; 2University of Colorado Cancer Center, Aurora, CO, USA; 3New York University Hospital, New York, NY, USA; 4Sarah Cannon Research Institute, Nashville, TN, USA; 5University of California at Los Angeles Hospital, Los Angeles, CA, USA; 6Vanderbilt University Medical Center, Nashville, TN, USA; 7University of Pennsylvania Abramson Cancer Center, Philadelphia, PA, USA; 8Sir Charles Gairdner Hospital, Perth, Australia; 9Peter MacCallum Cancer Center, Melbourne, Australia; 10Melanoma Institute Australia and Westmead Hospital, University of Sydney, Sydney, Australia; 11GlaxoSmithKline, Oncology, Philadelphia, PA, USA

GSK1120212, a selective, allosteric inhibitor of MEK1/MEK2, has previously demonstrated promising clinical activity. The objectives of this Phase II study were to determine the response rate (RR), safety, pharmacokinetics, progression-free survival (PFS), duration of response, and overall survival of GSK1120212 in metastatic BRAF-V600E or K mutant melanoma patients previously treated with a BRAF inhibitor (BRAFi; Cohort A) or with chemo- and/or immunotherapy (Cohort B). Patients with stable brain metastases were eligible. All patients were administered 2 mg GSK1120212 orally, once daily. Baseline demographics (all patients, n = 97) include median age 55 yr, stage IV (100%), M1c (74%), ECOG 1 (35%), BRAF-V600E (80%), and BRAF-V600K (13%). Rash, diarrhea, nausea, peripheral edema, and fatigue were the most frequent toxicities. Two events (2%) of central serous retinopathy were reported; both were reversible. There were no events of retinal vein occlusion. Dose reductions occurred in seven (7%) patients, most frequently for rash (3%). The median plasma trough concentrations were above the preclinical target in both cohorts. In Cohort A (n = 40), best unconfirmed response was one complete response (CR; RR = 3%) and 10 stable disease (SD; 25%). The median PFS was 1.8 months (95% CI 1.8–2.0). In Cohort B (n = 57), best unconfirmed response was two CRs (4%), 17 partial responses (PR; 30%), and 27 SD (47%) (preliminary RR = 33%; 95% CI 21.4–47.1). The disease control rate (CR + PR + SD) was 81%, and tumor reduction was observed in 64% of patients. Twenty-six patients were ongoing at data cutoff; PFS data are immature. These data confirm that the 2 mg once-daily dose of GSK1120212 is well tolerated and is clinically active in BRAF-V600E or K mutant metastatic melanoma patients previously treated with chemo- and/or immunotherapy. Minimal activity was observed in patients previously treated with a BRAFi, suggesting that BRAFi resistance mechanisms might also confer resistance to MEKi monotherapy.


Phase I/II expansion cohort of BRAF inhibitor GSK2118436 + MEK inhibitor GSK1120212 in patients with BRAF mutant metastatic melanoma who progressed on a prior BRAF inhibitor

K. Flaherty1, J. R. Infante2, G. S. Falchook3, J. Weber4, A. Daud5, O. Hamid6, R. Gonzalez7, D. Lawrence1, G. V. Long8,9, H. A. Burris2, K. B. Kim10, R. Kudchadkar4, A. Algazi5, P. Boasberg6, K. D. Lewis7, P. Sun11, A-M. Martin11, A. Allred11, S. Little11, P. Lebowitz11, K. Patel11, R. Kefford8,9

1 Massachusetts General Hospital and Dana-Farber Cancer Institute, Boston, MA, USA; 2Drug Development Unit, Sarah Cannon Research Institute, Nashville, TN, USA; 3Department of Investigational Cancer Therapeutics, Division of Cancer Medicine, U. T. MD Anderson Cancer Center, Houston, TX, USA; 4H. Lee Moffitt Cancer Center & Research Institute, Tampa, FL, USA; 5University of California San Francisco, San Francisco, CA, USA; 6Experimental Therapeutics/Immunotherapy, The Angeles Clinic and Research Institute, Los Angeles, CA, USA; 7University of Colorado Cancer Center, Aurora, CO, USA; 8Melanoma Institute Australia, and University of Sydney, NSW, Australia; 9Westmead Institute for Cancer Research, Westmead Millennium Institute, and Department of Medical Oncology, Westmead Hospital, Sydney, NSW, Australia; 10Department of Melanoma Medical Oncology, Division of Cancer Medicine, U. T. MD Anderson Cancer Center, Houston, TX, USA; 11GlaxoSmithKline Research and Development, Philadelphia, Pennsylvania and Research Triangle Park, NC, USA

BRAF inhibitors (BRAFi) have made significant advances in the treatment of BRAF-mutant metastatic melanoma, as evidenced by improvement in response rate and progression-free survival, although resistance to therapy can rapidly develop. Emerging clinical and preclinical evidence suggests that the combination of BRAF and MEK inhibition may be effective in delaying the emergence of resistance as well as preventing formation of BRAFi-induced proliferative skin lesions. To test these hypotheses, a Phase I/II study evaluating the combination of GSK2118436 (BRAFi) and GSK1120212 (MEK inhibitor [MEKi]) was initiated, including an expansion cohort of 24 patients who progressed following prior treatment with BRAFi at the recommended dose of 150 mg twice-daily GSK2118436 and 2 mg once-daily GSK1120212. Data for 18 patients are available at data cut-off: median age 46.5 yr, 56% male, ECOG performance status 0 (56%) or 1 (39%). Prior BRAFi included GSK2118436 (44%), vemurafenib (50%), and XL281 (6%). 72% of patients received prior BRAFi for ≥6 months, with median treatment duration of 8 months. For the 18 evaluable patients, responses to BRAFi/MEKi combination therapy were three (17%) partial response, nine (50%) stable disease, and six (33%) progressive disease. Minor responses (>10% to <30% tumor shrinkage) were observed in 7/9 patients (78%) with stable disease. The disease control rate was 67%. Clinical benefit was observed regardless of which prior BRAFi patients received. The most frequent treatment-related adverse events in patients who received any combination doses (n = 109) include pyrexia (31%), chills (22%), nausea (20%), diarrhea (17%), fatigue (16%), rash (13%), and vomiting (10%); cutaneous squamous cell carcinoma was reported in <1%. Based on responses observed in a subset of patients who progressed on prior BRAFi monotherapy, dual MAPK blockade may abrogate certain mechanisms of emergent BRAFi resistance. Updated safety and efficacy and preliminary biomarker data will be available at the time of meeting.


Benchmarks for evaluating phase II clinical trials in stage IV melanoma: the recent SWOG experience

V. Sondak1, M. Othus2, L. Flaherty3, B. Redman4, J. Moon2, K. Margolin5, C. Lao4, W. Carson6, A. Ribas7

1 Moffitt Cancer Center, Tampa, FL, USA; 2Fred Hutchinson Cancer Research Center and Southwest Oncology Group Statistical Center, Seattle, WA, USA; 3Karmanos Cancer Center, Detroit, MI, USA; 4University of Michigan, Ann Arbor, MI, USA; 5University of Washington, Seattle, WA, USA; 6Ohio State University, Columbus, OH, USA; 7University of California Los Angeles, Los Angeles, CA, USA

A metaanalysis involving 2100 stage IV melanoma patients from 70 phase II trial arms (Korn et al., JCO 2008;26:527–534) established benchmarks for 6-month progression-free survival (PFS6 = 15%) and 12-month overall survival (OS12 = 30%) that have been widely used for subsequent studies. However, those trials accrued patients between 1975 and 2005, and that metaanalysis did not incorporate serum lactate dehydrogenase (LDH) level as a prognostic factor. The relative importance and stability of PFS6 versus OS12 have also been questioned. To address these issues, we analyzed SWOG studies since 2005 and compared the results to the Korn data.

Five arms from four SWOG phase II studies involved 264 patients (26–68 patients per arm). Twenty-five patients with stage IV ocular melanoma on a trial of carboplatin-paclitaxel-sorafenib were analyzed separately. Four patients on two trials each were included in the analysis of only the earlier trial. Prognostic factors evaluated were age (<46, 46–60, >60), gender, prior treatment, performance status (0 versus ¡1; only one patient had PS2); AJCC stage; and LDH (normal versus >IULN). All studies (except the ocular trial) excluded patients with known brain metastases.

For the cutaneous melanoma trials overall, PFS6 = 14% (range 5–18%) and OS12 = 28% (19–35%). Patients on the ocular trial fared slightly but not significantly better (PFS6 = 28%; OS12 = 40%). All arms fell within the 95% confidence intervals of the Korn metaanalysis. In contrast to Korn, we found no significant trial-arm variability for PFS6 or OS12 after adjusting for all prognostic factors. On multivariate analysis, PS1 and increased LDH were significantly associated with shorter PFS and OS, while patients <46 yr old had significantly shorter PFS.

In conclusion, both benchmarks were consistent with the Korn data, with minimal between-arm variability. Both these benchmarks are appropriate for evaluating the results of modern phase II cooperative group trials in stage IV cutaneous melanoma, and potentially ocular melanoma as well.


Veliparib (ABT-888) plus temozolomide versus temozolomide alone: efficacy and safety in patients with metastatic melanoma in a randomized, double-blind, placebo-controlled trial

M. Middleton1, P. Friedlander2, O. Hamid3, A. Daud4, R. Plummer5, J. Richards6, N. Falotico7, R. Schuster7, B. Chyla7, E. McKeegan7, N. Mostafa7, X. Zhou7, J. Qian7, Y. Luo7, V. Giranda7, G. McArthur8

1 University of Oxford Department of Oncology, Churchill Hospital, Oxford, UK; 2The Mount Sinai Medical Center, New York, NY, USA; 3Experimental Therapeutics/Immunotherapy, The Angeles Clinic and Research Institute, Los Angeles, CA, USA; 4University of California, San Francisco, CA, USA; 5Newcastle University, Northern Institute for Cancer Research, Newcastle upon Tyne, UK; 6Oncology Specialists, SC, Park Ridge, IL, USA; 7Abbott Laboratories, Abbott Park, IL, USA; 8Divisions of Cancer Medicine/Cancer Research, Peter MacCallum Cancer Centre, Vic., Australia

The novel, orally bioavailable, small molecule, veliparib, inhibits PARP-1 (Ki 5 nM) and PARP-2 (Ki 3 nM). Preclinically, veliparib enhances antitumor activities of multiple cytotoxic therapies, including temozolomide (TMZ). This multicenter, double-blinded, placebo-controlled trial evaluated efficacy of veliparib + TMZ versus placebo + TMZ in prolonging PFS. Patients with unresectable Stage III/IV metastatic melanoma were randomized 1:1:1 to placebo BID + TMZ, veliparib 20 mg BID + TMZ, or veliparib 40 mg BID + TMZ. Efficacy endpoints included PFS, OS, ORR (RECIST). Assessments included tumor response/progression every 8 weeks (RECIST 1.0, reviewed centrally); safety (NCI-CTCAE v3.0); plasma samples for veliparib pharmacokinetics; and biopsies for potential biomarker associations with response.

346 patients were randomized to one of three treatments.

Neither veliparib group was statistically significantly better versus placebo in PFS, OS or, ORR, although veliparib was numerically superior for PFS and ORR. Toxicities in veliparib groups were as expected for TMZ with increased frequency of thrombocytopenia, neutropenia, leukopenia. BRAF and RAS mutation status had no impact on PFS or OS; patients retaining p16-protein expression in tumor tissue demonstrated treatment-specific increases in PFS and OS. Patients with lower expression of DNA-repair protein ERCC1 (tumor tissue and CTC) treated with veliparib had longer PFS. Other biomarkers are being evaluated. Veliparib exposures (Cmax, AUC) were dose proportional. Veliparib co-administration with TMZ appeared not to affect TMZ pharmacokinetics. Numerical increases in PFS (20 and 40 mg) and OS (40 mg) in veliparib-treated patients were not statistically significant. No new toxicity signals were identified.

 Veliparib BIDPlaceboP-valuea,b
20 mg + TMZ n = 11640 mg + TMZ n = 115BID + TMZ n = 11520 mg + TMZ40 mg + TMZ
  1. aVeliparib versus placebo. bStratified log-rank test for PFS, OS; stratified CMH test for ORR. cMedian, days (95% CI).

PFSc113 (92–168)110 (57–125)60 (57–111)P = 0.071P = 0.233
OSc327 (274–399)412 (346–483)390 (299–436)P = 0.995P = 0.162
ORR, % = 0.372P = 0.598
HR for OSa1.0090.790   


Molecular testing for BRAF V600 mutations in the Phase III trial of the selective BRAF inhibitor vemurafenib in metastatic melanoma: a comparison of the cobas® 4800 BRAF V600E mutation test and Sanger sequencing

J. Rueschoff1, K. Bloom2, S. Anderson3, C. Meldrum4, D. U. Vallera5, R. Schilling6, J. R.-J. Lee6, R. Langland6, H. Halait6, M. C. Dugan6, H. J. Lawrence6

1 Targos Molecular Pathology, Germany; 2Clarient, Aliso Viejo, CA, USA; 3Laboratory Corporation of America, Center for Molecular Biology and Pathology, Research Triangle Park, NC, USA; 4Peter MacCallum Cancer Centre, Australia; 5Quintiles, Westmont, IL, USA; 6Roche Molecular Systems, Pleasanton, CA, USA

The BRAF inhibitor vemurafenib (VEM) and its PCR-based companion diagnostic (cobas® 4800 BRAF V600 Mutation Test) were recently approved by the FDA for the treatment of patients with BRAFV600E-mutant metastatic melanoma, based on the improved overall survival observed in the global Phase III randomized trial of VEM versus dacarbazine. We report on the analytic performance of the cobas® test as compared with 2× bidirectional Sanger sequencing and quantitative massively parallel pyrosequencing (454) in a cohort of samples from the trial.

Specimens from 498 consecutive eligible screened patients were used to evaluate the positive (PPA) and negative (NPA) percent agreement of the cobas® test and Sanger sequencing for the detection of V600E (1799 T>A) mutations. Samples with discordant cobas® test and Sanger results were subjected to 454 sequencing. Of the 498 specimens, 2 (0.4%) had invalid cobas® results and 47 (9.4%) had invalid Sanger results. An agreement analysis between the cobas® and Sanger results was performed using the remaining 449 specimens. PPA with Sanger sequencing was 97% (216/222), and NPA was 85% (192/227). Thirty-five mutations were detected by cobas® which were not identified as V600E mutations by Sanger. Eight of these were wild type, 25 were V600K and two were other V600 mutations. 454 sequencing detected a V600E mutation in 7/8 specimens which were wild-type by Sanger. Six specimens were identified as wild-type by cobas® but V600E by Sanger; in 5/6 cases, 454 result was wild-type. The cobas® test detected 65.8% (25/38) of V600K specimens in the trial.

The cobas® test (i) had a lower failure rate than Sanger; (ii) was more sensitive in the detection of V600E mutations than Sanger, (iii) and detected a majority of V600K mutations. Robust, rapid and accurate molecular testing was achieved in this large international trial.


Tasisulam sodium versus paclitaxel as second-line treatment in patients with metastatic melanoma: a randomized phase III study

O. Hamid1, C. Garbe2, P. Wolter3, J. Richards4, M. Maio5, I. Conti6, C. Kaiser6, S. Saxman6, J. J. Grob7

1 The Angeles Clinic & Research Institute, Los Angeles, CA, USA; 2Center for Dermatooncology of the University of Tuebingen, Tuebingen, Germany; 3University Hospitals Leuven, Belgium; 4Oncology Specialists, Park Ridge, IL, USA; 5Medical Oncology and Immunotherapy, University Hospital of Siena, Siena, Italy; 6Eli Lilly and Company, Indianapolis, IN, USA; 7Timone Hospital, Marseilles, Marseilles, France

Tasisulam, a novel antineoplastic agent, has shown activity in second-line metastatic melanoma patients in a previous phase II study. The dose-limiting toxicity was myelosuppression. The primary objective of this randomized, open-label, phase III study was to compare the overall survival (OS) of tasisulam versus paclitaxel following first-line dacarbazine- or temozolomide-based chemotherapy. The accrual goal was 800 patients. Tasisulam was dosed by an albumin-based algorithm targeting a hypothesized therapeutic exposure range on day 1, and paclitaxel, 80 mg/m2, on days 1, 8, and 15, of a 28-day cycle. A total of 336 patients were randomized and 325 received treatment (164 tasisulam, 161 paclitaxel). Baseline characteristics were balanced between treatment arms. A preliminary safety review indicated an imbalance in possibly drug-related deaths between the study arms due to complications of myelosuppression, particularly fever and neutropenia, infection, and/or sepsis, among patients on the tasisulam arm. Because of this observed imbalance, the study was put on full clinical hold. The subsequent analysis of progression-free survival in all patients enrolled prior to the clinical hold did not meet the futility rule, indicating that tasisulam would unlikely be superior to paclitaxel if all the patients were followed until progression and/or death. Therefore, Lilly permanently closed the study. The most common possibly drug-related ≥grade 3 toxicities for tasisulam were thrombocytopenia, 18.9%, and neutropenia, 8.5%, versus 0.6% and 6.2%, respectively, for paclitaxel. The median follow-up time was 5.3 months. Median OS for randomized patients was not significantly different between arms (6.77 months for tasisulam and 9.36 months for paclitaxel; logrank P = 0.121). Overall, 13 (7.9%) patients died on study or within 30 days of treatment discontinuation from possibly drug-related events, all on the tasisulam arm. Five of these deaths were due to sepsis. Analyses are in progress to determine why complications of myelosuppression occurred at a higher frequency than anticipated.

Abstracts of the 8th International Congress of The Society for Melanoma Research November 9–11, 2011 Proffered Abstracts Selected for Poster Presentation

  1. Top of page
  2. 2011 International Melanoma Congress Advancement Through Collaboration joins the following four annual meetings in 2011:
  3. Congress Agenda
  4. Congress Proceedings Abstracts of the 8th International Congress of the Society for Melanoma Research November 9–11, 2011 Invited Faculty Abstracts
  5. Abstracts of the 5th Meeting of Interdisciplinary Melanoma/Skin Cancer Centres November 11–12, 2011 Invited Faculty Abstracts
  6. Abstracts of the 4th Melanoma Pathology Symposium of the International Melanoma Pathology Working Group November 13, 2011 Invited Faculty Abstracts
  7. Abstracts of the 8th International Congress of The Society for Melanoma Research November 9–11, 2011 Proffered Abstracts Selected for Oral Presentation
  8. Abstracts of the 8th International Congress of The Society for Melanoma Research November 9–11, 2011 Proffered Abstracts Selected for Poster Presentation
  9. Abstracts of the 5th International Meeting of Interdisciplinary Melanoma/Skin Cancer Centres November 11–12, 2011 Proffered Abstracts Selected for Poster Presentation


Upregulation of ERBB3/HER3 by FOXD3, an adaptive response promoting resistance to RAF/MEK-inhibitors

E. V. Abel, K. J. Basile, A. E. Aplin

Department of Cancer Biology and Kimmel Cancer Center, Thomas Jefferson University, Philadelphia, PA, USA

Targeting the B-RAF/MEK/ERK pathway has shown promise in the treatment of mutant B-RAF melanomas, but resistance to RAF/MEK-inhibitors has been observed. We have previously shown that the stem cell transcription factor FOXD3 is upregulated in mutant B-RAF melanoma cells upon RAF/MEK-inhibition, but the consequences of FOXD3 upregulation are unclear. Using microarrays and ChIP-sequencing, we identified ERBB3/HER3 as a direct target of FOXD3. FOXD3 enhanced ERBB3 expression and responsiveness to its ligand neuregulin-1 (NRG1). Interestingly, PLX4032 or AZD6244-treated cells displayed enhanced NRG1-dependent ERBB3 phosphorylation and downstream AKT activation. These effects were associated with NRG1 enhanced migration and growth of RAF/MEK-inhibitor treated cells in vitro. Phosphorylation of ERBB3 was found to be dependent on ERBB2/HER2, as NRG1/ERBB3-signaling was inhibited by ERBB2-targeting siRNA, the ERBB2-blocking antibody trastuzumab and the ERBB2-inhibitor lapatinib. Additionally, the combination of lapatinib and PLX4032 was more effective than either alone at inhibiting colony formation in the presence or absence of NRG1. Similarly, we found that combining lapatinib with PLX4720 was more effective than either drug alone at reducing tumor burden in melanoma xenografts. These results suggest that upregulation of ERBB3 by FOXD3 serves as an adaptive response to protect melanoma cells from the effects of RAF/MEK-inhibitors and promote resistance. In summary, targeting the NRG1/ERBB3 pathway in conjunction with RAF/MEK may provide a clinical benefit.


DCT and p53 induction in melanocytic cells and co-cultures upon MC1R activation

S. A. Ainger1, S. S. Wong1, J. H. Leonard2, R. A. Sturm1

1 Institute for Molecular Bioscience, University of Queensland, Queensland, Australia; 2Queensland Institute of Medical Research, Brisbane, Queensland, Australia

The melanocortin-1 receptor (MC1R) is highly polymorphic in human populations, with individuals carrying the variants R151C, R160W and D294H shown to be strongly associated with red hair, fair skin, inability to tan, and increased risk of developing skin cancers (RHC phenotype). Other low penetrant variants such as V60L, V92M and R163Q show intermediate responses to UVR when compared to wild type MC1R and the high penetrant RHC variants. Our studies have revealed that a potent MC1R agonist NDP-MSH and the cAMP activator drug forskolin increased expression of the pigmentation marker Dopachrome Tautomerase (DCT) and the dendricity of wild type MC1R melanocyte strains in co-culture with keratinocytes. Melanocytes homozygous for the R151C variant showed reduced dendricity increases and no changes in expression of DCT in response to NDP-MSH. To begin an assessment of the role DCT induction by MC1R ligand treatment plays in melanocyte cell function, siRNA knockdown experiments using melanoma cells and primary melanoblast cell monocultures have been carried out. DCT knockdown in melanoma cell lines reduced cell survival seen in MTT assays performed at 3 and 24 h, which was further reduced after 25 mJ/cm2 UVB exposure. Expression of p53 and pp53 increased after DCT siRNA treatment and UVB irradiation. A similar pattern of survival and protein expression was observed in melanoblast cell strains of WT MC1R genotype. DCT appears to affect DNA damage responses negatively via p53 in melanoma cell lines and primary melanoblast cell strains in monoculture. We intend to extend these studies of DCT function in melanocyte-keratinocyte co-cultures using a lentivirus shRNA delivery system.


Prolonged responses to vemurafenib in patients with BRAFV600-mutant melanoma with low tumor burden at baseline

R. Amaravadi1, K. Kim2, K. Flaherty3, P. Chapman4, I. Puzanov5, J. Sosman5, A. Ribas6, R. Lee7, K. Nolop8, G. McArthur9

1 Abramson Cancer Center University of Pennsylvania, PA, USA; 2The University of Texas M.D. Anderson Cancer Center, Houston, TX, USA; 3Massachusetts General Hospital Cancer Center, Boston, MA, USA; 4Memorial Sloan-Kettering Cancer Center, New York, NY, USA; 5Vanderbilt-Ingram Cancer Center, Nashville, TN, USA; 6Department of Medicine, David Geffen School of Medicine at UCLA, CA, USA; 7Hoffman La Roche, Nutley, NJ, USA; 8Plexxikon Inc., Berkeley, CA, USA; 9Department of Medical Oncology, Peter MacCallum Cancer Centre, Melbourne, Australia

While some patients treated with vemurafenib experience a transient response, others experience prolonged benefit and long-term survival. To determine clinical predictors of prolonged benefit with vemurafenib, we analyzed survival, baseline characteristics, and the impact of post-progression treatment in patients with short (<6 months) and prolonged (>12 months) PFS in all BRAFV600E melanoma Phase I patients (n = 48) treated with vemurafenib doses >240 mg bid. Clinical response was determined every 8 weeks using RECIST. Patients with progressive disease in an isolated site addressed by local therapy were allowed to continue vemurafenib post-progression if the remainder of disease was controlled, and until signs of multifocal progression occurred or clinical deterioration. Baseline characteristics of 19/48 (40%) patients who progressed early (PFS <6 months) versus the 15/48 (31%) with prolonged benefit (PFS >12 months) were: mean (SD) baseline LDH: 577 (720) versus 261 (117); % M1c (95% CI): 74 (49–91) versus 53 (27–79); mean (SD) sum of baseline target lesions: (16.8 (11.3) versus 6.5 cm (3.7); P = 0.002); % ECOG PS 1 (95% CI): 74 (49–91) versus 27 (7–55). Median KM estimate of overall survival was 16.9 months for the 48 patients, with 54% 1-yr and 45% 2-yr survival estimates. The median survival for 21/48 (44%) patients who either never progressed (3) or were treated post-progression for >30 days (18) with vemurafenib compared with those who were not treated post-progression was 27 versus 12 months, P = 0.03. These data suggest that long-term survival with vemurafenib can be achieved in patients with low baseline tumor burden. Treatment beyond progression may be associated with a survival benefit in patients with isolated disease progression. Additional analyses of the Phase II/III vemurafenib clinical trial data should be undertaken to investigate outcomes for continued vemurafenib treatment post-progression, in BRAFV600E melanoma patients.


Combined mTOR inhibition and autophagy inhibition: phase I trial of temsirolimus and hydroxychloroquine in patients with advanced solid tumors (dose escalation) and metastatic melanoma (expansion)

R. K. Amaravadi, R. Rangwala, Q. M. MCafee, L. Fecher, C. Chang, D. Torigian, A. Alavi, B. Saboury, J. Panosian, P. J. O’Dwyer, L. M. Schuchter

Abramson Cancer Center, University of Pennsylvania

A phase II trial of temsirolimus in melanoma reported a 0% stable disease (SD) rate. Therapy-induced autophagy is a key resistance mechanism that could limit clinical benefit with mTORi. In preclinical models, the autophagy inhibitor hydroxychloroquine (HCQ) enhances mTORi cytotoxicity. We conducted a 3 + 3 phase I trial of temsirolimus and HCQ in advanced solid tumors and melanoma. Patients were treated with single agent temsirolimus 25 mg IV for 1 week followed by weekly temsirolimus with daily HCQ. A 12 patient melanoma expansion would be treated at the maximal tolerated dose (MTD) or 1200 mg HCQ daily. Objectives included defining the MTD (primary), toxicity and response rates, measurement of therapy-induced autophagy and metabolic response in serial peripheral blood mononuclear cells (PBMC) and FDG-PET/CT scans, comparing the maximal standard uptake value (SUVmax) with total metabolic volumetric product (MVP). Fifteen of 29 evaluable patients had melanoma (14/15 BRAF WT). HCQ was escalated from 200 to 1200 mg daily in four cohorts. The MTD was not defined, but 1200 mg HCQ daily was the recommended phase II dose due to common and burdensome grade 2 toxicities including fatigue, anorexia, nausea, weight loss, and stomatitis. SD was achieved in 19/29 (65%) patients, including 11/15 (73%) with advanced melanoma. Significant therapy-associated autophagy inhibition was measured with HCQ 1200 mg in serial PBMC and tumor tissue and not with lower doses. A significant early therapy-associated decline in MVP and SUVmax was observed in patients with stable disease. In patients with progressive disease, MVP but not SUVmax increased significantly. In conclusion, the combination of temsirolimus and HCQ was safe and demonstrated significant biological and clinical activity in patients with metastatic melanoma. These results suggest that autophagy inhibition is a promising new therapeutic strategy in melanoma and a phase II trial of this combination is warranted.


Next-generation sequencing to search for melanoma susceptibility genes

L. G. Aoude1, K. Holohan2, Z. Z. Zhao1, V. Zismann2, M. Gartside1, S. Woods1, K. Dutton-Regester1, V. Bonazzi1, M. S. Stark1, J. M. Palmer1, H. Schmid2, J. Symmons1, S. MacGregor1, D. L. Duffy1, E. Holland2, D. C. Whiteman1, R. Kefford2, G. W. Montgomery1, N. G. Martin1, G. Mann2, J. Trent3, K. M. Brown4, N. K. Hayward1

1 Queensland Institute of Medical Research, Brisbane, Queensland, Australia; 2Westmead Institute for Cancer Research, Sydney, NSW, Australia; 3Translational Genomics Research Institute, Phoenix, AZ, USA; 4National Cancer Institute, National Institutes of Health, Gaithersburg, MD, USA

Using next-generation sequencing (NGS) technology, we sequenced a number of individuals from cutaneous melanoma families to identify novel melanoma susceptibility genes. The Illumina platform was used to sequence whole-genomes or exomes of 27 individuals from 13 Queensland families, with between one and three affected individuals sequenced in each family. Variants from NGS were aligned to the human reference genome, hg19, and then called using SAMtools and annotated using ANNOVAR. Any SNP listed in dbSNP or the 1000 Genomes Project was then removed from the analysis. From this more refined list, select variants that were found in rational melanoma candidate genes were studied further by (i) segregation analysis in the relevant family, to assess whether mutations were common to all affected family members and hence might be high penetrance mutations responsible for the increased melanoma risk in these families; and (ii) population-based case-control analysis. The latter was performed to determine if certain variants may be low penetrance risk alleles. This was done by genotyping 2372 Australian melanoma cases and 2104 unaffected controls using the Sequenom MassArray platform. To date, none of the mutations tested segregated in all affected members of the family in which it was discovered, though a number of interesting variants have been identified that are unique to the proband or a limited number of cases in certain families. These have not been found in any of the population-based melanoma cases or controls and their potential involvement in melanoma predisposition needs further investigation, possibly through functional studies. None of the more common variants tested showed significant differences in allele frequency between cases and controls, hence do not appear to contribute to melanoma susceptibility in the general population. Additional families are currently being screened and other unique variants followed up to extend this line of investigation.


Expression of isoforms of the transcription factor associated with microphthalmia (MITF) in peripheral blood of patients with malignant melanoma

N. Acosta1, N. Rangel2, A. F. Aristizabal2, S. Ramírez1

1 Group of Medical Basic Sciences Research, Faculty of Medicine, University of Rosario, Bogotá, DC Colombia, 2Group of Biomedical Research and Applied Human Genetics – GIBGA, Faculty of Medicine, University of Applied and Environmental Sciences U.D.C.A. Bogotá, DC Colombia

There are approximately 160 000 new cases of melanoma each year in the world and its mortality rate is higher than other types of skin cancer due to its high metastatic potential. The detection of micrometastases in peripheral blood by identifying tumoral biomarkers specific of melanoma cells can be a promising diagnostic tool of this type of cancer in the future. The transcription factor associated to microphthalmia (MITF) encodes a protein that is essential for the survival and development of melanocytes, and in melanocytic tumors an increase of expression and amplification of this gene has been found. MITF gene encodes 10 isoforms, being MITF-M specific of melanocyte a candidate for the detection of melanoma cells. Our aim was to determine the expression of three isoforms of MITF gene in peripheral blood of individuals with melanoma and compare it with people without any type of cancer. Blood samples from 61 patients with melanoma and from 61 controls were used for RNA extraction and analyzed by real time RT-PCR. Expression of the consensus region of MITF was detected in 58 cases and 60 controls with an increase of its expression level in the case group of 1.31 times compared with the control group, without finding statistically significant differences (P = 0.258). In all samples MITF-A and MITF-B isoforms were detected, and in the case group they presented an increase of their expression level of 1.9 and 1.8 times respectively. Statistically significant differences were found between the two groups (P = 0.002 in the two isoforms). Expression of MITF-M isoform was detected in 16 cases and in no control.

These results indicate that melanoma cells are being detected in peripheral blood and further studies are needed to confirm the proper use of gene MITF isoforms as biomarkers for the detection of this disease.


Polymorphisms of genes associated with response to UV radiation and its relation to the development of cutaneous malignant melanoma

A. F. Aristizabal1, N. Rangel1, C. Acosta2, S. Ramrez2

1 Group of Biomedical Research and Applied Human Genetics – GIBGA, Faculty of Medicine, University of Applied and Environmental Sciences U.D.C.A. Bogot, DC Colombia; 2Group of Medical Basic Sciences Research, Faculty of Medicine, University of Rosario, Bogot, DC Colombia

In Colombia, in 2007 skin cancer was the most common, with 895 new cases, of which there were 97 cases of cutaneous malignant melanoma (CMM), which ranked as the third type of skin cancer, after basal cell carcinoma (481 cases) and squamous cell carcinoma (134 cases), tumors that have very low mortality and rarely metastasize, contrary to what happens in cutaneous malignant melanoma where mortality and metastatic potential are very high. Since the increased exposure to ultraviolet radiation is considered as the environmental factor that contributes most significantly to the development of CMM, it's probably to think that XPD gene polymorphisms, involved in DNA damage repair caused by UVB and MC1R gene polymorphisms, which regulates pigmentation process, constitute allelic variants may be conferring susceptibility to the development of CMM. For this reason, the objective of this study was to determine allele and genotype frequencies of the MC1R Arg151Cys gene and XPD Lys751Gln gene polymorphisms using conventional PCR and RFLP in a Colombian population of 92 patients with cutaneous malignant melanoma and in a group control. These populations were compared to establish the association between the polymorphisms studied, clinicopathological characteristics (skin color, eye color, hair color, circumstances for which severe burns in the sun by the sun, skin type and age) and risk of developing cutaneous malignant melanoma.

There were no statistically significant differences (P < 0.05) using association tests χ2 and Fisher's exact test, between the presence of polymorphisms in XPD and MC1R genes and the development of cutaneous malignant melanoma. Likewise, there were no statistically significant differences (P < 0.05) between clinicopathological features and its association with the development of cutaneous malignant melanoma. The analysis between clinicopathological characteristics, the development of cutaneous malignant melanoma and the presence of polymorphisms also showed no significant differences.


Targeted therapies in melanoma – in vitro effects of single and combined drug treatment

A. Azimi, J. Hansson, C. Johansson, S. Egyhazi

Department of oncology & Pathology, Karolinska University hospital, Stockholm, Sweden

Malignant melanoma is one of the most rapidly increasing malignancies in Caucasian populations. So far, curative treatment is achieved only by surgical resection at an early stage, while the prognosis in cases with disseminated melanoma is poor, because existing system therapies are inefficient in the vast majority of patients. However, promising clinical trials of targeted therapies are ongoing, but it is already clear that drug resistance will be a clinical reality also for these new therapies. The aim of the study is to unravel molecular mechanisms and pathways underlying resistance to new therapies in melanoma and identification of proper drug combination(s) to overcome resistance problem. We are studying the effects of a set of MAPK and PI3K pathways inhibitors in melanoma cell lines that are wildtype or have different mutations/deletions in genes affecting these pathways (BRAF, NRAS, PTEN etc.). Cells are exposed to single drugs as well as combinations of inhibitors and/or chemotherapeutic agents. Our results have shown that there is a wide range of sensitivity to the BRAF inhibitor PLX4720 among BRAF mutated cell lines, with IC50 values varying from 0.3 to 30 μM. We are investigating the underlying mechanisms causing this variation in sensitivity to the drug. The expression of some of these genes/proteins affected by PLX4720, and their potential relevance for response to different treatments is currently tested in vitro.


Prognostic significance of tumor mitotic rate in T2 melanoma is independent of age

J. Baker1, A. Deal2, M. Meyers1, J. Frank2, K. Stitzenberg1, D. Ollila1

1 Division of Surgical Oncology, Department of Surgery, University of North Carolina, Chapel Hill, NC, USA; 2Lineberger Comprehensive Cancer Center, University of North Carolina, Chapel Hill, NC, USA

Prior work suggested that tumor mitotic rate (TMR) is an important prognostic variable in melanoma patients under the age of 45, but loses significance with increasing age. We sought to determine the association between age and the prognostic value of TMR in clinically node-negative (CNN) T2 melanoma.

A prospective IRB-approved database of cutaneous melanoma patients treated from 9 January 1997 to 3 January 2011 was used to identify patients with CNN T2 melanoma. TMR was reported in mitoses/mm2. Associations were evaluated using Fisher's Exact test.

We identified 313 CNN T2 patients. Twenty-four per cent were ≤45 yr, 37% 46–64 yr, and 39%≥65 yr. Primary site was head/neck (25%), trunk (37%), and extremity (38%). Nineteen per cent had ulceration, 11% had a positive sentinel node (SN), and 10% recurred. Forty-four per cent of patients had TMR ≥1/mm2. TMR ≥1/mm2 was significantly associated with recurrence; only 4% of those with TMR <1/mm2 developed a recurrence and 18% of those with TMR ≥1/mm2 recurred (P < 0.0001). Those with a positive SN had a recurrence risk of 26%, in SN negative patients with a TMR <1/mm2 and ≥1/mm2 recurrence risk was 2.5% and 15%. When controlling for other factors, TMR was significantly associated with recurrence in each age group (P = 0.01, 0.02, 0.03). TMR was associated with recurrence in non-ulcerated patients but not patients with ulceration (P < 0.0001 and 0.46), and SN negative but not SN positive patients (P < 0.0001 and 0.45).

TMR ≥1/mm2 is a significant risk factor for recurrence in all age groups with T2 melanoma. Although not equaling the recurrence risk as compared to SN-positive T2 patients, TMR ≥1/mm2 in SN-negative T2 patients still increases the recurrence risk 6-fold. As such, TMR appears to be an important variable that can be used to stratify follow-up regimens in SN negative T2 patients.


Antiproliferative effects of 1α-hydroxycholecalciferol in malignant melanoma in vitro and in vivo. The biological rationale for using vitamin D or active derivatives in melanoma adjuvant therapy

L. Spath1, L. Fidanza2, A. Ulivieri3, M. Carlesimo2, A. Narcisi2, B. Bucci3, M. Giubettini1, E. Luciani1, D. Pisani2, S. Sciacchitano3, A. Bartolazzi1,4 for the S. A. M. W. G. St. Andrea Melanoma Working Group

1 Pathology Research Laboratory, Sant’Andrea University Hospital, Rome, Italy; 2Dermatology Unit, Sant’Andrea University Hospital, Rome, Italy; 3Center for Experimental Research, San Pietro Fatebenefratelli Hospital, Rome, Italy; 4Molecular and Cellular Tumor Pathology Laboratory, Cancer Center Karolinska, Karolinska Hospital, Stockholm, Sweden

Malignant melanoma represents one of the most challenging issues in oncology. Early detection and surgery represent the mainstay of treatment for localized lesions but patients with high-risk melanomas (Breslow'thickness >0.75 mm) are still orphans of an effective adjuvant therapy. Unfortunately the metastatic condition is invariably associated with a dismal prognosis, despite any therapeutic effort. Recent epidemiological studies indicate that vitamin D insufficiency (10–30 ng/ml) is a common feature of cancer patients. Furthermore, preclinical data support a potential anticancer activity of vitamin D and its active metabolites in different tumor conditions.

Melanoma cell lines derived from patients with metastatic disease were used for studying the biological effects induced by vitamin D on melanoma cell growth in vitro. Experimental models of melanoma xenografts in nude mice were also considered to deeply investigate the long-term effects of a systemic vitamin D treatment on melanoma growth in vivo.

Finally, 46 melanoma patients with 25-hydroxyvitamin D insufficiency (10–30 ng/ml) were enrolled for an experimental adjuvant therapy with 1α-hydroxycholecalciferol 400–800 IU/day administered per os for at least 6–24 months. 25-hydroxyvitamin D, Calcium and PTH serum levels were monitored during follow-up.

Vitamin-D invariably impaired melanoma cell proliferation in vitro and tumor growth in vivo, driving melanoma cells towards an irreversible block in G1 or G2 (depending by the cell line). One of the vitamin D-mediated antiproliferative mechanisms involves over-expression of p27 and p21 and down-regulation of cyclin D1. Preliminary data derived from melanoma patients with vitamin D insufficiency show that 1α-hydroxycholecalciferol can be easily administered at 400–800 UI/die for at least 24 months, without significant side-effects, restoring the normal serum level of 25-hydroxyvitamin D.

These findings provide a biological rationale for using vitamin D or its active derivatives in melanoma adjuvant therapy, alone or in association with other therapeutic options.


Rab GTPases and dynamic cell cycle imaging in melanoma biology

K. A. Beaumont1, A. Anfosso1, I. Kinjyo1, O. Kanagawa2, R. A. Sturm3, J. L. Stow3, W. Weninger1,4,5, N. K. Haass1,4,5

1 Centenary Institute, Newtown, NSW, Australia; 2Research Center for Allergy and Immunology, RIKEN, Yokohama City, Japan; 3Institute for Molecular Bioscience, The University of Queensland, Brisbane, Queensland, Australia; 4Discipline of Dermatology, University of Sydney, NSW, Australia; 5Department of Dermatology, Royal Prince Alfred Hospital, Camperdown, NSW, Australia

Rab GTPases are a large family of proteins that regulate vesicular protein transport. Overexpression of Rabs and dysregulated protein trafficking causes increased proliferation and invasion in some epithelial cancers, and microarray studies have revealed expression of several Rabs correlate with melanoma cohorts of different metastatic potential. However their role in melanoma biology is poorly understood.

We show that Rab27a and Rab17 protein expression is higher in ‘proliferative’ melanoma cells lines, and that this expression correlates with levels of microphthalmia-associated transcription factor (MITF). Significantly, Rab27a upregulation was recently described as a driver of melanoma proliferation, though the underlying mechanism is unknown. Using Rab-GFP expression studies we have also shown that Rab11, which is a well-known recycling-endosome associated Rab, localises to plasma membrane protrusions and at the leading edge of migrating melanoma cells. We have developed a novel model that allows us to visualize individual cell cycle dynamics in real time in a 3D organotypic melanoma spheroid. Cells transfected with the fluorescence ubiquitination cell cycle indicator (FUCCI) plasmids appear red in G1, yellow in early S and green in S/G2/M phase with a gap in fluorescence after division. FUCCI-melanoma cells are grown as 3D spheroids and implanted into a collagen matrix in order to mimic tumour microenvironment. Using confocal microscopy we have determined that cells at the periphery of the spheroids derived from invasive (Rab17/Rab27a/MITFlow) cell lines are cycling and highly motile, invading into the collagen surrounding the spheroid, while cells in the interior of the spheroid are mostly arrested in G1 and slow moving. In contrast, proliferative (Rab17/Rab27a/MITFhigh) cell lines appear to show less invasion and a more random cell cycle progression pattern. Further studies on the effect of Rab knockdown or overexpression utilizing this model will allow us to determine the role of Rabs in melanoma proliferation or invasion.


The BRAFV600E kinase inhibitor vemurafenib induces apoptosis through a process involving induction of endoplasmic reticulum stress in melanoma cells

D. Beck1, H. Niessner1, K. Flaherty2, D. Kulms3, M. Schaller1, C. Garbe1, F. Meier1

1 Department of Dermatology, University of Tuebingen, Tuebingen, Germany; 2Cancer Center, Massachusetts General Hospital, Boston, MA, USA; 3Institute of Cell Biology and Immunology, University of Stuttgart, Stuttgart, Germany

In a previous study, we observed that the pan-RAF inhibitor sorafenib induces upregulation of endoplasmic reticulum (ER) stress-related genes and apoptosis in melanoma cells in vitro. In this study, we investigated whether vemurafenib, which selectively inhibits the BRAFV600E kinase and demonstrates potent antitumor activity in melanoma patients with the BRAFV600E mutation, induces ER stress-mediated apoptosis. The BRAFV600E kinase inhibitor vemurafenib inhibited growth, induced caspase-dependent apoptosis and upregulated the ER stress-related genes p8, CHOP, ATF4, ATF3 and TRB3 mRNA levels exclusively in BRAFV600E mutated melanoma cell lines. Apoptosis was correlated with the induction of the proapoptotic BH3-only protein Bim-particularly Bim short, which is linked to ER stress-mediated apoptosis. Western blot analysis showed that vemurafenib increases the protein levels of the ER stress marker CHOP in BRAF mutated but not in NRAS mutated melanoma cells. Treatment with vemurafenib resulted in a rapid increase in cytosolic calcium levels which is believed to reflect ER stress which can promote induction of the unfolded protein response. Furthermore, electron microscopy showed typical morphological signs of ER stress, in particular significant swelling of the endoplasmic reticulum lumen of BRAFV600E mutated melanoma cells treated with vemurafenib. siRNA inhibition of p8 diminished melanoma cell apoptosis induced by vemurafenib, overexpression of p8 significantly enforced melanoma cell apoptosis induced by vemurafenib. Furthermore, classical ER stress inducers such as thapsigargin and tunicamycin potently inhibited growth and induced apoptosis. Moreover, both thapsigargin and tunicamycin upregulated p8 and CHOP and induced apoptosis in vemurafenib-resistant melanoma cells. These data suggest that the BRAFV600E kinase inhibitor vemurafenib induces apoptosis in BRAFV600E mutated melanoma cells through upregulation of ER stress-related genes, and that melanoma cells which acquired resistance to vemurafenib may be sensitive to agents which induce ER stress-mediated apoptosis through a different mechanism.


New roles of the matrix metalloproteinase MT1-MMP in melanoma

K. Shaverdashvili, P. Wong, B. Bedogni

Department of Biochemistry, Case Western Reserve University, Cleveland, OH, USA

Melanomas are very aggressive and therapy resistant tumors that can very rapidly progress to metastasis. Widespread metastatic dissemination with relatively small (or undetectable) primary tumors is not an uncommon clinical occurrence in melanoma. One hypothesis to explain the highly aggressive nature of melanomas is that they possess an intrinsic migratory/invasive program that derives from the highly migratory neural crest cells (NCCs), embryonic precursors of melanocytes and melanoma cells. The migration process of NCC is regulated by the timely activation of a number of genes, among which is MT1-MMP (membrane type 1-MMP). MT1-MMP is a membrane associated matrix metalloproteinase that can directly cleaveextracellular matrix (ECM) components at the cellular leading edge, thus providing directionality to the migratory/invasive processes; and can enhance ECM degradation by activating soluble MMPs such as MMP2 and MMP13. By comparing the expression of various MMPs across a number of melanoma array data sets, MT1-MMP resulted consistently highly expressed. Furthermore, its expression correlates with progression from normal nevi to primary tumors and to fully metastatic cancer. Importantly, we have unraveled new functions of MT1-MMP in melanoma. MT1-MMP is not only critical for melanoma cell invasion, but it regulates cell growth and motility through the activation of a number of signalling pathways including PI3K/AKT, FAK/SRC and RhoA in a catalytically independent manner. Furthermore, MT1-MMP contributes to the reactivation of embryonic stem cell pathways by promoting the cleavage of Notch1. Together, these findings underline the complexity of MT1-MMP as a protein able to regulate multiple biological processes in melanoma cells by activating key signalling pathways involved in motility, invasion, growth and survival. In addition, they provide preliminary evidence that MT1-MMP may represent a new molecular target for melanoma therapy.


UV induces melanomagenesis in a nucleotide excision repair deficient mouse model dependent on the over-activation of endothelin-3

A. P. Benaduce, G. Grilo, D. Batista, L. Kos

Florida International University, Miami, FL, USA

Disruption of the Nucleotide Excision Repair (NER) pathway is a very significant step during ultraviolet radiation (UV)-induced melanomagenesis. Xeroderma Pigmentosum (XP) patients carry mutations in the XP complementation group of genes that play a role in NER. XP patients show high sensibility to UV light and a high propensity to developing skin cancer including melanoma. Xpa and Xpc deficient mice display phenotypes that are similar to those of XP patients but do not present with melanoma after UV exposure. In this study we aimed at developing a UV-induced mouse model of melanoma by combining deficiencies in Xpa or Xpc with the over-activation of the endothelin-3 (Edn3) pathway. The Edn3 signaling pathway has been linked to melanoma progression and its metastatic potential. Transgenic mice that express Edn3 under the control of the keratin 5 promoter (K5-Edn3) are hyperpigmented because of the accumulation of melanocytes in the skin, where they are not normally found. K5-Edn3 mice carrying targeted mutations in Xpa or Xpc were exposed to a single sub-erythemal neonatal dose of UV and monitored for melanoma development. Histomorphology, TRP1 and S100 immunostaining were used to verify primary skin tumors and metastases. Melanoma was only found in animals with the K5-Edn3 transgene. The penetrance of lesions was higher in animals that were homozygous for Xpa (67%) when compared to heterozygous (65%) or wild type (50%). Similarly, Xpc homozygous mice showed higher penetrance (100%) than heterozygous (83%) or wild type (75%) mice. Additionally, mice with disrupted NER tended to acquire their lesions at an earlier time. Animals with melanoma lesions presented enlarged and hyperpigmented lymph nodes that were diagnosed as local metastases. These results suggest that exposure to UV radiation, along with the over-activation of the Edn3 signaling pathway and disruption of the NER pathway can lead to melanomagenesis in mice.


The neuronal transcription factor Brn3a promotes melanocyte transformation and tumor growth in vivo

T. Hohenauer1, C. Berking1, D. Senft1, C. Kammerbauer1, S. Fraschka1, S. Haferkamp2, M. Irmler3, J. Beckers3,4, A. Schmidt5, S. Rothenfusser5, T. Ruzicka1, R. Besch1

1 Department of Dermatology and Allergology, Ludwig-Maximilian University, Munich, Germany; 2Department of Dermatology, University Hospital Würzburg, Würzburg, Germany; 3Helmholtz Zentrum München, German Research Center for Environmental Health, Institute of Experimental Genetics, Neuherberg, Germany; 4Technical University Munich, Center of Life and Food Sciences Weihenstephan, Freising, Germany; 5Division of Clinical Pharmacology, Department of Internal Medicine, Ludwig-Maximilian University, Munich, Germany

The neuronal transcription factor Brn3a (POU4F1) regulates cell survival in the developing nervous system. We observed that Brn3a is expressed in melanoma cells, but not in primary melanocytes. Inhibition of Brn3a in melanoma cells led to cell cycle arrest followed by apoptosis indicating an important role for melanoma cell cycle progression and survival. Microarray analysis identified the mitotic kinesin KIF11, also called Eg5, as a transcriptional target gene of Brn3a. KIF11 encodes for a motor protein required for assembly of the mitotic spindle and, therefore, promotes cell cycle progression. We now addressed the oncogenic role of Brn3a by expressing Brn3a via lentiviral transduction in primary non-malignant skin cells and in transformed Brn3a-negative cells. In primary melanocytes and fibroblasts, expression of Brn3a resulted in upregulation of KIF11 and promoted anchorage-independent growth as demonstrated by increased colony formation in soft agar. Next, the effect of Brn3a expression on tumor growth was tested in vivo: Primary human melanocytes stably expressing ectopic Brn3a did not form tumors in immunodeficient SCID/NOD mice. We, therefore, analyzed Brn3a-negative cells that have been engineered to become tumorigenic in mice. For this, RAS-transformed NIH3T3 fibroblasts and SV40/hTERT/RAS-transformed melanocytes were utilized. In both cell types, ectopic expression of Brn3a resulted in upregulation of KIF11 and in increased anchorage-independent growth. In vivo, Brn3a expression clearly accelerated tumor growth. Furthermore, increased levels of KIF11 in Brn3a-psoitive tumors were observed. Taken together, these data suggest that Brn3a contributes to cell proliferation by providing molecules that are required for high cell cycle progression, e.g., Kif11. This promotes tumorigenesis and leads to increased tumor growth in vivo.


New melanoma-associated copy number alterations identified by high resolution array CGH

S. Boi, L. Pasini, C. Cantaloni, V. Adami, M. Cristofolini, R. Micciolo, A. Quattrone

Pathology Department, S. Chiara Hospital, Trento, Italy

Although melanoma accounts for only 4% of all dermatologic cancers, it is responsible for 80% of deaths from skin cancer while commonly accepted morphological standards are often insufficient in providing prognostic and diagnostic accuracy. Misdiagnosis is relatively common and can result in inappropriate over- or under-treatment of patients. Recently, array comparative genomic hybridization (aCGH) analysis of genome copy number alterations (CNAs) has made the diagnosis of melanocytic lesions less challenging. In this study, we performed high-resolution aCGH analysis of twenty formalin-fixed paraffin-emended (FFPE) cutaneous melanoma samples ranging from 2 to 10 yr follow-up. Our data showed specific and non-random somatic CNAs. Most recurrent gains were in bands 6p21.1, 7p22.2-ter, 8q24-ter, 11q13.2, 16p13.3, 19p13.11, and 19q13.33 (>60% of samples). Novel genomic amplifications were detected at the level of 9q33.2-ter, and 20q13.33 bands. Importantly, amplifications of 11q13.1-13.2 and 12q14.1, harboring respectively the CCND1 and CDK4 loci. On the other side, most samples presented frequent losses of 5p14.2, 8p22, 9p21.1-ter, 14q12, 14q31.1-31.3, 16q21, 21q21.1. As expected, chromosome regions 9p21.3 (including CDKN2A) and 10q23.31 (including PTEN) were the major targets of mono- and bi-allelic deletions. Interestingly, we found that these two alterations were not mutually exclusive and strongly correlated with lymph node invasion. In this study, for the first time, we evaluated a subset of archival melanoma tumors with the most advanced aCGH technology, along with validation of gene copy number by real-time qPCR assays and correlation with clinical outcome. We also identified additional melanoma-specific alterations, including gains of 3p21.1-21.31 and 6p21.1-22.2 and losses of 5p14.2 and 14q12, which could potentially allow the identification of new melanoma-associated genes with crucial roles in tumor progression. Our study, besides confirming the most of somatic CNAs already identified, provides the highest resolution aCGH mapping analysis of cutaneous melanoma yet reported.


Cross-platform genome-wide array screening identifies COL1A2, THBS1, TNFRSF10D and UCHL1 as tumor suppressor genes frequently silenced by methylation in melanoma

V. F. Bonazzi1, D. J. Nancarrow1, M. S. Stark1, R. J. Moser2, G. M. Boyle3, L. G. Aoude1, C. Schmidt4, N. K. Hayward1

1 Oncogenomics Laboratory, Queensland Institute of Medical Research, Herston, Brisbane, Queensland, Australia; 2SEQUENOM, Asia Pacific Office, Queensland Institute of Medical Research, Herston, Brisbane, Queensland, Australia; 3Drug Discovery Group, Queensland Institute of Medical Research, Herston, Brisbane, Queensland, Australia; 4Cancer Immunotherapy Group, Queensland Institute of Medical Research, Herston, Brisbane, Queensland, Australia

Epigenetic regulation of tumor suppressor genes (TSGs) has been shown to play a central role in melanomagenesis. By integrating data from gene expression and methylation array analyses we identified novel candidate TSGs frequently methylated in melanoma. We validated the methylation status of some of the most promising TSGs using the highly sensitive Sequenom Epityper assay in a large panel of melanoma cell lines and resected melanomas, and compared the findings with that from cultured melanocytes. We found transcript levels of UCHL1, COL1A2, THBS1 and TNFRSF10D were inversely correlated with promoter methylation. The effect of this methylation on expression was confirmed at the protein level for UCHL1 and THBS1. Characterization of how silencing of these candidate TSGs is related to melanoma development will increase our understanding of the etiology of this cancer and may provide tools for its early diagnosis.


Downstream effects of reduction in nucleotide excision repair in response to cisplatin treatment in melanoma

N. A. Bowden1,2,3, K. A. Ashton1,2,3, K. A. Avery-Kiejda4, X. D. Zhang4, P. Hersey4, R. J. Scott1,5

1 Centre for Information Based Medicine, University of Newcastle, Newcastle, NSW, Australia; 2Discipline of Medical Genetics, School of Biomedical Sciences, Faculty of Health, University of Newcastle, Newcastle, NSW, Australia; 3Hunter Medical Research Institute, Newcastle, NSW, Australia; 4Oncology and Immunology Unit, Calvary Mater Newcastle Hospital, Newcastle, NSW, Australia; 5Division of Genetics, Hunter Area Pathology Service, John Hunter Hospital, Newcastle, NSW, Australia

There is overwhelming evidence that one of the primary risk factors for melanoma is exposure to ultraviolet radiation (UVR). UV-mimetic agents such as cisplatin, are used in the treatment of many types of cancer, however their efficacy in the treatment of melanoma is limited. The removal and repair of large helix distorting DNA damage, induced by both UVR and cisplatin is orchestrated by the precise DNA repair pathway, nucleotide excision repair (NER). We recently reported limited induction of the GGR DNA damage recognition arm of NER in melanoma cell lines 24 h after cisplatin treatment. Despite growing evidence that GGR may play a role in cisplatin resistance, the underlying cause of this has not been investigated. Similarly, the downstream effects on transcription of reduced GGR and subsequent NER have also not been thoroughly investigated. The aims of the present study were to examine changes in mRNA transcript levels of regulators of GGR after cisplatin treatment and to investigate the downstream effect of low GGR response in melanoma cell lines. We have confirmed that the GGR regulators, BRCA1 and PCNA, are induced in normal cellular response to cisplatin induced DNA damage, but there is complete absence of induction of these regulators in melanoma cell lines. In addition, we identified highly significant overlap of transcript expression in melanoma cell lines after cisplatin treatment with transcripts highly correlated with the DNA double-strand break repair genes BRCA1, BRCA2, ATM and CHEK2. Finally, we confirmed complete lack of induction of XPB in melanoma cell lines and a significant overlap of transcript expression with XPB deficient cells after UVR. Taken together, these findings provide further support for the role of NER deficiency in melanomagenesis and resistance to cisplatin treatment.


Targeting the defective decatenation checkpoint in melanoma

K. M. Brooks, M. Ranall, S. Pavey, P. Mukhopadhyay, M. Wigan, N. Giles, B. Gabrielli

University of Queensland Diamantina Institute, Princess Alexandra Hospital, Brisbane, Queensland, Australia

In Australia melanomas account for 10% of all reported cancers and 3% of cancer mortalities. Current treatments offer <25% response rate and <8 months improved survival. Cancers frequently demonstrate defects in the cell cycle, particularly the checkpoint mechanisms. Loss of these protective mechanisms may provide opportunities for selectively targeting cancer cells. One G2 phase checkpoint, the decatenation checkpoint, ensures that entangled sister chromatids can be segregated with high fidelity during cellular replication. Topoisomerase II is responsible for resolving these tangles. Using ICRF-193, an inhibitor of Topoisomerase II, a panel of melanoma cell lines were investigated for decatenation checkpoint function. This analysis revealed that 17/25 melanoma cell lines possessed a defective decatenation checkpoint, suggesting it to be a common defect. Treatment with ICRF-193 caused checkpoint defective cells to undergo an aberrant mitosis, resulting in accumulation of cells exhibiting abnormal nuclear morphologies, while checkpoint functional cells remained relatively unaffected. Comparative gene expression analysis revealed differential expression of cell-division and cell-death associated genes, suggesting adaptive changes compensating for checkpoint failure. A strong gene-expression signature for the checkpoint defect was also observed in this analysis and potential bio-markers are being investigated for use identifying this defect in tumours. A synthetic lethality siRNA screen of the kinome (~700 genes) was carried out to identify genes/pathways crucial to the viability of cell lines with a defective decatentation checkpoint. This screen identified 10 potential hits, many involved in PI3Kinase signalling. We have validated this pathway as a selective target. This research has shown that the defective decatenation checkpoint is a viable target in melanoma. Using targets identified by the synthetic lethality screens and bio-markers to the defect in tumours, it should be possible to identify tumours most likely to be sensitive to drugs targeting the checkpoint defect, providing effective and selective therapeutics for the treatment of metastatic melanoma.


Expression of PAFR as part of a pro-survival response to chemotherapy: a novel target for combined therapy in melanoma

A. C. Onuchic1, C. M. L. Machado1, R. F. Saito1, S. Jancar2, R. Chammas1

1 Laboratório de Oncologia Experimental, Faculdade de Medicina da Universidade de São Paulo and Instituto do Câncer do Estado de São Paulo, Brazil; 2Departamento de Imunologia, Instituto de Ciências Biomédicas da Universidade de São Paulo, Brazil

Platelet activating factor (PAF) is an endogenous bioactive lipid with potent proinflammatory properties. The binding of PAF to its receptor (PAFR) is known to induce the expression of cytokines and angiogenic factors, and has been shown to play an important role in tumor progression and metastasis. Previous data from our group revealed that a PAFR antagonist, WEB2170, inhibits tumor growth in a B16F10 murine melanoma model, improving the survival rate of melanoma-bearing mice when combined with chemotherapy. The investigation of PAFR was extended to human melanoma cell lines, and data obtained by western blotting and flow cytometry showed the presence of PAFR in SKmel37 cells, revealing an augmented expression of PAFR in cells treated with cisplatin in vitro. Based on these findings, nude mice were inoculated with SKmel37 cells and treated with cisplatin and a similar PAFR antagonist, WEB2086. Notably, animals treated with both agents showed an important decrease in tumor growth, when compared to the control group and groups treated with only one agent. The treatment of SKmel37 cells in vitro suggested a slight but non-significant increase in cell death when submitted to treatment with cisplatin and WEB2086, but not enough to account for the results obtained in vivo. In this context, a role of PAFR in the tumor microenvironment, inducing a pro-tumorigenic setting, gains importance, suggesting that the modulation of tumor response to chemotherapy by PAFR may be dependent on interaction with microenvironmental elements, in addition to tumor cells. The present results help unravel PAFR-dependent mechanisms in tumor progression, supporting the concept of PAFR antagonists as a promising strategy in cancer therapy, especially in combination with chemotherapy.

Supported by FAPESP and CNPq, Brazil.


Hypoxia and MITF control metastatic behaviour in mouse and human melanoma cells

Y. Cheli1,2, S. Giuliano1,2, N. Fenouille1,2, M. Allegra1,2, V. Hofman3, P. Hofman2,3, P. Bahadoran1,2,4, J.-P. Lacour2,4, S. Tartare-Deckert1,2,4, C. Bertolotto1,2,4, R. Ballotti1,2,4

1 INSERM, U895, Centre Méditerranéen de Médecine Moléculaire (C3M), Equipe 1, Biology and pathologies of melanocytes, Equipe Labellisée par la Ligue Contre le Cancer, Nice, Cedex, France; 2Université de Nice-Sophia Antipolis, UFR médecine, Nice, France; 3ERI21/EA 4319 and Human Biobank, Nice, France; 4CHU NICE, department of dermatology, Cedex, NICE, France

Melanomas are very aggressive neoplasms with notorious resistance to therapeutics. It was recently proposed that the remarkable phenotypic plasticity of melanoma cells allows for the rapid development of both resistance to chemotherapeutic drugs and invasive properties. Indeed, the capacity of melanoma cells to form distant metastases is the main cause of mortality in melanoma patients. Therefore, the identification of the mechanism controlling melanoma phenotype is of paramount importance. In the present report, we show that deletion of MITF, the master gene in melanocyte differentiation, is sufficient to increase the metastatic potential of mouse and human melanoma cells. MITF silencing also increases fibronectin and Snail, two mesenchymal markers that might explain the increased invasiveness in vitro and in vivo. Furthermore, ablation of this population by forskolin-induced differentiation or MITF forced expression dramatically decreases tumour and metastasis formation, suggesting that eradication of low-MITF cells might improve melanoma treatment. Moreover, we demonstrate that a hypoxic micro-environment decreases MITF expression through an indirect, HIF1α dependent transcriptional mechanism and increases the tumorigenic and metastatic properties of melanoma cells. We identified Bhlhb2, a new player in melanoma biology, as the mediator of hypoxia/HIF1a inhibitory effect on MITF expression. Our results reveal a hypoxia-HIF1α-BHLHB2-MITF cascade controlling the phenotypic plasticity in melanoma cells and favouring metastasis development. Targeting this pathway might be helpful in the design of new anti-melanoma therapies.


Stathmin 1 is a potential novel oncogene in melanoma

J. Chen, X. Yang, X. Zhang, H. E. Feilotter, V. A. Tron

Department of Pathology and Molecular Medicine, Queen's University, Kingston, ON, Canada

In previously published studies, we demonstrated that miR-193b expression is reduced in melanoma relative to benign nevi, and also that miR-193b represses Cyclin D1 and Mcl-1 expression. We suggested that stathmin 1 (STMN1) might be a target of miR-193b. STMN1 normally regulates microtubule dynamics through either sequestering free tubulin heterodimers or promoting microtubule catastrophe. Overexpression of STMN1 has been observed in a variety of human malignancies, but its association with melanoma is unknown. We now report that STMN1 is up-regulated during development and progression of melanoma relative to benign nevi, and that STMN1 is directly regulated by miR-193b. Using an experimental cell culture approach, overexpression of miR-193b by synthetic miRNAs was seen to repress STMN1 expression, while inhibition of miR-193b by anti-miR oligos increased STMN1 expression in melanoma cells. The use of a luciferase reporter assay confirmed that miR-193b directly regulates STMN1 by targeting the 3’UTR of STMN1 mRNA. STMN1 knockdown by STMN siRNA or by overexpression of miR-193b in melanoma cells drastically repressed cell migration and proliferation rate, whereas ectopic expression of STMN1 using lentivirus increased cell migration and proliferation rate.

Taken together, our study suggests that downregulation of miR-193b may contribute to increased STMN1 expression in melanoma, which consequently promotes migration and proliferation of tumor cells.


Molecular mechanisms of riluzole-mediated DNA double-stranded breaks in human melanoma cells

B. A. Wall1,2, J. Wangari-Talbot2,3, L. J. Yu2, A. Khan4, S. Chen2,4

1 The Joint Graduate Program in Toxicology, Rutgers University and UMDNJ-Robert Wood Johnson Medical School, Piscataway, NJ, USA; 2Susan Lehman Cullman Laboratory for Cancer Research, Ernest Mario School of Pharmacy, Rutgers University, Piscataway, NJ, USA; 3Graduate School of Biomedical Science at UMDNJ-Robert Wood Johnson Medical School, Piscataway, NJ, USA; 4Cancer Institute of New Jersey, New Brunswick, NJ, USA

With an incidence of approximately 170 000–200 000 cases per year, brain metastases are an exceedingly common event in the progression of many human malignancies with melanoma ranking the 4th most common. The survival rate for patients with advanced metastasis ranges from 2 to 8 months with prolonged survival rates of 5% after 5 yr. Results obtained using the current single or multi drug chemotherapy regiments are modest at best and new strategies urgently need to be identified that can provide rational approaches in improving survival rates and success of treatment. Riluzole is an FDA approved drug used for the treatment of amyotrophic lateral sclerosis (ALS) and has been shown to modulate glutamatergic signaling by decreasing the amount of available ligand (glutamate), however, the precise mechanisms have not been fully delineated. Preliminary data has revealed that treatment of melanoma cells in vitro using relatively low concentration of Riluzole resulted in the accumulation of a substantial fraction of melanoma cells in the G2/M phase of the cell cycle followed by the build-up of cells in the sub G1 phase, indicating apoptotic cell death. It is well known that the G2/M phase is considered a radiosensitive phase of the cell cycle, in addition, Riluzole readily crosses the brain-blood barrier. Based on these observations, we have begun to study the effects of Riluzole on glutamatergic signaling in our system and its implications as an agent that sensitizes melanoma cells to the effects of ionizing radiation in an attempt to test it as a potential therapy to use in conjunction with ionizing irradiation in melanoma brain metastases. We found that combining Riluzole with γ-irradiation results in a decrease of xenograft tumor growth in vivo.


Protein kinase C is a therapeutic target for melanomas with GNAQ or GNA11-mutations

X. Chen1, V. Ding2, K. G. Griewank1, B. C. Bastian1

1 Human Oncology and Pathogenesis Program, Memorial Sloan Kettering Cancer Center, New York, NY, USA 2Department of Dermatology, University of California, San Francisco, CA, USA

Uveal melanoma (UM) is the most common primary intraocular malignant tumor in adults. There currently is no effective treatment. About 80% of UMs contain either GNAQ or GNA11 activating mutation. Understanding the downstream effector pathways activated by mutant GNAQ/11 is of considerable interest as it may help to identify opportunities for targeted therapy. We introduced oncogenic GNAQ/11 into 293T cells, immortalized mouse and human melanocytes. In all settings, overexpression of GNAQQ209L or GNA11Q209L induced activation of protein kinases C (PKC) and the MAPK pathway. ShRNA mediated knock-down of endogenous GNAQ oncoproteins in human melanoma cell lines with GNAQ mutations or melan-a cells that stably express oncogenic GNAQ abrogates both PKC and MAPK activation. The PKC inhibitors AEB071 and AHT956 selectively inhibited growth of all six human melanoma cell lines with GNAQ/11 mutations tested with IC50 values ranging from 56 to 467 nM for AEB071, from 4 to 123 nM for AHT956, while six melanoma cell lines without these mutations showed no effects at doses up to 1 μM. Similar effects were seen in mouse melan-a cells stably transduced with mutant GNAQ or GNA11 mutation, whereas the compounds had no effects on cells transduced with the wild type counterparts. In cells harboring GNAQ/11 mutations, AEB071 and AHT956 inhibit PKC signaling and the MAPK pathway in a concentration-dependent manner, placing MAPK downstream of PKC in this setting. In vivo antitumor activity of AEB071 and AHT956 is currently under investigation. Our data show that human melanoma cell lines with GNAQ/11 mutation are selectively sensitive to PKC inhibitors, offering an opportunity for rational therapeutic intervention for patients with this type of cancer.


Phosphatidylinositol 4,5-Bisphosphate 5-Phosphatase A regulates PI3K/Akt signaling in human melanoma cells

Y. Ye1,2,3,4, L. Jin1,5, J. Wilmott1,6, W. L. Hu7, R. F. Thorne8, L. Dong1,2,3, C. de Bock8, C. C. Jiang1,2,3, H. Rizos1,9, M. Wu7, L. J. Zhang4, R. Scolyer1,6, P. Hersey1,5, X. D. Zhang1,2,3

1 Melanoma Institute Australia, Sydney, NSW, Australia; 2Priority Research Center for Cancer Research, School of Medicine and Public Health, University of Newcastle, NSW, Australia; 3Cancer Research Program, Hunter Medical Research Institute, NSW, Australia; 4Department of Immunology, Anhui Medical University, Hefei, Anhui, China; 5Kolling Institute of Medical Research, University of Sydney, NSW, Australia; 6Department of Pathology, University of Sydney, Camperdown, Sydney, Australia; 7Hefei National Laboratory for Physical Sciences at Microscale and School of Life Sciences, University of Science and Technology of China, Hefei, Anhui, China; 8Cancer Research Unit, University of Newcastle, Callaghan, NSW, Australia; 9Westmead Institute for Cancer Research, Westmead Hospital, University of Sydney at Westmead Millennium Institute, Westmead, NSW, Australia

The PI3K/Akt pathway is constitutively activated in up to 70% of melanomas. This is closely associated with melanoma development, progression, and resistance to therapeutics including mutant BRAF inhibitors. Inappropriate activation of the PI3K/Akt pathway has been attributed to various mechanisms, in particular, down-regulation or loss of PTEN. However, no study has yet examined the role of inositol polyphosphate-5-phosphatases (5-phosphatases), another class of negative regulators of PI3K/Akt signaling, in activation of the pathway in melanoma. We show here that one of the 5-phosphatases, PIB5PA (phosphatidylinositol 4, 5-bisphosphate 5-phosphatase A), is frequently down-regulated, or lost in metastatic melanomas, which is correlated with activation of Akt, and contributes to proliferation, survival, migration, and invasion of melanoma cells. The expression of PIB5PA was generally low in melanoma cell lines and fresh melanoma isolates. This was mirrored by studies on melanocytic tumor tissue sections. Induction of PIB5PA in melanoma cell lines that conditionally expressed exogenous PIB5PA in response to 4-hydroxytamoxifen (4-OHT) inhibited cell proliferation, migration, and invasion, and enhanced apoptosis induced by various stimuli. These were due to inhibition of PI3K/Akt, as they were abolished when an active form of Akt (myr-Akt) was co-introduced into the cells. In addition, ectopic expression of PIB5PA retarded melanoma growth in a xenograft mouse model. Down-regulation of PIB5PA in melanoma cells appeared to be, at least in part, due to histone hypoacetylation mediated by histone deacetylase (HDAC) 2 and HDAC3 as shown by studies using selective HDAC inhibitors and siRNA knockdown. Collectively, these results identify PIB5PA deficiency or loss as a novel driver of activation of PI3K/Akt signaling in melanoma cells that is important for melanoma progression and resistance to treatment.


The effects of endothelin 3 in a melanoma mouse model

N. Chin, J. C. Gallegos, R. Cruz, R. Garcia, A. Gonzalez, D. Fernandez, L. Kos

Florida International University, Miami, FL, USA

Endothelin receptor b (Ednrb) and its ligand Endothelin 3 (Edn3) have been implicated in melanoma. Several ‘in vitro’ studies show an upregulation in Ednrb and Edn3 at both the protein and mRNA levels, as melanoma becomes more aggressive. This study investigated the putative role played by Ednrb and Edn3 over-expression in melanoma progression in vivo. We crossed Tg(Grm1)Epv transgenic mice that aberrantly express metabotropic glutamate receptor 1under the Dopachrome tautomerase promoter, leading to spontaneous melanocytic lesions in the ears and tails that do not metastasize, with transgenics that over-express Edn3 under the Keratin 5 promoter (K5-Edn3). In the Tg(Grm1)Epv/K5-Edn3 mice, tumors appeared 2–3 months earlier and grew at a rate 3–4 times faster when compared to Tg(Grm1)Epv mice. These mice also showed lesions in the dorsal aspect where they are not found in Tg(Grm1)Epv mice. Approximately eighty-eight percent of Tg(Grm1)Epv/K5-Edn3 mice had pigmented lesions in distant organs such as the lung, spleen and salivary glands. These lesions were positive for the melanocyte specific marker TRP-1, as well as melanoma marker S100. Considering the faster tumor growth rate in the triple transgenic mice, we investigated differences in the angiogenic response in tail tumors taken from both the triple transgenic mice and the control population. Immunofluorescence analysis with the endothelial cell marker CD31 showed that there were more endothelial cells per area in the triple transgenics than the controls. Real Time PCR analysis of tail tumors also showed higher expression levels of angiogenic related genes such as Hif-1α in the triple transgenics. Based on our results, over-expression of Edn3 in the local environment is sufficient to alter the kinetics of melanocytic tumors’ progression and lead them to a fully malignant state. Our preliminary data also suggests that Edn3 may be increasing tumor angiogenesis leading to faster melanoma tumor development.


Automated platform (SnapPath™) to generate functional signaling profiles of live melanoma samples

D. P. Clark1, E. Jeffries1, A. Schayowitz1, J. Villanueva2, M. Herlyn2, C. Cabradilla1

1 BioMarker Strategies, Baltimore, MD, USA; 2The Wistar Institute, Philadelpha, PA, USA

Insight into signal transduction networks in metastatic melanoma is critical to understand mechanisms of resistance to targeted therapy. We have utilized an automated platform (SnapPath™) to expose live cellular suspensions of melanoma cell lines to growth factors (EGF) and/or BRAF inhibitors (PLX4720; SB-590885) in order to modulate their signal transduction networks. This modulation was then characterized by a multiplexed analysis (BioPlex) of phosphoproteins under different conditions to generate functional signaling profiles. Cell lines were divided into two groups: BRAF inhibitor sensitive (451Lu; Mel1617; SKMEL-28; SKMEL-31; COLO829; Malme 3M) and resistant (451Lu-R; Mel1617-R; RPMI-7951; SKMEL-2; SKMEL-3; A2058) based on PLX4720 IC50 values.

Exposure of melanoma cells to BRAF inhibitors revealed significant differences in MAPK pathway suppression between sensitive versus resistant cell lines, as measured by inhibition of pMEK1 and pERK1/2 in PLX4720-treated versus untreated cells, and by pERK levels in PLX4720 exposed cells. Interestingly, two PLX4720 resistant cell lines that appeared to have significant MAPK pathway suppression also had high levels of pAKT, suggesting an intrinsic resistance mechanism through PI3K/AKT pathways.

Analysis of two models of acquired resistance to BRAF inhibitors revealed significantly different functional signaling profiles of resistant lines compared with the sensitive parental lines. Specifically, the resistant lines displayed significantly less activity of PLX4720 against the MAPK pathway, higher levels of pERK1/2 in the presence of PLX4720, and higher levels of pAKT than sensitive parental lines.

These results support the utility of this automated platform in the generation of functional signaling profiles to inform preclinical or clinical drug development, and as predictive tests on human samples to guide targeted therapy.


Maintenance of melanophore morphology and survival depends on autophagy and divalent cations in vps11 zebrafish mutants

L. F. Clancey1, A. J. Beirl1, T. H. Linbo2, C. D. Cooper1

1 Department of Molecular Biosciences, Washington State University Vancouver, Vancouver, WA, USA; 2Department of Biological Structure, University of Washington, Seattle, WA, USA

We have characterized a zebrafish pigment cell mutant, the melanophore integrity mutant, which displays defects in black and silver pigment cell survival (melanophores and iridophores, respectively). Mapping and candidate gene analysis links the integrity mutant mutation to vacuolar protein sorting 11 (vps11w66). vps11w66 mutants show a dramatic reduction in melanophore and iridophore number that is specification and differentiation independent. TUNEL analysis and treatment with the caspase inhibitor, zVAD-fmk, indicates that vps11w66 pigment cell death is not caused by apoptosis. While an increase of intracellular calcium (using the L-type calcium channel agonist, BayK8644) exacerbates pigment cell loss, an increase in extracellular magnesium is protective. Treatment of vps11w66 larvae with Bafilomycin A1, a vacuolar ATPase and autophagy inhibitor, increases melanophore number and restores melanophore morphology. Last, transmission electron microscopy analysis of mutant epidermis plasma membrane indicates a general loss of plasma membrane integrity, possibly due to necrosis or another caspase independent mode of death. Taken together, these data suggest a connection between vps11 function, autophagy and divalent cation concentration in the maintenance of zebrafish pigment cell survival.


Inhibition of CRD-BP sensitizes melanoma cells to chemotherapeutic agents

E. Craig, V. Spiegelman

Department of Dermatology, University of Wisconsin School of Medicine and Public Health, Madison, WI, USA

We previously reported that malignant melanomas express high levels of coding region determinant-binding protein (CRD-BP). This molecule binds mRNA and is important for the activation of anti-apoptotic pathways, a mechanism often linked to insensitivity to therapeutics. However, it is not known whether CRD-BP plays a role in the resistance of melanomas to anti-cancer treatment. Here we demonstrate that knock-down of CRD-BP with a specific sh-RNA enhances the effect of dacarbazine, temozolomide, vinblastine and etoposide on both primary and metastatic melanoma cell lines. CRD-BP down-regulation contributes to cell sensitization by increasing apoptosis and diminishing melanoma cell growth in response to chemotherapeutic agents. Furthermore, inhibition of CRD-BP decreases MITF expression and reintroduction of MITF partially compensates for the absence of CRD-BP. These findings suggest that high expression of CRD-BP in melanoma cells confers resistance to chemotherapy and that these CRD-BP responses are mediated, at least in part, by MITF.


Regulation of HERV-K by MEK-ERK and p16-CDK4 in melanoma cells

J. Dong, J. Z. Li, T. Sheng, X. Wan, T. Liu, H. Wu

Department of Pathology and Sealy Center for Cancer Cell Biology, University of Texas Medical Branch, Galveston, TX, USA

Deregulation of BRAF-MEK-ERK and p16-CDK4-RB pathways is a common occurrence in cancer biology. We have reported that blocking CDK4 enhances the therapeutic efficacy of BRAF-MEK inhibitors in melanoma cells. Human endogenous retrovirus K (HERV-K) is activated in melanoma cells but not in normal melanocytes, and HERV-K inhibition using RNAi can slow down the growth of melanoma mouse xenograft. The growth characteristics of melanoma cells that can be modified by HERV-K activation (e.g., changes in cell shape, loss of melanin, anchorage-independent growth) overlap with those that can be blocked by BRAF-MEK and CDK4 inhibitors, which prompted us to hypothesize that HERV-K is regulated by MEK-ERK and p16-CDK4 pathways.

We examined the expression of HERV-K GAG and ENV, the active form of ERK (p-ERK), and p16 in a panel human melanocytic specimens including 38 benign nevi and 34 melanomas using immunohistochemistry. We also investigated the regulation of HERV-K by specific MEK and CDK4 inhibitors in melanoma cells. We found that HERV-K GAG and ENV immunoreactivity is significantly more frequent in melanomas than in nevi; the expression of the GAG and ENV proteins is positively correlated with p-ERK and negatively correlated with p16 expression; and that HERV-K expression in melanoma cells can be blocked by MEK and CDK4 inhibitors, especially when used in combination. Our results support the notion that HERV-K may be activated by MEK-ERK and p16-CDK4 pathways in melanoma cells. Importantly, if HERV-K drives tumor progression downstream of RAF-MEK-ERK and p16/CDK4, when HERV-K is already turned on, cells may escape the inhibitory effects of therapies targeting RAF-MEK-ERK and p16/CDK4. Triple therapy of targeting HERV-K, MEK and CDK4, may be necessary to produce more effective and long-lasting therapeutic effects.


Role of RAF proteins in a NRAS-induced melanoma mouse model

S. Druillennec1, C. Dorard1, A. Valluet1, M. Larcher1, F. Beermann2, M. Baccarini3, L. Larue1, A. Eychène1

1 Institut CURIE-Recherche, INSERM U1021-CNRS UMR3347, Centre Universitaire, Orsay, France; 2Swiss Institute for Experimental Cancer Research (ISREC), Lausanne, Switzerland; 3Center for Molecular Biology, University of Vienna, Max F. Perutz Laboratories, Vienna, Austria

Deregulation of the RAS/RAF/MEK/ERK pathway is a common phenomenon in cutaneous melanoma (CMM), which drives melanocyte transformation and tumoral progression. Activating mutations of BRAF or NRAS occur at a high frequency in about 50% and 15% of CMM patients, respectively. The recent development of pharmacological inhibitors such as vemurafenib (PLX4032) specifically targeting mutated BRAFV600E raises new hopes in the treatment of CMM. However, such compounds cannot be used for half of melanoma patients, including those mutated on NRAS, due to paradoxical effects of BRAF inhibition by the Vemurafenib on the MAPK pathway. In addition, the contribution of BRAF and its closely related kinase CRAF downstream of RAS during tumoral progression appears complex. In vitro studies have shown that, when NRAS is mutated in human melanoma cell lines, the cells switch their signaling from BRAF to C-RAF to activate the MEK/ERK pathway.

In order to study the respective role of RAF kinases in vivo, we crossed conditional BRAF and/or CRAF knockout mice with a melanoma mouse model inferred by the expression of the mutated form of human NRASQ61K in the melanocyte lineage. The skin of transgene-expressing animals displays exacerbated pigmentation and spontaneously develops a large number of nevi, which sometimes evolve to melanomas. Single or compound ablation of BRAF and CRAF is achieved upon conditional Cre expression in the melanocyte lineage. These models allow us to investigate the role of both kinases at each step of tumoral progression, from tumor initiation (formation of benign nevi) to metastasis. Our preliminary results show that BRAF is not required for tumoral maintenance.


Identification of TFG (TRK fused gene) as a putative metastatic melanoma tumour suppressor gene

K. Dutton-Regester1,2, L. Aoude1, D. J. Nancarrow1, M. S. Stark1, L. O’Connor3, C. Lanagan3, G. M. Pupo4,5, V. Tembe4,5, C. D. Carter5, M. O’Rourke6, R. A. Scolyer5,6,7, G. J. Mann4,5, C. Schmidt2, A. Herington2, N. K. Hayward1

1 Queensland Institute of Medical Research, Oncogenomics laboratory, Brisbane, Qld, Australia; 2Faculty of Science and Technology, Queensland University of Technology, Brisbane, Qld, Australia; 3Queensland Institute of Medical Research, Cancer Immunotherapy Laboratory, Brisbane, Qld, Australia; 4University of Sydney at Westmead Millennium Institute, Westmead, NSW, Australia; 5Melanoma Institute of Australia (formerly the Sydney Melanoma Unit), North Sydney, NSW, Australia; 6Tissue and Pathology and Diagnostic Oncology, Royal Prince Alfred Hospital, Camperdown, NSW, Australia; 7Discipline of Pathology, Sydney Medical School, the University of Sydney, NSW, Australia; 8Morris Tower, 149 Wickham Tce Brisbane, Qld, Australia

High density SNP arrays can be used to identify DNA copy number changes in tumours such as homozygous deletions of tumour suppressor genes and focal amplifications of putative oncogenes. Illumina Human CNV370 Bead chip arrays were used to assess the genome for unbalanced chromosomal events occurring in 39 stage III metastatic melanoma cell lines. A number of genes previously recognised to have an important role in the development and progression of melanoma were identified including homozygous deletions of CDKN2A (13 of 39 samples), CDKN2B (10 of 39), PTEN (3 of 39), PTPRD (3 of 39), TP53 (1 of 39) and amplifications of MITF (2 of 43), CCND1 (2 of 43), MDM2 (1 of 43) and NRAS (1 of 43). In addition, a number of novel focal homozygous deletions potentially targeting novel melanoma tumour suppressor genes were identified. FAS, CH25H, BMPR1A, ACTA2 and TFG were further investigated in a larger cohort of melanomas by sequencing to identify non-synonymous somatic mutations. Non-synonymous mutations were identified in BMPR1A (1 of 43), ACTA2 (2 of 43) and TFG (5 of 103). A number of potentially important mutation events occurred in TFG including the identification of a mini mutation ‘hotspot’ at amino acid residue 380 (P380S and P380L) and the presence of multiple mutations in two melanomas. Although the biological function of TFG is as yet poorly characterised, the TFG gene is known to be involved in a number of gene fusions that occur in a variety of cancers and is thought to affect the NF-κB and MAPK pathways. Mutations in TFG may represent an alternative mechanism to activate these pathways and therefore may have important clinical relevance for current therapeutic strategies to treat metastatic melanoma.


Proteomics approach to identify novel markers for response to chemotherapy in cutaneous melanoma

A. Azimi, M. Pernemalm, J. Hansson, J. Lehtio, C. Johansson, S. Egyhazi

Department of Oncology–Pathology, Karolinska Institutet, Stockholm, Sweden

Malignant melanoma is one of the most rapidly increasing malignancies in Caucasian populations. So far, curative treatment is achieved only by surgical resection at an early stage, while the prognosis in cases with disseminated melanoma is poor, and we lack predictive markers to identify the few patients who respond to existing chemotherapy. Proteomic studies were performed on pre-treatment lymph node metastases, comparing the protein expression profile of five tumors from patients who responded to dacarbazine (DTIC)/temozolomide (TMZ) with nine tumors from non-responders. The majority of the tumors have already been analyzed by gene expression microarray, thus, allowing for a direct comparison of the proteome and the transcriptome in the same tumors. The samples have been matched for gender, age and histopatologic type. Briefly, digested peptides were labeled with iTRAQ labeling and separated by isoelectric focusing, followed by highthroughput tandem mass spectrometry (HPLC-MS/MS).

Univariate and multivariate data analyses were performed using SAM, principle component analysis (PCA) and O-PLS (by SIMCA), followed by pathway analysis with Ingenuity. Several identified candidates had also been detected by gene expression microarray. Validation of the top candidates for therapy response is performed on the same sample set as well as on a larger independent second set of tumors, using immunoblotting or immunohistochemistry methods. Proteins of interest will be further investigated by functional studies in a panel of melanoma cell lines.


Integrative phospho-proteomic and genomic analyses identify AXL as a potential biomarker and therapeutic target for NRAS-mutated melanoma

I. V. Fedorenko1, K. H. T. Paraiso1, B. Fang1, J. Koomen1,2, K. L. Nathanson3,4, B. Wubbenhorst3,4, R. Mathew5, J. Messina5, K. S. M. Smalley1

1 The Department of Molecular Oncology, The Moffitt Cancer Center & Research Institute, Tampa, FL, USA; 2Proteomics Core, The Moffitt Cancer Center & Research Institute, Tampa, FL, USA; 3Medical Genetics, University of Pennsylvania School of Medicine, Philadelphia, PA, USA; 4The Abramson Cancer Center, University of Pennsylvania School of Medicine, Philadelphia, PA, USA; 5Department of Pathology and Cell Biology, University of South Florida College of Medicine, Tampa, FL, USA

Currently, few therapeutic options exist for the 15–20% of patients whose melanomas harbor activating mutations in NRAS. Here, we have used comprehensive phospho-proteomic and genomic analysis to characterize the patterns of intracellular signaling in a panel of NRAS and BRAF-mutated melanoma cell lines, with the goal of identifying new biomarkers and therapeutic targets for NRAS mutated melanomas. Our preliminary analyses showed BRAF mutated melanomas to have less diverse intracellular signaling than the NRAS group, which tended to have constitutive phosphorylation in a wide-range of receptor tyrosine kinases (RTKs), such as c-MET, EGFR, HER2 and c-Abl. Although these patterns of RTK activity were heterogeneous across the NRAS mutated melanoma cell line panel, all of the NRAS mutated cell lines examined showed constitutive activity in the receptor tyrosine kinase Axl. A role for Axl in the oncogenic behavior of NRAS mutated melanomas was suggested by array CGH studies showing genomic amplification of Axl in a significant fraction of the NRAS mutated melanoma cell lines and by Western Blot experiments demonstrating increased expression of Axl and its ligand Gas6. Mass spectrometry analysis and immunoprecipitation revealed Axl to be constitutively phosphorylated at the Tyr702 autophosphorylation site in NRAS mutated melanoma cell lines only. Functionally, siRNA knockdown and pharmacological inhibition of Axl impaired the survival and invasion of NRAS mutated melanoma cell lines grown under 3D organotypic cell culture conditions. Studies are ongoing to confirm the role of Axl in the initiation and progression of melanomas harboring NRAS mutations and determine whether Axl constitutes a therapeutic target for this melanoma sub-group.


Cancer drug treatment is unnatural selection

E. H. Flach1, I. Fedorenko2, K. Paraiso2, K. S. M. Smalley2, A. R. M. Anderson1

1 Integrated Mathematical Oncology, Moffitt Cancer Center, Tampa, FL, USA; 2Molecular Oncology, Moffitt Cancer Center, Tampa, FL, USA

Targeted drug treatment reduces the tumour volume, but there is almost always recurrence even under chronic treatment. We show that the tumour population is heterogenous. Then the drug treatment is a selection process, targeting specific subpopulations. If treatment is stopped, phenotypic drift causes reversion towards the original wild-type population. Our model is a discrete population of cells, the individual equivalent of a population model. The cells each have a distinct phenotype. This phenotype determines their fitness. The fitness changes under drug conditions: we define a fitness landscape for both drug and drug-free conditions. Experimentation shows evidence of only partial reversion to wild-type. We extend the complexity of the fitness landscape to multiple fitness ‘wells’. Reversion after drug treatment only fills one of the wells. The overall behaviour matches experimental observations.

Our model concept extends to considering alternative treatments. Temporal variation appears unhelpful but well-chosen combination therapies could be effective. This approach gives a quantitative prediction of treatment strategies.


Demethylation of DNA contributes to Brn3a expression in melanoma

S. Fraschka1, M. Heppt1, T. Hohenauer1, R. Kappler2, T. Ruzicka1, C. Berking1, R. Besch1

1 Department of Dermatology and Allergology, Ludwig-Maximilian University, Munich, Germany; 2Department of Pediatric Surgery, Research Laboratories, Ludwig-Maximilian University, Munich, Germany

During embryogenesis, the transcription factor Brn3a (POU4F1) is expressed in the neural crest and is involved in survival and differentiation of precursors of neuronal cells. We observed that Brn3a is highly expressed in human melanoma cells, but not in primary human melanocytes or fibroblasts. Loss-of-function experiments showed that Brn3a is required for melanoma cell proliferation and survival. However, the mechanisms that cause Brn3a up-regulation in melanoma are not known.

In adult tissue, long-term silencing of embryonic genes often relies on epigenetic mechanisms, e.g., methylation of CpG-rich DNA regions. Analysis of the Brn3a promoter region revealed two large CpG islands suggesting that Brn3a expression may be silenced in melanocytes via DNA methylation. We tested whether DNA-demethylating reagents cause re-expression of Brn3a. Treatment of melanocytes with 5-aza-2′-deoxycytidine resulted in 3–6-fold increase in Brn3a expression. Treatment of primary fibroblasts led to a lower increase compared to melanocytes, suggesting that other melanocyte-specific factors contribute to Brn3a expression. Furthermore, upregulation of Brn3a in melanocytes by 5-aza-2′-deoxycytidine did not reach the levels observed in melanoma cells indicating that further tumor-specific epigenetic or genetic mechanisms promote Brn3a expression in melanoma. Next, the genetic regulation of Brn3a was investigated by analyzing the 5′ region of the Brn3a gene. Reporter assays of different fragments of the 5′ region revealed that the proximal 1000 bp region is required to promote full promoter activity in melanoma cells. No activity was observed in Brn3a-negative fibroblasts. Subsequently, transcription factors that putatively bind in this region were identified by in silico analysis.

In conclusion, the data show that demethylation of the Brn3a gene locus contributes to Brn3a expression in melanoma. In addition, lineage-specific factors as well as other tumor-specific factors are involved Brn3a expression.


Beta-catenin signaling reduces migration in the melanocyte lineage

S. J. Gallagher, M. Y. Kumasaka, D. Champeval, V. Delmas, L. Larue

Institut Curie, Developmental Genetics of Melanocytes, CNRS UMR3347, INSERM U1021, Orsay, France

The ability of melanoma to readily metastasize is especially concerning due to the lack of effective treatments for melanoma once it has progressed. Activation of the Wnt/β-catenin pathway occurs in about 30% of melanoma, and activated β-catenin expression suppresses senescence of murine melanocytes via reduction of p16INKA transcription. During mouse development, proliferating melanoblasts migrate dorso-laterally from the neural crest to the belly. An activated form of β-catenin (bcat*) was expressed specifically in the melanocyte lineage by using the tyrosinase enhancer-promoter. bcat* impaired the migration of melanoblasts during embryogenesis, leading to a white belly spot. Furthermore, chemical and genetically enforced canonical beta-catenin signaling reduced the migration of both melanocytes and melanoma cells in vitro. In human melanoma cells, we show that β-catenin signaling reduced migration via MITF-M. However, such inhibition of migration was partially independent of MITF-M. Altogether, these results show that in contrast to other cancers such as colorectal cancer, beta-catenin signaling reduces cell migration in melanoblast derivatives, thus therapeutics targeting the Wnt pathway may not produce expected outcomes in melanoma.


Analysis of gene expression profile of primary and metastatic melanoma

R. Essner1, K. W. Gong1, B. Chmielowski1, R. S. Finn1, W. D. Tap2, D. Slamon1

1 Division of Hematology & Oncology, Tranlational Oncology Research Labs, Department of Medicine, David Geffen School of Medicine at UCLA, Los Angeles, CA, USA; 2Memorial Sloan-Kettering Cancer Center, New York, NY, USA

Metastatic malignant melanoma is largely refractory to existing therapies and has a very poor prognosis with a median survival rate of 6 months and 5-yr survival rate of <5%. Better understanding of melanoma biology, especial the gene expression profile between primary and metastatic melanoma, may provide useful information for developing new therapeutic approaches to this disease. RNA was extracted from 40 primary melanoma and 49 metastatic melanoma tissue. RNA was submitted to microarray. The gene expression profile was analyzed with MATLAB software. 8261 genes were differentially expressed between primary and metastatic melanoma with P < 0.01 and false discovery rate 5%. Panther Pathway classification showed that these genes were enriched with genes which were involved in Cytoskeletal regulation by Rho GTPase, Inflammation mediated by chemokine and cytokine signaling pathway, Notch signaling pathway, p53 pathway and Angiogenesis. Interrogation of genes involved in cytoskeletal regulation by Rho GTPase showed that the mRNA level of LIM domain kinase 1 (LIMK1) was significantly higher in metastatic melanoma than that in primary melanoma. The expression of LIMK1 was knockdown in melanoma cell line M238 with siRNA. mRNA level of LIMK1 was reduced to 70–80% at day 2 after siRNA treatment. The protein level of LIMK1 was reduced ~50% at day 6. Knockdown of LIMK1 at protein level ~50% did not have any effect on the growth of M238. However it significantly reduced the invasive ability of M238 in vitro.

In summary, primary melanoma and metastatic melanoma have different gene expression profiles. It may provide useful information for finding therapeutic targets and biological marker. LIMK1 may not involve cell proliferation. Knockdown of LIMK1 expression may significantly reduce the invasive ability of melanoma cell line M238.


Investigation of the HSP90 inhibitor XL-888 as a treatment for NRAS mutant melanoma

E. Haarberg1, K. H. T. Paraiso1, J. Koomen1, K. S. M. Smalley1,2

1 Department of Molecular Oncology, The Moffitt Cancer Center & Research Institute, Tampa, FL, USA; 2Department of Cutaneous Oncology, The Moffitt Cancer Center & Research Institute, Tampa, FL, USA

Activating mutations in either BRAF or NRAS are found in the majority of melanoma cases (~70%), and mutations in the two oncogenes appear to be mutually exclusive. The BRAF inhibitor vemurafenib was recently approved by the FDA for treatment of melanoma harboring BRAF mutations. Unfortunately, treatment of NRAS mutant cell lines with BRAF inhibitors results in an activation response, thus unique strategies are needed for treatment of melanoma harboring NRAS mutations. Heat shock protein 90 (HSP90) is important for proper folding and stability of a large proportion of cellular proteins, including oncogenes such as BRAF-V600E, ERB-B2, and BCR-ABL. Here we describe an investigation of the HSP-90 inhibitor XL-888 (Exelixis) in melanoma cell lines with mutations in NRAS. XL-888 inhibited proliferation of a panel of ten NRAS mutant melanoma cell lines at concentrations between 30 and 300 nM. Treatment of four of these cell lines with 300 nM XL-888 for 72 h caused partial G2 cell cycle arrest and 60–90% of the cells to become apoptotic. Long term treatment with XL-888, in the order of 4–6 weeks, completely eradicate the cultured melanoma cells at concentration between 10 and 30 nM. Inhibition of HSP-90 affected a wide range of proteins including several involved in regulating proliferation, cell cycle, and apoptosis. Levels of ARAF, cdc2 and Mcl1 decreased, while the pro apoptotic proteins BAK and BiM increased. siRNA mediated knockdown of Cdc2 caused an increase in the fraction of cells in the G2 phase, although weaker than the effect of XL-888, while Mcl-1 knockdown resulted in an apoptotic response similar to that observed with XL-888 treatment. We therefore suggest that inhibition of HSP90 may be a potential therapeutic strategy for the treatment of melanomas with NRAS mutations.


MITF expression dictates the subcompartment-specific distribution of differentially cycling tumor cells in melanoma

N. K. Haass1,2,3, A. Anfosso1, K. A. Beaumont1, D. S. Hill1, P. Mrass1, I. Kinjyo1, O. Kanagawa4, W. Weninger1,2,3

1 Centenary Institute, Newtown, NSW, Australia; 2Discipline of Dermatology, University of Sydney, NSW, Australia; 3Department of Dermatology, Royal Prince Alfred Hospital, Camperdown, NSW, Australia; 4Research Center for Allergy and Immunology, Riken, Yokohama City, Japan

Dysregulated tumor cell proliferation is a cancer hallmark. This study aims to uncover the cell cycle dynamics of individual melanoma cells within their complex microenvironment. We have developed a model to visualize melanoma cell cycle dynamics in real-time in vitro and in vivo. Cells transfected with fluorescence ubiquitination cell cycle indicator (FUCCI) plasmids appear red in G1, yellow in S and green in S/G2/M-phase with a fluorescence gap during cytokinesis. FUCCI-melanoma cells were grown as 3D-spheroids and implanted into a collagen matrix to mimic tumor architecture and microenvironment, or as xenografts in NOD/SCID mice.

In our 3D-spheroids, initially the ratio of red:green melanoma cells was roughly equal and distributed randomly. Within hours the interior cells became slow-moving and arrested in G1, while peripheral cells were cycling and highly motile, invading cell cycle phase-independently the collagen away from the spheroid. We sorted and cultured the interior cells in 2D. Live-cell imaging revealed that the G1 arrested population re-entered the cell cycle, however with a slight lag compared to the ‘outer’ population. This novel FUCCI-melanoma spheroid model has allowed us to mimic conditions that occur in tumors in vivo and determine their effects on cell cycle dynamics.

Intravital multiphoton microscopy of FUCCI-melanoma xenografts visualizing cell motility relative to intact tumor vasculature in live mice revealed that the tumors can be divided into two groups: Xenografts derived from MITFhigh melanoma lines proliferate heterogeneously throughout the tumor, while MITFlow lines proliferate predominantly at the tumor periphery and in close proximity to capillaries, while most cells further away from oxygen and nutrient supply arrest in G1.

Our data suggest that MITF expression dictates the subcompartment-specific distribution of differentially cycling tumor cells in melanoma, which may result in differential sensitivity to apoptosis and therefore may contribute to the resistance of melanoma to therapy.


UV induces adult keratinocyte-derived factors that regulate melanocyte phenotype

M. Heppt1, A. Santiago-Walker1, G. Zhang1, C. Perlis2, N. Soussi-Yanicostas3, G. Saintigny4, C. Mahé4, M. Herlyn1, M. Fukunaga-Kalabis11The Wistar Institute, Philadelphia, PA, USA; 2Fox Chase Cancer Center, Philadelphia, PA, USA; 3Centre de Recherche de l’Institut du Cerveau et de la Moëlle épinière, Paris, France; 4Chanel Parfum Beauté, Neuilly/Seine, France

Major phenotypes of melanocytes including melanin production, proliferation, migration and adhesion are regulated by adjacent keratinocytes. Elucidating the crosstalk between keratinocytes and melanocytes is key for our understanding of pigmentation in human skin. In this study, we hypothesized that upon UV exposure, human adult skin-derived keratinocytes control melanocytes in a manner distinct from neonatal skin-derived keratinocytes. We isolated keratinocytes from human adult skin and neonatal skin and performed global gene expression analyses after UVB irradiation. Utilizing EDGE software, we selected molecules whose patterns of gene expression were different between adult and neonatal keratinocytes over a time course after the UVB irradiation. Among those molecules, GM-CSF was significantly upregulated after the UVB irradiation in adult keratinocytes compared to neonatal keratinocytes. We observed that the recombinant GM-CSF treatment induced morphological changes and increased tyrosinase expression in human neonatal melanocytes in vitro. Furthermore, we confirmed that the expression of the protein anosmin-1, encoded by the KAL1 gene, was upregulated in adult keratinocytes after UVB irradiation. Anosmin-1 is known to regulate proliferation and migration of neuronal cells during CNS development. We then generated a lentiviral vector for overexpression of anosmin-1 in keratinocytes in order to characterize the role of the protein in human skin. In a transwell assay, melanocytes increased migration toward conditioned medium from anosmin-1-expressing keratinocytes compared to condition medium from control keratinocytes. Co-culturing melanocytes with anosmin-1-expressing keratinocytes increased dendrite formation in melanocytes. Our study identified novel adult keratinocyte-derived molecules involved in controlling melanogenesis upon UV exposure.


Optimization of endoplasmic reticulum (ER) stress-inducing therapies for melanoma using real-time cell cycle imaging

D. S. Hill1, A. Anfosso1, I. Kinjyo1, O. Kanagawa2, P. E. Lovat3, W. Weninger1,4,5, N. K. Haass1,4,5

1 Centenary Institute, Newtown, NSW, Australia; 2Research Center for Allergy and Immunology, RIKEN, Yokohama City, Japan; 3Dermatological Sciences, Newcastle University, Newcastle upon Tyne, UK; 4Discipline of Dermatology, University of Sydney, Sydney, NSW, Australia; 5Department of Dermatology, Royal Prince Alfred Hospital, NSW, Australia

We have demonstrated that targeting the endoplasmic reticulum with fenretinide (synthetic retinoid) or velcade (26S proteosome inhibitor) induces cell cycle arrest and apoptosis of metastatic melanoma cells in vitro and in vivo. This study aims to investigate the effect of ER stress-inducing agents on the dynamics of cell division and cell death of individual melanoma cells within the complex tumor microenvironment.

We have developed a novel model to visualize cell cycle dynamics in melanoma cells, in real-time, in vitro and in vivo. Cells transfected with fluorescence ubiquitination cell cycle indicator (FUCCI) plasmids appear red in G1, yellow in S and green in S/G2/M phase with a fluorescence gap during cytokinesis. FUCCI-melanoma cells were grown as 3D spheroids and implanted into a collagen matrix to mimic tumor architecture and microenvironment, or as xenografts in NOD/SCID mice. Using confocal and intravital multiphoton microscopy respectively, we observed in 3D spheroids in vitro and in xenografts in vivo, that melanomas are composed of differentially cycling tumor cells in a subcompartment-specific distribution, which we hypothesis may result in differential sensitivity to apoptosis. Flow cytometry and confocal microscopy indicated that treatment of FUCCI-melanoma cells with fenretinide or velcade induced G1 or G2 accumulation respectively, in 2D culture over the course of 24 h. Interestingly, while spheroids treated with low concentrations of velcade accumulated in G2 phase after 24 h, by 72 h the majority of cells appeared to be in G1 phase, suggesting cell cycle synchronization and indicating that cells which do not undergo apoptosis can recover and re-enter the cell cycle. Furthermore, combined treatment caused synergistic cytotoxicity in 2D and in 3D culture. Subsequent experiments utilizing our established in vivo FUCCI-model will optimize the fenretinide/velcade combination treatment for metastatic melanoma.


A Phase Ib/II study of lenvatinib (E7080), a VEGFR and FGFR tyrosine kinase inhibitor (TKI), in combination with dacarbazine (DTIC) versus DTIC alone as first-line therapy in patients with stage IV melanoma: Phase Ib safety and efficacy results

E. Calvo1, C. Becerra2, M. Maio3, E. Marshall4, P. Lorigan5, J. Stephenson6, M. Middleton7, P. Nathan8, M. Del Vecchio9, P. Ascierto10, A. Testori11, H. Glen12, A. M. Arance13, P. Lopez Criado14, J. C. Hassel15, W. Harth16, E. Meek17, L. Carmes18, J. P. Hodge19

1 START Madrid, Centro Integral Oncologico Clara Campal, Madrid, Spain; 2Sammons Cancer Center, Dallas, TX, USA; 3Medical Oncology and Immunotherapy, University Hospital, Siena, Italy; 4Clatterbridge Centre for Oncology, Wirral, UK; 5The Christie, Manchester, UK; 6Hematology and Oncology Associates of SC, Greenville, SC, USA; 7Churchill Hospital, Oxford Cancer Center, Oxford, UK; 8Mount Vernon Hospital, Northwood, UK; 9Fondazione IRCCS Istituto Milano, Italy; 10Istituto Nazionale Tumori, Napol, Italy; 11IEO Istituto Europeo di Oncologia, Milano, Italy; 12Beatson West of Scotland Cancer Center, Glasgow, UK; 13HU Clinic i Provincial de Barcelona, Barcelona, Spain; 14Centro Oncologico MD Anderson, Madrid, Spain; 15Universitaetsklinikum Heidelberg, Heidelberg, Germany; 16Vivantes Klinikum Spandau, Berlin, Germany; 17Quintiles Transnational, Livingston, West Lothian, UK; 18Quintiles Transnational, Paris, France; 19Quintiles Transnational, Research Triangle Park, NC, USA;

E7080 is an oral tyrosine kinase inhibitor (TKI) that targets VEGFR1–3, FGFR1–4, RET, KIT, and PDGFRβ. In Phase I studies of E7080, 5/14 (35.7%) patients with metastatic melanoma receiving E7080 QD had tumor reductions of 20–85%, including two confirmed PRs. Median PFS was 7 months (range 3.5–12.2). Based on these results, the current trial was initiated to evaluate E7080 and dacarbazine (DTIC) in first-line metastatic melanoma. This is a two-part study: Phase Ib to determine safety, tolerability, and MTD [safety run-in with three doses (16, 20, 24 mg QD) of E7080 combined with DTIC (1000 mg/m2) Q 21 days]; and a randomized Phase II portion to assess safety and preliminary efficacy (MTD of E7080 + DTIC versus DTIC alone). In Phase Ib, 16 patients (males: 56%; median age = 64 years; 100% Caucasian) with metastatic melanoma received E7080 at 16 mg QD (n = 3), 20 mg QD (n = 7), and 22 mg QD (n = 6) with DTIC. Two patients reported DLTs of Grade 3 hypertension (20 and 22 mg); one patient had febrile neutropenia/Grade 4 thrombocytopenia at 22 mg. The MTD was 20 mg QD E7080 with 1000 mg/m2 DTIC. Most common AEs were hypertension (75%), nausea (69%), and fatigue (56%). Most common Grade ≥3 AEs were hypertension (44%) and neutropenia (38%). Median PFS at MTD = 7.1 months (two patients ongoing; range 1.4–8.8); best response was PR in 25% of patients. Tumor DNA will be evaluated for mutations (e.g. BRAF, NRAS, KRAS, PI3KCA) and compared with patient outcomes.

E7080 with DTIC in patients with first-line metastatic melanoma demonstrated a tolerable and predictable safety profile, achieving PRs and prolonged PFS. In Phase II, the magnitude of treatment effect of E7080 + DTIC versus DTIC alone will be determined, and molecular features of responsive melanoma patients will be evaluated.

Partial funding for this study was provided by Eisai Inc.


CDK4 and 6 as therapeutic targets in human melanoma – just redundant proteins?

C. Briand, W. Jerney, S. Blunder, N. Schicher, H. Pehamberger, C. Hoeller

Department of Dermatology, Division of General Dermatology, Medical University of Vienna, Vienna, Austria

The cyclin dependant kinases (CDKs) are proteins that promote cell cycle progression. CDK4 and CDK6, active at the G1-checkpoint, were mostly seen as redundant proteins but recent evidence points to independent and important functions in various experimental models. The aim of this study was to identify the specific contribution of CDK4 and CDK6 to growth and invasiveness of melanoma and to explore a possible role for inhibition by the specific CDK4/6 inhibitor PD033299 for melanoma.

We first tested the influence of increasing concentration of PD0332991 on growth (MTS assays) and viability of 518A2 (BRAFV600E), M24 (NRAS Q61R) and SKMel28 (BRAF V600E and CDK4 R24C) melanoma cells in vitro. We observed IC50 values at approximately 11 nM after 24 and 48 h, and a growth stop at ~10 nM PD033299.

Following treatment with si- or shRNAs a significant downregulation of the protein expression of CDK4 or CDK6 was observed. siRNA treatment led to a significant decrease in cell proliferation after downregulation of each CDK. Furthermore, the migration of cells in scratch assays as well as in transmigration assays was clearly reduced indicating that the loss of one of these G1-kinases is not compensated by the other, as has been indicated in the literature.

Further validation in a human melanoma xenotransplantation mouse model using shRNA transduced melanoma cells, as well as PD033299, is being performed. Promising preliminary results demonstrate that downregulation of either CDK is associated with reduced tumor growth and regression of tumors at an early stage. This supports a possible role of CDK4/6 inhibition as part of a treatment strategy for metastatic melanoma.


Identification of differentially expressed genes in matched formalin-fixed paraffin-embedded primary and metastatic melanoma tumor pairs

R. Jewell1, A. Mitra1, C. Conway1, J. Iremonger2, C. Walker1, F. de Kort3, M. Cook4, A. Boon5, V. Speirs2, J. Newton-Bishop1

1 Section of Epidemiology and Biostatistics, Leeds Institute of Molecular Medicine (LIMM), Leeds, UK; 2Section of Pathology and Tumour Biology, LIMM, Leeds, UK; 3ServiceXS, B. V. Leiden, The Netherlands; 4Royal Surrey County Hospitals NHS Trust, Surrey, UK; 5Leeds Teaching Hospitals NHS Trust, UK

Comparison of gene expression profiles between a primary melanoma and a matched early metastatic specimen will provide essential biological insight into early metastatic processes. In this study, the DASL (cDNA mediated annealing, selection, extension and ligation) assay has been used to identify the lowest limits of reliable gene expression detection in very small formalin-fixed sentinel node biopsy (SNB) melanoma samples. Furthermore, gene expression by nodal metastases was compared with that in the matched formalin-fixed primary tumor. Tissue was sampled from twenty-five melanoma deposits within the sentinel node, using laser capture microdissection. Expression of 502 genes was assessed using the Human Cancer panel. The mean number of genes detected using DASL was used as a measure of quality and was lower in SNB samples than when using a core of primary melanoma tumor (242 versus 434 genes). Area of tissue microdissected, RNA concentration and qRT-PCR quality control did not predict performance of samples on the array but age of sampled tissue negatively correlated with number of genes detected (P = 0.01). For samples that performed successfully with the assay, matched primary samples were available for 10 samples. Gene expression profiles correlated between all matched tumor pairs (Spearman's rho 0.15–0.80, P < 0.01), although a number of genes were differentially expressed between nodal and primary tumors in all tissue pairs, including fibroblast growth factor 3 (FGF3) and FGF8.

The DASL assay can be used to generate gene expression data from small formalin-fixed samples, but not consistently. Increasing tissue age reduced the number of genes detected. Differentially expressed genes were identified across 10 matched primary and nodal tumor pairs suggesting that the DASL assay could be used to derive essential biological information about early metastasis from primary and SNB samples.


XL184: c-Met inhibition is effective in a mouse xenograft model of metastatic uveal melanoma

I. Yeh, K. Griewank, V. Ding, B. Bastian

Departments of Pathology and Dermatology, University of California, San Francisco, CA, USA

Oncogenic mutation of GNAQ is an early event in uveal melanoma development. Identification of critical downstream effectors of GNAQ could allow for targeted therapy of this deadly form of melanoma. We identified c-Met upregulation as a downstream effect of GNAQ Q209L mutation by analyzing expression changes in immortalized mouse melanocytes transduced with GNAQ Q209L.

XL184 is a small-molecule tyrosine kinase inhibitor that has activity against c-Met. In vitro assays of XL184 demonstrate activity against human uveal melanoma cell lines (OMM1.3 and Mel202) with GNAQ activating mutations. Intravenous injection of 1 × 106 OMM1.3 cells into immunocompromised NSG mice results in the development of multiple liver tumor nodules after 6–8 weeks, recapitulating liver tropism seen in humans. We treated mice starting 4 weeks after OMM1.3 intravenous injection with XL184. After 4 weeks of treatment, the treated mice demonstrated decreased liver tumor size with central tumor necrosis on histologic examination. The activity of XL184 is likely through its inhibition of c-Met signaling as we found little expression of the other targets of XL184 in OMM1.3. In conclusion, XL184 demonstrates activity in a mouse xenograft model of metastatic uveal melanoma, and has potential to be active against hepatic metastases of uveal melanoma, the major cause of mortality in this rare cancer.


Characterization of Notch signaling pathway in dermal stem cells

M. Fukunaga-Kalabis, D. Hristova, M. Heppt, L. Li, M. Herlyn

The Wistar Institute, Philadelphia, PA, USA

Recently, multipotent stem cells with neural-crest-likecharacteristic have been identified in the dermis of human skin. These dermal stem cells (DSCs) are capable of forming spheres, they expressthe neural crest marker NGFRp75, display extensive self-renewal capacity anddifferentiate into several derivatives of neural crest, includingmelanocytes and neural cells. In the 3D skin reconstruct model, intradermally embedded DSCs can differentiate into melanocytes,which migrate to the basal epidermis. These findings suggest that stem cells in the dermis constitutea reservoir for epidermal melanocytes and they may be a potential target of melanomagenesis. So far, little is known about the mechanisms of self-renewal of DSCs, their migration from the dermis to the epidermis, and differentiation to melanocytes. Here, we examined the role of Notch signaling on self-renewal of DSCs since the activation of Notch signaling has been associated with melanoma development and the survival of melanocyte stem cells and neural stem cells. In DSCs, high level of activated Notch was observed similarly to melanoma spheres. Among Notch target genes, HES1 and HES4 were highly expressed in DSCs. The level of expression of these genes were rapidly downregulated during melanocyte differentiation. Inhibition of Notch signaling by gamma-secretase inhibitor reduced the formation of DSC spheres. These data indicate that Notch signaling contributes to maintenance of DSCs.


Altering melanocyte and melanoma lineage identity for differentiation therapy

C. K. Kaufman1, C. Mosimann2, K. Lawson2, B. Wang2, L. I. Zon1,2

1 Department of Medical Oncology, Dana Farber Cancer Institute, Boston, MA, USA; 2Stem Cell Program and Division of Hematology/Oncology, Children's Hospital Boston, Howard Hughes Medical Institute, Harvard Medical School, Boston, MA, USA

The specification of melanocytes from neural crest precursor cells requires the activity of the transcription factor micropthalmia-associated transcription factor (mitf). In addition to directing formation of melanocytes, the normal functions of mitf are also linked as a lineage oncogene to the abnormal properties of melanoma. We are working to better understand how to alter melanocyte and melanoma lineage character and test the hypothesis that mechanisms that disfavor formation and survival of melanocytes will also inhibit melanoma formation and growth. Transgenic zebrafish, Danio rerio, that express the BRAFV600E oncogene under the control of the mitf promoter develop melanoma in a p53−/−background. As expected, zebrafish lacking mitf (termed nacre mutants) fail to form melanocytes and thus melanoma. By rescuing mitf expression with a mitf minigene and linking this rescue construct to a second transgene in cis, we are able to guarantee misexpression of our chosen second test transgene in all melanocytes that then develop. Using this approach, we are misexpressing neural crest transcription factors to attempt to skew neural crest cells away from the melanocyte lineage or reprogram committed melanocytes to a different but related neural crest lineage (e. g. glial/Schwann cell) and determine the effects on melanoma formation. Initial results indicate that misexpression of glial or ‘anti-melanocyte’ transcription factors delays the onset of de novo melanoma tumors in our zebrafish model. We are working to show evidence of cell identity alterations in these melanocytes/melanomas and to determine the effects of transcription factor misexpression on previously formed melanomas. We anticipate that this approach of differentiation therapy will add an important additional element to traditional immunologic and oncogene-targeted therapies for melanoma.


Primary cutaneous Hgf-Cdk4R24C mouse melanomas escape T-cell-mediated immunosurveillance through reversible dedifferentiation in an inflammatory microenvironment

J. Kohlmeyer, J. Landsberg, M. Renn, T. Bald, M. Cron, S. Mikus, A. Sporleder, T. Tüting

Laboratory of Experimental Dermatology, Department of Dermatology and Allergy, University of Bonn, Bonn, Germany

Antigen-specific T-cells can cause regression of primary melanomas but tumors frequently recur and patients die of progressive metastatic disease. According to the cancer immunosurveillance theory immune cells initially recognize and impair malignant transformation, but eventually fail to control tumor growth due to the emergence of immune escape variants. Here we experimentally studied this process of immunoediting in primary melanomas of Hgf-Cdk4R24C mice that grow progressively after a period of immunosurveillance by adoptively transferred TCRtg pmel-1 T-cells recognizing the melanosomal protein gp100. Cohorts of carcinogen-exposed Hgf-Cdk4R24C mice bearing small primary melanomas in the skin received naïve pmel-1 T-cells that were activated in vivo with the recombinant adenovirus Ad-gp100. Repetitive booster vaccinations maintained T-cell effector functions and delayed tumor progression for several weeks leading to significantly prolonged survival. Detailed histopathologic investigations of melanomas developing after long-term T-cell-mediated immunosurveillance revealed the appearance of hypomelanotic tumor areas with increased immune cell infiltration. Experiments with a transplantable Hgf-Cdk4R24C melanoma cell line recapitulated key features of primary cutaneous parental tumors in vivo. Importantly, these tumors also grew progressively after prolonged T-cell-mediated immunosurveillance and developed pronounced depigmentation with infiltration of myeloid cells. Hgf-Cdk4R24C melanoma cells from depigmented tumors showed a dedifferentiated phenotype with a strongly reduced ability to stimulate pmel-1 T-cells in vitro. Unexpectedly, this dedifferentiated phenotype was reversed upon transplantation in vivo and could again be induced by adoptively transferred T-cells and long-term control of tumor growth. In vitro experiments are ongoing to address the hypothesis that exposure of pigmented Hgf-Cdk4R24C melanoma cells to inflammatory mediators decreases the expression of melanocytic antigens and the ability to stimulate pmel-1 T-cells. Taken together, our results suggest that primary cutaneous Hgf-Cdk4R24C mouse melanomas escape pmel-1 T-cell-mediated immunosurveillance through reversible dedifferentiation in an inflammatory microenvironment. This establishes a novel, previously not described principle of immune escape for primary melanoma.


Health related quality of life (HRQL) of patients receiving ipilimumab with dacarbazine as first-line treatment for unresectable stage III/IV melanoma

S. Kotapati1, S. Francis2, B. Sherrill3

1 Bristol-Myers Squibb, Wallingford, USA; 2Bristol-Myers Squibb, Braine.l’Allued, Belgium; 3RTI Health Solutions, Research Triangle Park, NC, USA

Ipilimumab (IPI) is a monoclonal antibody which improved survival with minimal negative impact on HRQL in patients with previously treated unresectable stage III/IV melanoma. The current Phase III study compared DTIC (a global standard) with or without IPI (10 mg/kg) in untreated patients with advanced melanoma and also showed survival improvement. This analysis summarizes the HRQL outcomes during the current trial.

Patients were randomized to receive DTIC + IPI (n = 250) or DTIC + placebo (n = 252) for 12 weeks, followed by 12 weeks of DTIC only. Patients eligible for the maintenance phase continued to receive IPI or placebo (pla) dosing. EORTC-QLQ-C30 scores were assessed at baseline, week 4, 7, 12 then every 12 weeks until disease progression. Change in scores were considered as ‘no change’ (0–5), ‘a little’ (5–10 points), ‘moderate’ (10–20 points) and ‘very much’ (>20).

At week 12, unadjusted mean changes from baseline in both treatment groups were ‘no change’ to ‘moderate’ for all HRQL domains including symptom scores. Both groups experienced a little decline in average global health status (GHS) scores (DTIC + pla = −6.5; IPI + DTIC = −10.0). At week 24, change in GHS was −6.5 in the DTIC + pla group (n = 83) and −4.4 in the IPI + DTIC group (n = 65); at week 36, the change was −5.4 in the DTIC + pla group (n = 36) and +3.0 in the IPI + DTIC group (n = 45); and at week 48, the change was −1.8 in the DTIC + pla group (n = 26) and +2.2 in the IPI + DTIC group (n = 30). Longitudinal analyses using repeated measures modelling did not detect statistically significant differences between treatment groups for any HRQL scales or symptom scores (GHS P = 0.19).

During the first 12 weeks of dosing, small to moderate decline in HRQL scores were observed in both treatment arms. Most HRQL scales and symptom scores were similar to baseline for patients that continued to provide HRQL assessments at week 24, 36 and 48.


Targeting ERK in BRAF-V600E mutant melanomas

C. Krepler1, A. Samatar2, Y. Chen1, M. Halloran1, M. Samanta1, X. He1, B. Wubbenhorst3, A. Vultur1, K. L. Nathanson3, M. Herlyn1, J. Villanueva1

1 The Wistar Institute, Philadelphia, PA, USA; 2Merck Laboratories, Kenilworth, NJ, USA; 3Division of Medical Genetics and The Abramson Cancer Center, University of Pennsylvania School of Medicine, Philadelphia, PA, USA

The discovery of activating mutations in BRAF, mostly BRAFV600E, in approximately 50% of melanomas has led to the development of targeted inhibitors of the BRAF/MAPK/ERK pathway. The first inhibitor of this class, vemurafenib (Zelboraf®), to be approved by the FDA for the treatment of advanced melanoma has shown promising results in clinical trials highlighting this pathway as a suitable target for melanoma therapy. A host of alternative targeted agents against kinases in this pathway are currently in pre-clinical or clinical development. It is becoming increasingly clear, that although ~ 50% of patients with an activating BRAF mutation respond to selective BRAF inhibitors, the other 50% do not; furthermore, all patients who respond to BRAF inhibitors eventually develop resistance. In the present study, we report that melanoma cell lines with a BRAFV600E mutation but with intact expression of PTEN (PTEN+) are sensitive to a novel ATP-competitive MAPK/ERK inhibitor in vitro and in vivo. On the other hand, BRAFV600E mutant melanoma cell lines that do not express PTEN (PTEN-) are less sensitive to MAPK/ERK inhibitors. PTEN- melanoma cell lines also exhibit higher levels of phospho-AKT, and combination regimes of PI3K and ERK inhibitors do indeed lead to decreased viability. These findings support the mounting evidence that PTEN status can have a profound influence on the impact of MAPK pathway inhibitors in malignant melanoma. Therefore, it might be advantageous to determine the presence or absence of functional PTEN in tumors of patients treated with BRAF/MEK/ERK inhibitors. Moreover, it is essential to develop effective combination strategies to treat patients refractory to inhibitors of the MAPK pathway.


Using protein arrays and system biology to identify targets of leelamine

O. F. Kuzu1, R. Gowda1,2, S. V. Madhunapantula1,2,3, A. Sharma1,2,3, G. P. Robertson1–7

1 Department of Pharmacology, The Pennsylvania State University College of Medicine, Hershey, PA, USA; 2Penn State Melanoma Center, The Pennsylvania State University College of Medicine, Hershey, PA, USA; 3Penn State Melanoma Therapeutics Program, The Pennsylvania State University College of Medicine, Hershey, PA, USA; 4Department of Surgery, The Pennsylvania State University College of Medicine, Hershey, PA, USA; 5Department of Pathology, The Pennsylvania State University College of Medicine, Hershey, PA, USA; 6The Foreman Foundation for Melanoma Research, The Pennsylvania State University College of Medicine, Hershey, PA, USA; 7Department of Dermatology, The Pennsylvania State University College of Medicine, Hershey, PA, USA

Melanoma is the deadliest and most metastatic form of skin cancer. Vemurafenib, a selective mutant B-Raf inhibitor that inhibits specifically Erk signaling pathway, is just approved by the FDA for the treatment of late stage metastatic melanoma. Vemurafenib has an 81% response rate, but drug resistance develops few months after treatment. Thus, there is an urgent need for the discovery of new therapeutic agents, which would inhibit multiple disease driving pathways to reduce the potential of drug resistance. To identify compounds acting in this manner, a natural compound library was screened through which leelamine, was identified as a potential agent for advanced stage melanoma. In this study, we unraveled the signaling pathways that are targeted by leelamine in melanoma cells via protein array analysis followed by systems biology approaches. Data demonstrated that leelamine mediated tumor inhibition was through inhibition of multiple major signaling pathways involved in melanoma development, including AKT3, STAT3 and ERK. Further analyses with receptor tyrosine kinase arrays identified that leelamine reduced the activity of several receptor tyrosine kinases but most notably platelet derived growth factor receptor (PDGFR). Analyses with GeneGo's MetaDrug tool identified 5-lipoxygenase as a potential target for leelamine. Furthermore, baicalein, an inhibitor of lipoxygenase pathway, rescued melanoma cells from leelamine induced cell death. This pathway inhibits PDGFR stimulated cell proliferation and G1 to S transition via modulation of the PI-3-kinase pathway. Therefore, we speculate that leelamine might be acting through regulation of lipoxygenases. Since, pathways that are regulated by leelamine are extensively involved in melanoma development, this drug has the potential to overcome resistance that quickly develops when an agent targets a single protein or pathway.


Cadherin switch: a prognostic marker in melanoma

J. Lade-Keller1, R. Riber-Hansen1, P. Guldberg2, H. Schmidt3, S. J. Hamilton-Dutoit1, T. Steiniche1

1 Institute of Pathology, Aarhus University Hospital, Aarhus, Denmark; 2Institute of Cancer Biology, Danish Cancer Society, Copenhagen, Denmark; 3Department of Oncology, Aarhus University Hospital, Aarhus, Denmark

Cadherins are important cell adhesion molecules involved in both cell-cell adhesion and cell signaling. Several studies have proposed a shift in melanoma cell expression from E- cadherin to N-cadherin during tumor development, as a result of which melanoma cells lose contact with keratinocytes and adhere to stromal fibroblasts and endothelial cells. Our objective was to examine the expression of E-cadherin and N-cadherin in a large melanoma cohort and to correlate the results with clinical data.

Immunohistochemical expression of E-cadherin and N-cadherin was examined using tissue microarrays of 404 melanomas (median Breslow thickness, 1.14 mm; 18% ulcerated tumors) with a median follow-up of 5 and 8 years, for distant-metastasis-free survival and overall survival, respectively.

The percentage of positive cells was estimated semi-quantitatively and grouped as low or high expression (E-cadherin: <50% = low, ≥50% = high; N-cadherin: <5% = low, ≥5% = high). In addition, a cadherin profile was created grouping low E-cadherin and high N-cadherin expression (loE/hiN) against any other combinations of the two cadherin reaction patterns. The loE/hiN profile was strongly associated with increasing primary tumor thickness (P = 0.004) and the presence of ulceration (P = 0.009). In univariate analyses, overall survival (HR: 1.9; 95%CI: 1.1–2.8; P = 0.012), melanoma-specific survival (HR: 2.1; 95%CI: 1.0–4.2; P = 0.039) and distant-metastases-free survival (HR: 2.1; 95%CI: 1.1–4.1; P = 0.033) were significantly lower in patients with the loE/hiN profile compared to patients without this profile. Multivariate analyses with adjustments for primary tumor thickness and ulceration showed the loE/hiN profile to be an independent prognostic factor for distant-metastases-free survival (HR: 2.0; 95%CI: 1.0–3.9; P = 0.049).

Our findings identify cadherin expression profile (loE/hiN) as a predictor for distant- metastasis-free survival, but not for overall relapse-free survival. Analysis of factors that determine lymphatic and hematogenous metastatic routes in melanoma may be a fruitful area for future study.


On the interplay of telomere-related genes, nevi and the risk of melanoma

C. Bodelon1, R. M. Pfeiffer1, D. Calista2, M. C. Fargnoli3, G. Bianchi-Scarra4, P. Ghiorzo4, M. T. Landi1

1 Division of Cancer Epidemiology and Genetics, National Cancer Institute, Bethesda, MD, USA; 2M. Bufalini Hospital, Cesena, Italy; 3University of L’Aquila, L’Aquila, Italy; 4University of Genoa, Genoa, Italy

Although common melanocytic nevi and dysplastic nevi are strong melanoma risk factors, only a low proportion of subjects carrying multiple nevi develop melanoma. Allelic variation in genes regulating telomere function may increase the risk of progression from nevi to melanoma, after bypassing the senescence state.

We conducted a pooled analysis from three case-control studies and one family study from Italy, totaling 794 melanoma cases and 770 unaffected individuals, to examine the association between 517 SNPs mapping to 39 telomere-related genes and melanoma risk. In addition, we explored whether these genes were related to nevi count or the presence of dysplastic nevi in controls. Genotyping was performed using Illumina Infinium custom arrays. SNP association analysis was performed using logistic regression adjusting for age and sex. The role of body mass index was also explored. Meta-analysis of all studies was performed using a random effect model. Gene-based P-values were computed using an adaptive combination of SNP-based P-values.

Preliminary results suggest that RECQL4, a gene involved in genome stability, is associated with melanoma risk (gene-based P-value = 0.002). The C allele of the most significant SNP in this gene, rs2721173, increased the risk of melanoma (Odds Ratio (OR): 1.36, 95% Confidence Interval (CI): 1.13–1.63). RTEL1, a gene regulating telomere elongation, was associated with the presence of dysplastic nevi in unaffected subjects (gene-based P = 0.017), with a two-fold increased risk for those carrying the G allele of the most significant SNP in this gene, rs6011002 (OR = 2.19, 95% CI: 1.32–3.65). We are currently examining whether variation in telomere-related genes modifies the association between nevi count and/or dysplastic nevi with melanoma risk.

In conclusion, we found suggestive evidence that genes regulating telomere are associated with melanoma risk and dysplastic nevi. If confirmed, these results may identify subgroups of subjects who would benefit from more intensive screening.


PKCɛ regulates ATF2 nuclear localization and availability to alter mitochondrial permeability following genotoxic stress

E.Lau1, H. Kluger2, T. Varzano1, David Rimm2, Z. A. Ronai1

1 Signal Transduction Program, Sanford-Burnham Medical Research Institute, La Jolla, CA, USA; 2Department of Medicine, Yale University, New Haven, CT, USA

Activating Transcription Factor 2 (ATF2) is an AP-1 family transcription factor that requires phosphorylation by JNK/p38 and by ATM for transcriptional and DNA damage response activities, respectively. ATF2 exhibits primarily nuclear localization in malignant melanoma, contrary to its more cytoplasmic localization in non-malignant squamous/basal skin carcinomas (SCC/BCC). Intriguingly, genetic inactivation of ATF2 in melanocytes inhibits melanoma development, whereas its inactivation in keratinocytes results in increased skin papillomas, indicating a tumor suppressor function in non-malignant skin tumors. Here, we report that genotoxic stress promotes ATF2 translocation from nucleus to mitochondria. In contrast to primary melanocytes, keratinocytes, and SCC cells, such mitochondrial translocation of ATF2 is impaired in malignant melanoma cells. We identified a novel PKCɛ phosphorylation site (threonine 52) on ATF2, which causes its nuclear localization. Inhibition of PKCɛ or expression of non-phosphorylatable phosphomutant (T52A) causes its nuclear export and localization at the mitochondria. Biochemical and immunocytochemical analyses confirmed ATF2 association with Hexokinase-1 (HK1) and Voltage-Dependent Anion Channel 1 (VDAC) – two mitochondrial outer membrane, identified by mass spectrometry as ATF2 interactors, implicated in control of mitochondrial membrane potential. Notably, expression of ATF2T52A perturbs HK1:VDAC1 complexes and mitochondrial membrane integrity, whereas ATF2T52E confers resistance to genotoxic stress and PKCɛ inhibition. Similarly, melanoma cells expressing relatively high PKCɛ and pT52-ATF2 levels exhibit greater resistance to genotoxic stress. Melanoma tumor tissue microarray (TMA) revealed correlation between PKCɛ expression and poor prognosis in metastatic cohorts. Our findings identify the role of PKCɛ in regulation of ATF2 subcellular localization and its availability to contribute to genotoxic stress induced cell death.


A conditional mutation implicates zebrafish mitfa in melanoma development and maintenance

J. A. Lister1,2, J. Richardson3,4, Z. Zeng3,4, M. Mathers3,4, E. E. Patton3,4

1 Massey Cancer Center; 2Department of Human and Molecular Genetics, Virginia Commonwealth University, Richmond, VA, USA; 3MRC Human Genetics Unit, Institute of Genetics and Molecular Medicine; 4University of Edinburgh Cancer Research UK Centre, Edinburgh, UK

The microphthalmia-associated transcription factor, MITF, has been described as the master regulator of the melanocyte lineage from which melanoma arises. The contribution of MITF to the development and progression of melanoma has only recently begun to be fully appreciated, and much about this involvement is still unclear. Recent evidence suggests that modulation of MITF activity at different stages of melanoma may play a key role in the progression of this cancer. While expression of a BRAF (V600E) transgene in the zebrafish melanocyte lineage cooperates with tp53 loss of function to generate melanoma, we have found that hypomorphic mutations in zebrafish MITF (mitfa) can cooperate with the BRAF (V600E) transgene to induce melanoma formation in the absence of germline tp53 mutations. Moreover, the histopathology of mitfa;Tg-BRAF (V600E) tumors is distinct from that of tp53;Tg-BRAF (V600E) tumors. By exploiting a temperature-sensitive mitfa allele to vary levels of MITF activity, we find that melanomagenesis is inhibited in fish raised at a restrictive temperature (low MITF activity), and that melanomas that arise in fish at permissive temperature (moderate MITF activity) arrest or regress when these fish are shifted to restrictive temperature. These data suggest that phenotype switching through modulation of MITF activity may be involved in melanoma initiation, and provide in vivo support for the role of MITF/mitfa as a lineage-survival oncogene.


MAGE transcription factors promote oncogenesis through multiple mechanisms

B. J. Longley1,2, T. Z. Xiao1, N. Bhatia1, Q. Meng3, G. A. Lomberk4, R. Urrutia4

1 Department of Dermatology, University of Wisconsin School of Medicine and Public Health, Madison, WI, USA; 2Paul P. Carbone Comprehensive Cancer Center, University of Wisconsin School of Medicine and Public Health, Madison, WI, USA; 3McArdle Laboratory, University of Wisconsin School of Medicine and Public Health, Madison, WI, USA; 4Department of Molecular Neuroscience, Department of Biochemistry and Molecular Biology, and Gastroenterology Research Unit, Mayo Clinic, Rochester, MN, USA

Class I MAGE proteins (MAGE) are normally expressed only in developing germ cells but are aberrantly expressed in melanomas and many other cancers, making them near perfect therapeutic targets. MAGE expression increases with melanoma progression and correlates negatively with relapse free survival, independent of Breslow's thickness, mitotic rate, and ulceration. In a series of experiments to investigate mechanisms for MAGE effects we first found that knockdown of MAGE by doxycycline induced, lentiviral encoded shRNA increases p53 activity, apoptosis, and sensitivity to DNA double strand breaks (DSBs) in low passage human melanoma cultures, and decreases in vitro proliferation and growth of these cells in athymic nude mice. Second, we studied DSB repair, which requires ATM kinase dependent phosphorylation of KAP1, an ubiquitin E3 ligase that also binds MAGE. We found that MAGE increases ATM phosphorylation of KAP1 and increases repair of endonuclease induced DNA double strand breaks. Finally, because KAP1 is an obligate binding partner for KRAB zinc finger transcription factors (KZFTFs), which guide it to specific chromatin sites where it represses genes by inducing localized heterochromatin characterized by histone 3 lysine 9 trimethylation, we studied MAGE effects on KZFTF mediated gene repression. We found MAGE proteins control KZFTF-KAP1 binding and KZFTF poly-ubiquitination and degradation. Furthermore, ChIP and mRNA analysis showed that MAGE regulate KAP1 localization, thereby affecting local chromatin structure and transcription of ~3000 genes, controlling cascades that include INK4a/ARF, b-RAF,AP-2γ, c-MYC, and the tumor suppressor ID1. Particularly, MAGE increases b-RAF transcription by de-repressing b-RAF, and decreases tumor suppression by de-repressing ID1. Thus MAGE may affect p53, apoptosis, and DNA DSB repair as well functioning as master controllers of KZFTF-KAP1 mediated gene repression. We conclude that MAGE proteins contribute to oncogenesis through multiple mechanisms including and that interference with MAGE expression and function are important therapeutic goals.


Validation of RAB22A as target for miR-211 and its expression pattern in melanocytes and melanoma

N. Maddodi, S. Madhavan, S. Devi, V. Setaluri

Department of Dermatology, School of Medicine and Public Health, University of Wisconsin, Madison, WI, USA

MicroRNA-211 is an intronic miRNA derived from the intron of TRPM1, a gene that codes for transient receptor potential family member calcium channel. miR-211 expression correlates with expression of TRPM1, which is involved in melanocyte calcium homeostasis and melanin pigmentation. Expression of TRPM1and miRNA-211 is significantly lower in rapidly proliferative melanocytes compared to slow growing, differentiated melanocytes. Overexpression of miRNA-211 in melanocytes suppresses the growth of rapidly growing melanocytes. Similarly, overexpression of miR-211 in melanoma cell inhibits their growth and reduces their migration. Although RAB22A (member of the RAS family of small GTPases), with a target predictions score that is the highest or a score that is among the highest, is predicted to be a target for miR-211, it has not been validated as miR-211 target in melanocytes and pattern of its expression between melanocytes and melanoma cells has been investigated. In this study, using RAB22A 3′-UTR luciferase reporter and site-directed mutagenesis, we show that RAB22A is a miR-211 target and show that antagomir-mediated suppression of miR-211 decreases RAB22A mRNA expression in melanocytes. Surprisingly, although human melanoma cells lines show several hundred-fold lower miR-211 expression, RAB22A expression is also significantly low in melanoma cells. We investigated the effect of overexpression of RAB22A in melanoma cells based on our preliminary data we propose that the unexpected pattern of expression of a melanocyte-selective miR and its target(s) in melanoma should be taken into consideration in understanding the role miR-211 in melanoma biology.


Senescent cells develop a PARP-1 and nuclear factor-{kappa}B-associated secretome (PNAS)

M. Ohanna, S. Giuliano, C. Bonet, V. Imbert, V. Hofman, J. Zangari, K. Bille, C. Robert, B. Bressac-de Paillerets, P. Hofman, S Rocchi, J. F.Peyron, J. P. Lacour, R. Ballotti, C. Bertolotto

INSERM, U895, Equipe 1, Biologie et Pathologies des Mélanocytes de la Pigmentation Cutanée au Mélanome, Equipe labellisée Ligue Nationale contre le Cancer, Nice, France

Melanoma cells can enter the process of senescence, but whether they express a secretory phenotype, as reported for other cells, is undetermined. This is of paramount importance, because this secretome can alter the tumor microenvironment and the response to chemotherapeutic drugs. More generally, the molecular events involved in formation of the senescent-associated secretome have yet to be determined. We reveal here that melanoma cells experiencing senescence in response to diverse stimuli, including anti-melanoma drugs, produce an inflammatory secretory profile, where the chemokine ligand-2 (CCL2) acts as a critical effector. Thus, we reveal how senescence induction might be involved in therapeutic failure in melanoma. We further provide a molecular relationship between senescence induction and secretome formation by revealing that the poly(ADP-ribose) polymerase-1 (PARP-1)/nuclear factor-κB (NF-κB) signaling cascade, activated during senescence, drives the formation of a secretome endowed with protumoral and prometastatic properties. Our findings also point to the existence of the PARP-1 and NF-κB-associated secretome, termed the PNAS, in nonmelanoma cells. Most importantly, inhibition of PARP-1 or NF-κB prevents the proinvasive properties of the secretome. Collectively, identification of the PARP-1/NF-κB axis in secretome formation opens new avenues for therapeutic intervention against cancers.


Loss of Noxa delays melanomagenesis in a predisposed, genetically-defined mouse melanoma model

N. Mohana-Kumaran1, K. M. Lucas1, B. Roediger1, R. A. Scolyer2, W. Weninger1,3,4, J. D. Allen1, N. K. Haass1,3,4

1 Centenary Institute, Newtown, NSW, Australia; 2Tissue Pathology and Diagnostic Oncology, Royal Prince Alfred Hospital, Melanoma Institute Australia, NSW, Australia; 3Discipline of Dermatology, University of Sydney, NSW, Australia; 4Department of Dermatology, Royal Prince Alfred Hospital, Camperdown, NSW, Australia

Melanoma drug resistance is often attributed to defective apoptosis pathways. Links between the intrinsic apoptosis pathway and melanomagenesis have not been established in vivo. In this study we utilized a mouse melanoma model in which the Cdkn2a locus is knocked-out and RAS is constitutively active (Cdkn2a−/−,Tyr-HRAS; ‘control’). This mouse line was crossed to either a Noxa or a Puma-knockout line resulting in Cdkn2a−/−, Tyr-RAS, Noxa−/− (‘Noxa-KO’) or Cdkn2a−/−, Tyr-RAS, Puma−/−(‘Puma-KO’).

Three groups of 30 mice per genotype (control, Puma-KO and Noxa-KO) were compared for tumor latency, penetrance, location and growth. Fifteen mice of each group were UVB irradiated when 3 days old.

Irrespective of UVB irradiation or not, Noxa-KO significantly delayed melanomagenesis (102 days, UVB; 134 days, no UVB) compared to control (90 days, UVB; 88 days, no UVB) and Puma-KO (75 days, UVB; 81 days, no UVB). This resulted in 94% (UVB) and 81% (no UVB) tumor penetrance in Noxa-KO compared to 100% penetrance in controls and Puma-KO. Once tumors were established, Noxa-KO accelerated tumor growth compared to Puma-KO. Interestingly, controls and Puma-KO developed more tumors on the ears than Noxa-KO. Whereas limited tumor numbers on the ear were observed in non-UVB Noxa-KO, none developed in UVB irradiated Noxa-KO mice. Tumor invasion was studied by generating in vitro three-dimensional melanoma spheroids from cells isolated from the tumors. In this model spheroids derived from Noxa-KO melanomas showed inhibited spheroid invasion compared to control and Puma-KO spheroids.

Noxa-KO delayed melanomagenesis, decreased tumor penetrance and spheroid invasion but accelerated tumor growth. We speculate that loss of Noxa may be rescued by induction of other BH3-only pro-apoptotic proteins e.g. Puma, which could be responsible for the delayed melanomagenesis and limited spheroid invasion. This suggests that, once tumors are established, Noxa may not be required for melanoma progression.


A novel, natural BRAFV600E/CDK4 dual inhibitor for melanoma

S. Bhaskaran1, M. Klausner2, J. L. VandeBerg1, H. B. Nair1

1 Southwest National Primate Research Center, Texas Biomedical Research Institute, San Antonio, TX, USA; 2MatTek Corporation, Ashland, MA, USA

Challenges in targeting multiple cell signaling pathways and the emergence of drug resistance to current therapies in melanoma warrant the need for novel drug candidates that can act on multiple targets and pathways to alleviate melanoma progression. The pentahydroxyl flavone, gossypin (GP), has been reported to inhibit osteoclastogenesis through inhibition of TGF beta-activated kinase1 (TAK-1) activated nuclear factor kappa-B (NFK-B). In this study we have demonstrated that GP induces cell-cycle arrest at G1/S-phase in melanoma cell lines by inhibiting cyclin dependent kinase 4 (CDK4). We have demonstrated also that GP treatment significantly induces p21 and P27 expression, both at mRNA and protein levels. An in vitro CDK assay revealed that GP inhibits CDK4 kinase activity. The inhibitory effect was found to be superior to that of the existing pan-CDK kinase inhibitor, Roscovitine. Furthermore, GP inhibits BRAFV600E kinase activity in vitro. This conclusion was further validated by evaluation of downstream target proteins MEK and ERK. GP inhibited phosphorylation of MEK and ERK in a dose-dependent manner and down-regulated cyclin D1. Additionally, we have investigated the anti-invasive/metastatic potential of GP in a three-dimensional skin reconstruction culture consisting of A375 human melanoma cells mixed with normal human keratinocytes embedded within a collagen-constricted human fibroblast matrix. We have documented that GP is able to inhibit the invasiveness or metastatic potential of A375 cells in a dose-dependent manner to a greater extent than Roscovitine. GP inhibited cell proliferation in a variety of human melanoma cell types (A375, 1205Lu, WM793B and WM1552C, SK-MEL-31, and SK-MEL-28). Interestingly, the effective dose for inhibiting cancer cell proliferation was nontoxic to normal human melanocytes (HPM). In light of these results, GP could be a potential therapeutic molecule as a dual inhibitor for targeting BRAFV600E and CDK4 in melanoma therapy.


Novel quantification of MART1 verified Ki67 proliferation rates by automated image analysis accurately separates melanomas from nevi

P. S. Nielsen1, R. Riber-Hansen1, J. Raundahl2, T. Steiniche1

1 Department of Pathology, Aarhus University Hospital, Aarhus, Denmark; 2Visiopharm A/S, Hoersholm, Denmark

Although the Ki67 proliferation rate already may aid the diagnosis of melanoma and correlation with clinical outcome has been demonstrated, much is gained by combining Ki67 with the melanocytic marker MART1. Hereby, Ki67-positive melanocytic cells can be accurately distinguished from proliferating lymphocytes, stromal, and epithelial cells. Particularly for prognostic purposes, such improved accuracy may be essential. Yet, Ki67 quantification by manual counting is time-consuming and cumbersome. Instead, automated image analysis (AIA) enables rapid and novel quantification methods for entire tumor sections.

Our objective was to design an AIA protocol that computes the proliferation rate of Ki67/MART1 double stained melanomas and nevi and to compare automated rates with previously published manual estimates.

Whole slide images were captured from 48 melanomas and 77 nevi stained with an immunohistochemical cocktail against Ki67 and MART1. Proliferation rates were established by AIA using different output equations based on number or area.

The differences between mean indices of melanomas and nevi were significant (P < 0.0001) for all equations. And, their misclassification of melanomas and nevi were similar and comparable with manual estimates. Yet, automated rates were significantly higher than manual rates (P < 0.0001). The most favorable automated method misclassified two melanomas and ten nevi; equal to the ROC-curve area 0.9570 (0.9254–0.9886). Excluding lesions with few melanocytic cells, extremely sensitive to small variations in the Ki-67 rate, one of 42 melanomas and four of 53 nevi were misclassified; ROC-curve area 0.9906 (0.9788–1.0000).

In conclusion, excellent discrimination between melanomas and nevi was demonstrated by AIA of Ki67/MART1 double stains with diagnostic performances equal to manual estimations. Testing different definitions of automated proliferation rates, no single definition stood out; thus, a variety of definitions may be utilized. Future research into the diagnostic and prognostic capabilities of Ki67/MART1 double stains will be simplified by the novel opportunity to estimate proliferation rates automatically.


Inhibition of the PI3K-AKT signalling pathway to overcome therapy resistance in melanoma-derived brain metastasis

H. Niessner1, A. Adam2, A. Bornemann3, J. B. Honegger4, L. Quintanilla-Fend2, C. Busch1, J. Bauer1, L. Just5, K. Hoek6, C. Garbe1, F. Meier1

1 Department of Dermatology, University of Tübingen, Tübingen, Germany; 2Department of Pathology, University of Tübingen, Tübingen, Germany; 3Institute of Brain Research, Universtity of Tübingen, Tübingen, Germany; 4Department of Neurosurgery, University of Tübingen, Tübingen, Germany; 5Institute of Anatomy, University of Tübingen, Tübingen, Germany; 6Department of Dermatology, University Hospital of Zürich, Zürich, Switzerland

Brain metastases occur in over 70% of patients with metastatic melanoma and are the most common cause of death. Current therapy options are neurosurgery, radiosurgery, whole brain radiation, chemotherapy and supportive care. The median survival time for melanoma patients with brain metastasis ranges from 0.7 to 5 months. Therefore, new therapy strategies are mandatory.

In melanoma, activation of the RAF-MEK-ERK and PI3K-AKT-mTOR signaling pathways makes a decisive contribution to tumor progression and treatment resistance. Clinical studies suggest a transient effect of BRAF inhibitors in melanoma brain metastases. We asked if inhibition of these pathways would be a promising strategy for the treatment of melanoma brain metastases. We blocked both pathways at different levels and investigated the effects on viability/proliferation and survival/apoptosis of >10 newly isolated cell lines derived from melanoma brain metastases. Furthermore, immunohistochemical analyses of brain metastases from >10 melanoma patients including matched extracerebral metastases for p-ERK, ERK, p-AKT, AKT and PTEN were performed.

Growth inhibition was most pronounced with PI3K inhibitors achieving growth inhibition rates of up to 80%. Moreover, PI3K inhibitors potently induced apoptosis in cerebral metastatic melanoma cells. Immunohistochemically, p-ERK was seen predominantly at the tumor periphery of both cerebral and extracerebral metastases. Interestingly, most melanoma brain metastases were highly positive for activated AKT, whereas matched extracerebral metastases in the same patients were weakly positive or negative for activated AKT. PTEN expression was downregulated in brain metastases. PTEN expression and AKT activation in the brain appeared to be independent of BRAF and NRAS mutation status.

Together, these findings suggest that activation of AKT is relevant for the survival and growth of melanoma cells in the brain parenchyma and that inhibition of PI3K-AKT signaling may be a suitable strategy to enhance and/or prolong the antitumor effect of BRAF inhibitors in melanoma brain metastases.


Modulation of the Wnt/ROR signaling axis in melanoma by hypoxia

M. P. O’Connell, A. T. Weeraratna

The Wistar Institute, Philadelphia, PA, USA

Melanoma cells have been shown to be capable of a switch in phenotype from a proliferative epithelial phenotype, to an invasive, mesenchymal phenotype. Previous data from our laboratory implicates Wnt5A in the switch to an invasive phenotype. Wnt5A overexpression results in an increase in metastatic capacity of melanoma cells, and a decrease in melanoma differentiation antigens e.g., MART1 and gp100. These effects are modulated through the orphan tyrosine kinase receptors, ROR1 and ROR2. The role and function of ROR2 in increasing metastatic potential has already been reported.

The current study demonstrates that overexpressing ROR2 results in a loss of pigmentation in melanoma cells, a decrease in MART1 expression, a change in morphology that mimics an EMT, and an increase in invasion. Further, we demonstrate that ROR1, a close homolog of ROR2, has the opposite effect on melanoma cells, decreasing their invasive ability. ROR1 knockdown by siRNA demonstrates decreases in MART1 and GP100, and motility, and an increase in the invasive capability of melanoma cells. Treatment with recombinant Wnt5A decreased ROR1 expression, implying that downregulation of ROR1 may be an important event in increasing malignancy in melanoma. In addition, ROR2 knockdown in cells that have high levels of Wnt5A, but low levels of ROR1, increases ROR1 expression.

It is unclear what causes the switch from a proliferative, epithelial, ROR1 positive phenotype to an invasive, mesenchymal, ROR2 positive phenotype, but our data suggest that contributing factors include hypoxia. Hypoxic treatment increases Wnt5A levels, decreases MITF expression, and increases metastasis of melanoma. Furthermore, upon exposure to hypoxia, ROR1 expression is decreased in proliferative phenotype cells, and ROR2 expression is increased in invasive cells. These data suggest that changes in the ROR receptor repertoire may be initiated by hypoxic changes in the microenvironment.


Age-dependent significance of MC1R variants as risk factor for actinic damage on the back

J.Wendt, O. Schanab, S. Rauscher, M. Binder, H. Pehamberger, I. Okamoto

Division of General Dermatology, Department of Dermatology, Medical University of Vienna, Währinger Gürtel, Vienna, Austria

Only recently, we have identified site dependent signs of actinic damage (freckling, wrinkling and lentigo) as independent risk factor of melanoma. We now aimed to elucidate causal factors of actinic damage on different sites of the body in a central European population. Two thousand two hundred and fifty participants were recruited between 2008 and 2010 at the Department of Dermatology at the Medical University of Vienna. Our study revealed differences between associated factors for actinic damage on the face and on the trunk. The only risk factor that remains consistently significant for all three facial skin alterations was age, whereas alterations on the back were additionally associated with male sex, sunburns, holidays and the presence of MC1R variants. Particularly for dorsal freckling and wrinkling on the neck (R/R: OR 4.07, P = 0.004; R/r: OR 3.33, P = 0.001; R/0: OR 2.44, P = 0.001; r/r: OR 2.07, P = 0.049; versus 0/0 for freckling and R/R: OR 4.61, P = 0.001; R/r: OR 3.81, P < 0.001; R/0: OR 3.65, P < 0.001; r/r: OR 2.87, P = 0.003; r/0: OR 2.45, P < 0.001 versus 0/0 for wrinkling), the associations with MC1R variants were pronounced. Moreover, the association of MC1R variants with dorsal actinic damage was age dependent with the strongest association in the youngest group (≤50 years) compared with older groups. Our data support our previous finding that actinic alterations on the back are more indicative for past sun exposure associated with melanoma risk.


Prevalence of BRAF and NRAS mutations in primary tumours and different types of metastasis among melanoma patients

M. Colombino1, M. Capone2, A. Lissia3, A. Cossu3, C. Rubino4, V. De Giorgi5, D. Massi5, M. Maio6, S. Staibano7, O. Nappi8, E. Pagani9, A. Manca1, MC. Sini1, G. Botti2, C. Caracò2, N. Mozzillo2, P. A. Ascierto2, G. Palmieri1

1 Unit of Cancer Genetics, Institute of Biomolecular Chemistry, CNR, Sassari; Italy; 2National Cancer Institute ‘Fondazione G. Pascale’, Naples; Italy; 3Institute of Pathology, Azienda Ospedaliero Universitaria, Sassari; Italy; 4University of Sassari, Sassari Italy; 5University of Florence, Italy; 6Oncology, ‘Santa Maria alle Scotte’ Hospital, Siena; Italy; 7University ‘Federico II’ of Naples; Italy; 8Institute of Pathology, ‘Cardarelli’ Hospital, Naples; Italy; 9Istituto Dermatologico dell’Immacolata (IDI), Rome; Italy

Mutations of NRAS and BRAF genes have been identified with high frequency in nevi, cutaneous melanomas, and melanoma metastases. Prevalence of such mutations during the disease progression phases and among the different types of metastasis still remains inconclusive. A total of 275 tumour tissues from 116 melanoma patients were screened for mutations; among them, paired samples of microdissected primary melanomas (N = 92) and synchronous or asynchronous metastases (N = 156) from same patients were included. Tissue samples underwent mutation analysis by automated DNA sequencing. Secondary lesions were from: lymph nodes (LM; N = 77), skin (SM; N = 36), visceral (VM; N = 23) and brain (BM; N = 44) sites. Overall, mutations were identified in 56/95 (59%) primary melanomas [43%BRAF– 16%NRAS], 49/77 (64%) lymph nodes [48%BRAF– 16%NRAS], 22/36 (61%) subcutaneous metastases [53%BRAF– 8%NRAS], 13/23 (57%) visceral metastases [43%BRAF– 13%NRAS], and 31/44 (70%) brain metastases [48%BRAF– 23%NRAS]. Overall, a slight and not significant increase in mutation frequency after progression from primary melanoma was observed in our series: 115/180 (64%) mutated metastases [48%BRAF– 16%NRAS]. Considering the paired samples from the same patients, LM (92% consistency) and VM (96% consistency) presented a highly similar prevalence and distribution of BRAF/NRAS mutations in comparison with primary melanomas, whereas a discontinuous pattern of mutations was detected in BM (80% consistency) or, mostly, SM (75% consistency). Occurrence of distinct mutation distribution between primary melanomas and correspondent metastases suggests that independent subclones have been generated in a limited fraction of patients. Our results provide further clues about the impact of NRAS and BRAF mutations among the different stages of melanoma progression.


The heat shock protein-90 inhibitor XL888 induces apoptosis in melanoma cells with diverse BRAF inhibitor resistance mechanisms

K. H. T. Paraiso1, E. Wood1, V. W. Rebecca1, Y. Xiang1, H. E. Haarberg1, A. Sarnaik2, J. John1, J. M. Koomen1, K. S. M. Smalley1,2

1 The Departments of Molecular Oncology, The Moffitt Cancer Center & Research Institute, Magnolia Drive, Tampa, FL, USA.; 2Cutaneous Oncology, The Moffitt Cancer Center & Research Institute, Magnolia Drive, Tampa, FL, USA

The clinical use of BRAF inhibitors is being hampered by the acquisition of drug resistance whose means are varied or in some cases unexplored. Here we demonstrate that the HSP90 inhibitor XL888 overcomes BRAF inhibitor resistance mediated through PDGFR-B, acquisition of a de novo NRAS mutation, overexpression of COT, cyclin D1 amplification and those mediated through as yet unknown resistance mechanisms. In each case, XL888 treatment led to potent inhibition of cell growth and induction of apoptosis with similar results in 3D cell culture and long-term colony formation assays, while significant tumor regression was observed in an in vivo upregulated PDGFR-B model. Following XL888 treatment of the vemurafenib resistant cell line panel, quantitative increases in the HSP90 inhibition biomarker, HSP70, were also observed by liquid chromatography multiple reaction monitoring mass spectrometry (LC-MRM). Further, LC-MRM based HSP screening was shown to be a tool that could be potentially exploited in the clinical setting as it was highly sensitive for the detection of basal levels of HSP90 from patient fine needle aspirate (FNA) derived samples. Mechanistically, XL888 treatment led to degradation of the proteins known to be responsible for therapy escape (PDGFR-B, COT, CRAF, cyclin D1) as well as those involved in resistance (JNK, Mcl-1, 26S proteasome) and was associated with a significant upregulation of BIM expression. In most cases the increase in BIM expression and apoptosis induction following XL888 treatment was superior to that seen following dual MEK/PI3K inhibition. Taken together, these results show that HSP90 inhibition may be a highly effective strategy at managing the diverse array of resistance mechanisms being reported to BRAF inhibitors and may be more effective at reactivating BIM expression than combined MEK/PI3K therapy.


Functional classification of cellular proteome profiles support the identification of drug resistance signatures in melanoma cells

V. Paulitschke1, V. Haudek-Prinz2, J. Griss2, C. Gerner2, H. Pehamberger1, R. Kunstfeld1

1 Department of Dermatology, Medical University of Vienna, Vienna, Austria; 2Department of Medicine I, Medical University of Vienna, Vienna, Austria

Drug resistance is a major obstacle in melanoma treatment. Recognition of specific resistance patterns, the understanding of the patho-physiology of drug resistance and identification of remaining options for individual melanoma treatment would greatly improve therapeutic success. We analyzed drug resistance signatures by proteome profiling and a recently established bioinformatic tool chain.

The A375 melanoma cell line was identified as sensitive to cisplatin and the M24met melanoma cell line was resistant to cisplatin. In addition we make use of one established melanoma resistant cell line TMFI and the known sensitive cervix carcinoma Hela cell line. Proteome profiling shotgun analysis of cytoplasmic, nuclear and secreted protein fractions resulted in the identification of more than 2200 proteins in A375 and M24met melanoma cell line and of more than 1600 in TMFI and Hela leading to two comparable groups of resistant versus sensitive carcinoma cell lines. Classification according to GO of cell organelles and functional groups revealed in both comparison groups a major difference in the lysosomal fraction and a higher expression of proteins involved in Ca2+ ion binding, cell adhesion, MAP kinase signaling and protein transport. To discriminate resistance profiles we compared constitutively expressed proteins in the resistant with induced proteins in the sensitive cancer cell line. This strategy identified about fourty proteins which can be clearly assigned to specific processes including Ca2+ ion binding, protein transport, nucleotide binding, mRNA processing, metabolic/enzymatic processes and mitochondrial metabolism.

Performing these comparisons we created a resistance list which now will be reevaluted for the possible predictiveness for the response to cisplatin of other tumor cells.

In conclusion, proteome profiling combined with functional classification of melanoma contributes significantly to our understanding of drug resistance mechanism and might lead to a patient tailored therapy of melanoma.


First evidence for tumor cell-leucocyte fusion as an event in human cancer progression: a melanoma brain metastasis with a donor-patient hybrid genome in a bone marrow transplant patient

R. Lazova1,2, G. S. LaBerge3,4, E. Duvall4, N. Spoelstra5, V. Klump1, M. Sznol2,6, D. Cooper2,6, R. Spritz3, J. M. Pawelek1,2

1 Department of Dermatology, School of Medicine, New Haven, CT, USA; 2The Yale Cancer Center, Yale School of Medicine, New Haven, CT, USA; 3Human Medical Genetics, University of Colorado Health Sciences Center, Denver, CO, USA; 4Denver Police Crime Lab, Denver, CO, USA; 5Department of Medicine, Universioty of Colorado Health Sciences Center, Denver, CO, USA; 6Medical Oncology, Yale School of Medicine, New Haven, CT, USA

Tumor cell fusion with motile bone marrow-derived cells (BMDCs) has long been posited as a mechanism for metastasis. While there is much support for this from in vitro and animal studies it has yet to be confirmed in human cancer, as tumor and marrow-derived cells from the same patient cannot be distinguished genetically. To test the fusion possibility we have analyzed donor and patient genes in human cancers arising fortuitously after allogeneic bone marrow transplant (BMT). In two such previous studies donor genes were detected in tumor cells but there was no provision to detect patient genes and fusion could not be fully corroborated. Here we used forensic short tamden repeat (STR) length-polymorphisms to distinguish two or more individuals in the same sample. We report our first findings from a human melanoma brain metastasis arising post-allogeneic BMT. The tumor was sectioned end-to-end. Tumor cells were isolated free of leucocytes by laser microdissection. DNA was isolated from tumor cells and normal pre-BMT lymphocytes and analyzed at 14 autosomal loci and the X and Y chromosomes. While lymphocytes showed normal diploidy with balanced alleles, tumor cells showed widespread aneuploidy, with donor and patient alleles present at some loci and only donor or patient alleles at others. With few exceptions, virtually identical aberrant genetic profiles were present in different areas throughout the tumor. This signified a clonal origin of the tumor from a single fusion event between a donor BMDC and a patient cancer cell. While from only one case, the findings indicate for the first time that fusion played a role in the metastasis of a human cancer. We anticipate that the techniques used herein will pave the way for future studies on the extent and impact of BMDC-tumor fusion.


An impaired cell senescence program in nevus melanocytes likely contributes to risk for malignant progression

J. S. Pawlikowski1, T. McBryan1,2, H. Wu2, P. D. Adams1

1 Institute of Cancer Sciences, CR-UK Beatson Labs, University of Glasgow, Glasgow, UK; 2Fox Chase Cancer Center, Philadelphia, PA, USA

Cellular senescence is characterized by an irreversible arrest of cell proliferation and is initiated by various stressors such as activated oncogenes, oxidative stress and shortening of telomeres. Senescence acts as a tumor suppression mechanism to arrest the development of cancer. Of particular interest, benign human nevi are clonal hyperproliferations of melanocytes that often contain an activated BRAF or N-RAS oncogene. Research has subsequently shown that nevi contain senescent melanocytes. Consequently, it is now widely accepted that cellular senescence contributes to the barrier that prevents malignant progression of nevus melanocytes to melanoma. Nonetheless, it is thought that roughly 30% of melanomas arise from a pre-existing nevus, suggesting that senescence in nevus melanocytes is not a perfect tumor suppressor mechanism.

Like many other cancers, melanomas sometimes display constitutive activation of the Wnt signalling pathway. Moreover, we have previously shown that activated Wnt signaling can bypass cellular senescence. Therefore, we set out to examine the status of Wnt signaling in senescent primary human melanocytes in vitro and in benign human nevi. Expression profiling showed that in senescent human melanocytes in vitro, proliferation-promoting Wnt targets, such as cyclin D1 and c-myc, are typically repressed, whereas other Wnt targets are expressed. Like senescent melanocytes in vitro, benign human nevi also exhibited activated Wnt signaling, but, crucially, in nevi some proliferation-promoting Wnt targets, such as cyclin D1, are also expressed. Using a murine model, we also found that the activation of Wnt induces massive proliferative expansion of oncogene expressing melanocytes in vivo.

Based on these data we propose that the melanocytes within benign human nevi are not completely senescent, but have already partially bypassed the senescence program due to a re-activation of proliferation-promoting Wnt target genes. We suggest that this impaired senescence program is a major contributor to the risk of melanoma associated with human nevi.


Influence of immune related factors in the modulation of melanoma stemness

M. Perego, A. Tuccitto, M. Tazzari, V. Beretta, F. Rini, L. Rivoltini, C. Castelli

Department of Experimental Oncology, Fondazione IRCSS, Istituto Nazionale dei Tumori, Milan, Italy

In our previous studies we were able to select melanoma cells growing in vitro as melanospheres endowed with strong tumor initiating capacity and displaying a heterogeneous phenotype for stem associated markers. However, several evidence suggest that indeed CSC is not a static compartment but rather stemness features can be acquired by tumor cells in response to environmental signals. Thus our research is focused on dissecting the reciprocal interaction between melanoma initiating cells and the immune system with the aim of defining any role of immune-related factors in shaping the fraction of tumor cells responsible for tumor maintenance. Analysis of the secretion profile of melanospheres revealed that melanospheres display a secretory capacity quantitatively and qualitatively different from that of melanoma adherent cells or melanocytes. We focus our attention on IL-10, produced by one of our melanosphere culture and on IL-6 that was selectively produced by melanoma cells when cultured as adherent cells and by melanocytes. Since melanospheres express the receptors for these cytokines, we tested their effect in modulating melanosphere proliferation and self-renewal. Specifically, IL-6 exerted its effects both reducing the number of live cells when added to melanosphere culture medium and diminishing their self-renewal and clonogenic ability, conversely, IL-10 strongly promoted melanosphere self-renewal. Interestingly, only melanospheres were responsive to these soluble factors, while adherent cells were not influenced by the exposure to the same cytokines.

These data underline a strong relation between the immune system and melanoma stem cells, and future experiments will be aimed at fully addressing the effects of cytokines in regulating melanoma heterogeneity, plasticity and in vivo tumor maintenance.


SPRY4-IT1, a novel melanoma-upregulated long noncoding RNA modulates apoptosis, motility and invasion

D. Khaitan1, M. E. Dinger2, J. Mazar1, J. Crawford2, M. A. Smith2, J. S. Mattick2, R. J. Perera1

1 Sanford Burnham Medical Research Institute, Orlando, FL, USA; 2Institute for Molecular Bioscience, University of Queensland, St. Lucia, Australia

The identification of cancer-associated long non-coding RNAs (lncRNAs) and the investigation of their molecular and biological functions are important to understand the molecular biology of cancer and its progression. Although the functions of lncRNAs and the mechanisms regulating their expression are largely unknown, recent studies are beginning to unravel their importance in human health and disease. Here, we report that a number of lncRNAs are differentially expressed in melanoma cell lines in comparison to melanocytes and keratinocyte controls. One of these lncRNAs, SPRY4-IT1 (GenBank accession ID AK024556), is derived from an intron of the SPRY4 gene and is predicted to contain several long hairpins in its secondary structure. RNA-FISH analysis showed that SPRY4-IT1 is predominantly localized in the cytoplasm of melanoma cells, and SPRY4-IT1 RNAi knockdown results in defects in cell growth, differentiation and higher rates of apoptosis in melanoma cell lines. Differential expression of both SPRY4 and SPRY4-IT1 was also detected in vivo, in 30 distinct patient samples, classified as primary in situ, regional metastatic, distant metastatic, and nodal metastatic melanoma. The elevated expression of SPRY4-IT1 in melanoma cells compared to melanocytes, its accumulation in cell cytoplasm, and effects on cell dynamics, including increased rate of wound closure upon SPRY4-IT1 overexpression, suggest that the higher expression of SPRY4-IT1 may have an important role in the molecular etiology of human melanoma. SPRY4-IT1 molecular function, transcriptional regulation, cellular compartmentalization, interaction with other RNAs and proteins will be discussed.


Tumor cell-selective roles of cytoplasmic polyadenylation regulators in melanoma progression

E. Pérez-Guijarro1, E. Cañón1, D. Olmeda1, D. Alonso1, T. G. Calvo1, R. Méndez2, M. S. Soengas1

1 Melanoma Laboratory, Molecular Pathology, Spanish National Cancer Research Centre (CNIO), Madrid, Spain; 2Translational Control of Cell Cycle and Differentiation Laboratory, Molecular Medicine, Institute for Research in Biomedicine (IRB), Barcelona, Spain

Melanoma development is long-known for being associated with a plethora of changes in mRNA gene expression profiles. Whether these alterations result from the sum of individual events targeting discrete transcription factors, and/or from the action of pleiotropic RNA modulators is still unclear. Mining databases for novel tumor-associated factors, we identified a particularly intriguing upregulation of transcripts coding for cytosolic polyadenylation binding proteins (CPEBs) in melanoma specimens. We considered CPEBs interesting for functional analyses as cytoplasmic mRNA polyadenylation and deadenylation can have a direct impact on mRNA half life, intracellular distribution and ultimately, time- and context-dependent translational control. Validation analyses confirmed a significantly higher expression of CPEB4 protein in melanoma cells than in normal melanocytes, offering a putative window for therapeutic intervention. Interestingly, depletion of CPEB4 by RNA interference abrogated melanoma cell proliferation. This inhibitory effect was associated with massive changes in cell morphology, most notable at the nuclear level. Real-time imaging of cell cycle progression revealed defects in cytokinesis and mitotic progression. Roles of CPEB4 in cell division were tumor-cell selective as normal melanocytes did not display signs of mitotic catastrophe. The characterization of CPEB4 targets will also be presented. Altogether our data support the hypothesis that cytoplasmic mRNA polyadenylation modulators may constitute a new class of pro-oncogenic factors which can impinge on multiple aspects of melanoma cell proliferation and maintenance.


C-Kit mediated signaling to MITF in melanocytes: pathways and modifications

B. Phung1, J. Sun1, A. Schepsky2, E. Steingrimsson2, L. Rönnstrand1

1 Experimental Clinical Chemistry Department of Laboratory Medicine, Lund University, Malmö, Sweden; 2Department of Biochemistry and Molecular Biology, Faculty of Medicine, University of Iceland, Reykjavík, Iceland

The receptor tyrosine kinase c-KIT and the basic-helix-loop-helix-leucine zipper transcription factor MITF are both essential for melanocyte development. C-KIT and MITF are known to play a role in malignant melanoma and recent work has suggested that MITF may act as a rheostat which switches melanoma cells from a proliferative to an invasive phenotype. Thus, regulating MITF activity is a key step in melanocyte development as well as in melanomagenesis.

Activation of MITF results in a characteristic electrophoretic mobility shift of the protein. We took advantage of this shift to identify new signaling pathways downstream of c-KIT that directly activate MITF. Our study shows that c-KIT stimulation activates MITF specifically through the c-KIT phosphorylation sites Y721 (PI3 kinase binding site), Y568 and Y570 (Src binding site). This not only confirms the involvement of the Ras-Raf-Erk signaling pathway in the activation of MITF, but also establishes that Src kinase binding to Y568 and Y570 of c-KIT is required. Using selective inhibitors we observe and verify that c-KIT induced activation of MITF is dependent on PI3-, Akt-, Src-, p38- and Mek kinases. Moreover, the proliferative effects of c-KIT are dependent on MITF in HEK293T cells. In contrast, c-KIT Y568F and Y721F mutants are less effective in driving cell proliferation, compared to wild type c-KIT. In order to further dissect the mobility shift of MITF, we used phosphorylation specific antibodies that identify phospho-Ser73 and phospho-Ser409 of MITF to show that the shift is caused by the phosphorylation of Ser73. In addition, our results show that the phosphorylation of Ser73 is dependent on Ser409 being phosphorylated. Moreover, phosphatase-treatment suggests the involvement of additional phosphorylation sites which currently are being mapped.

Our results reveal novel mechanisms by which c-KIT signaling regulates MITF, with zimplications for understanding both melanocyte development and melanoma.


The M-SKIP Project: an international pooled-analysis on melanocortin-1 receptor (MC1R) variants and skin carcinogenesis

S. Gandini1, M. C. Fargnoli2, P. Maisonneuve1, M. Kayser3, R. Kumar4, E. Nagore5, J. Han6, J. Hansson7, P. A. Kanetsky8, P. Ghiorzo9, N. A. Gruis10, T. Dwyer11, R. F. de Misa12, W. Branicki13, T. Debniak14, B. K. Armstrong15, L. D. Marrett16, N. Morling17, M. T. Landi18, G. Palmieri19, G. Ribas20, S. B. Gruber21, A. Stratigos22, R. C. Millikan23, L. Cornelius24, T. Motokawa25, S. Anno26, P. Helsing27, H. A. Culver28, C. B. Begg29, T. H. Wong30, S. Rosso31, R. P. Gallagher32, M. Berwick33, P. Autier34, J. C. García-Borrón35, J. Little36, J. Newton-Bishop37, F. Sera38, S. Raimondi1 for the M-SKIP Study Group

1 European Institute of Oncology, Milan, Italy; 2University of L’Aquila, Italy; 3Erasmus University Medical Center, Rotterdam, The Netherlands; 4German Cancer Research Center, Heidelberg, Germany; 5Instituto Valenciano de Oncologia, Valencia, Spain; 6Brigham and Women's Hospital and Harvard Medical School, Boston, USA; 7Karolinska Institutet, Stockholm, Sweden; 8Perelman School of Medicine at the University of Pennsylvania, Philadelphia, USA; 9University of Genoa, Italy; 10Leiden University Medical Center, The Netherlands; 11Murdoch Childrens Research Institute, Victoria, Australia; 12Hospital Universitario Nuestra Senora de Candelaria, Santa Cruz de Tenerife, Spain; 13Institute of Forensic Research, Krakow, Poland; 14Pomeranian Medical University, Polabska, Poland; 15The University of Sidney and The Cancer Council New South Wales, Sidney, NSW, Australia; 16Cancer Care Ontario, Toronto, ON, Canada; 17Department of Forensic Genetics, University of Copenhagen, Copenhagen, Denmark; 18National Cancer Institute, NIH, Bethesda, USA; 19Istituto di Chimica Biomolecolare, CNR, Sassari, Italy; >20Fundacion Investigation Hospital Clinico Universitario de Valencia, Spain; 21University of Michigan, Ann Arbor, MI, USA; 22University of Athens, Andreas Sygros Hospital, Athens, Greece; 23University of North Carolina, Chapel Hill, NC, USA; 24Washington University, St. Louis, MO, USA; 25POLA Chemical Industries, Yokohama, Japan; 26Shibaura Institute of Technology, Tokyo, Japan; 27Oslo University Hospital, Norway; 28University of California, Irvine, USA; 29Memorial Sloan-Kettering Cancer Center, New York, USA; 30University of Edinburgh, UK; 31Centro per la Prevenzione Oncologica, Torino, Italy; 32British Columbia Cancer Agency, Vancouver, BC, Canada; 33University of New Mexico, Albuquerque, USA; 34International Prevention Research Institute, Lyon, France; 35University of Murcia, Spain; 36University of Ottawa, Ottawa, ON Canada; 37University of Leeds, UK; 38MRC Centre of Epidemiology for Child Health, UCL Institute of Child Health, London, UK

MC1R, located at 16q24.3, is the major gene involved in skin pigmentation. Although MC1R variants are robust markers of increased melanoma risk, the specific contribution of each MC1R variant is unclear. The M-SKIP Project aims to perform a pooled-analysis on the role of MC1R variants in carcinogenesis having sufficient power to look at individual variants. Main endpoints are skin cancers and phenotypic characteristics involved in skin carcinogenesis: hair, eye and skin color, skin type, freckles, solar lentigines, common and atypical nevi.

Out of 43 contacted investigators, 31 (72%) joined the M-SKIP Project and sent anonymous data on 7740 melanoma cases, 2052 non-melanoma skin cancer (NMSC) cases, and 16 048 controls.

In initial analysis to test the association between any MC1R variant and (i) skin cancer overall, (ii) melanoma and NMSC, and (iii) phenotypic characteristics, the pooled data were fitted with fixed-effects logistic regression models.

Our preliminary results showed that carrying at least one MC1R variant increases the risk of skin cancer (OR:1.70;95%CI:1.54–1.87), with similar ORs for melanoma (OR:1.71;95%CI:1.53–1.91), and NMSC (OR:1.72;95%CI:1.42–2.08). When we investigated the role of each of the nine most prevalent MC1R variants (V60L, D84E, V92M, R142H, R151C, I155T, R160W, R163Q, D294H), we found significant associations with skin cancer for all variants except p.R163Q, with ORs ranging from 1.12 (95%CI: 1.01–1.24) for V92M to 2.03 (95%CI:1.45–2.84) for D84E. MC1R variants were associated with all the studied phenotypic characteristics except light eye color; the strongest association was found for carriage of any MC1R variant with red hair color (OR: 5.59; 95%CI: 3.84–8.12). Refined statistical analyses taking into account heterogeneity amongst studies, gene-phenotype interactions, and adjustment for covariates are ongoing.

This data-bank represents a reference and guide for all investigators working in this research field. Collaboration among investigators will improve the knowledge and accelerate our understanding of the role of MC1R variants in skin carcinogenesis.


miR-125b affects migratory and invasive potential of melanoma cells

F. Rambow, L. Larue

Institut Curie, Developmental Genetics of Melanocytes, CNRS UMR3347, INSERM U1021, Orsay, France

MicroRNAs (miRNAs) are small noncoding RNAs that extensively regulate gene expression. Generally, miRNAs inhibit mRNA translation and stability by imperfect base-pairing. One miRNA might have hundreds of predicted mRNA targets. miRNAs are involved in the regulation of many biological processes, such as development, differentiation, cell cycle, and apoptosis amongst others. Increasing evidence shows an important role for miRNAs in initiation and progression of cancer. In melanoma, profiling studies in cell lines and tumors identified various miRNA signatures linked to in vitro characteristics or clinical data. However, only a few miRNAs have been shown to affect directly the aggressive features of the disease such as enhanced motility and invasion. After screening the expression of various miRNAs of potential interest in a series of cell lines having different migratory potential, it appeared that miR-125b was upregulated in highly migratory melanoma cells. Silencing of miR-125b expression in these cells reduced migratory and invasive potentials, according to wound scratch and matrigel invasion assays. Interestingly, among various potential mRNA targets of miR-125b, MITF and related genes were found to be inversely expressed in melanoma cell lines.


Oncogenic addiction of melanoma cells to the p62/sequestosome1 protein

E. Riveiro-Falkenbach1,2, D. Alonso-Curbelo1, D. Olmeda1, M. Cifdaloz1, P. Ortiz-Romero2, J. L. Rodríguez-Peralto3, M. S. Soengas1

1 Melanoma Laboratory, Molecular Pathology Programme, Spanish National Cancer, Research Centre (CNIO), Madrid, Spain; 2Departments of Pathology, Hospital Universitario 12 de Octubre, Madrid, Spain; 3Dermatology, Hospital Universitario 12 de Octubre, Madrid, Spain

Aggressive cancers, and melanoma is no exception, are characterized by an exacerbated cell metabolism. This feature, which can favor tumor initiation and progression, also provides a tractable platform for rational drug design. Thus, tumor cells are significantly more dependent on the continuous clearance of aggregate-prone proteins or damaged organelles than their normal counterparts. Specifically, the basal rate of autolysosomal and ubiquitin (Ub)-mediated proteolysis can considerably differ between benign and malignant cell compartments. However, how this putative ‘Achilles’ heel’ is controlled in melanoma, and whether it can be exploited therapeutically is not clear. Here we will report that the p62/Sequestosome 1 (an adaptor protein that can both recruit ubiquitinated proteins and modulate macroautophagy), can represent a novel point of vulnerability for melanoma. Tissue microarrays revealed a notable upregulation of p62 in melanoma specimens, particularly at metastatic stages. In contrast, melanocytic nevi were virtually negative for p62 expression. Comparative genomic hybridization and expression analyses in a large panel of melanoma cell lines support a transcriptionally-associated control of p62 without amplification of its genomic locus. Interestingly, genetic depletion of p62 invariably led to a senescence-like phenotype in melanoma cells, but not in their normal melanocytic counterparts. In mouse xenograft models, this inhibition of cell proliferation translated into a significant blockage of tumor progression. p62-deficient melanoma cells showed an accumulation of cytosolic aggregates consistent with defective intracellular degradation programs. However, the nuclei of p62 downregulated cells were also markedly altered, suggesting additional roles of this protein in cell division. In fact, cDNA arrays identified Survivin and a series of E2F1- dependent modulators of DNA replication and mitotic progression as unexpected targets of p62 in melanoma cells. Here we will present data in tissue specimens validating this hypothesis and supporting a therapeutically relevant function of p62 as a novel melanoma oncogen.


Metformin inhibits melanoma development through autophagy and apoptosis mechanisms

T. Tomic1,2, T. Botton1,2, M. Cerezo1,2, C. Bertolotto1,2,3, S. Tartare-Deckert1,2,3, P. Bahadoran1,2,3, R. Ballotti1,2,3, S. Rocchi1,2,3

1 INSERM, U895, teame 1 Nice, France; 2Université de Nice Sophia Antipolis, UFR de Médecine, IFR50, Nice, France; 3Service de Dermatologie, CHU Nice, France

Metformin is the most widely used antidiabetic drug due to its proven efficacy and its limited secondary effects. Interestingly, recent studies have reported that metformin can block the growth of different tumor types. Here, we show that metformin exerts antiproliferative effects on melanoma cells, while normal human melanocytes are resistant to these metformin-induced effects. To better understand the basis of this antiproliferative effect of metformin in melanoma, we characterized the sequence of events underlying metformin action. We showed that 24h metformin treatment induced a cell cycle arrest in G0/G1 phases, while after 72h, melanoma cells underwent autophagy as demonstrated by electron microscopy, immunochemistry, and by quantification of the autolysosome-associated LC3 and Beclin1 proteins. In addition, after 96 h post metformin treatment we observed robust apoptosis of melanoma cells. Interestingly, inhibition of autophagy by knocking down LC3 or ATG5 decreased the extent of apoptosis, and suppressed the antiproliferative effect of metformin on melanoma cells, suggesting that apoptosis is a consequence of autophagy. The relevance of these observations were confirmed in vivo, as we showed that metformin treatment impaired the melanoma tumor growth in mice, induced autophagy and apoptosis markers. Taken together our data suggest that metformin has an important impact on melanoma growth, and may therefore be beneficial in patients with melanoma.


CDKN2A unclassified variants (UVs) in familial malignant melanoma: the erratic behavior of different substitutions at the same amino acid position

M. C. Scaini1, G. Funari1, L. Elefanti1, C. Menin2, E. D’Andrea1,2

1 Section of Oncology, Department of Oncology and Surgical Sciences; University of Padova, Padova, Italy; 2Immunology and Molecular Oncology Unit, Istituto Oncologico Veneto (IOV), IRCCS, Padova, Italy

CDKN2A is the major gene involved in familial malignant melanoma. Germline mutations are found in melanoma-prone families and multiple primary melanoma patients, most are missense mutations and occur throughout the entire coding region. Thus, sequence variants of the encoded protein (p16INK4A) can disrupt protein conformation as well as the association with its targets, CDK4/6. A growing number of sequence changes are being identified but, unless their pathogenic role can be demonstrated, they cannot be used for identification of carriers at risk. In the evaluation of these UVs, it is common practice to define as pathogenic the site, rather than the different type of amino acid (aa) substitution in the same position. Aim of this study is the characterization of p16INK4A variants with aa changes in specific, highly conserved domains, to establish which substitutions are tolerated without impairment of the protein function. Preliminarly, we mutagenized glycine-35 and obtained five p16INK4A variants, we expressed them in a human p16-null osteosarcoma cell line (U2-OS) and tested the ability of the mutants to block cell cycle in G1-phase, as the p16wt is able to do. Transfected U2-OS were monitored for several days and cell counts were performed at different time points. Alternatively, 10 days after seeding at low density, total colony number was determined. Each aa variant was evaluated for its ability to hamper p16INK4A function and classified as ‘loss of function’, ‘wt-like’ or ‘intermediate behavior variant’.

In conclusion, variants obtained by different aa substitutions at the same position showed different effects on cell cycle regulation; therefore the aa position per se, is not sufficient for UVs classification. Moreover, discrimination between mutant and neutral variants might add not only valuable knowledge to our understanding of critical protein residues, but also it will improve the information during genetic counselling sessions.


The viral chemokine receptor US28 mediates its anti-tumorigenic effect in melanoma via13

S. Joshi1, C. Wels1, M. Fukunaga-Kalabis2, M. Herlyn2, H. Schaider1

1 Cancer Biology Unit, Department of Dermatology, Medical University of Graz, Graz, Austria; 2The Wistar Institute, Philadelphia, PA, USA

Reports investigating on the biology of the human cytomegalovirus encoded G-protein coupled receptor US28 in different cell types led to the understanding that it has potential to impart differential effects like apoptosis, tumorigenesis, migration etc. on target cells, most probably depending on the host and cellular microenvironment. Here we report that US28 expression in melanoma cell lines results in reduced proliferation and caspase mediated apoptosis. The observation that the signaling mutant R129A shows reduced effects compared to wild type US28 while the C-tail truncated mutant has an enhanced effect due to its higher surface level retention indicated that modulation of intracellular signaling by US28 is responsible for the observed effects. As US28 is known for activation of different signaling pathways via multiple G-protein interactions, the involvement of different G-proteins was determined. We found that Gα13 is selectively executing the apoptosis inducing effects of US28. Interestingly, endogenous expression levels of Gα13 vary drastically in the analyzed panel of melanoma cell lines with a gradual loss at advanced stage of disease. In vivo reduced tumor growth for US28 compared to GFP control expressing melanoma cells was observed. To summarize, our study reports on a mechanism how US28 is inducing apoptosis in melanoma cells and indentified Gα13 as the main G-protein in mediating this effect.


TGF-β1 abrogates whereas TNF-α enhances resistance to ERK inhibitors in BRAF mutant melanoma cells through Twist1

D. R. Menon1, S. Joshi1, C. Wels1, B. Rinner2, J. M. Brandner3, H. Schaider1

1 Cancer Biology Unit, Department of Dermatology, Medical University of Graz, Graz, Austria; 2Center for Medical Research, Medical University of Graz, Graz, Austria; 3Department of Dermatology, University Hospital Hamburg-Eppendorf, Hamburg, Germany

Melanomas along with many other cancer types harbor BRAF mutations which constitutively activate the MAPK-ERK1/2 pathway. PLX4032 (vemurafenib) a BRAF inhibitor, showed approximately 80% response during clinical trials, but in turn gave rise to acquired resistance in cancer patients. There was no good explanation given for the different responses of melanoma patients towards this therapy. We found that TGF β1 signaling could act as a factor predicting the response, as it enhanced the susceptibility of BRAF mutant melanoma cells towards exposure of PD98059 (MEK1 inhibitor) and PLX4032 by almost three times. Inhibition of TGF β1 signaling by using an SB431542 (ALK5 inhibitor) almost completely abolished the apoptotic effect of PLX4032 exposure in vitro. Over expression of Twist1, an epithelial – mesenchymal transition regulator induced a resistance in BRAF mutant cells to a combination of PLX4032 and TGF beta1. In BRAF mutant cell lines exposure to PLX4032 abrogated the stability of Twist1, whereas exposure to TNF alpha reestablished Twist1 in these cell lines leading to a resistance towards a combined PLX4032 and TGF beta1 exposure. To summarize, our studies suggest that TGF β1 with PLX4032 could be a better strategic approach, when compared to PLX4032 alone, but the stability of Twist1 inflicted by TNF alpha may contribute to acquired resistance towards inhibitors targeting BRAF – ERK signaling.


Targeting of MAPK and PI3K pathways simultaneously improves anti-tumor activity against metastatic melanoma

R. Serrano, R. Calero, E. Morchón, A. Ocaña

AECC Cancer Research Unit, Albacete University Hospital Spain

Identification of druggable targets in metastatic melanoma is a main objective given the bad prognosis of this disease in the metastatic setting. Several oncogenic alterations have been described in melanoma, including abnormal activation of the MAPK pathway secondary to mutations in the NRAS or B-RAF gene. The phosphatidylinositol 3-kinase/Akt pathway can be additionally activated by the dysregulation of tyrosine kinase receptors or by an activating loop when mutations in BRAF or NRAS are present. Thus, several small kinase inhibitors to target the BRAF/MAPK and PI3K/Akt pathways are in clinical development.

In the present work we have studied the phospho-proteomic profile of membrane tyrosine kinase receptors and the activation status of downstream kinases in a panel of human melanoma cell lines representative for NRAS or BRAF mutations. Furthermore, MTT proliferation assays, flow cytometry and western blot experiments were performed to investigate the antitumor activity of the HSP90 inhibitor 17AAG and the PI3K/mTOR inhibitor BEZ235 in these lines.

Our results show that activation of several tyrosine kinase receptors (i.e EGFR, IGF-IR, PDGFR) in addition to BRAF and NRAS mutations lead to an activated state of MAPK and Akt pathways. The administration of 17AAG at very low doses (<0.1 μM) efficiently reduced cell growth and decreased the effective dose of the BRAF inhibitor sorafenib to inhibit BRAF signalling. However, the PI3K/Akt pathway remained activated. Interestingly, concomitant administration of low doses of the PI3K/mTOR inhibitor BEZ235 and 17AAG showed a synergistic activity decreasing melanoma cell growth. In addition, this combination is needed to target simultaneuosly the MAPK and PI3K/Akt pathways in BRAF and NRAS-driven melanoma cells.

These results demonstrate that the combination of 17AAG and BEZ235 could be an alternative therapeutic strategy to overcome resistance to BRAF inhibitors in melanoma tumors.


The pan-ErbB receptor tyrosine kinase inhibitor canertinib promotes apoptosis of malignant melanoma in vitro and displays anti-tumour activity in vivo

E. A. Severinsson1, C. Trinks1, H. Gréen2, A. Abdiu3, A.-L. Hallbeck1, O. Stål1, T. M. Walz1

1 Division of Oncology, Department of Clinical and Experimental Medicine, Faculty of Health Sciences, Linköping University, Linköping, Sweden; 2Clinical Pharmacology, Division of Drug Research, Department of Medical and Health Sciences, Faculty of Health Sciences, Linköping University, Linköping, Sweden; 3Department of Plastic Surgery, Hand Surgery and Burns, University Hospital, Linköping, Sweden

The ErbB receptor family has been suggested to constitute a therapeutic target for malignant melanoma. Here we investigate the effect of the irreversible pan-ErbB tyrosine kinase inhibitor canertinib on cell growth and survival in human melanoma cells in vitro and in vivo. Canertinib significantly inhibited growth of cultured, RaH3 and RaH5, melanoma cells in a dose-dependent manner as determined by cell counting. Half-maximally growth inhibitory dose (IC50) was approximately 0.8 μM and by 5 μM both cell lines were completely growth-arrested within 72 h of treatment. Incubation of exponentially growing RaH3 and RaH5 cells with 1 μM canertinib resulted in an accumulation of cells in the G1-phase of the cell cycle within 24 h of treatment without induction of apoptosis as determined by flow cytometry. Immunoblot analysis showed that 1 μM canertinib markedly inhibited ErbB1-3 receptor phosphorylation with a concomitant decrease of Akt-, Erk1/2- and Stat3 activity in both cell lines. In contrast to the cytostatic effect observed at ≤5 μM canertinib, higher concentrations induced apoptosis as demonstrated by the Annexin V method and Western blot analysis of PARP cleavage. Furthermore, canertinib significantly inhibited growth of RaH3 and RaH5 melanoma xenografts in nude mice. Pharmacological targeting of the ErbB receptor tyrosine kinases by canertinib may prove successful in the treatment of patients with metastatic melanoma.


CAVATAKTM (Coxsackievirus A21) displays potent oncolytic activity in BRAFV600E mutant melanoma cells resistant to selective BRAF kinase inhibitors

D. R. Shafren1,2, M. Farrelly3, A. J Croft3, B. Davies2, J. Stewart2, R. Ingham2, M. Y. Quah1, G. Au1,2, P. Hersey3

1 The Picornaviral Research Unit, The School of Biomedical Sciences, Faculty of Health, University of Newcastle, Newcastle, NSW, Australia; 2Viralytics Ltd., Suite 1B, Pymble, NSW, Australia; 3Department of Oncology, University of Newcastle, David Madison Building, Newcastle, NSW, Australia

CAVATAKTM is a naturally occurring Picornavirus, which induces mild upper respiratory symptoms during natural infection of humans. CAVATAKTM also displays potent oncolytic activity against both in vitro cultures of human cancer cells and against in vivo xenografts of human cancers in mouse models of melanoma, prostate cancer, breast cancer and multiple myeloma, all which exhibit high surface intercellular adhesion molecule-1 (ICAM-1) expression. The cellular targeting receptor for CAVATAKTM is ICAM-1. CAVATAKTM is currently in Phase II clinical evaluation in patients with late stage melanoma. Vemurafenib (PLX4032) is a potent inhibitor of a mutant form of BRAF (BRAFV600E) that is found in approximately 50% of late stage melanoma patients. In Phase III clinical studies vemurafenib significantly increases progression free and overall survival in previously untreated late stage melanoma patients. Despite this significant anti-melanoma activity, the disease of a majority of treated patients eventually become resistant to the activity of vemurafenib. In the present study, two in vitro melanoma cell cultures expressing the BRAFV600E mutation were sequentially propagated in the presence of increasing concentrations of PLX4720 (1.0–10 μM) until drug resistant cell populations were selected. Dose–ranging challenge of mutant BRAF melanoma in vitro cultures with CAVATAKTM yielded comparable levels of specific viral mediated oncolysis as evidenced by viral replication (qPCR) and cell lysis (MTT assay) analysis. Furthermore, in studies on cell cultures generated directly from tumour biopsies of late stage melanoma patients taken prior to treatment and following significant clinical disease progression while on vemurafenib treatment CAVATAKTM displayed comparable levels of potent oncolytic activity independent of whether the cell cultures were generated from biopsies taken prior to or following progression on vemurafenib treatment. Overall, these preliminary findings suggest that systematic melanoma treatment regimes including both vemurafenib and CAVATAKTM may have favourable clinical outcomes and as such deserve further evaluation.


CD133-positive melanoma subset contributes to tumorigenesis through vasculogenic mimicry-dependent vascular niche morphogenesis

C. Y. Lai1, J. H. Sheen2, Q. Zhan1, G. F. Murphy1, M. Y. Hsu1,2

1 Department of Dermatology, Boston University School of Medicine, Bosoton, MA, USA; 2Program in Dermatopathology, Department of Pathology, Brigham and Women's Hospital, Harvard Medical School, Boston, MA, USA

Subsets of melanoma cells expressing putative stem cell markers and possessing tumor initiation/self-renewal capacities have been identified. However, the tissue localization of such melanoma subpopulations, so-called ‘niche’, and its structural-functional significance in tumorigenesis remians largely unexplored. Using in situ hybridization and double-label immunofluorescence, we observed that the CD133+, as well as the ABCB5+, melanoma subsets are co-localized to vascular niches in close proximity to the vasculogenic mimicry (VM)-like channels lined by CD144 (VE-cadherin)+ melanoma subpopulation. To investigate the significane of the CD133 melanoma subset, loss-of-function was achieved in WM1617 metastatic melanoma cells using lentiviral shRNA. CD133 knockdown siginificantly retarded tumorigenicity in mice. Analysis of the CD133 knockdown WM1617 melanoma cells using qRT-PCR showed reduced expressions of the niche-associated markers (CD144 and ABCB5) in vitro. Accordingly, CD133 knockdown xenografts exhibited markedly attenuated CD144+ VM-like channel formation, and depletion of ABCB5+ cells in the vascular niches in vivo. Taken together, our study showed that: 1) the CD133+/ABCB5+ subsets of melanoma cells reside in complex vascular niches that are spatially associated with anastomosing VM-like channels containing CD144+ melanoma cells, and 2) the CD133 melanoma subset contributes to tumor growth through VM-dependent vascular niche morhognesis.


Glutamate/cystine antiporter, xCT, regulates proliferation of melanoma via metabotropic glutamate receptor 1 (GRM1)

S.-S. Shin1, B.-S. Jeong1, B. Wall2, S. Rosenberg1, S. Neglio1, D. Milgraum1, S. Chen2, J. Goydos1

1 Cancer Institute of New Jersey, Robert Wood Johnson University Hospital, New Brunswick, NJ, USA; 2Rutgers, The State University of New Jersey, Piscataway, NJ, USA

Within the last years there are accumulating evidences suggesting that glutamate is an inducer of cellular signalling in various non-neuronal systems including skin, bone, and lymph node. Previously we have demonstrated the oncogenic role of metabotropic glutamate receptor 1 (GRM1) in melanomagenesis via glutamate signalling. We showed that disruption of glutamate signalling in melanoma cells by the presence of a glutamate release inhibitor, riluzole, significantly reduced the proliferation of melanoma cells in vitro, tumor progression in vivo as well as in phase 0 and phase 2 clinical trials with orally administered riluzole that stabilized late stage melanoma patients. Based on these results, we focused on identifying glutamate transporters involved in autocrine activation of GRM1. By microarray screening, we identified a glutamate/cystine antiporter, xCT, as a candidate that mediate excess glutamate release. An inverse correlation was observed between xCT expression with clinical responses of patients from the phase 2 trial. We further examined basal levels of xCT expression in normal skin (n = 2), nevi (n = 2) and melanoma biopsies (n = 20), we found a link between xCT expression in both protein and mRNA level with the progression of melanoma while normal skin biopsies did not show detectable xCT expression. In addition, IHC of xCT in melanomas from different stages of the disease suggested that xCT expression is a progression-related event. Elevated levels of xCT expression were detected in several cultured melanoma cells except for Skmel2 compared to normal primary melanocytes. Stable xCT-Skmel2 clones exhibited 4-folds extracellular glutamate release and enhanced cell proliferation in comparison to vector-control Skmel2 clones. Treatment of xCT-Skmel 2 clones with either riluzole or Bay 36-7620 (GRM1 antagonist) led to a decrease in the number of viable cells compared to vector control. Taken together, these results indicate that xCT may be involved in GRM1 induced melanomagenesis via autocrine glutamate signalling.


The role of tumor suppressor INPP4b in melanoma

A. Slipicevic1,2, A. Kaur1, V. A. Florenes2,3, M. Herlyn1

1 The Wistar Institute, Philadelphia, PA, USA; 2Pathology Clinic, Norwegian Radium Hospital, Oslo University Hospital, Montebello, Oslo, Norway; 3Faculty of Health Sciences, Oslo University College

Activation of the phosphoinositide 3-kinase (PI3K)/Akt signaling pathway is a key event in melanoma development and progression due to its effects on cell survival, proliferation and metastasis. Furthermore, activation of PI3/Akt signaling pathway contributes to development of resistance to a variety of chemotherapeutic agents. Therefore, proteins of the PI3K/Akt signaling pathway represent attractive therapeutic targets in melanomas. Also, mechanisms regulating Akt phosphorylation need to be unraveled in detail in order to improve design of novel therapeutic strategies.

Inositol polyphosphate 4-phosphatase-II (INPP4b) acts as a tumor suppressor in epithelial carcinomas by regulating PI3K/Akt signaling pathway. INPP4b loss of heterozygosity (LOH) is detected in human cancers including breast and ovarian cancer as well as melanoma. However, the role of INPP4b protein in melanoma development and progression is still unclear. We report here that INPP4b is highly expressed in normal human melanocytes but in melanoma cell lines both mRNA and protein levels of INPP4b are decreased. INPP4b loss is also observed in phosphatase and tensin homolog (PTEN)-null cell lines. Furthermore, these cell lines have higher levels of activated, phosphorylated Akt compared to INPP4b+/PTEN- cell lines suggesting that loss of both genes is necessary for full activation of the pathway. Preliminary results show that INPP4b knockdown in melanocytes affects morphology, proliferation and anchorage-independent growth properties. These data suggest that INPP4b functions as a tumor suppressor through regulation of the PI3K/Akt signaling pathway, and that loss of INPP4b protein might contribute to melanoma development.


Neural crest stem cells and melanoma formation: a likely connection

O. Shakhova1, G. Civenni1, L. Sommer1

1 Cell and Developmental Biology, Institute of Anatomy, University of Zurich, Switzerland;

During embryonic development, neural crest cells generate a variety of cell types, including melanocyte stem cells and eventually differentiated melanocytes. We have recently shown that human melanoma is composed of distinct cell types reminiscent of neural crest derivatives. This heterogeneity is established by multipotent cells expressing the neural crest stem cell (NCSC) marker CD271. When isolated from solid tumors using a method that leaves intact cell surface epitopes, these NCSC-like cells, but not CD271-negative cells, form tumors upon xenotransplantation that mirror the heterogeneity of the parental melanoma. Moreover, even upon transplantation into fully immunocompromized mice, presence of CD271-positive cells is required for long-term tumor expansion in vivo. These data indicate a role of stem cell-like cells in tumorigenesis. To further address this issue, we made use of a mouse melanoma model, in which NrasQ61K-oncogene expression in the melanocytic lineage consistently leads to formation of melanoma. Similar to human melanoma, these tumors are composed of various cell types expressing markers of neural crest derivatives. Importantly, tumor initiation in these mice is associated with melanocyte stem cell expansion and emergence of melanoblasts from the stem cell niche, resulting in skin hyperpigmentation and tumorigenesis. Moreover, in vivo manipulation of genes known to regulate normal NCSCs affects melanoma formation. Thus, oncogene-mediated activation of melanocyte stem cells is a crucial event in melanoma formation.


Melanoma whole-exome sequencing reveals novel driver mutations

M. S. Stark1, K. Dutton-Regester1,2, S. L. Woods1, M. Gartside1, V. F. Bonazzi1, L. G. Aoude1,3, D. Muzny4, D. Chow5, J. J. Ellis6, J. Reid4, V. Zismann5, S. Tyagi1, I. Newsham4, Y.-Q. Wu4, J. M. Palmer1, T. Pollak1, D. Youngkin5, B. R. Brooks8, C. Sereduk5, C. Lanagan1, C. W. Schmidt1, B. Kobe6, H. Yin5, K. M. Brown5,7, R. Gibbs4, J. Trent5,8, N. K. Hayward1

1 Queensland Institute of Medical Research, Brisbane, Qld, Australia; 2Queensland University of Technology, Brisbane Qld, Australia; 3School of Medicine, University of Queensland, Brisbane, Qld, Australia; 4Baylor College of Medicine, Houston, TX, USA; 5Translational Genomics Research Institute, Phoenix, AZ, USA; 6School of Chemistry and Molecular Biosciences, Institute for Molecular Biosciences and Centre for Infectious Diseases Research, University of Queensland, Brisbane, Qld, Australia; 7National Cancer Institute, MD, USA; 8Van Andel Research Institute, Grand Rapids, MI, USA

Utilising whole-exome capture, we sequenced eight melanoma cell lines and their matched normal lymphoblastoid cell lines (LCLs) using two next-generation sequencing platforms. Overall, 3215 somatic alterations were identified; range 243–523 per sample. Of these, 2139 were predicted to alter protein structure (range 175–326 per sample). We found a total of 1925 missense, 122 nonsense, 32 splice-site and 64 small insertion/deletion mutations, along with 1076 synonymous (silent) mutations. Of the 1740 genes found to have protein-altering changes, 446 were reported to be mutated in a recent exome analysis of melanoma and 166 have mutations documented in the COSMIC database. The overlap between our dataset and these two other source lists revealed a total of 58 genes commonly mutated in melanoma, suggesting that many of these genes are likely ‘drivers’ in melanoma pathogenesis. Additionally, mutations were found in 22 genes lying in previously described regions of homozygous deletion, including PTPRD, a putative tumor suppressor gene for melanoma and glioblastoma, which had five mutations in four samples. Next-generation sequencing technologies have provided the ability to identify cancer-associated mutations in an unbiased manner. These catalogues of mutations have enormous potential to inform on the molecular etiology of the disease and to identify novel targets for therapeutic intervention.


Identification and characterisation of the MITF-interactome

T. Strub, D. Koludrovic, I. Davidson

Institut de Génétique et de Biologie Moléculaire et Cellulaire, 1 Rue Laurent Fries, Illkirch, France

The transcription factor MITF plays a pivotal role in melanoma. We previously performed profiling of MITF genomic occupancy by ChIP-seq in human 501Mel cells and identified MITF regulated genes by RNA-seq following siRNA-mediated MITF knockdown. This approach showed that loss of MITF leads to both down- and up-regulation of target genes involved in DNA replication, repair and mitosis, invasion and metastasis, indicating that it acts as a transcriptional activator or repressor in a promoter-specific manner. The identification of MITF partner proteins and characterisation of their role in MITF-mediated gene regulatory networks may explain how it exerts this ‘dual function’. We have generated 501mel cells expressing Flag-HA tagged MITF and performed tandem affinity purification followed by mass-spectrometry of both the soluble nuclear and chromatin associated MITF fractions. Using this approach, we identified many known partners (for example b-catenin and subunits of the BRG1 chromatin remodeling complex). We also show that MITF forms heterodimers with TFE3, TFEB and TFEC in melanoma cells suggesting that MITF does not occupy all of its genomic sites as a homodimer. In addition, we identify a set of other MITF interactors including transcription cofactors, components of the DNA replication machinery and components of the DNA repair machinery. Our data suggest that MITF forms multiple complexes and that it may act not only as a transcription factor in melanoma cells, but that it may influence other nuclear processes.


Pharmacological profile of LGX818 : a potent and selective RAF kinase inhibitor with excellent PK/PD, efficacy and toxicological profiles in preclinical models

D. D. Stuart1, N. Li2, D. Poon1, K. Aardalen1, S. Kaufman1, H. Merritt1, F. Salangsang1, E. Lorenzana1, A. Li2, M. Ghoddusi1, G. Caponigro3, F. Sun2, S. Kulkarni4, S. Kakar4, N. Turner1, R. Zang1, J. Tellew2, N. Pryer1

1 Novartis Institutes for Biomedical Research, Emeryville, CA, USA; 2Genomics Institute of the Novartis Research Foundation, San Diego, CA, USA; 3Novartis Institutes for Biomedical Research, Cambridge, MA, USA; 4Novartis Pharmaceuticals Corporation, East Hanover, NJ, USA

Results from recent clinical studies indicate that selective RAF inhibitors have significant activity in patients with metastatic melanoma whose tumors express BRAFV600E. However, not all patients respond equally well to treatment and the longevity of response is typically limited. Our goal was to develop a potent and selective RAF inhibitor with excellent pharmacological properties that would provide sustained and strong RAF/MEK/ERK pathway inhibition. An inhibitor with such a profile should maximize the degree and duration of patient response.

LGX818 is a potent BRAF inhibitor with selective anti-proliferative and apoptotic activity in cells expressing BRAFV600E. In the A375 (BRAFV600E) human melanoma cell line LGX818 suppresses phospho-ERK (EC50 = 3 nM) leading to potent inhibition of proliferation (EC50 = 4 nM). No significant activity was observed against a panel of 100 kinases (IC50 > 900 nM) and LGX818 did not inhibit proliferation of >400 cell lines expressing wild-type BRAF. LGX818 is orally bioavailable with good pharmacokinetic properties in preclinical species. Single dose PK/PD studies using human tumor xenografts indicate LGX818 causes strong (>80%) and sustained (>24 h) inhibition of the RAF/MEK/ERK pathway. The sustained target inhibition is in part a result of LGX818 having a very long off-rate from BRAFV600E (>30 h) resulting in extended pathway inhibition following drug clearance. Maximum tumor regression is achieved in the A375 tumor xenograft model at a dose of 5 mg/kg bid and excellent tolerability is observed at doses ≥300 mg/kg bid (linear increase in exposure). Furthermore, at therapeutically-relevant doses, LGX818 does not induce epithelial hyperplasia as has recently been described for other RAF inhibitors.

LGX818 is a potent and selective RAF kinase inhibitor with unique biochemical properties that contribute to an excellent pharmacological profile. A Phase I clinical trial in patients with BRAF mutant melanoma has recently been initiated in the US and Europe.


Role of NEDD9 in melanoma tumorigenesis

H. Sung1, F. Gonzalez1, L. Chin2, M. Kim1

1 Comprehensive Melanoma Research Center, Moffitt Cancer Center, Tampa, FL, USA; 2Department of Medical Oncology, Dana-Farber Cancer Institute, Boston, MA, USA

We have previously identified NEDD9 (Neural precursor cell Expressed, Developmentally Down-regulated 9) as a novel melanoma metastasis gene based on cross-species comparison of human and mouse melanomas. We have shown that NEDD9 cooperates with the RAS/RAF/MAPK pathway activation via oncogenic HRASV12G or BRAFV600E expression to enhance proliferation and to drive invasion in vitro and lung seeding in vivo of melanoma cells.

To explore the biological roles of NEDD9 in melanoma tumorigenesis, we have generated tet-inducible NEDD9 mouse model that carries a doxycycline-responsive, melanocyte-targeted NEDD9. We also introduced a tyrosinase-driven luciferase reporter to enable bioluminescence imaging to track metastasis. Two independent transgenic founder lines were obtained and crossed onto the non-metastatic Tyr-RAS* Ink4a/Arf−/− melanoma model to establish a colony of compound mutant mice designated as ‘iNEDD9’ (e.g. Tyr-RAS*::Tyr-rtTA/Tet-NEDD9::Tyr-Luc+::Ink4a/Arf−/−). Melanocyte and melanoma cell lines derived from iNEDD9 model showed doxycycline-regulated NEDD9 expression. Although small in cohort sizes, when we turned on NEDD9 expression in melanocytes via doxycycline administration at weaning age, we observed cooperation of NEDD9 with RAS/RAF/MAPK pathway activation, leading to decreased tumor latency and increased tumor initiation. Kaplan–Meier melanoma free survival curve shows decreased median survival of iNEDD9 mice (109 (line A, P = 0.0303) and 110 (line D, P = 0.0322) days) compared to control mice without NEDD9 (163 days). However, iNEDD9 mice succumb to Ink4a/Arf deficiency related lymphomas and sarcomas before lymph node or distal metastasis can be documented. To avoid this, we replaced Ink4a/Arf allele with conditional cInk/Arf::Tyr-CreERT2 lines. Utilizing this model, we are evaluating the effect of NEDD9 upregulation at molecular and histological level in order to understand molecular basis for the role of NEDD9 in melanoma tumorigenesis.

This study will generate a body of knowledge for the in vivo roles of NEDD9 in melanoma tumorigenesis and an invaluable melanoma model driven by genetic lesions observed in human melanoma patients.


Several changes in the transcriptome differentiate melanocytes from melanoma cells

R. Swoboda1, R. Gupta1, L. Hughes2, R. V. Davuluri1, M. Herlyn1

1 The Wistar Institute, Philadelphia, PA, USA; 2Fox Chase Cancer Center, Philadephia, PA, USA

It has been shown that non-coding RNAs influence the expression of mRNAs. Several of these non-coding RNAs play a role in tumorigenesis. For example trans-acting HOTAIR RNA is involved in metastasis of breast tumors. But there are also direct correlations between antisense RNAs and the corresponding complementary mRNA, e.g. p53 and WRAP53. Our working hypothesis is that selected genes involved in tumorigenesis are regulated by their cis-acting antisense RNA. Therefore melanocytes and melanoma may differ in the expression levels of mRNAs due to the levels of antisense RNAs. To test our hypothesis we determined the transcriptome of one melanocyte (FOM191-1) and one melanoma cell line (WM793). We used rRNA-depleted total RNA for directional sequencing of 72bp reads on the Illumina platform. Sixty five million reads of WM793 and 53 million reads of FOM191-1 were available for analysis. The filtered data (e.g. removal of rRNA contaminations) were aligned to a custom built transcriptome library and the published human genome. The majority of the reads could be aligned to the human genome, although a higher percentage of the melanoma reads could not be aligned most likely due to the increased mutation rate in tumors. Although there were no differences in the levels of antisense and sense RNAs in most of the genes we found between melanocytes and melanomas for several hundred genes changes in the ratios of sense to antisense RNAs. Besides the changes in antisense expression we discovered splice variations in several genes and also differences in expression levels of microRNAs further adding additional gene expression changes. Currently we are sequencing additional melanocyte and melanoma samples to distinguish common and cell line specific hits and to determine the most common genes regulated by antisense expression in melanoma samples.


Epicutaneous TPA treatment promotes the spontaneous metastatic spread of cutaneous Hgf-Cdk4R24C melanomas through TLR4-dependent inflammatory responses in the tumor microenvironment

T. Bald1, J. Landsberg1, M. Renn1, S. Mikus1, J. Kohlmeyer1, E. Gaffal1, A. Limmer2, T. Quast3, T. Tüting

1 Laboratory of Experimental Dermatology, Department of Dermatology and Allergy, University Hospital Bonn, Bonn, Germany; 2Institutes of Molecular Medicine and Experimental Immunology, University Hospital Bonn, Bonn, Germany; 3Life and Medical Sciences Institute, Laboratory of Molecular Immunology, University of Bonn, Bonn, Germany;

The stimulation of toll receptors (TLR) represents a double-edged sword in cancer development. It activates anti-tumor immunity but also contributes to inflammatory responses that promote angiogenesis and tumor progression. Here we experimentally investigated how endogenous TLR4 ligands released in the skin following treatment with the tumor promoter 12-O-Tetradecanoylphorbol-13-acetate (TPA) affect the pathogenesis of cutaneous melanoma in the genetically engineered Hgf-Cdk4R24C mouse model. In our initial investigations we confirmed that two epicutaneous applications of TPA effectively stimulated epidermal hyperplasia, immune cell infiltration and induction of inflammatory genes in the skin of Hgf-Cdk4R24C mice. In subsequent experiments we treated cohorts of Hgf-Cdk4R24C mice with TPA twice weekly for 25 weeks. This induced a diffuse infiltrative expansion of melanoma cells locally in the skin and systemically in draining lymph nodes and lungs. Interestingly, long-term TPA treatment of Hgf-Cdk4R24C mice previously exposed to the carcinogen 7,12-dimethylbenzanthracene (DMBA) selectively increased the infiltrative and metastatic growth of melanoma cells without affecting the development of primary nodular melanomas. Repetitive epicutaneous applications of TPA also promoted the development of spontaneous metastases in lymph nodes and lungs of wild type mice bearing serial transplants of Hgf-Cdk4R24C melanomas in the skin. Exposure of cultured Hgf-Cdk4R24C melanoma cells to TPA and TLR4 ligands directly activated the expression of inflammatory genes and stimulated migration in vitro. However, the prometastatic effect of TPA treatment on melanoma transplants was completely absent in TLR4-deficient mice in vivo. Importantly, the presence of TLR4 on recipient mice was required for efficient TPA-dependent induction of inflammatory alterations associated with melanoma metastases. Taken together, these results provide experimental evidence that TLR4-dependent inflammatory responses in the microenvironment of cutaneous Hgf-Cdk4R24C melanoma contribute to the infiltrative growth and spontaneous metastatic spread of tumor cells.


Epigenetic gene inactivation in cutaneous malignant melanoma metastases

R. Tuominen1, S. Egyházi1, M. Frostvik Stolt1, D. Lindén2, C. Johansson1, C. Chang1, J. Hansson1

1 Department for Oncology and Pathology, Karolinska Institutet, Stockholm, Sweden; 2Karolinska University Hospital Solna, Solna, Sweden

Cutaneous malignant melanoma (CMM) is a treatment resistant disease, and we lack predictive markers to identify patients who respond to chemotherapy.Inactivation of the DNA repair protein O6-methylguanine-DNA methyltransferase (MGMT) through promoter hypermethylation increases significantly temozolomide (TMZ) response in glioblastoma. We are investigating whether MGMT can be used as a predictive marker also in melanoma but are also focusing on epigenetic inactivation of genes involved in cell cycle regulation and apoptotic response. We have access to two cohorts of metastatic CMM tumours (from Karolinska University Hospital, Stockholm, Sweden (cohort 1) and Leuven University Hospital, Leuven, Belgium (cohort 2)). Cohort 1 derives from DTIC/TMZ first-line monotreatment patients whereas the cohort 2 consists of DTIC + carboplatinum or cisplatinum treated tumors with multiple tumours available/patient. MGMT hypermethylation is screened using PCR, high resolution melt point analysis and pyrosequencing and compared to mRNA and protein expression for the genes of interest.

Our results regarding cohort 1 the promoter hypermethylation of MGMT show a differential methylation frequency in treatment responders versus non-responders (9/21, 42.9% versus 10/50, 20.0%, P = 0.08, Fisher's Exact Test, 86.6% successful analyses).

Cohort 2 (combination treated patients DTIC+ any of carboplatin/cisplatin) revealed also pronounced overrepresentation of MGMT promoter hypermethylation among responders versus non-responders (5/20, 25.0% versus 5/49, 10.2%, P = 0.14, Fisher's Exact Test, 92.1% successful analyses), but this cohort does not allow identification of the DTIC responders. Analysis of multiple tumors from same patients (20/24 patients with multiple tumours had concordant methylation status in all tumours), confirms MGMT promoter hypermethylation being a host characteristic as reported previously (Hassel JC et al 2010 Br J Cancer 103:820–826).

The CHEMORES project is funded by a European Union 6th framework programme grant. This work is also supported by grants from The Cancer Research Funds of Radiumhemmet.


Identifying epigenetic regulators of melanoma formation using zebrafish

E. van Rooijen1, G. van de Hoek1, L. I. Zon2

1 Division of Hematology/Oncology, Children's Hospital, Harvard Medical School, Boston, MA, USA; 2Stem Cell Program and Division of Hematology/Oncology, Children's Hospital and Dana Farber Cancer Institute, Howard Hughes Medical Institute, Harvard Stem Cell Institute, Harvard Medical School, Boston, MA, USA

Melanoma is the most aggressive and deadliest form of skin cancer. While genetic changes have been well characterized, the epigenetic contribution (changes in gene expression without alterations of the DNA sequence) to melanoma development has only recently come into focus. Many chromatin factors have been associated with misexpression during melanoma development and metastasis; however, it is currently unclear whether these represent driver or passenger lesions. We are using a unique zebrafish model to address the role of epigenetic factors in modulating melanoma. Zebrafish expressing human BRAFV600E under a melanocyte-specific promoter mitfa, develop invasive melanoma in a p53-deficient background. When mitfa loss-of-function alleles are introduced into Tg (mitfa:BRAFV600E)/p53−/− zebrafish, no melanocytes are formed and melanoma cannot develop. When melanocytes are rescued by reintroducing a minigene for mitfa, melanoma can again arise from these cells. Using a transposon-based expression vector (miniCoopR), we have developed a pooling approach to functionally test pools of five full-length human chromatin factor cDNAs in our zebrafish melanoma model. Thus far we have functionally tested 74 cDNAs and have identified six pools that severely accelerate melanoma onset. We have injected single factors to determine candidate oncogenes. These will be further evaluated using human melanoma cultures, and human melanoma tissue microarrays. The powerful combination of using an efficient pooling-based chromatin factor over-expression screen in zebrafish, with validation and further functional analyses in human cell-based assays and melanoma tissue samples, will lead to important novel insights into epigenetic regulators of melanoma. Diagnostic and/or prognostic markers could be identified, that may aid in the prevention and/or treatment of melanoma.


BRAF and NRAS mutational status are prognostically important in thick and locally advanced cutaneous melanoma

R. E. Vilain1, S. G. Braye1, L. Ashman2, R. J. Scott1,2

1 Hunter Area Pathology Service, Hunter New England Health Service, Newcastle, NSW, Australia; 2School of Biomedical Sciences and Pharmacy, University of Newcastle, Newcastle, NSW, Australia

The evolution in the understanding of kinase oncogene addiction as the driver of melanomagenic MAPK cascade activation has encouraged the molecular characterization of melanoma on the basis of potentially druggable kinase mutations. Currently, the three best-characterized targets are BRAF, KIT and NRAS. Mutations in these genes largely occur in a mutually exclusive manner, with the former two kinases offering tantalizing therapeutic opportunities, and are currently the subject of phase II and III trials. Notwithstanding the potential therapeutic opportunities offered by some of these mutations, there is conflicting data on the potential prognostic information conferred by the knowledge of the NRAS and BRAF mutational status. In this study, we have genotyped 192 cases of thick and/or metastatic melanoma, arising in a population exposed to a high degree of ambient UV radiation, and analysed the mutations in relation to the degree of cutaneous sun damage, and to clinical outcome. We report BRAF and NRAS mutations are, respectively, associated with impaired and enhanced survival outcomes. Furthermore, this effect is mediated by a differential metastatic rate to locoregional sites, and subsequently upon distant sites.


Targeting protein trafficking pathways increases melanoma sensitivity to multiple agents

Z.-M. Huang1, L. Zhong2, R. Scherzer2, S. Demetrios2, M. P. Patel3, G. Millhauser3, D. H. Oh1,2, J. E. Cleaver1, M. L. Wei1,2

1 University of California, San Francisco, CA, USA; 2Veterans Administration Medical Center, San Francisco, CA, USA; 3University of California, Santa Cruz, CA, USA

A critical barrier impeding effective therapy of metastatic melanomas is the high rate of tumor resistance to a wide variety of therapies. Novel therapies and improvement in existing therapies are urgently needed. Here we demonstrate the efficacy of targeting protein trafficking pathways in vitro in melanoma cells, using methods independent of mutational status, and demonstrate the ability to up-regulate and down-regulate tumor treatment sensitivity. We demonstrate a marked increased sensitivity of melanoma cells to cis-diaminedichloroplatinum II (cDDP, cis-platin); carboplatin; dacarbazine; or the poly(ADP ribose) polymerase (PARP) inhibitor ABT888 together with temozolomide; achieved by targeting genes that regulate both protein trafficking and the formation and function of the melanosome, an intracellular organelle unique to pigmented cells such as melanoctyes and melanoma cells. Human melanoma cells depleted of either of the protein trafficking regulators VPS33A or cappuccino/CNO, with an increased sensitivity to cDDP, had no difference in nucleotide excision repair rates or baseline apoptosis but had increased nuclear localization of fluorescently labeled cDDP, increased DNA damage by platination, and increased apoptosis after cDDP incubation. Depleted cells exhibited a decreased proportion of mature melanosomes compared to undepleted cells. Melanoma cell surface signaling by binding the melanocortin 1 receptor (MC1R) with the antagonist agouti signaling peptide (ASIP) altered protein trafficking, decreased the proportion of mature melanosomes formed and increased cDDP sensitivity, while receptor binding with the agonist melanocyte stimulating hormone (MSH) resulted in differentially altered protein trafficking, an increased proportion of mature melanosomes formed and increased resistance to cDDP. These results indicate that targeting protein trafficking molecules markedly increases melanoma treatment sensitivity to multiple agents, likely by influencing the degree of sequestration of the treatment agent into mature melanosomes.


GRM1 signal transduction activates pro-angiogenic pathways in human melanoma

Y. Wen1, J. Li1, S. Shin1, S. Chen2, J. Goydos1

1 Cancer Institute of New Jersey, New Brunswick, NJ, USA; 2Susan Lehman Cullman Laboratory for Cancer Research, Rutgers University, Piscataway, NJ, USA

We recently reported the results of Phase 0 and Phase II clinical trials targeting GRM1 signaling using a glutamate release inhibitor, riluzole, in patients with advanced melanoma. We demonstrated reduction in tumor size in 34% of patients treated on these trials. To examine the function of GRM1 signal in melanoma, human melanoma cell lines over-expressing GRM1 (G2 and G4) were constructed using the low-GRM1-expressing UACC903 human melanoma cell line. We found that G2 and G4 cells produced much larger tumors in nude mice than vector control V1 cells and did not demonstrate hemorrhage necrosis found in the smaller V1 xenografts. Using high-molecular-weight FITC-dextran injected to the left ventricle of the animals, we found significantly more abundant tumor blood vessels in G2 and G4 xenografts than V1. We also found that conditioned medium from G2 and G4 cells, but not from V1 cells, stimulated endothelial cell migration. We next examined whether increased GRM1 expression resulted in the release of angiogenesis factor using an angiogenesis antibody array. We found that the conditioned medium from G2 had higher levels of IL-8 and VEGF compared with V1. We confirmed the increased IL-8 and VEGF in the conditioned medium of G2 and G4 by ELISA. We obtained pre- and post-treatment tumor samples from patients on our recently completed Phase II trial of single agent riluzole in patients with advanced melanoma and performed Western blotting on protein lysates from these tumor samples. We found that IL-8, VEGF, and the vascular marker CD31 were all decreased in the post-treatment samples from patients whose tumors responded to riluzole but not in those samples from patients who did not respond. We conclude that over-expression of human GRM1 activates angiogenesis, and renders cells more susceptible to the inhibitory effects of riluzole. Further researches on angiogenesis mechanism of GRM1 are under way.


Activation of apoptotic pathways by attenuated vaccinia viruses in melanoma

X. Wu1,2, P. Dai2, W. Wang1,2, T. Feldman3,4, X. Jiang3,4, A. C. Halpern1,4, A. N. Houghton4,5, S. Shuman2,4, T. Merghoub4,5, J. Wolchok4,5, L. Deng1,4

1 Dermatology Service, Department of Medicine, Memorial Sloan-Kettering Cancer Center, New York, NY, USA; 2Molecular Biology Program, Memorial Sloan-Kettering Cancer Center, New York, NY, USA; 3Cell Biology Program, Memorial Sloan-Kettering Cancer Center, New York, NY, USA; 4Weill Medical College of Cornell University, New York, NY, USA; 5Immunology Program, Memorial Sloan-Kettering Cancer Center, New York, NY, USA

Advanced melanoma is largely refractory to conventional therapies, including chemotherapy and radiotherapy. Immunological approaches to melanoma treatment have been proven to be beneficial in animal models and human clinical trials and deserve to be actively explored. Poxviruses are large cytoplasmic DNA viruses that hold promises as vectors for immunotherapy and as oncolytic agents. Poxviruses encode many genes that evade the host immune system; chief among them is vaccinia E3 virulence gene. Mutant vaccinia with E3 deletion (E3L) is non-pathogenic in mice. We have previously shown that E3L infection of melanoma cells induces apoptosis and intratumoral injection of E3L into intradermally implanted poorly immunogenic murine B16 melanomas results in the regression of the tumors at the injection sites as well as rejection of tumor challenge at the contralateral side. In this study, we focus on the mechanisms through which vaccinia E3L or MVAE3L (deletion of E3L from a highly attenuated Modified vaccinia virus Ankara strain) induces apoptosis in melanoma cells and fibroblasts. Our results indicate that attenuated vaccinia viruses have potentials to be developed as novel therapeutics for melanoma.


Vitamin D status and VDR expression patterns at presentation with CMM in relation to histological prognostic markers

C. Wyatt1, C. Hurst1, R. Lucas2, M. Kimlin1

1 Queensland University of Technology, Brisbane, Queensland, Australia; 2Australian National University, Canberra, Australia

Queensland has the highest rate of cutaneous malignant melanoma (CMM) in the world. Exposure to solar ultraviolet radiation (UVR) is the main risk factor for CMM in fair-skinned populations, however there is growing evidence that sun exposure is paradoxically protective against mortality from melanoma. CMM is found more frequently on areas of the body habitually covered by clothing and studies report that CMM is more common among indoor workers than outdoor workers and in those with intermittent rather than chronic patterns of UVR exposure. One postulated explanation of this paradox is that higher levels of sun exposure are associated with higher vitamin D status, and vitamin D has known anti-tumour effects, through several pathways. However few studies have been conducted to date to investigate this hypothesis. We aimed to identify the relationship between the vitamin D status and vitamin D receptor (VDR) expression patterns of patients with newly diagnosed melanoma and surrogate histological markers associated with prognosis. Using a comprehensive multi-method approach, 100 CMM patients were recruited immediately prior to the wider excision of their CMM, from two plastic surgeons based in Brisbane, Queensland. Blood was taken pre-operatively to measure vitamin D status, a questionnaire provided data on demographics, sun exposure history, sun protection habits, skin cancer history and lifetime places of residence, and a clinical examination recorded number of naevi and skin colour (by spectrophotometry). Tissue from the tumour block was stained to examine VDR expression patterns and to quantify levels of solar elastosis, while the outcome measures of Breslow thickness, Clark level, ulceration and mitotic rate were retrieved from the histopathology reports. The results of analyses of the relationships between markers of melanoma prognosis and VDR expression and vitamin D status at the time of diagnosis, adjusted for possible confounding factors, will be presented.


Treatment targeting hypoxic tumor cells enhances therapeutic effect of sunitinib in melanoma

S. J. Liu, X. Xu

Department of Pathology and Laboratory Medicine, University of Pennsylvania, Philadelphia, PA, USA

Neoangiogenesis is a critical step during tumor progression. Antiangiogenic therapy has shown great promise in cancer treatment. However, the delay in tumor progression associated with antiangiogenic therapy is often not translated into increased overall survival time. It has been postulated that antiangiogenic therapy may lead to emergence of more aggressive cancer phenotype through increased hypoxia in the solid tumor. Tumor hypoxia is well recognized to promote tumor progression. TH-302, a 2-nitroimidazole triggered hypoxia-activated prodrug (HAP), releases bromo-isophosphoramide mustard (Br-IPM), which induces DNA crosslinking under hypoxic condition. TH-302 has been shown in human tumor xenograft models to target the hypoxic compartment selectively and reduce tumor volume. However, subcutaneous xenografts are often poorly vasculized with spontaneous tumor necrosis and different from tumors in human. Here, we report a series of studies to investigate whether TH-302 by itself or in combination with antiangiogenic therapy may increase the survival of melanoma bearing transgenic mice (Tyr::CreER; BRAFCa/+; Ptenlox/lox). To visualize hypoxia within the tissue, we injected hypoxyprobe 1 h before sacrificing the mice, and immunohistochemical stain was performed to detect the amount of hypoxyprobe in the tumor tissues. We found that hypoxyprobe was present in the cutaneous melanoma and sunitinib significantly increased the amount of hypoxyprobe in the tumor as well as the expression level of HIF-1α and HIF-2α in the tumor tissue by western blot analysis. Treatment with TH-302 alone for 2 weeks significantly increased survival of the melanoma bearing transgenic mice. Treatment with Sunitinib or Sunitinib plus TH-302 for 1 week did not change survival of the melanoma bearing mice. On the contrary, long-term treatment (4 weeks) of Sunitinib starting 1 day after melanoma induction or when the tumor size reached approximately 100–150 mm2, moderately increased the survival of the melanoma bearing mice. In addition, treatment with combination of Sunitinib and TH-302 further prolonged the lifespan of the melanoma bearing transgenic mice in both regimens. Interestingly, when we combined temozolomide with TH-302, there was no effect on mouse survival, suggesting that TH-302 may only enhance the therapeutic efficacy of certain treatment agents. These studies provide a translational rationale for combining TH-302 with antiangiogenics to increase the treatment benefit of antiantiogenics in melanoma.


Hsa-miR-200c inhibits melanoma progression through down-regulation of Bmi-1 and up-regulation of E-cadherin

S. J. Liu, M. T. Tetzlaff, X. Xu

University of Pennsylvania, Philadelphia, PA, USA

MicroRNAs (miRNAs) are short noncoding RNAs that post-transcriptionally regulate gene expression. Hsa-miR-200c has been shown to be involved in epithelial cancer progression. Here we show that hsa-miR-200c is down-regulated in metastatic melanomas compared to primary melanomas. Over-expression of miR-200c in melanoma cells resulted in significantly decreased cell proliferation, self-renewal capacity, drug resistance and migratory capacity. Mechanistically, miR-200c over-expression induced significant down-regulation of Bmi-1 expression with a concomitant increase in E-cadherin expression. The effects of miR-200c on E-cadherin expression and melanoma cell proliferation, self-renew and migration can be rescued by over-expression of Bmi-1 in these cells. Furthermore, miR-200c over-expression resulted in significantly decreased melanoma growth and metastasis in vivo. The xenografts also showed Bmi-1 down-regulation and E-caherin up-regulation. In summary, miR-200c inhibits melanoma proliferation and metastasis through suppression of the Bmi-1 expression and increasing E-cadherin expression. Our results suggest that miR-200c may as a target for melanoma treatment.


Fascin is critical for transforming growth factor beta induced migration and invasion in melanoma and other tumour cells

J. Sun1,2,*, H. He1,2,*, Y. Xiong3, S. Lu1,2, J. Shen1,2, A. Cheng1,2, J. M. Lancaster3, M. Kim1,4, S. Yang1,2

1 Comprehensive Melanoma Research Center, H. Lee Moffitt Cancer Center & Research Institute, Tampa, FL, USA; 2Department of Tumor Biology, H. Lee Moffitt Cancer Center & Research Institute, Tampa, FL, USA; 3Women's Oncology, H. Lee Moffitt Cancer Center & Research Institute, Tampa, FL, USA; 4Molecular Oncology, H. Lee Moffitt Cancer Center & Research Institute, Tampa, FL, USA

Fascin has been associated with poor prognosis, shorter survival and more metastatic diseases in multiple tumors. It is believed that fascin facilitates tumor metastasis by promoting the formation of invasive membrane protrusions. However, the mechanisms by which fascin is overexpressed in tumors are not clear. TGFβ is a cytokine secreted by tumor and mesenchymal cells and promotes metastasis in many late stage tumors. The pro-metastasis mechanisms of TGFβ remain to be fully elucidated. Here we demonstrated that TGFβ induced fascin expression in melanoma and other tumor cells through the canonical Smad-dependent pathway. Fascin was critical for TGFβ promoted filopodia formation, migration and invasion. More importantly, fascin expression significantly correlates with TGFβ1 and TGFβ receptor I levels in a cohort of tumor patients. Our results indicate that elevated TGFβ level in tumor microenvironment may be responsible for fascin overexpression in some of the metastatic tumors. Our data also suggest that fascin could play a central role in TGFβ-promoted tumor metastasis.


Clinicopathologic analysis and BRAF and KIT mutation in 29 amelanotic acral melanomas

S. J. Yun1, S. M. Chun1, Y. D. Choi2, S. A. Jin1

1 Department of Dermatology, Chonnam National University Medical School, Gwangu, South Korea; 2Department of Pathology, Chonnam National University Medical School, Gwangu, South Korea

Amelanotic acral melanoma (AAM) is rare, and often difficult to diagnose AAM because of its wide range of clinical appearances and lack of pigmentation. The prognosis of AAM may be poor owing to misdiagnosis and delay in diagnosis. Although biopsy of amelanotic melanoma may typically yield the correct diagnosis, immunohistiochemical stains such as S100 and HMB45 frequently provide diagnostic information in cases histologically challenging. Sometimes, HMB45 stain is focally positive or absent. In this study, we conducted a restrospective case review of AAM from 2003 to 2011 in Chonnam National University Hospital. Total 29 cases were found and analyzed by clinical and demographic data, such as sex, age, the clinical features, and the site of presentation. In addition, pathologic parameters including pathologic subtype, cell type, ulceration, Breslow thickness, and mitotic counts were observed. There were 20 complete types and nine incomplete types. Eighteen cases had nodular melanoma, eleven had acrallentiginous melanoma, and there were no superficial spreading melanoma. Melanoma cell types of 26 cases were epithelioid (19), spindle (8), small cell (5) or rhabdoid (4) types. All 25 cases had an ulcer on the surface except only four cases. Fourteen cases had subungual melanomas. Immunohistochemical stains were performed in 24 of 29 cases. Fourteen of 24 cases were positive for HMB45, ten were negative or focal positive. Especially, eight of 10 cases with negative or focal positive for HMB-45 were complete types. BRAF and KIT mutation was performed in 25 of 29 cases. BRAF mutation was discovered in only one of 25 cases. KIT mutation was discovered in five of 25 cases. One case with BRAF mutation also had KIT mutation. These results suggest that AAM show nodular, subungual melanomas with various cell types. HMB45 stain is variable, and KIT mutation is detected in 20.0%.


Targeting anti-apoptotic mechanisms for reversal of resistance to BRAF inhibitors in melanoma

P. Hersey1–4, D. Wroblewski1,3,4, F. Lai1,3,4, C. C. Jiang1,3,4, X. D. Zhang1,3,4

1 Melanoma Institute Australia, Sydney, NSW, Australia; 2Kolling Institute of Medical Research, University of Sydney, NSW, Australia; 3Priority Research Center for Cancer Research, School of Medicine and Public Health, University of Newcastle, NSW, Australia; 4Cancer Research Program, Hunter Medical Research Institute, NSW, Australia

We have recently demonstrated that prolonged exposure to PLX4720 at concentrations that induce high levels of apoptosis in most BRAFV600E melanoma cell lines results in BRAFV600E melanoma cells resistant to PLX4720-induced apoptosis. These cells can proliferate, albeit with reduced growth rate, in the presence of PLX4720, but regain high growth rate once the inhibitor is withdrawn. This mirrors closely the observation in clinical trials that mutant BRAF melanomas frequently regress in response to specific inhibitors, but rebound in growth after limited durations. These results strongly suggest that induction of cell death is a major determinant of long-term responses of mutant BRAF melanomas to specific inhibitors. We show here that targeting anti-apoptotic mechanisms by the Bcl-2 inhibitor ABT737 or the histone deacetylase inhibitor (HDACi) SAHA reverses resistance of mutant BRAF melanoma cells to PLX4720-induced apoptosis. PLX4720 in combination with ABT737 or SAHA synergistically induced apoptosis in treatment naïve mutant BRAF melanoma cell lines and fresh melanoma isolates. Strikingly, the combinations similarly synergized in induction of apoptosis in short-term cultured mutant BRAF melanoma cell lines established from melanomas relapsed after in vivo administration of vemurafenib. The synergistic effect between PLX4720 and ABT737 appeared to be mediated by the BH3-only protein Bim, in that siRNA knockdown of Bim significantly blocked induction of apoptosis, although there was no further increase in the Bim expression by co-treatment with the compounds. In contrast, Bim was further up-regulated by the combination of PLX4720 and SAHA, but its inhibition by siRNA had only limited effect on induction of apoptosis, suggesting that redundant apoptosis-initiating mechanisms were activated by the combination, as Bim is known to mediate apoptosis induced by either SAHA or PLX4720 alone. Regardless, our results identify Bcl-2 inhibitors and HDACi as promising agents to enhance therapeutic efficacy of mutant BRAF inhibitors in the treatment of melanoma.


Therapy-induced senescence in human melanoma cells as an integrated form of stress response

G. Zhang1,2, W. Zhi3, M. Herlyn1

1 The Wistar Institute, Philadelphia, PA, USA; 2Graduate Program in Bioengineering, University of Pennsylvania, Philadelphia, PA, USA; 3Department of Computer Science, New Jersey Institute of Technology, Newark, NJ, USA

Oncogene-induced senescence (OIS) is a well-documented phenomenon both in vitro and in vivo as a tumor suppressive mechanism. Premalignant cells quickly develop a sophisticated yet not well-understood mechanism to overcome and subsequently repress OIS to facilitate its malignant transformation. Here we have focused on therapy-induced senescence by inhibiting Aurora kinase through Aurora kinase(s) inhibitors in a panel of human melanoma cell lines. By implementing an integrated genomics approach, we have uncovered the necessity of autophagy/lysosome, unfolded protein response (UPR), and mTOR signaling pathways for therapy-induced senescence as an integrated stress response. Particularly, the mTOR inhibitor Rapamycin suppresses the senescence phenotype by converting senescence to quiescence, whereas autophagy inhibitor Bafilomycin A1 (BafA1) or Hydroxychloroquine (HCQ) can induce cell death in melanoma cells when combined with Aurora kinase inhibitors, suggesting that autophagy is one of survival mechanisms for melanoma cells undergoing therapy-induced senescence. Staining cells with autophagy marker LC3B has indicated that some but not all cells up-regulate autophagy when undergoing therapy-induced senescence. siATG5 can partially inhibit therapy-induced senescence in melanoma cells, arguing other unknown mechanisms can also underlie this stress response. Aurora kinase inhibition mediated senescence is terminal with few exceptions in specific cell lines. In contrast, selective B-RAF inhibitors initially induce ‘pseudo senescence’ in B-RAF mutant melanoma cells, as those stressed cells can adapt to and quickly recover upon the withdraw of drugs. Combination of Aurora kinase inhibition with BH3-mimetic drugs can efficiently induce cell death. Our study can potentially lead to a framework of utilizing the pro-senescence therapy followed by converting senescence to cell death to eliminate all melanoma cells as a novel clinic approach toward melanoma intervention.


Not playing by the rules: the tumor-stroma game in melanoma progression

E. Kim1, D. Basanta1, V. Rebecca2, J. Messina3, R. Mathew3, K. S. Smalley1,2,4, A. R. A. Anderson1

1 Integrated Mathematical Oncology Department, H. Lee Moffitt Cancer Center and Research Institute, Tampa, FL, USA; 2Department of Cutaneous Oncology, H. Lee Moffitt Cancer Center and Research Institute, Tampa, FL, USA; 3College of Medicine Pathology and Cell Biology, University of South Florida, Tampa, FL, USA; 4Molecular Oncology, H. Lee Moffitt Cancer Center and Research Institute, Tampa, FL, USA

Melanoma development is known to involve genetic changes in melanocytes as well as the disruption of both cell-cell and cell-microenvironment interactions. However, the mechanisms by which the deregulated interactions lead to melanoma development have not yet been fully understood. It is our view, that we must first model normal skin form and the regulatory mechanisms that maintain skin homeostasis before we can model cancer initiation. To this end, we have developed a hybrid multiscale mathematical model of normal skin (virtual skin). The model focuses on key cellular (melanocytes, keratinocytes, fibroblasts) and microenvironmental variables (growth factors, ECM) that regulate normal skin homeostasis.

We examine how changes in microenvironmental factors will affect skin structure and interactions between keratinocytes, melanocytes, and fibroblasts. We first systemically perturbed the virtual skin model by maximizing growth factor concentration in the domain during several periods of time. The simulations show that the virtual skin system recovers from small changes but fails to compensate for large perturbations. We also incorporated senescent fibroblasts, which produce multiple growth factors, matrix and proteases. Simulations show these constitutively activated fibroblasts stimulate nearby fibroblasts to proliferate, that then feedback on melanocyte growth.

Successful melanoma initiation requires preexisting mutations within the melanocyte population, allowing them to exploit stromal activations. To test this hypothesis we incorporated mutant melanocytes into the model. Model simulations provide a series of virtual skin pathologies that readily recapitulate a spectrum of true aberrant clinical pathologies.

Collectively, the interplay between melanocytes and stromal activation play a significant role on driving melanoma development. Based on our simulations combined with clinical data, we speculate that constitutively activated fibroblasts may create a pro-oncogenic environment that synergizes with mutations to drive melanoma initiation and progression.


Human melanoma tumors contain cancer stem cells with a high aldehyde dehydrogenase activity and a distinct molecular signature

Y. Luo1, K. Dallaglio1,2, Y. Chen3, S. Robinson4, M. D. McCarter5, R. Gonzalez4, K. D. Lewis4, D. R. Roop1,2, V. Vasiliou3, W. A. Robinson4, M. Fujita1,2

1 Department of Dermatology, University of Colorado Denver, Aurora, CO, USA; 2Charles C Gates Regenerative Medicine and Stem Cell Biology, University of Colorado Denver, Aurora, CO, USA; 3Department of Pharmaceutical Sciences, University of Colorado Denver, Aurora, CO, USA; 4Division of Medical Oncology, Department of Medicine, University of Colorado Denver, Aurora, CO, USA; 5Department of Surgery, University of Colorado Denver, Aurora, CO, USA

Although cancer stem cell concept is accepted for many tumors, the existence of cancer stem cells has been the subject of debate in human melanoma because of high frequency of tumor initiation in highly permissive xenotransplantation condition using NOD/SCID interleukin-2 receptor γ-chain-null (NSG) mice. However, very recent studies revealed evidence for melanoma stem cells even in NSG mice. Here we identified melanoma stem cells with a high aldehyde dehydrogenase (ALDH) activity from melanoma tumor specimens and characterized their gene expression. ALDH+ melanoma cells were 18-fold more tumorigenic than ALDH- cells in NOD/SCID mice (P < 1e−12). ALDH+ cells self-renewed and differentiated in serial xenotransplantion. When NSG mice were used as recipients, we found a trend toward increased tumorigenesis in ALDH- cells (10-fold increase), whereas the frequency of tumor initiation from ALDH+ cells was not affected in NSG mice compared with NOD/SCID mice, implicating immunoevasion mechanism of ALDH+ cells in NOD/SCID mice. Furthermore, ALDH+ cells were more tumorigenic than ALDH- cells even in NSG mice (P = 0.0159). Microarray analysis and gene expression profiling of ALDH+ cells identified 99 upregulated genes and 48 downregulated genes including genes associated with stem cell function. Among them, FOXP1, CDC42, USH1C, and ADAR were further confirmed. ALDH1A1 and ALDH1A3 were predominant ALDH isozymes contributing to the high ALDH activity in human melanoma, and knockdown of ALDH1A1 or ALDH1A3 by siRNA led to the decreased cell viability associated with induction of cell cycle arrest and apoptosis. Finally, we showed that melanoma tumor growth was inhibited by shRNA knockdown of ALDH1A1 or ALDH1A3 in vivo. The study confirms the presence of melanoma stem cells in human melanoma tumors and identifies gene signatures, which would be potential targets for effective therapy of human melanoma.


UV radiation interacts with Endothelin-3 and loss of the Cdkn2a locus to promote melanomagenesis

A. Saldana-Caboverde, J. Ortega, L. Kos

Florida International University, Miami, FL, USA

Exposure to ultraviolet (UV) radiation is the most important environmental risk factor associated with melanoma but it is unclear how it leads to its development. Melanoma risk is also influenced by the incidence of inherited mutations in melanoma susceptibility loci, such as the CDKN2A locus. Recently, the receptor for the Endothelin-3 (Edn3) peptide, the G-coupled endothelin receptor b (Ednrb), was found to be upregulated in melanoma metastases. We hypothesize that the over-activation of endothelin signaling interacts with the melanoma inducing properties of UV radiation to bring about melanoma development and progression and that loss of the Cdkn2a locus enhances Edn3-induced effects to trigger melanomagenesis at a more rapid rate. We aim to develop a UV-dependent mouse model of melanoma driven by the over-activation of Edn3. For this purpose, K5-Edn3 transgenic mice, which have increased melanocyte numbers in the epidermal-dermal junction, were crossed with Cdkn2a-/-mice to examine the role of this locus in UV-induced melanomagenesis. Preliminary results show that upon neonatal exposure to UV radiation, skin tumors presenting histopathological features that resemble nodular melanoma arise in K5-Edn3 transgenic mice. These tumors developed more frequently and rapidly in K5-Edn3 mice having null deletions of the Cdkn2a locus. Furthermore, exposure to a second dose of UV radiation during adulthood resulted in increased tumor penetrance (from 23% to 77%) and decreased time to tumor development. Local lymph node metastases have been found in 20% of animals harboring nodular melanomas. Our results suggest that the over-activation of endothelin signaling interacts with the loss of the Cdkn2a locus and the melanoma inducing properties of UV radiation to elicit melanoma development. This study may provide the first UV-inducible animal model of melanoma based on the over-activation of a G-coupled receptor.


Identification and functional validation of therapeutic kinase targets downstream of V600EBRAF

A. Sharma1,2,3, S. Madhunapantula1,2,3, R. Gowda1,2,3, R. Neves1,2,3, G. P. Robertson1–6

1 Department of Pharmacology, The Pennsylvania State University College of Medicine, Hershey, PA, USA; 2The Melanoma Center, Hershey, PA, USA; 3The Melanoma Therapeutics Program, 500 University Drive, Hershey, PA, USA; 4Department of Surgery, The Pennsylvania State University College of Medicine, Hershey, PA, USA; 5Department of Pathology, The Pennsylvania State University College of Medicine, Hershey, PA, USA; 6Department of Dermatology, The Pennsylvania State University College of Medicine, Hershey, PA, USA

B-Raf is the most mutated gene in melanoma with ~60% of patients having mutant V600E protein. Targeted therapies such as PLX4032, known clinically as vemurafenib, has been developed that preferentially binds to V600EB-Raf protein to inactivate the constitutively active kinase. Vemurafenib has been shown to have a high clinical response rates in patients with stage IIIC and IV BRAF-mutant melanoma and has been approved for clinical use by the FDA. Although this represents a major advancement in the application of targeted therapeutics for melanoma, this treatment is not curative and responses have been relatively short lived. Pharmacological agents targeting the MAP kinase pathway, such as vemurafenib or Mek-1/2 inhibitors, show promise in the clinic but development of resistance or more aggressive invasive disease suggests other members of the same cascade might need to be targeted in combination with vemurafenib or sequentially to more effectively treat this disease. However, the identity and mechanism of action of many of these proteins remain to be established. To identify kinase targets downstream of V600EB-Raf, a siRNA-based screen was undertaken. Aurora kinase B (AURK B) and Wee1 were identified and signaling validated as lying downstream of the V600EB-Raf signaling cascade. AURKB and Wee1 protein levels were overexpressed in 83% and ~45% of melanoma patient tumors respectively and in cell lines with highest amounts found in those derived from advanced disease. Targeting AURKB or Wee1 using siRNA or the pharmacological agent VX-680 or Wee1 Inhibitor II in cultured cells decreased cellular proliferation inducing a G2/M block and increasing the subG0/G1 population, which had the consequence of decreasing tumor development by 30–80%. siRNA or pharmacological agent (PLX4032, UO126) mediated inhibition decreased expression and activity, thereby demonstrating the protein could be used as a biomarker for assessing efficacy of drugs targeting the V600EB-Raf siganling pathway. These results validate these genes as potentially important therapeutic targets lying downstream of V600EB-Raf in melanoma cells.


LC3B expression, a marker for autophagy, is a common feature of melanoma and other solid tumors and associated with proliferation, metastasis and poor outcome

R. Lazova1,2, R. L. Camp2,3, V. Klump1, S. F. Siddiqui3, R. K. Amaravadi4,5, J. M. Pawelek1,2

1 Department of Dermatology, Yale University School of Medicine, New Haven, CT, USA; 2Yale Cancer Center, Yale University School of Medicine, New Haven, CT, USA; 3Department of Pathology, Yale University School of Medicine, New Haven, CT, USA; 4Yale University School of Medicine, New Haven, CT, USA; 5Abramson Cancer Center and Department of Medicine, University of Pennsylvania, Philadelphia, PA, USA

Measurement of autophagy in melanoma and other cancers and correlation with histopathologic grading or clinical outcomes has been limited. Accordingly, we investigated LC3B as an autophagosome marker by analyzing nearly 1100 melanomas, 600 breast carcinimas and tissue specimens from other 20 types of cancer, focussing on LC3B staining and clinical outcomes in melanoma and breast cancer. IHC and IF protocols were developed for automated quantitative analysis (AQUA) using antibodies versus LC3 isoform B (LC3B) and Ki-67. Multiple tumor microarrays, including clinically annotated breast and melanoma TMA's were employed. An AQUA program was developed to quantitate LC3B distribution in punctate and diffuse compartments of the cytoplasm. LC3B staining was moderate to high in 87% of melanomas and 84% of all tumors in general. The area occupied by punctate LC3B was elevated three to five fold in tumors with the highest LC3B intensities versus the lowest. LC3B and Ki-67 showed strong correlations (P < 0.0001) and mitotic figures were most often seen in high versus low LC3B tumors, together indicating an association of autophagy with proliferation. In melanoma, LC3B levels were elevated in lymph nodes, and in breast cancers in node-positive versus node-negative primaries and associated with increased nuclear grade and shortened survival. Thus, high LC3B was associated with proliferation, metastasis, high nuclear grade and worse outcome. The results indicate a surprisingly common expression of LC3B in melanoma and other malignancies and strongly support emerging evidence that autophagy plays a significant role in cancer progression. It is thus becoming apparent that autophagy presents an important new target of therapeutic vulnerability in malignancy.


Using organotypic skin cultures as a platform for screening novel therapeutic agents with anti-melanoma activity

M. Tiago, C. A. Brohem, E. M. Oliveira, R. D. Paes, A. Campa, S. B. M. Barros, S. S. Maria-Engler

Department of Clinical Chemistry and Toxicology, Faculty of Pharmaceutical Sciences, University of São Paulo, Sao Paulo, Brazil

There is a growing demand for the use of reconstructed skin and dermis in the laboratory for in vitro assays of cytotoxicity, viability, cell growth, irritability, and to properly evaluate the interactions between cells and their extracellular matrix. The development of genetically controlled and well characterized skin models have important implications, not only for scientists and physicians, but also for manufacturers, consumers, governing regulatory boards and also animal welfare organizations. It has long been suggested that cells cultured on 2D substrates such as culture plates, lose a myriad of important signals, key regulators, as well as the underlying tissue phenotype. Strategies that allow the reconstruction of artificial human skin equivalents in a 3D setting, including both dermal and epidermal components are expected to provide important insights into skin physiology and pathophysiology that cannot be answered solely in the context of monolayer tissue culture. Skin models can be broadly categorized into (i) those containing only epidermal components, (ii) grafts consisting of dermal components alone, and (iii) full-thickness composite grafts containing both epidermal and dermal components. Using the dermal equivalent model we characterized the mechanisms of cell viability, invasion and inflammation of human metastatic melanoma cell lines, with the goal of using this to investigate new therapeutic approaches for the treatment of melanoma. Our results show that this model accurately recreates pathophysiology of melanoma, with the cells showing an increased resistance to death mechanisms and an increase in the expression of metalloproteinases and cytokines that favor tumor progression. In our ongoing studies, we aim to use this dermal melanoma model as a platform for the screening and evaluation of novel compounds with anti-melanoma activity.


The complex role of sun exposure on melanoma progression

S. Gandini1, R. B. Buffon1, P. Autier2, G. Tosti1, F. Baldini1, G. Spadola1, F. Cataldo1, A. Testori1

1 European Institute of Oncology, Milan, Italy; 2International Prevention Research Institute, Lyon, France

Sunlight is recognized as the main environmental cause of skin cancer, and in 2009 the WHO classified the full spectrum of ultraviolet radiation (UV) as a human carcinogen. However few studies have showed that UV exposure prior to diagnosis may be associated with improved melanoma (MM) survival. The aim of this analysis is to verify the hypothesis that sunny vacations are associated with a reduced risk of disease progression in patients with a diagnosis of MM.

We investigated the relationship between UV exposure and MM prognostic factors, taking into account potential confounding factors. We reviewed the clinical database of patients who had received a histological diagnosis between 1995 and 2009 at the Division of Melanoma of the European Institute of Oncology (Milan, Italy). We analyzed 446 melanoma cases: median age 44, 46% were men and 75% were persons with high educational level. Forty-one percent of cases had sunny vacations in the last 2 years before diagnosis. Two hundred and forty-two (54%) had Breslow thickness >1 mm (the ‘thick’ MM). Thick melanoma were more frequent among older patients (65% at age>60) than among younger (51%; P = 0.02), in men than in women (63% versus 47; P = 0.001), in patients with a lower educational level (70% versus 49%; P = 0.0001), and in patients who had vacations in tropical countries (61% versus 53%; P = 0.15). After adjusting for gender, education level, histology and cutaneous site, patients who had had sunny vacations before diagnosis were found with significantly lower tumor thickness (P = 0.05). In contrast, sunny vacation after the diagnosis of melanoma was not associated with Breslow thickness.

Significant inverse association between sun exposure and Breslow suggest that some factors associated with sun exposure may play a more complex role in the progression of melanoma than previously thought.


European ancestry, host factors and polymorphisms in DNA repair genes modify the risk of cutaneous melanoma: a case-control study in a high UV index region in Brazil

G. Francisco1, F. T. Gonçalves2, O. C. Luiz3, S. P. de Souza3, C. Festa-Neto4, J. A. Sanches4, R. Chammas1, G. J. F. Gattas2, J. Eluf-Neto3

1 Laboratório de Oncologia Experimental, Departamento de Radiologia e Oncologia and Instituto do Câncer do Estado de São Paulo; 2Departamento de Medicina Legal; 3Departamento de Medicina Preventiva, Faculdade de Medicina da Universidade de São Paulo; 4Departamento de Dermatologia, Faculdade de Medicina da Universidade de São Paulo

UV radiation is the major environmental factor related to development of cutaneous melanoma. Besides sun exposure and the influence of latitude, some host characteristics such as skin phototype, hair and eye color are also risk factors to melanoma. Polymorphisms in DNA repair genes could also be good candidates for susceptibility genes for melanoma, mainly in global regions exposed to high solar radiation. We carried out a hospital-based case-control study in Brazil, to evaluate the contribution of host factors and genetic polymorphisms of DNA repair to the risk of melanoma. A total of 412 (202 patients with melanoma and 210 controls) were analyzed regarding host characteristics for melanoma risk as well as to 11 polymorphisms in DNA repair genes. We found an association of some host characteristics with melanoma development, such as light eye and hair color, fair skin, presence of nevi, sunburns in childhood and adolescence and also European ancestry (OR = 4.68; 95% CI = 2.66–8.24). Regarding DNA repair gene polymorphisms, we found an association for the XPG 1104 His/His genotype (OR = 0.32; 95% CI = 0.13–0.75) and for three polymorphisms in XPC gene (PAT+; IV-6A and 939Gln), which represent a haplotype for XPC. Melanoma risk was higher to individuals carrying the complete XPC haplotype than each individual polymorphism (OR = 3.64; 95% CI = 1.77–7.48). Association of the host characteristics with XPC haplotype also contributed to melanoma risk. Our data indicate that host factors, European ancestry and XPC polymorphisms contributed to melanoma risk in a region exposed to high sun radiation, such as Brazil.

Supported by FAPESP and CNPq.


Prohibitin protects human metastatic melanoma cells from cisplatin-induced cell death

T. Tortelli Jr, A. Otake, R. F. Saito, N. Schwarz, R. Chammas

Laboratório de Oncologia Experimental, Faculdade de Medicina da Universidade de São Paulo, Instituto do Câncer do Estado de São Paulo, Brazil

Melanoma incidence is increasing worldwide and represents a clinical challenge, as its treatment outcome is still poor. A possible cause to the failure of melanoma treatment is the development of chemoresistance, whose molecular bases are still poorly understood. Previously, a proteomic approach was conducted to identify differences in protein expression or accumulation using cisplatin in LB373 cell line and prohibitin (PHB) was among the molecules which accumulated upon cisplatin treatment. Here, we have investigated prohibitin, a mitochondrial chaperone and an E2F inhibitor, accumulation and subcellular compartmentalization in cisplatin-induced cell death in human metastatic melanoma cell lines (Mel 85, SKMel37 and LB373). Prohibitin was observed co-localized to the mitochondria or within nuclear compartments co-localized to minichromosome maintenance complex 7 (MCM7). Prohibitin expression was also induced by cisplatin in others melanoma cells and its overexpression is associated to cisplatin-induced resistance in the LB373 and Mel 85 melanoma cell line. Other cellular stress system, like cell starvation and the endoplasmic reticulum stress, could also induce prohibitin accumulation showing its correlation to the unfolded protein response (UPR) at the endoplasmic reticulum and others mechanisms of cellular stress. Our data suggest that Prohibitin may be involved in melanoma resistance, not only for its overexpression, but also due to its nuclear compartmentalization. Nuclear reorganization, as evidenced by prohibitin compartmentalization, accompanies the cellular response to cisplatin towards survival. Also, prohibitin may be part of other stress mechanisms and can be a potential target for melanoma treatment.

Supported by FAPESP and CNPq.


BRAFV600E transforms skin neural crest stem cells via bypassing oncogene-induced senescence

S. M. Kumar1, H. Yu1, R. Yang1, K. Nathanson2, S. Liu1, X. Xu1

1 Departments of Pathology and Laboratory Medicine, University of Pennsylvania School of Medicine, Philadelphia, PA, USA; 2Department of Medicine, University of Pennsylvania School of Medicine, Philadelphia, PA, USA

Adult stem cells are multipotent cells that persist in a small number in adult tissues throughout the lifespan of an organism. Unlike differentiated cells, adult stem cells are intrinsically resistant to senescence. It is unclear how adult stem cells in solid organs respond to oncogenic stimulation and whether these cells play a role in cancer initiation. We report here that expression of BRAFV600E in human skin neural crest stem cells (hNCSCs) did not result in oncogene-induced senescence (OIS) but increased their proliferation potential. hNCSCs with BRAFV600E acquired anchorage independent growth capacity and formed small subcutaneous tumors in vivo. Mechanistically, the transformed cells expressed elevated level of CDK2 and CDK4. Unlike BRAFV600E inducing upregulation of p16ink4a in human melanocytes, BRAFV600E did not induce expression of p16ink4a in hNCSCs, and this was accompanied by increased expression of phosphorylated retinoblastoma (RB) protein. BRAFV600E increased expression of MITF and EST1/2 and resulted in melanocytic differentiation of hNCSCs. When MITF was overexpressed in hNCSCs harboring BRAFV600E, it dramatically increased their tumorigenic capacity and the tumor cells acquired the ability to invade surrounding tissues. In conclusion, skin hNCSCs bypass BRAFV600E-induced senescence through lack of p16ink4a induction and resulted in early transformation of hNCSCs. MITF potentiates the oncogenic effect of BRAFV600E in transformation. Our data indicate that adult stem cells respond to oncogenic stimulation distinctively different from differentiated cells and suggest that skin NCSCs are a potential target for oncogene-induced transformation.


An oncogenic MITF germline mutation impairs sumoylation and predisposes to melanoma and renal carcinoma

C. Bertolotto1, F. Lesueur2, S. Giuliano1, T. Strub3, M. Lathrop4, I. Davidson3, M. F. Avril5, F. Demenais6, R. Ballotti1, B. Bressac-de Paillerets2

1 INSERM, U895, équipe 1, Université of Nice Sophia-Antipolis, Nice, France; 2Service de Génétique, Institut de Cancérologie Gustave Roussy, INSERM, U946, Villejuif, France; 3Institut de Génétique et de Biologie Moléculaire et Cellulaire, CNRS, INSERM, Université de Strasbourg, Illkirch, France; 4Commissariat à l’Energie Atomique, Centre National de Génotypage, Evry, France; 5AP-HP, Hôpital Cochin-Tarnier, Service de Dermatologie et Faculté Paris Descartes, Paris, France; 6INSERM, U946, Genetic Variation and Human Diseases Unit, Paris, France

To date, no common environmental and/or phenotypic factor has been associated with melanoma and renal cell carcinoma (RCC). The known risk factors for melanoma include sun exposure, pigmentation, and nevus phenotypes; risk factors associated with RCC include smoking, obesity, and hypertension. A recent study of coexisting melanoma and (RCC) in the same patients supports a genetic predisposition underlying the association between these two cancers. The microphthalmia associated transcription factor (MITF) was proposed to act as a melanoma oncogene; it also stimulates the transcription of hypoxia inducible factor (HIF1A), whose pathway is targeted by kidney cancer susceptibility genes. We therefore hypothesized that the MITF might play a role in conferring a genetic predisposition to co-occurring melanoma and RCC. We identified a germline missense substitution (Mi-E318K) in MITF that occurred at a significantly higher frequency in genetically-enriched patients affected with melanoma, RCC or both cancers, when compared with controls. Overall, Mi-E318K carriers had a higher than fivefold increased risk of developing melanoma, RCC or both cancers (odds-ratio = 5.55, 95% confidence interval = 2.59–12.91). Codon 318 is located in a small-ubiquitin like modifier (SUMO) consensus site (ΨKXE) and Mi-E318K severely impaired sumoylation of MITF. Mi-E318K enhanced MITF protein binding to the HIF1A promoter and increased its transcriptional activity compared to wild type MITF. Further, we observed a global increase in Mi-E318K occupied loci. In an RCC cell line, gene expression profiling identified a Mi-E318K signature related to cell growth, proliferation, and inflammation. Finally, the mutant protein enhanced melanocytic and renal cell clonogenicity, migration, and invasion, consistent with a gain-of-function role in tumorigenesis. Hence, our data provide insights on the link between transcription, sumoylation and cancer.


Acquired resistance to BRAF inhibition can confer cross-resistance to combined BRAF/MEK inhibition

K. Gowrishankar, S. Snoyman G. Pupo, T. M. Becker, R. F. Kefford, H. Rizos

Westmead Institute for Cancer Research and Melanoma Institute of Australia, University of Sydney at Westmead Millennium Institute, Westmead Hospital, Westmead, Australia

Activating mutations in the serine/threonine kinase BRAF are found in approximately 50% of metastatic melanomas. Targeted therapy using BRAF inhibitors has shown tremendous success in clinical trials and is expected to play the major role in the treatment of the disease. However, acquisition of resistance is common and greatly diminishes durability of the clinical efficacy of these drugs.

Here we discuss the analysis of melanoma cell clones with acquired resistance to the GSK-BRAF inhibitor GSK2118436. We found that all drug resistant (DR) melanoma sub-clones displayed re-activation and often enhanced MAPK signalling, as measured by increased p-ERK in western blot assays, in the presence of GSK2118436. Intriguingly although these DR clones were derived from a single parent, reactivation of MAPK signalling was driven by several distinct mechanisms including the acquisition of activating N-RAS mutations (detected by Sequenom's Oncocarta screening for acquired mutations), increased accumulation of COT1 and the upregulation of tyrosine kinase receptor, PDGFRß, as detected by western blots. These drivers of resistance have been described for vemurafenib (PLX4032), and this is the first report demonstrating analogous mechanisms in acquired GSK-BRAF inhibitor resistance. Sensitivity of the DR clones to GSK-MEK inhibitor GSK1120212 was also analysed. The majority of these BRAF inhibitor resistant sub-clones showed weaker sensitivity to MEK inhibition and to the combination of MEK and BRAF inhibitors, compared to the parental clone.

We have shown that a single cell line can develop heterogeneous mechanisms of resistance to BRAF inhibition, and this suggests that melanoma tumours from a single patient may develop heterogeneous mechanisms of resistance many of which may confer resistance to multiple MAPK inhibitory therapies. Thus, sequential treatment with various MAPK inhibitors may not be as effective as combination therapies that target multiple nodes of oncogenic signaling pathways.


Lack of p16INK4a triggers DNA hyper-replication and initiates genetic instability in human melanocytes

C. Fung, G. Pupo, R. F. Kefford, H. Rizos

Westmead Institute for Cancer Research and Melanoma Institute of Australia, University of Sydney at Westmead Millennium Institute, Westmead Hospital, Westmead, Australia

Genetic alterations affecting the p16INK4atumour suppressor sequence are among the most frequent events in human cancer. Specific inactivation of p16INK4a occurs in the majority of human melanomas, and more than 60 distinct germline p16INK4amutations have been identified in over 190 melanoma-prone kindreds worldwide. The tumor suppressor functions of p16INK4a are well recognized and yet the impact of p16INK4a deficiency on cellular proliferation, lifespan and oncogene-induced senescence remain ambiguous. For instance, loss of p16INK4a function did not consistently enhance the proliferation of primary human cells, and p16INK4a-deficiency led to increased expression of p53 and its downstream transcriptional target p21Waf1 in melanocytes but not in fibroblasts

In this study, we applied a p16INK4a RNA interference strategy to examine the precise role of p16INK4a in regulating the proliferation and replicative capacity of primary human melanocytes. Our results confirm that p16INK4a deficiency enhanced the proliferation and extended the replicative lifespan of melanocytes. The hyper-replication of melanocytes was associated with the deregulation of replication initiation factors, the accumulation of single-strand DNA lesions and the activation of DNA damage and p53 checkpoint pathways. Strikingly, p53 activity was not sufficient to arrest p16INK4a-depleted melanocytes, and co-depletion of p53 did not alter the proliferation of these melanocytes. Moreover, benign nevus cells lacking p16INKa expression accumulated increased levels of foci containing the DNA damage marker, g-H2AX.

Thus, p16INK4a activity is critical in maintaining the replicative capacity and genome integrity of human melanocytes and we suggest that p16INK4a loss facilitates melanomagenesis by promoting the proliferation and enabling the selection of genetically unstable transformed melanocytes.


Developing chemotherapeutics which disable the actin cytoskeleton of melanoma cells

J. R. Stehn1, N. K. Haass2,3, M. Desouza1, V. B. Sequeira1, H. Treutlein4, J. Zeng4, P. R. B. B. Nascimento2, T. Butler5, A. McCluskey6, S. Palmer1, E. Hardeman1, D. Winlaw5, W. Weninger2,3, P. Gunning1

1 School of Medical Sciences, University of NSW, Sydney, NSW, Australia; 2Centenary Institute, Newtown, NSW, Australia; 3Discipline of Dermatology, University of Sydney, Camperdown, NSW, Australia; 4Qubist Molecular Design Pty Ltd and Monash Institute of Pharmaceutical Sciences, Monash University, Parkville, Vic., Australia; 5Kids Heart Research, The Children's Hospital at Westmead, Westmead, NSW, Australia; 6Chemistry, School of Environmental & Life Sciences, University of Newcastle, Callaghan, NSW, Australia

To treat aggressive cancers such as melanoma it is imperative that new chemotherapeutic targets and agents be developed. The actin cytoskeleton, due to its role in biological processes fundamental to cellular transformation, has always been a sought after chemotherapeutic target. However, despite many years of research, attempts to develop actin targeting drugs for chemotherapy remains unsuccessful due to their lack of specificity which ultimately causes unacceptable cardiac and respiratory toxicity. We have previously shown that tropomyosin, an integral component of the actin cytoskeleton, defines functionally distinct populations of actin filaments. We have identified a specific tropomyosin isoform common to all tumour cells tested to date which regulates cell proliferation and have designed a new class of compounds to target this filament population. The lead compound, TR100, was shown to be effective against a panel of melanoma cell lines with an average EC50 of 2–3 μM. When tested in 3D melanoma spheroid models, which more accurately mimic the tumour microenvironment, TR100 inhibited melanoma cell growth and motility. This effect translated to a reduction in tumour cell growth in vivo in a NOD/SCID melanoma xenograft mouse model. In vitro data using isolated rat cardiomyocytes demonstrated that TR100 had minimal impact on contractile function. In vivo data from the drug treated animals also showed no evidence of cardiac damage as measured by blood Troponin I levels and no changes in the intraventricular septum thickness of isolated hearts. These results demonstrate that it is possible to target distinct actin filament populations based on the tropomyosin composition. Next generation anti-tropomyosin compounds with improved efficacy and specificity have now been developed. Preliminary data demonstrate that these compounds exhibit increased selectivity for transformed cells and show efficacy against the SKMEL28 melanoma cell line. This class of compounds has enormous chemotherapeutic potential for the treatment of melanoma.


Specific binding of heparin-reactive peptides with melanoma in vitro and in vivo

J. S. Wall1,2, A. K. LeBlanc2, T. Richey1, A. Stuckey2, E. Martin1, S. Macy1, R. Donnell3, S. J. Kennel1,2

1 Department of Medicine, University of Tennessee Graduate School of Medicine, Knoxville, TN, USA; 2Department of Radiology, University of Tennessee Graduate School of Medicine, Knoxville, TN, USA; 3Department of Pathobiology, University of Tennessee College of Veterinary Medicine, Knoxville, TN, USA

The ability to specifically target tumors and their metastases is critically important for the accurate diagnosis, disease staging and treatment of patients with melanoma. The most commonly used technique for tumor imaging is positron emission tomography (PET) using 2-deoxy-2-(18F)fluoro-D-glucose (18FDG) which accumulates at sites of hypermetabolic activity, such as tumors. However, 18FDG also accumulates at sites of inflammation, within the myocardium, and brain. As a tumor-specific alternative certain peptides have been used successfully to target melanocytic neoplasms, and in contrast to 18FDG these reagents afford an opportunity for imaging and therapy.

Herein we describe our preliminary observation that a heparin-binding peptide, designated p5 and certain variants, bind melanoma tumors in tissue sections and in vivo, using the B16F10 murine model of lung colonization. We hypothesize that the peptides are targeting hypersulfated heparan sulfate proteoglycans on the surface of melanoma tumor cells. Importantly, this form of heparan sulfate is not observed within healthy murine organs.

Peptides were labeled with biotin and used for histochemical staining. These reagents were shown to bind preferentially to human and canine melanoma cells in formalin-fixed paraffin-embedded tissue sections. Furthermore, radioiodinated peptides were used to detect, by using small animal SPECT/CT imaging, pulmonary colonies of B16F10 cells in C57Bl/6 mice. Co-localization of 125I-labeled p5 peptides, but not control peptide with tumor lesions by confirmed using microautoradiography. Binding of the 125I-labeled p5 peptide was not uniform throughout the tumor mass; however both pulmonary and metastatic nodal tumor cells were targeted. In addition, a variant of peptide p5 was shown in similar experiments to associate specifically with tumor vasculature.

These studies demonstrate that heparin-binding peptides, such as p5, can bind to melanocytic neoplasms of human, canine, and murine origin. Although preliminary, these data support the further investigation of these and similar reagents for the radiodetection of these tumors.


The immune synapse between melanoma-specific CD8+ T cells and the tumor produces a signature of effector competence and turns the reacting lymphocytes into secondary antigen presenting cells

R. Uzana, G. Eisenberg, T. Peretz, A. Machlenkin, M. Lotem

Sharett Institute of Oncology, Hadassah Medical Organization, Jerusalem, Israel

Immune synapse is a contact-secretion process that occurs between antigen-specific lymphocytes and antigen presenting cells. In melanoma, membrane acquisition by melanoma-cognate CD8+ T cells, ‘trogocytosis’, occurs directly from tumor cells.

We set to correlate the intensity of trogocytosis with the functional capacity of CD8+ cytotoxic T cells and elucidate the effect of membrane tagging on T cells.

MART-1:26-35 CD8+ T cell clones which differed in their trogocytosis capacity were generated from melanoma patients. Functional evaluation of the clones showed that the percentage of trogocytosis-capable T cells closely paralleled each clone's IFN-γ and TNF-α production, lysosome degranulation and lysis of melanoma cells. The highly cytotoxic clone displayed the highest TCR peptide binding affinity, but when peptide-affinity was bypassed, the clones still reflected different intensity of activation. In the search for molecules associated with the functional difference of T cells with the same antigen specificity, we identified NTB-A, a homotypic receptor of the SLAM family, as such a determinant.

In a second set of experiments the MART-1:26-35 CD8+ T cell clones were stained with antibodies used to identify melanoma. We found that following incubation with the tumor, T cells were clearly stained by melanoma, arkers. When CD8-cytotoxic T cells were isolated immediately following incubation and then re-incubated with T cells of the same specificity or with T cells against a different melanoma antigen (gp100:154–162), they induced substantial secretion of IFN-γ and TNF-α by the newer subsets.

Thus, following encounter with melanoma cells, CD8+, anti-melanoma T cells can be identified using basic staining against melanoma antigens. By carrying melanoma antigens on their surface, CD8+ T cells act as secondary antigen presenting cells, and induce cytokine secretions in other lymphocytes who did not encounter a tumor. This is a newly identified mechanism which enhance T cell function by inter- and intra-clonal interactions.

Abstracts of the 5th International Meeting of Interdisciplinary Melanoma/Skin Cancer Centres November 11–12, 2011 Proffered Abstracts Selected for Poster Presentation

  1. Top of page
  2. 2011 International Melanoma Congress Advancement Through Collaboration joins the following four annual meetings in 2011:
  3. Congress Agenda
  4. Congress Proceedings Abstracts of the 8th International Congress of the Society for Melanoma Research November 9–11, 2011 Invited Faculty Abstracts
  5. Abstracts of the 5th Meeting of Interdisciplinary Melanoma/Skin Cancer Centres November 11–12, 2011 Invited Faculty Abstracts
  6. Abstracts of the 4th Melanoma Pathology Symposium of the International Melanoma Pathology Working Group November 13, 2011 Invited Faculty Abstracts
  7. Abstracts of the 8th International Congress of The Society for Melanoma Research November 9–11, 2011 Proffered Abstracts Selected for Oral Presentation
  8. Abstracts of the 8th International Congress of The Society for Melanoma Research November 9–11, 2011 Proffered Abstracts Selected for Poster Presentation
  9. Abstracts of the 5th International Meeting of Interdisciplinary Melanoma/Skin Cancer Centres November 11–12, 2011 Proffered Abstracts Selected for Poster Presentation


Melanoma recognition by young doctors in the UK

A. Carter

Foundation 2 Doctor, A&E Department, William Harvey Hospital, Ashford, Kent, UK

Despite having a greater incidence of skin melanomas in Australia compared to the United Kingdom, the percentage of those who will go on to have a skin melanoma cited as their main cause of death is proportionally far greater in the UK. Junior Doctors in their first three years of work in the UK were surveyed as to their knowledge of malignant melanomas. It was found that most junior doctors felt their level of knowledge was low and did not feel confident at both recognising and further investigating suspicious lesions. Given that early diagnosis and treatment is necessary in decreasing mortality rates, it is essential that UK medical schools and postgraduate education programmes reconsider the importance and reassess the amount of teaching on melanomas to young doctors.


Cutaneous melanoma without a junctional component undergoing sentinel node biopsy: outcomes and comparison with primary cutaneous melanoma

C. Cook, J. Browning, B. Walls, J. Zager, V. Sondak, K. Egan, J. Messina

H. Lee Moffitt Cancer Center, Tampa, FL, USA

The appropriate workup and treatment of patients who present with cutaneous melanoma lacking a junctional component (‘dermal melanoma’), and no evidence of systemic disease, is debated. In these cases, the absence of melanoma in situ makes the distinction between traumatized/regressed primary melanoma and metastatic melanoma a challenge. When there is no clinical evidence of metastatic disease, patients with solitary dermal melanoma may be offered sentinel lymph node biopsy and wide local excision (SLNB/WLE). We conducted a retrospective chart review to evaluate clinicopathologic parameters, surgical outcomes, and overall survival in patients with and without a junctional/in situ melanoma component in the initial skin biopsy on which the diagnosis was established. Of 991 consecutive patients with clinically localized cutaneous melanoma undergoing SLNB / WLE at the Moffitt Cancer Center in Tampa, FL, a total of 37 (3.7%) were found to have a dermal melanoma. A total of 149 melanoma-related deaths were documented, 12 among patients with dermal melanoma (median follow up among surviving patients: 5.2 years). Kaplan–Meier survival rates were 80% and 92%, 3 years after SLN biopsy/ WLE, in patients with dermal and primary melanoma, respectively (log rank P value = 0.0004). Breslow thickness was significantly greater in dermal than primary melanomas (median: 4.4 versus 1.7 mm, P < 0.001) whereas groups were similar in patient age and gender. The prevalence of SLN metastasis was greater in primary (19.7%) than dermal (10.8%) melanoma, though the difference was nonsignificant (P = 0.18). In multivariate regression, patients with dermal melanoma had a significant 2.2-fold excess mortality from melanoma (hazard ratio: 2.2; 95% CI: 1.2–4.0; P = 0.009). These findings document reduced survival despite lower SLN+ rates in patients with dermal melanoma. While SLNB is still informative in this setting, findings suggest that a proportion of these patients represent stage IV melanoma.


Quality assurance of data collection of neurocognitive function testing in a pilot whole brain radiotherapy clinical trial for metastatic melanoma

M. Gonzalez1, M. Cronin1, A. Hong1, G. Fogarty1, A. Grierson1, H. Dhillon2, J. F. Thompson1,3, J. Vardy4

1 Melanoma Institute Australia, University of Sydney; 2Centre for Medical Psychology & Evidence-based Decision-Making, Central Clinical School, University of Sydney, Psycho-Oncology Cooperative Research Group, School of Psychology, University of Sydney; 3Melanoma Institute Australia, Discipline of Surgery, Sydney Medical School, University of Sydney; 4Sydney Medical School, University of Sydney, Sydney Cancer Centre, Sydney

Melanoma Institute Australia (MIA) is the leading recruiter to the Melanoma Whole Brain Radiotherapy (WBRT) clinical trial, having contributed 20/55 patients currently randomised. A major study secondary endpoint is neurocognitive function (NCF) measured by 6 standardised, validated assessments. Of the 55 patients randomised, 40 will participate in the NCF component of the trial. The aim of this project is to assess NCF data collected for quality and completeness which is essential for arriving at a valid study endpoint. The protocol stipulates that NCF testing continue in all subjects until death, distant intracranial failure or withdrawal of consent. However, NCF testing in this population is difficult as patients have stage IV melanoma and have had significant intracranial treatment. To enhance completion rates, MIA clinical research coordinators attained competence in the administration of the NCF test by undergoing a training session with a trial monitor and viewing of a purpose-made DVD.

The clinical research coordinator administers the NCF test at baseline and every 2 months thereafter. The protocol mandated window for assessments is ± 2 weeks for each test. Data were collected about completion rates, the timing of assessments and the completeness of test batteries.

NCF tests have been completed in 73/79 (92%) expected instances. The completion rate within protocol mandated assessment windows is 65/79 (82%). Of the tests submitted, 99% were complete. Those variables detracting from completion rates included: poor performance status, disease progression, and geographic distance between patient and site.

Our study shows assessing NCF in a RCT is feasible and compares favourably with other trials. However, while total compliance is satisfactory and test completeness is on target, completion rates of NCF testing within protocol timelines at MIA could improve. We have learnt that patients who present with variables that detract from completion rates should be better supported and selected.


Oral nicotinamide for skin cancer prevention

D. L. Damian1, D. Surjana1, A. J. Martin2, G. M. Halliday1

1 Dermatology, Sydney Cancer Centre, Royal Prince Alfred Hospital, University of Sydney, Sydney, NSW, Australia; 2NHMRC Clinical Trials Centre, University of Sydney, Sydney, NSW, Australia

Nicotinamide (vitamin B3) prevents photocarcinogenesis in mice and is highly protective against both UVA and UVB-induced immunosuppression in humans. Nicotinamide is a precursor of NAD, which is centrally involved in cellular metabolism. Repair of photodamaged DNA in the skin requires high levels of cellular energy, which is depleted by UV exposure. Nicotinamide prevents this UV-induced cellular energy loss and glycolytic blockade, upregulates enzymes involved in energy production and enhances repair of both direct and oxidative DNA damage in human keratinocytes and ex vivo skin. We assessed the effectiveness of oral nicotinamide, 500 mg twice daily, in reducing premalignant actinic keratoses (AKs). The 35 heavily sun damaged participants in this randomised, double-blinded placebo controlled trial were encouraged to continue their usual sun protection measures. AKs were counted at baseline, 2 and 4 months. By 4 months, AK counts were 35% lower with nicotinamide compared to placebo (P < 0.0001). In a second trial using nicotinamide 500 mg once daily (n = 41), AK counts were 29% lower with nicotinamide compared to placebo (P = 0.005). We also counted new, histologically confirmed skin cancers during these 4 month trials. The placebo groups (n = 37) developed a total of 20 new nonmelanoma skin cancers, compared to only 4 new skin cancers in the 37 participants taking nicotinamide (relative rate adjusted for number of previous cancers = 0.24, 95% CI 0.08–0.71, P = 0.01). Nicotinamide is non-toxic, inexpensive and widely available, and holds great promise for skin cancer chemoprevention.


Do vitamin A serum levels moderate the protective effect of vitamin D on melanoma relapse?

S. Field, F. Elliott, J. H. Barrett, D. T. Bishop, J. A. Newton-Bishop

Section of Epidemiology and Biostatistics, Leeds Institute of Molecular Medicine, University of Leeds, Leeds, UK

Low serum levels of vitamin D have been reported to be associated with thicker tumors and poorer outcome for melanoma patients. The transcriptional effects of vitamin D are mediated by binding to the vitamin D receptor, which forms a heterodimer with RXRα to which metabolites of vitamin A bind.

Sun avoidance after melanoma diagnosis is likely to reduce vitamin D levels further so that supplementation is increasingly common. Many sources of vitamin D however also contain vitamin A, which could theoretically reduce the protective effects of supplementary vitamin D. This study was therefore designed to study the effects of vitamin A on tumor thickness and outcome.

Vitamin A and D (25-hydroxyvitamin D3)levels were available on 795 patients recruited to the Leeds Melanoma Cohort. Median vitamin A was 1.84 μM (range 0.47–4.06).

Vitamin A levels were not correlated with Breslow thickness (rho = −0.0026, P = 0.94). High serum vitamin A levels (≥2.2 μM, 75th percentile) conferred a non-significant increased risk of relapse compared to <2.2 μM (unadjusted hazard ratio (HR) 1.20, 95%CI 0.88–1.63, P = 0.25). Increasing vitamin D levels (per 20 nM) were significantly protective for relapse in univariate (HR 0.85, 95%CI 0.74–0.97, P = 0.02) and multivariate models including vitamin A (HR for vitamin D 0.86, 95%CI 0.75–0.99, P = 0.03, HR for vitamin A 1.25 (95%CI 0.91–1.73, P = 0.17), (adjusted for age, sex, BMI, tumor thickness, site).

Stratifying by vitamin A level and adjusting for age, sex, Breslow and BMI, a protective effect of increasing vitamin D levels was seen in both groups with very similar HRs (vitamin A ≥ 2.2: HR = 0.88, 95%CI 0.68–1.15, P = 0.36, vitamin A < 2.2: HR = 0.85, 95%CI 0.72–1.00, P = 0.05).

In summary, we saw no significant evidence that vitamin A levels moderated the protective effect of vitamin D on relapse or overall survival, although we cannot exclude a small effect or an effect of particularly high levels of vitamin A.


Osteopontin (OPN) as a potential prognostic biomarker in cutaneous malignant melanoma

A. Filia1, T. Wind2, F. Elliot1, D. T. Bishop1, R. Banks2, J. A. Newton-Bishop1

1 Section of Epidemiology & Biostatistics, Leeds Institute of Molecular Medicine, Leeds, UK; 2Section of Oncology & Clinical Research, Leeds Institute of Molecular Medicine, Leeds, UK

Although serum LDH level has prognostic information in stage IV melanoma (and is therefore included in the AJCC staging system), there are no serological markers of clinical utility as prognostic biomarkers in earlier disease. Our group has recently reported that increased gene expression of osteopontin (OPN) in primary tumours is of independent prognostic value for melanoma, adding to immunohistochemical evidence previously reported by Rangel et al 2008.

OPN can be readily detected in blood and increased levels have been reported in the blood of patients with a number of different cancers. The aim of this study was therefore to investigate whether OPN levels in the blood act as a prognostic biomarker in cutaneous malignant melanoma. OPN levels were measured in 120 stored plasma samples from the Leeds Cohort study using an ELISA assay. Plasma was taken in all cases soon after diagnosis (median time = 7.4 months) but samples were randomly selected from early and late relapsers in this pilot.

Increased levels of OPN were associated with thicker primaries (as indicated by greater Breslow thickness) (P = 0.056). Higher plasma levels of OPN were not predictive of positive sentinel nodes (P = 0.95), although previous studies of OPN expression in the associated primary tumours showed that higher expression was predictive of nodal metastases. Higher OPN levels were strongly associated with overall survival (HR = 8.1 for highest versus lowest quartile, P = 0.002) when data were adjusted for age, sex, BMI, Breslow thickness, tumour site and vitamin D level but only a modest association was observed between OPN levels and relapse free survival (HR = 2.4 for highest versus lowest quartile, P = 0.05) in the adjusted data.


Primary cutaneous melanoma arising in a long-standing irradiated keloid

L. Fish1, K. Gray1, J. Bell1, L. Duncan2, D. Shupp3, J. Lewis1

1 University of Tennessee Graduate School of Medicine, Department of Surgery; 2University of Tennessee Graduate School of Medicine, Department of Pathology; 3Penn State Hershey Dermatology

Ionizing radiation has been used to treat a variety of clinical conditions, including tinea capitis and treatment of prominent keloids. The use of ionizing radiation in the treatment of these benign conditions is controversial and is associated with an increased risk of developing cutaneous malignancies. Most of these are basal cell carcinomas and very few are melanoma. Keloids, however, are rarely associated with malignancy.

We report the case of a 57 yr-old Caucasian male with an extensive lifetime history of sun exposure and long-term smoking, who presented for evaluation of a recurrent cutaneous chest wall keloid. The keloid initially developed secondary to a varicella zoster infection, for which he received treatment with external beam irradiation of the chest wall at 9 yr-old. The keloid initially resolved, however, 18 months prior to re-evaluation his keloid began to grow and despite corticosteroid injections, would not regress. Punch biopsy revealed primary desmoplastic melanoma amidst the keloid scar.

Clinically, it is important to evaluate suspicious lesions and changing keloids that fail to respond to conventional treatments, particularly in those with a history of prior irradiation. A tissue diagnosis is necessary to guide appropriate treatment of the patient's condition. We believe this to be the first case of melanoma arising in an irradiated keloid.


Meta-analysis on artificial light and skin cancer: update

S. Gandini1, M. Boniol2, P. Autier2, P. ter Boyle2

1 European Institute of Oncology, Milan, Italy; 2International Prevention Research Institute, Lyon, France

Sunlight is recognized as the main environmental cause of skin cancer, and in 2009 the WHO classified the full spectrum of ultraviolet radiation (UV) as a human carcinogen. Sunbed use represents an increasingly frequent source of artificial UV exposure in light-skinned populations.

In a previous meta-analysis on artificial light and skin cancer we showed that the risk of cutaneous melanoma was increased by 75% for people who have used tanning beds before 35 years of age compared to those who have never used them. The aim of this work is to update the available evidence.

A comprehensive bibliographic search was conducted to identify relevant studies on sunbed use. Estimates were extracted from 24 studies published before March 2011, for a total of 10 003 cases. Random effects models were used to obtain summary risk estimates (SRR) for ever and high versus lowest exposure; dose-response model were used to evaluate the risk for increase number of sunbed session.

Ever-use of sunbeds was positively associated with melanoma: SRR = 1.36; 95% CI, 1.16–1.61; I2 = 62. No indication of publication bias was found.

In order to decrease the influence of possible biases, we also calculated summary estimates including only cohorts and population-based case–control studies. The summary relative risk was very similar (SRR = 1.35; CI, 1.12–1.63).

We evaluated also the frequency of use comparing the highest exposure versus no exposure, a significant 50% increase risk was found (SRR = 1.51; CI, 1.19–1.92; I2 = 40). Considering dose-response calculations of lifetime sessions we found an increase risk of 4% every ten sessions (SRR = 1.004; CI, 0.999–1.009).

First exposure to sunbeds before 35 years of age doubled the risk of melanoma, with no indication of heterogeneity (SRR = 2.05; 95% CI, 1.47–2.85; I2 = 0; based on 11 informative studies).

Sunbeds exposure is confirmed to significantly increase the risk of melanoma.


A randomized multicentre trial comparing 2 versus 4-cm surgical excision margins for thick (>2 mm) primary cutaneous melanoma

P. Gillgren1, K. T. Drzewiecki2, M. Niin3, H. PGullestad4, H. Hellborg5, E. Mnsson-Brahme6, C. Ingvar7, U. Ringborg8

1 Department of Clinical Science and Education and Department of Oncology and Pathology Karolinska Institutet, Department of Surgery, Stockholm Söder Hospital, Stockholm, Sweden; 2Department of Plastic Surgery, Breast Surgery & Burns, University Hospital Rigshospitalet, Copenhagen, Denmark; 3Department of Oncology and General Surgery, North Estonian Regional Hospital, Tallin, Estonia; 4Department of Surgery, Oslo University Hospital Radiumhospital, Oslo, Norway; 5Oncological Centre, Karolinska University Hospital, Stockholm, Sweden; 6Department of Oncology and Pathology Karolinska Institutet, Karolinska University Hospital, Stockholm, Sweden; 7Department of Surgery, Lund University Hospital, Lund, Sweden; 8Department of Oncology and Pathology Karolinska Institutet, Karolinska University Hospital, Stockholm, Sweden

There are still doubts and information is lacking regarding optimal surgical resection margins for patients with localized cutaneous melanoma (CM) more than 2 mm thick.

In a randomised controlled parallel trial (RCT), with nine participating European centres, patients with CM were allocated to either a 2-cm or a 4-cm resection margin. Patients with a CM >2 mm thick, clinical stage IIA-C, were randomised in a 1:1 allocation ratio to one of the two groups on a parallel basis and stratified by geographic region. Random allocation sequence (by sealed envelope or by computer generated lists) and permuted blocks procedures were used. It was hypothesized that there would be no differences between the treatment arms in overall survival (OS), being the primary endpoint, or recurrence free survival (RFS), being the secondary endpoint. The median follow-up time was 6.7 years. In a separate long term subgroup analysis on OS in the Swedish cohort the median follow-up time was 11.8 years.

The trial enrolled 936 patients (465 were treated with a 2-cm resection margin, 471 patients with a 4-cm resection margin). There were 644 patients in the Swedish, 180 in the Danish, 80 in the Estonian and 32 in the Norwegian part of the study. There were no harm-related issues. Patients were recruited from January 22 1992 to May 19 2004. There was no statistically significant difference in OS between the two treatment groups, the survival-rate at 5 years from randomisation was 65% in both groups (Kaplan–Meier estimate, Log Rank P = 0.69). There was no difference when comparing RFS, 55% of the patients were free of recurrence in both groups at 5 years (P = 0.82). There were 134 deaths due to CM in the group treated with a 2-cm resection margins as compared to 138 in the group treated with a 4-cm margins (hazard ratio, 0·99; 95% CI, 0·78–1.26; P = 0.95). There were 194 recurrences (as first event) in the group treated with a 2-cm resection margin, as compared to 200 in the group treated with a 4-cm resection margin (hazard ratio, 0·98; 95% CI, 0·80–1·19; P = 0·80).


Mapping of difficult lentigo maligna with in vivo confocal microscopy: outcomes

P. Guitera1,2, F. Moloney2, R. Scolyer1,3, M. Quinn1, J. Stretch1, A. Hong1, G. Fogarty1, S. W. Menzies2

1 Melanoma Institute Australia, and The University of Sydney, Sydney, NSW, Australia; 2Department of Dermatology and Sydney Melanoma Diagnostic Centre, Royal Prince Alfred Hospital; 3Tissue Pathology and Diagnostic Oncology, Royal Prince Alfred Hospital, Discipline of Pathology, Sydney Medical School, The University of Sydney

Recently a method to distinguish lentigo maligna (LM) from benign macules of the face with in vivo reflectance confocal microscopy (RCM) has been reported (1).

We sought to determine whether RCM mapping of difficult LM has changed management and outcomes.

Thirty seven patients (26 females, 11 males), aged from 47 to 88 years old (average 71) have been imaged with RCM at the Sydney Melanoma Diagnostic Centre and Melanoma Institute Australia to map margins. Thirty five had been diagnosed with LM and 2 with LMM. Fifteen had a recurrent episode, nine multiple recurrences. Nine were amelanotic (comprising eight non visible and one pink colored), nine were light colored and 18 were partially pigmented.

Of 29 patients with visible lesions, 17 (59%) had subclinical margins superior to the 5 mm recommended in the Australasian guide lines. RCM mapping has changed the management in 73% of patients. Eleven had changes in the surgical management; sixteen were offered a medical treatment (radiotherapy or imiquimod) because surgery was too difficult. Examples will be detailed and the management of difficult LM will be discussed.


Melanoma and non-melanoma skin cancer diagnosis using in vivo confocal microscopy: analysis of 710 consecutive equivocal cases using a two step method

P. Guitera1, S. W. Menzies1, C. Longo2, A. M. Cesinaro3, R. A. Scolyer4, G. Pellacani2,5

1 Sydney Melanoma Diagnostic Centre, Royal Prince Alfred Hospital, Discipline of Dermatology, The University of Sydney and Melanoma Institute Australia, NSW, Australia; 2Department of Dermatology, Arcispedale S Maria Nuova, Reggio Emilia, Italy; 3Department of Pathology, University of Modena and Reggio Emilia, Italy; 4Tissue Pathology and Diagnostic Oncology, Royal Prince Alfred Hospital, Discipline of Pathology, Sydney Medical School, The University of Sydney, and Melanoma Institute Australia, Sydney, NSW, Australia; 5Department of Dermatology, University of Modena and Reggio Emilia, Italy

In vivo reflectance confocal microscopy (RCM) allows the visualisation of the upper layers of the skin at cellular resolution. Our aim was to define on lesions excised in two specialized skin cancer clinics a model to accurately detect basal cell carcinoma (BCC) and melanoma (MEL).

The study comprised 710 lesions (216 MEL; 266 nevi; 119 BCC; 67 benign macules of the face (differential diagnosis of lentigo maligna); 33 actinic keratoses, Bowen's disease and squamous cell carcinomas and nine dermatofibromas). The frequency of 47 features was recorded for each lesion. The accuracy of the BCC model containing eight independently significant features defined on multivariate analysis of the training set (50%) and tested on the rest of series (50%) was 100% sensitivity (52/52 BCC); 88.5% specificity; AUC = 0.998 (CI95%: 0.979–0.997). The eight features were: Polarized elongated structures in the superficial layer, telangiectasia and convoluted vessels, basaloid nodules, epidermal shadow or ‘clefting’ (due to hyporeflective stroma). There were three negative features: papillae non-visible; disarrangement of the epidermal layer and cerebriform nests were more specific of MEL. Multivariate discriminant analysis based on the training set, excluding BCC, identified seven independently significant features for the diagnosis of MEL. The accuracy of the MEL model on the test set was 87.6% sensitivity (92/105 MEL); 70.8% specificity (226/319 others without BCC), AUC = 0.854 (CI95%: 0.810–0.899)

In conclusion, BCC diagnosis was accurate with RCM (similar to pathology assessment) but MEL diagnosis was difficult on these highly selected cases chosen by dermoscopy experts. The four invasive MEL that were misdiagnosed by RCM have been reviewed and were nevoid. Spitz nevi and sun-damaged macules of the face demonstrated the difficulties that pathologists and RCM experts share.


Should a positive deep melanoma biopsy margin be an indication for sentinel lymph node biopsy?

T. J. Hieken1,2,3, V. Krishnamurthy1,2

1 NorthShore University HealthSystem Skokie Hospital, Skokie, IL, USA; 2Rush University Medical Center, Chicago, IL, USA; 3Rush Medical College, Chicago, IL, USA

Current guidelines state that a positive deep biopsy margin should be considered as an indication for sentinel lymph node biopsy (SLNB) for patients with thin cutaneous melanoma. However intuitively logical this recommendation appears, there is little objective supporting data. Thus, we undertook this study to evaluate how often melanoma patients with positive deep biopsy margins developed an indication for SLNB after wide local excision (WLE), final histopathology and tumor staging. We evaluated 458 of 503 consecutive melanoma cases without clinical evidence of metastatic disease and with sufficient data for analysis. Statistical analysis was done with SAS software. Patients were 50% male, 97% Caucasian, mean age 65 ± 0.7 years. Anatomic site was trunk (36%), lower extremity (29%), upper extremity (19%), head/neck (16%). Histology included superficial spreading (58%), lentigo maligna (26%), nodular (11%), acral lentiginous (2%). Final T-stages were Tis (28%), T1 (42%), T2 (14%), T3 (9%), T4 (7%). Mean tumor thickness was 1.6 ± 0.1 mm. Biopsy type was excisional (48%), shave (30%), punch (18%), incisional (4%). The likelihood of a secondary recommendation for SLNB was nil for patients with negative biopsy margins, 2% for patients with an isolated deep or lateral margin+, 6% including all patients with a deep or unspecified margin+ and 13% for those patients with both deep and lateral margins+, P < 0.001. Overall, 126/458 (28%) underwent SLNB with 21 LN+ (17%). Of 336 patients with T1/Tis melanoma by biopsy, 37 underwent SLNB, including 7/135 with negative, 4/35 with +unspecified, 2/27 with +deep only, 10/96 with +lateral only, and 14/43 with +deep and lateral biopsy margins. No patient with Tis/T1 melanoma on biopsy who actually underwent SLNB for +biopsy margins had a +SLN. Considering both potential morbidity and cost, our data support revision of current treatment guidelines, excepting the subset of patients with both deep and lateral positive biopsy margins.


Prognostic significance of BRAF mutations in patients with stages IIIB/C cutaneous melanoma

R. W. Joseph1, J. Jakob2, P. Hwu2, M. A. Davies2, J. Gershenwald2

1 Mayo Clinic, Jacksonville, FL, USA; 2The University of Texas at MD Anderson Cancer Center (MDACC), Houston, TX, USA

Recent studies indicate that melanoma patients with a BRAF mutation have shorter overall survival (OS) from the diagnosis of stage IV disease. Since therapies targeting the MAPK pathway are likely to be explored in the adjuvant setting, it is important to determine if BRAF mutations correlate with distant relapse-free (DRFS) or OS in patients with stage III melanoma. Under an IRB approved protocol, we determined the BRAF mutation status of 121 stage IIIB/C non-acral cutaneous melanomas among patients who underwent lymphadenectomy at MDACC from 11/1988- 10/2009. DRFS and OS were calculated from the time of stage IIIB/C diagnosis. Stage IIIB patients had improved DRFS (1.42 versus 1.02 years, P = 0.04) and OS (5.19 versus 2.17 years, P = 0.01) versus stage IIIC in this cohort (Table). BRAF mutation rates did not differ between stage IIIB (42%) and stage IIIC patients (43%). BRAF mutation status did not correlate with RFS or OS in either stage. Overall, these data provide important information for the design and analysis of clinical trials involving patients with BRAF mutations.

Table Table. 
 Number (%)DRFS (years)P valueOS (years)P value
  1. aIIIB versus IIIC (log-rank test)

  2. bBRAF mut versus BRAF wt (log-rank test).

IIIB31 (26)1.420.63b5.190.35b
BRAF mut13 (42)1.356.50
BRAF wt18 (58)1.794.90
IIIC90 (74)1.020.14b2.170.80b
BRAF mut39 (43)0.782.17
BRAF wt51 (57)1.261.95


An automated multi-imaging registration method for the detection and quantification of morphological changes across pigmented skin lesions

S. Kacenjar1, M. Zook2, A. Ostrow1, E. Ng1

1 Lockheed Martin Corporation, Moorestown, NJ, USA; 2Fox Chase Cancer Center, Rockledge, PA, USA

This paper describes an image processing system that automatically aligns and detects changes in the shape and boundary characteristics of skin lesions on a patient's skin. It consists of three functional elements: (i) coarse alignment of time-sequenced imagery (TSI), (ii) refined alignment of local skin topographies, and (iii) assessment of local lesion pigmentation changes. During the coarse alignment process, a 2D correlation process is used that performs in a rough alignment of the patient's skin topography. Since the skin is a deformable membrane, this 3D alignment process only approximates and still must be compensated for non-linear local stretching of the skin. To achieve this compensation, a new algorithm leveraging our Department of Defense (DoD) expertise, is used to refine the local morphological mapping. Here the optimization process is driven using the minimization of entropy between the multiple TSIs. Once the camera models are corrected for local skin deformations, the images are digitally compared for spatial changes across the aligned images. Limits on the detectability of lesion changes are established by the fidelity to which the algorithm corrects for local skin deformation and background alterations, which was developed by leveraging our DoD Discrimination expertise. Ultimately these factors will provide limits of detectability of malignant growth.

Key to this work is the refinement of an algorithm to perform the fine adjustment in local skin topography. These topographies are modeled as local translational, magnification and rotational changes as manifested within the local camera model. Testing the robustness of this alignment algorithm is achieved using both simulated data as well as pilot data of large-scale imagery of patients’ backs. Results show that the capture-limits between TSI can perform well with substantial translational offsets (±20 pixels) and local rotational mismatches between images (±10 degrees), and not limited to single mole inspections.


Cost-effectiveness of a fluorescence in situ hybridization (FISH) assay for the diagnosis of melanoma in the United States

A. R. Kansal1, A. J. Shaul1, S. Stern1, K. Busam2, C. A. Doucet3, D. B. Chalfin3

1 United BioSource Corporation, Modeling and Simulation, Bethesda, MD, USA; 2Memorial Sloan-Kettering/Cornell Dermatopathology Program at the Memorial Sloan-Kettering Cancer Center in New York, New York, NY, USA; 3Abbott Molecular, Des Plaines, IL, USA

This study evaluated cost-effectiveness of a FISH assay in melanoma diagnosis from the perspective of US third-party payers.

A decision-analytic model simulated melanoma diagnosis using microscopic assessment alone versus addition of FISH (sensitivity = 92%; specificity = 94%). We simulated a clinical setting in which initial biopsy microscopic assessment (sensitivity = 73%; specificity = 78%) was followed by expert pathologist assessment (sensitivity = 89%; specificity = 79%) for inconclusive results. Diagnostic accuracy can vary widely by setting, so initial and expert assessment sensitivities and specificities ranging to 100% were evaluated. Scenarios 1 and 2 evaluated addition of FISH to initial or expert assessments, respectively, versus diagnosis without FISH. We assumed biopsy was curative for 68% of local melanomas. Patients could receive treatment and were followed throughout their lifetime in a Markov model. Clinical and cost parameters were taken from SEER, literature, and CMS's fee schedule. Outcomes included incremental cost per quality adjusted life-year (cost/QALY) gained and per misdiagnosis avoided.

8.5%, 3.4%, and 2.5% of melanomas required treatment after missed diagnosis for initial, expert and FISH assessments, respectively. In Scenarios 1 and 2, the cost/QALY was $15 243 and $47 674, respectively, when addition of FISH was compared to no FISH. Incremental cost per misdiagnosis avoided was $3348 and $4082, respectively. FISH remained below a $100 000/QALY threshold for a range of parameter values until non-FISH sensitivity and specificity were both ≥89%. Cost-effectiveness was also influenced by stage, with FISH most cost-effective for populations with local or regional melanoma.

In specific clinical settings, a FISH assay could be cost-effective for melanoma diagnosis.


Clinical significance of the destruction of proximal nail plate in subungual melanoma

S.-J. Lee1, Y. H. Choi1, C. H. Song1, B.-S. Kim2, M.-B. Kim2, W. J. Lee1, D. W. Kim1

1 Department of Dermatology, Kyungpook National University School of Medicine, Daegu, Korea; 2Department of Dermatology, College of Medicine, Pusan National University, Busan, Korea

Subungual melanoma is an uncommon subtype of melanoma, constituting only 1–3% in the Western reports, but, relatively more common in Asians. Though, little is known about clinical significance of some features like destruction of nail plate which is commonly observed.

Forty-nine patients with subungual melanoma who were diagnosed at two university hospitals were included in present study. We divided these patients into four groups according to Breslow thickness (1 mm threshold) and existence of proximal nail plate destruction (PNPD); we regarded a melanoma as a thick melanoma if there was an evidence of complete pathological regression. PNPD was defined as visible loss of full thickness of nail plate, occurred at proximal part of the plate touching eponychium. We evaluated the relationship between PNPD and thickness of subungual melanoma.

In this study, average Breslow thickness was 3.18 mm in PNPD group and 1.74 mm in non-PNPD group. In addition, 17 of 20 patients (85.0%) in PNDP group group revealed more than 1 mm thickness, instead, only 14 of 29 patients (48.3%) in non-PNPD did. PNPD was significantly frequent in thick melanoma group than thin melanoma group (P = 0.009).

Conclusively, there was a strong statistical correlation between PNPD and tumor thickness. So, sufficient consideration of sentinel lymph node biopsy and careful preparation of surgical specimen to reveal thicker area than biopsy should be undertaken in the patients with PNPD.


Clinical significance of the destruction of proximal nail plate in subungual melanoma

S.-J. Lee1, Y. H. Choi1, C. H. Song1, B.-S. Kim2, M.-B. Kim2, W. J. Lee1, D. W. Kim1

1 Department of Dermatology, Kyungpook National University School of Medicine, Daegu, Korea; 2Department of Dermatology, College of Medicine, Pusan National University, Busan, Korea

Subungual melanoma is an uncommon subtype of melanoma, constituting only 1–3% in the Western reports, but, relatively more common in Asians. Though, little is known about clinical significance of some features like destruction of nail plate which is commonly observed.

Forty-nine patients with subungual melanoma who were diagnosed at two university hospitals were included in present study. We divided these patients into four groups according to Breslow thickness (1 mm threshold) and existence of proximal nail plate destruction (PNPD); we regarded a melanoma as a thick melanoma if there was an evidence of complete pathological regression. PNPD was defined as visible loss of full thickness of nail plate, occurred at proximal part of the plate touching eponychium. We evaluated the relationship between PNPD and thickness of subungual melanoma.

In this study, average Breslow thickness was 3.18 mm in PNPD group and 1.74 mm in non-PNPD group. In addition, 17 of 20 patients (85.0%) in PNDP group group revealed more than 1 mm thickness, instead, only 14 of 29 patients (48.3%) in non-PNPD did. PNPD was significantly frequent in thick melanoma group than thin melanoma group (P = 0.009).

Conclusively, there was a strong statistical correlation between PNPD and tumor thickness. So, sufficient consideration of sentinel lymph node biopsy and careful preparation of surgical specimen to reveal thicker area than biopsy should be undertaken in the patients with PNPD.


Identifying individuals at high risk of melanoma: a simple tool

C. Fortes1, S. Mastroeni1, L. Bakos2, G. Antonelli3, L. Alessandroni1, M. A. Pilla3, M. Alotto1, A. Zappala4, T. Mannooranparampil1, R. Bonamigo5, P. Pasquini1, F. Melchi3

1 Clinical Epidemiology Unit, IDI-IRCCS, Rome, Italy; 2Department of Dermatology, Hospital de Clínicas de, Porto Alegre, Federal University of Rio Grande do Sul, Porto Alegre (UFRGS); 3VIII Dermatology Unit, IDI-IRCCS, Rome, Italy; 4Oncology Unit, IDI-IRCCS, Rome, Italy; 5Department of Dermatology, Fundação Faculdade Federal de Cie^ ncias, Me′ dicas, Porto Alegre, Brazil

Simple and reliable tools for identifying patients at high risk for melanoma with preventive measures have important public health implications. An individual risk score for cutaneous melanoma was constructed and externally validated. With the summary coefficients of the risk factors for cutaneous melanoma, derived from a meta-analysis, a melanoma risk score was tested in an Italian population and externally validated in a Brazilian population. Common nevi, skin and hair color, freckles, and sunburns in childhood were the variables included in the final predictive model. The discriminatory ability of the models was assessed by the receiver operating characteristic (ROC) curve. The performance of the model was also evaluated by conducting an external validation. The area under the curve (AUC) of the candidate model was 0.79 (95% confidence interval: 0.75–0.82). The same model, when applied in the Brazilian population, presented an AUC of 0.79 (95% confidence interval: 0.70–0.86). At the cut-off level of 3 and more, 89 and 80% of the melanoma cases were correctly classified as ‘at risk for melanoma’ in the Italian and in the Brazilian populations, respectively. The risk model is a simple tool that identifies patients for preventive measures and may be used with reasonable confidence in different populations. The risk model may help family doctors in referring patients to dermatological clinics and thus improve early diagnosis.


Challenges of early diagnosis of melanoma in Serbia

D. Nikolic1,2, A. Nikolic1,3, V. Stanimirovic4

1 Faculty of Medicine, University of Belgrade; 2UMC Bezanijska kosa, Belgrade; 3Cardiovascular Institute Dedinje, Belgrade; 4Medicines and Medical Devices Agency of Serbia

Melanoma is one of the most common cancers with the worst prognosis and growing incidence. The death rate has gone up more then 50% in 10 years. Early diagnosis of melanoma guarantees cure in 99% of patients. Digital epiluminescence microscopy (DELM) makes sub-surface structures of skin accessible for in vivo microscopic examination. Digital dermatoscopy, capable of photographing the pigmented lesion, process the images and produce an exact diagnosis.

During the period of 5 years, we examined 11.352 patients with the total number of evaluated images of 354.785 DELM images. Of the 1022 cutaneous pigmented lesions removed and histological examined, 232 (22.7%) were non-melanocytic lesions and 790 (77.3%) were melanocytic lesions. Sensitivity and specificity of DELM in the analysis of melanocytic lesions with a pigment network were both very high (99.8% and 98.7%, respectively). Forty-seven new case of thin CMM (Clark I and II, Breslow I) were identified from one hundred seventy six new cases of all types of CMM (26.7%). A remarkable proportion of melanoma in situ (21 out of 176; 11.9%) was diagnosed exclusively by DELM.

The incorporation of dermatoscopic techniques has greatly enhanced the diagnosis of this cutaneous tumour. The prognosis of melanoma can be improved if the tumor is recognized and treated in its early phase. Despite its high incidence, the rate of survival from a melanoma is high, thanks to the melanoma prevention campaigns in Serbia much more in the last 10 years which enable the early detection of the melanoma and, thus, guarantees the efficacy of surgical treatment of the tumour.


Double Whammy: Sarcoidosis and rapidly enlarging liver hemangiomas induced by interferon in malignant melanoma treatment

W. B. Ooi, S. Dahiya

Baystate Medical Center/Tufts University School of Medicine, MA, USA

A 54 year old gentleman otherwise healthy presented to an academic tertiary hospital with acute post-prandial crampy abdominal pain. Computed tomography (CT) scan with oral contrast of his abdomen revealed a swirled dilated appearance of his proximal small bowel consistent with volvulus. He underwent emergent exploratory laparatomy with resection of 5 cm tumor involving small bowel. The tumor stained positive for S-100 and Melan-A, confirming malignant melanoma. Exploratory laparatomy also revealed cystic appearing lesions in his liver. He underwent a follow-up CT scan with intravenous contrast of his abdomen 5 days later which revealed a 7 cm lesion and two smaller lesions of <2 cm in his liver with contrast enhancement suggestive of cavernous hemangiomas. No primary site of cutaneous melanoma was identified. Positron emission tomography (PET) did not reveal any areas of hypermetabolic activity. Given risk for disease recurrence, patient was commenced on high dose intravenous interferon alpha for a month and transitioned to subcutaneous interferon alpha injections three times per week for 5 months. At treatment end, follow up abdominal CT scan revealed multiple new hemangiomas and enlargement of previous lesions up to 9 cm and 4 cm each. CT scan of the chest revealed multiple new pulmonary nodules up to 1.2 cm in size as well as bilateral hilar and mediastinal adenopathy. Patient underwent transbronchial needle aspiration biopsy of the mediastinal lymph nodes which revealed non-necrotizing granulomatous inflammation which stained negative for S-100, Mart-1 and Ziehl-Neelsen, confirming pulmonary sarcoidosis. Three months after treatment completion, patient underwent further imaging which showed resolution of the pulmonary nodules and hilar and mediastinal adenopathy. Liver hemangiomas were unchanged in size. This case report serves to illustrate the potential adverse effects of interferon therapy in malignant melanoma. While interferon induced sarcoidosis is increasingly recognized, enlarging hemangiomas have only been reported in Hepatitis C treatment.


Whole brain radiotherapy following local treatment of 1–3 intracranial metastases of melanoma – a phase III trial (ANZMTG 01/07; TROG 08/05)

G. Fogarty1,2, E. Paton3, A. Hong2, B. Shivalingam2,4, K. D. Jacobsen5, B. Burmeister6, J. Thompson2

1 Mater Sydney Radiation Oncology, Sydney, NSW, Australia; 2Melanoma Institute Australia, Sydney, NSW, Australia; 3Australia and New Zealand Melanoma Trials Group, Sydney, NSW, Australia; 4Royal Prince Alfred Hospital, Sydney, NSW, Australia; 5Radium Hospital, Oslo, Norway; 6Princess Alexandra Hospital, Brisbane, Australia

Whole Brain Radiotherapy (WBRT) following local treatment of intracranial metastases of melanoma with neurosurgery and/or stereotactic irradiation is controversial. The WBRT in Melanoma Trial (WBRT Mel) is being conducted to investigate this scenario. Previous WBRT trials in other histologies have been plagued by problems with accrual, accentuated by a reluctance to randomise by both opponents and proponents of WBRT.

A phase III randomised controlled trial is open at 15 hospitals across Australia and overseas (Fogarty et al, 2011). The pilot phase finished on the 30 June2011 and the full study has commenced (involving 200 randomised patients). The primary endpoint is distant intracranial failure as assessed by regular MRI scanning. Secondary endpoints include the effect of WBRT on neurocognitive function and overall survival. Extensive quality assurance activities have been implemented, including an MRI audit process performed by a neuro-radiologist.

Fifty-eight patients have been accrued to date including 16 patients from international sites. The ANZMTG 01.07 WBRTMel trial has the highest recruitment rate of a WBRT trial to date. Preliminary demographic and disease specific data will be presented.

The ANZMTG 01.07 WBRTMel trial is accruing well and should answer the question of whether WBRT should be used in this scenario. Supporting international collaboration has been crucial to successfully meeting recruitment targets.


The epidemiology and clinical and demographic characteristics of advanced melanoma patients in the US: evidence from a retrospective medical chart review

F. J. Malinoski1, G. Matthias Prinz1, P. Wolthoff1, C. P. Macahilig1, A. S. Yang2, J. R. Penrod2

1 Medical Data Analytics, Parsipanny, NJ, USA; 2Bristol-Myers Squibb, Plainsboro, NJ, USA

Advanced melanoma (AM) includes both incident and recurrent cases of unresectable (UR) Stage III and metastatic Stage IV disease. In the US, the staging distribution and other clinical and demographic characteristics of new AM patients (pts) are not well documented.

A randomized, retrospective, observational medical chart review was conducted in two parts. A nationally representative random stratified quota sample of 113 oncologists was selected. In Part I, which focused on the epidemiology of AM, physicians selected randomly 1261 medical charts of pts newly diagnosed with AM between 2004 and 2008. In Part II, data from a stratified random sample of 752 medical charts of pts newly diagnosed with AM between 2004 and 2008 were collected, including a comprehensive set of pts’ clinical and demographic characteristics.

Part I data showed that new pts were more likely to be Stage IV (77.2%) than UR Stage III (22.8%). In the Part II sample, the mean age was 62.3 years (median 64), 57.4% were male, and 91.2% were Caucasian. At AM diagnosis, most pts were assessed by their physician to have a poor prognosis (poor/unfavorable, 89.4%; good/favorable, 10.6%). Primary tumors were located on the extremities (35.1%), trunk (34.1%), neck (10.2%) and head (9.2%). For Stage IV pts, reported metastatic sites (median = 2) included the lungs (62.0%), liver (41.3%), brain (28.1%), bone (27.1%), and distant lymph nodes (22.2%). The ECOG performance status at the start of first therapy was: 0–1, 67.7%; 2, 21.6%; 3–4, 10.7%. 47.9% were retired from the work force. 42.3% had private insurance as their primary medical coverage.

AM imposes a heavy burden on pts and society as the majority of pts are of working age (<65), are diagnosed with Stage IV disease, and have a poor prognosis.


Impact of metastatic site on costs and survival in advanced melanoma

C. M. Reyes1, S. D. Byfield2, S. Satram-Hoang3, R. Linke1, A. Teitelbaum2

1 Genentech, Inc., San Francisco, CA, USA; 2Innovus, Eden Prairie, MN, USA; 3Health Outcomes Consulting, Roseville, CA, USA

Background:  Metastatic melanoma (mM) has a poor prognosis, with poor long-term survival. Few studies report survival and healthcare costs by metastatic sites.

Methods:  A retrospective analysis utilized claims data from a national commercial health insurer from 1/2007–3/2010. Patients included had ≥2 melanoma claims (ICD-9-CM 172.xx) ≥30 days apart and ≥2 metastatic disease claims (ICD-9-CM 196.xx-198.xx) ≥30 days apart or ≥1 cancer-related treatment claim with a melanoma diagnosis. The diagnosis date of metastasis was the index date. Patients were continuously enrolled for 6 months prior to and ≥3 months after the index date. Those with other primary cancers were excluded. Cohorts were created based on metastatic sites: lymph node only (LN), 1–3 non-LN, or >3 non-LN. Mortality and all-cause per-patient per-month (PPPM) costs were estimated. Multivariate proportional hazards (PH) model examined non-LN mets survival.

Results:  There were 431 mM patients: 59 with claims for LN only, 219 with 1–3 non-LN and 153 with >3 non-LN, with a mean age of 52, 59 and 57 years, respectively (P < 0.01). Most were male (65%) with mean baseline Charlson comorbidity index of 3.52. While 182 patients (42%) died during the study period, mortality varied by cohort: 3% in LN only; 37% in 1–3 non-LN, and 64% in >3 non-LN (P < 0.001). Unadjusted mean PPPM healthcare costs were $6773, $10 999, and $15 762 for the LN only, 1–3 non-LN and >3 non-LN cohorts respectively (P < 0.001). In multivariate analysis, patients with brain mets (versus no brain mets), liver mets (versus no liver mets) or lung mets (versus no lung mets) had higher risk of death: brain (HR = 2.816; 95%CI 2.043–3.881), liver (HR = 1.875; 95%CI 1.384–2.541) and lung (HR = 1.397; 95%CI = 1.009–1.934).

Conclusions:  Melanoma patients with multiple metastastic sites had significantly higher costs and worse survival compared to those with LN only metastasis. Presence of brain metastasis contributes significantly to poorer survival.


Treatment patterns among patients with metastatic melanoma treated in the community oncology setting

M. S. Walker1, C. Reyes2, J. Kerr1, S. Satram-Hoang3, E. J. Stepanski1

1 ACORN Research, Memphis, TN, USA; 2Genentech, Inc. South San Francisco, CA, USA; 3Outcomes Research Consulting, Roseville, CA, USA

The 5 year survival rate for metastatic melanoma (MM) is <20%. We report treatment patterns and outcomes among patients treated in 1st and later lines in community oncology settings.

Retrospective review of medical records of 202 patients enrolled with MM diagnosed 1/1/2006 to 9/1/2010, and treated in community oncology practices. Patients were ≥ 18 years at diagnosis of MM, with no treatment of other cancers within 5 years. Treatment regimens, demographic, disease and clinical characteristics were collected. Kaplan–Meier analysis and Cox regression were used to compare progression free survival (PFS) and overall survival (OS) across regimen groups.

Among patients treated systemically, PFS in this study was 3.25 months in first line, 2.3 months in second line, and 1.84 months in third line. Median OS was 7.66 months. There was no evidence that PFS varied significantly across regimens, although impaired performance status and presence of brain metastasis were associated with shorter PFS in first line (HR = 2.93, P = .0009; HR = 1.95, P = .001, respectively). At each disease progression (from MM diagnosis and following), decreasing percentages were treated systemically. At MM diagnosis, 102/202 (50.5%) patients got systemic therapy. At first through third progressions: 76/165 (46.1%) patients, 47/110 (42.7%) patients, and 16/56 (28.6%) patients, respectively, received systemic therapy. Cox regression analysis of OS showed that shorter survival was associated with diagnosis at stage I – III (versus stage IV; HR = 1.79, P = .011), prior non-use of interferon (HR = 1.75, P = .015), presence of brain and liver metastasis (HR = 1.98, P = 0.0002; HR = 1.49, P = 0.03, respectively), and impaired performance status (HR = 2.23, P = .001).

Overall findings show that PFS in patients with MM is short, and declining over lines of treatment, with little evidence of differential efficacy of the available therapies during this period.


Remote cutaneous photography for public dermal screenings

V. Sams1, G. Samaras2, L. Fish1, M. Jones1, J. Bell3, K. Gray3, J. Lewis3

1 University of Tennessee Medical Center Knoxville, Department of Surgery; 2University of Tennessee Medical Center Knoxville, UT Cancer Institute; 3Division of Surgical Oncology, University of Tennessee Medical Center Knoxville

The incidence of cutaneous malignancy continues to increase worldwide. No randomized studies demonstrate an overall benefit from public cutaneous screenings, however, the benefit to targeted individuals may be substantial. We hypothesized a cutaneuous photographic screening program would direct individuals with photographically suspicious lesions toward intervention.

Participants who attended community screening venues between 09/2010 and 05/2011 were offered the opportunity to participate in photographic dermal screening. Targeted lesions were digitally photographed after obtaining informed consent and personal history, stored on a secure database, and graded by a physician. The grading system evaluated the nature of the lesion and the quality of the photograph. Lesions were given a skin-rads (SR) score of 0–3. SR0's are inevaluable, SR1's benign, SR2's suspicious, and SR3's were in need of biopsy. Participants with SR2/3 lesions were referred to dermatologists and contacted for follow-up results. Image quality was considered good (g), intermediate (i), or poor (p). Participants with poor quality photos were ecouraged to seek clinical evaluation.

Four hundred and thirty-one participants were screened, while 159 (36.9%) consented to participate in the study. Of the 297 lesions photographed, 165 (55.6%) were SR1, 88 (29.6%) SR2, 27 (9.1%) SR3, and 18 (5.6%) SR0. Of the photos, 191 (64.3%) were graded good, 77 (25.9%) intermediate, and 29 (9.8%) poor quality. Seventy-seven participants had SR2 or SR3 lesions of which we were able to contact 41 (53%). Eighteen (44%) sought medical attention and 11 (26.8%) underwent biopsy. Per participant report, biopsy results included four non-melanoma carcinoma, two melanoma, three benign, and four were unknown.

Public Photographic Dermal Screening (PPDS) has the potential to spare individuals unnecessary expense if lesions are benign and identify patients with suspicious lesions who need further evaluation. This process continues to be refined and studied at our institution.


Reflectance confocal microscopy aids to differentiate benign and malignant melanocytic lesions with similar dermoscopic features

M. I. Sanchez, A. L. Ross, S. Hu, H. Ravinobitz, M. Oliveiro, J. M. Grichnik

Department of Dermatology and Cutaneous Surgery, University of Miami Miller School of Medicine

Early melanoma detection is key as it allows for identification and removal of melanoma at highly curable stages. At great advance for clinical detection has been dermoscopy. This surface microscope reveals patterns of growth not visible to the naked eye and has been shown to increase the accuracy of melanoma diagnosis and reduce unnecessary biopsies. However, not all dermoscopic patterns are clear cut and additional technologies are necessary to facilitate an accurate diagnosis on these difficult cases. Reflectance confocal microscopy (RCM) is a noninvasive tool that allows for the in vivo evaluation of cellular structures of the skin from the stratum corneum on the surface through to the papillary dermis. RCM allows for the identification of retractile melanocytic cells with nuclear atypia, in abnormal locations (pagetoid cells), and with extensive sheets of dendrites, all features associated with melanoma. Herein, we present matched dermoscopic cases from our database that we were not able to discriminate by dermoscopic criteria alone. We show that although similar dermoscopic features are present, confocal microscopic features clearly delineated the benign and malignant lesions. Confocal microscopy is a valuable clinical tool and facilitates diagnosis in lesions otherwise difficult to differentiate through clinical and dermoscopic criteria.


Quality of care in a specialist melanoma centre: a patient survey conducted at Melanoma Institute Australia

R. P. M. Saw1–4, J. Winstanley1,2, B. Burgess1,2, P. Webster1,2, L. Visintin1,4,5, A. Hong1–4, M. J. Quinn1,3,4, K. F. Shannon1,3,4, A. J. Spillane1,2,3,4,6, J. R. Stretch1–4, J. F. Thompson1–4

1 Melanoma Institute Australia, North Sydney, Australia; 2The University of Sydney, Australia; 3Mater Hospital, North Sydney, Qld, Australia; 4Royal Prince Alfred Hospital, Camperdown, Vic., Australia; 5Westmead Hospital, Westmead, NSW, Australia; 6Royal North Shore Hospital, St Leonards, NSW, Australia

A patient's perception that they are receiving high quality care is an important component of negotiating their cancer journey to achieve the best compliance and psychosocial outcome. Determining patient perception of quality of care provided by a medical facility is one measure of quality assurance that can lead to improvements in clinical systems and better use of resources.

Clinicians at Melanoma Institute Australia (MIA) see approximately 5000 patients per year (both new and those in follow-up). Assessment of quality of care has not previously been documented.

The aim of this study was to document quality of care (from the patient's perspective) at MIA. Three validated questionnaires covering quality of care and factors influencing quality of care were given to patients to complete following their clinic visit. The questionnaires were collated with patients’ clinical data.

Surveys were returned by 299 patients in a 5-week period in September/October 2009. Eighty percent of these patients were being treated for melanoma – with 65% of these having stage I melanoma (others were seen for non melanoma skin cancer or breast disease). Fifty-seven patients were male with an average age of 59 years.

Eighty-four percent of patients reported their current physical health, and 91% reported their current psychological well-being as quite or very good. Patients were satisfied with the quality of care received at MIA, with 100% of patients mostly agreeing they had received the best possible medical care (as far as they could tell). Clinicians appeared to understand patients’ experiences, were respectful and cared for their patients (in 97–99% of cases). Seventeen percent of patients reported being partly or completely dissatisfied that their medical care was determined by staff procedures, rather than their own requests/needs,

In conclusion, MIA patients were mostly satisfied with the quality of care they received. Staff procedures will be reviewed to try to improve care.


Different dosing schedules positively influence the effect of the EGF-receptor inhibition in combination with chemotherapy on malignant melanoma

N. Schicher, S. Blunder, C. Briand, W. Jerney, V. Paulitschke, H. Pehamberger, C. Hoeller

Division of General Dermatology, Department of Dermatology, Medical University of Vienna, Vienna, Austria

The aim of this study was to identify if the combination of erlotinib with various cytotoxic drugs is superior to each single drug and if different dosing schedules would influence a possible synergistic effect. It is well established that erlotinib induces a G1/S phase cell cycle arrest with reduced biosynthetic activity in tumor cells. We speculated that treating melanoma cells with erlotinib concomitantly with cytotoxic drugs could result in a reduced cytotoxic effect and that this could be overcome by sequential dosing schedules.

To test our hypothesis we used two different melanoma models in 6–8 week old female SCID mice. The first in vivo model focused on tumor growth inhibition and thereafter the most effective combination was chosen for a spontaneously metastasizing melanoma model. Animals were treated with solvent control, erlotinib, chemotherapy, erlotinib continuously or discontinuously with a 5 days dose interruption prior to either dacarbazine, cisplatin or paclitaxel. Tumors and lungs were analyzed immunohistochemically for proliferation, apoptosis and melanoma metastasis with Ki-67, TUNEL and vimentin staining, respectively.

Each of the drugs tested showed activity in this model system. The addition of erlotinib to chemotherapy substantially added to the reduction of tumor volume. The immunohistochemical analysis substantiated these findings as assessed through Ki-67 and TUNEL staining. Most interestingly, the discontinuous erlotinib/cisplatin treatment showed a significantly higher reduction in tumor volume compared all other treatment arms. This combination proved to be superior in reducing melanoma metastasis too. The combination with paclitaxel tended to be more efficient in the discontinuous treatment group while in combination with DTIC different dosing schedules had no effect.

Our data indicate that different dosing schedules of a TKI in combination with chemotherapy result in diverse treatment responses and that it needs to be evaluated for each drug separately to identify the ideal schedule for potential synergism.


HSP27 is essential for melanoma cell proliferation, migration and chemo resistance

S. Blunder, C. Briand, W. Jerney, H. Pehamberger, C. Hoeller, N. Schicher

Division of General Dermatology, Department of Dermatology, Medical University of Vienna, Vienna, Austria

Heat shock proteins (Hsps) are molecular chaperones which are up regulated upon a broad variety of cell stress. Recent studies have shown that the up regulation of HSP27 is common in different cancers and is linked to poor prognosis and metastasis. HSP27 interferes with apoptosis in the intrinsic as well as in the extrinsic pathway. Hence, it can confer resistance to apoptosis elicited through chemotherapeutic drugs, resulting in decreased chemo- and radiosensitivity of cancer.

The aim of this study was to identify the relevance of HSP27 in melanoma using a specific antisense oligonucleotide (ASO). Cell proliferation, migration and apoptosis were used as readout-parameters.

518A2 (BRAF V600E), M24 (NRAS Q61R), MelJuso (NRAS Q61L), CHL1 (wt) and 607B (NRAS Q61K) melanoma cells were transfected with ASO or a scrambled control oligonucleotide (SCO) for two consecutive days. HSP27 protein expression status was determined by Western Blotting. MTS assays were used to assess cell death and proliferation; scratch assays were performed to test the impact on migration. Apoptosis was determined by Annexin-FACS staining and Caspase-3 Cleavage activity. The determined IC50 for the ASO was used to assess a potential sensitizing effect of ASO treatment towards chemotherapeutic agents.

Treatment of melanoma cells with HSP27 ASO is an effective tool to downregulate HSP27 protein levels at low nanomolar concentrations. Further, ASO treatment clearly reduces cell proliferation and induces cell death through apoptosis, irrespective of the genetic background. In scratch assays, ASO transfection impairs migratory capabilities of melanoma cells. Upon ASO treatment melanoma cells show increased chemosensitivity towards cytotoxic drugs.

These preliminary data underline the role of HSP27 in melanoma. Further validation in a xenotransplant melanoma mouse model is being performed. Correlation of HSP27 expression levels and disease stage of melanoma will be important to further assess the potential role of HSP27 directed agents in melanoma.


Surgery in stage IV melanoma patients: results from a single institution.

E. Pennacchioli1, S. Gandini2, F. Verrecchia1, G. Tosti1, G. Spadola1, F. Baldini1, M. Mosconi1, P. Ferrucci1, A. Testori1

1 Melanoma and Sarcoma Unit, European Institute of Oncology, Milan, Italy; 2Epidemiology Division, European Institute of Oncology, Milan, Italy

There is no consensus regarding the appropriate management of melanoma patients at stage IV and resectable metastases. We conducted a systematic review and meta-analysis to examine the effect of surgery on overall survival (OS) in patients with resectable metastases. The systematic review was performed by searching MEDLINE, EMBASE, Cochrane and ISI Web of Science. The meta-analysis was performed using time-to-event data from which hazard ratios (HRs) and 95% confidence intervals (CIs) of OS were estimated and summary estimates were obtain through random effects model. Heterogeneity and publication bias were investigated by sensitivity analyses and funnel plot regression. Thirty-three retrospective studies evaluating survival from resected and unresected metastases were found from 1978 until 2010. In resected patients median follow-up is 1.5 yr and 2-yr OS is 27% whereas in unresected the median follow-up is 6 months and 2-yr OS is 12%. Meta-analysis was carried out on the seven studies reporting information on visceral metastases and including 1165 patients. We included also 93 patients at stage IV treated at the European Institute of Oncology. Surgery has shown an improvement in survival of patients managed with palliative resection of their visceral metastases: the summary estimate suggests a significant decrease in the risk of mortality of 40% for patients who undergo surgery, with no indication of heterogeneity [HR = 0.6 (0.45–0.78); I2 = 0]. However we found indication for publication bias (P = 0.03) suggesting that small studies published only mainly if they found significant results. Our results suggest the benefit of surgical resection for advanced-stage melanoma. Patients with limited sites and numbers of metastases should be considered for curative resection regardless of the location of the disease.


Evaluation of incomplete sentinel node biopsy procedures and sentinel node positivity rates as surgical quality assurance parameters in melanoma patients

A. J. Spillane1–5, L. E. Haydu1,2, N. C. Lee5, R. F. Uren2,6, J. R. Stretch1–4, K. F. Shannon1,3,4, M. J. Quinn1,3,4, R. P. M. Saw1–4, W. H. McCarthy1,2, J. F. Thompson1–4

1 Melanoma Institute of Australia, North Sydney, NSW, Australia; 2The University of Sydney, Sydney, NSW, Australia; 3Mater Hospital, North Sydney, NSW, Australia; 4Royal Prince Alfred Hospital, Camperdown, NSW Australia; 5Royal North Shore Hospital, St Leonards, NSW, Australia; 6Nuclear Medicine and Diagnostic Ultrasound, RPAH Medical Centre, Sydney, Australia

Sentinel node biopsy (SNB) is a valuable staging technique for melanoma patients. There is very little literature on quality assurance (QA) validation of individual surgeon's ability to perform SNB. Incomplete SNB (I-SNB) is defined as fewer lymph nodes retrieved than identified on preoperative lymphoscintigraphy (LSG). This study evaluates I-SNB and SNB positivity rates as potential QA parameters.

A prospective database identified 2874 patients with primary melanoma and SNB performed in a single lymphatic field. The patients were managed by seven surgeons at a specialist melanoma center. Matching LSG data were obtained from a separate database. LSG was performed with small particle colloid allowing lymphatic channels entering sentinel nodes to be visualized on early dynamic scanning.

The overall rate of I-SNB was 17.7% (including neck 23.3%, axilla 7.8% and groin 28.8%). The individual surgeons’ percentage of cases in each SNB region varied significantly (P < 0.001). The individual surgeons’ overall I-SNB rate varied significantly (P < 0.001). The frequency of I-SNB in the axilla varied significantly between surgeons from 3–16% (P < 0.001) with median 6%; the groin I-SNB rate varied significantly from 21–41% (P = 0.002) with median 26% and neck I-SNB rate varied from 19–43% (P = 0.374) with median 22%. The respective axillary, groin and neck SNB positivity rates for I-SNB patients were 10%, 23% and 18% compared to complete SNB patients 14%, 19% and 14%. There were no significant differences between surgeons’ SNB positivity rates.

In conclusion, I-SNB rates vary between lymph node regions and between surgeons. The SNB positivity rates do not vary commensurate with the rates of I-SNB. Thus, I-SNB is not a differentiator of surgical performance in the ranges set by these experienced surgeons. However, surgeons falling outside these ranges may vary in their ability to detect SNB positivity. Centers where LSG with small particle radiocolloid is not available may have different I-SNB rates.


Incomplete sentinel node biopsy is not clearly related to survival or regional recurrence in cutaneous melanoma patients

A. J. Spillane1–5, N. C. Lee1, T. Pang1, L. E. Haydu2,3, R. F. Uren2,6

1 Royal North Shore Hospital, St. Leonards, NSW, Australia; 2Sydney Medical School, The University of Sydney, Sydney, NSW, Australia; 3Melanoma Institute Australia, Sydney, NSW, Australia; 4The Mater Hospital, North Sydney, NSW, Australia; 5Royal Prince Alfred Hospital, Camperdown, NSW, Australia; 6Nuclear Medicine and Diagnostic Ultrasound, RPAH Medical Centre, Camperdown, NSW, Australia

In melanoma patients, we define incomplete sentinel node biopsy (I-SNB) as when fewer lymph nodes are removed during sentinel node biopsy (SNB) than are identified on preoperative lymphoscintigraphy (LS). This study quantifies the frequency of I-SNB and seeks to evaluate any correlation with outcome in patients who had negative SNB at the time of primary melanoma diagnosis.

We undertook evaluation of a prospective database of consecutive patients having LS and negative SNB from 1996 to 2006. Additional LS information was obtained from a nuclear medicine database. All statistical analyses were performed using the IBM SPSS Statistic 19.0 software package.

I-SNB occurred in 20% of the cohort (n = 2007). For axillary (n = 895), groin (n = 569) and neck/axial patients (n = 334) I-SNB occurred in 12%, 26% and 28% of cases respectively (P < 0.001). On univariate analysis, there was a significant association between I-SNB and worse disease-free survival (DFS), P = 0.007 and trend towards worse melanoma-specific survival (MSS), P = 0.056. I-SNB was not associated with worse regional recurrence-free survival (RRFS), P = 0.144. There was no relationship between I-SNB and worse DFS, RRFS or MSS on multivariate analysis. Sentinel node region (axilla better than groin and neck/axial) had a significant association with RRFS (P = 0.039) on univariate analysis and DFS on univariate (P = 0.009) and multivariate analysis. Significantly worse outcomes for MSS, DFS and RRFS were seen with male gender, increasing age, high mitotic count, ulceration, and increasing Breslow thickness.

In conclusion, this study demonstrates no statistically significant relationship between I-SNB and patient outcomes when adjusting for known prognostic factors. These data do not exclude the possibility that I-SNB may have a weak association with worse outcomes.


Inguinal or ilio-inguinal dissection for metastatic melanoma in groin lymph nodes – a randomized trial is still required

A. J. Spillane1–5, L. E. Haydu1,2, R. P. M. Saw1–4, J. F. Thompson1–4

1 Melanoma Institute of Australia, North Sydney, NSW, Australia; 2The University of Sydney, Sydney, NSW Australia; 3Mater Hospital, North Sydney, NSW, Australia; 4Royal Prince Alfred Hospital, Camperdown, NSW, Australia; 5Royal North Shore Hospital, St Leonards, NSW, Australia

Regional lymph node dissection (RLND) is currently the most effective therapy for metastatic melanoma in groin lymph nodes (LN). Whether patients have inguinal dissection (ID) or ilio-inguinal dissection (I-ID) varies between institutions as may the thoroughness of surgery. The preferred operation remains contentious with multiple publications in 2011 reaching conflicting conclusions. This study aims to analyse predictors for survival outcomes for ID and I-ID with a focus on extent of surgery.

A prospective database identified 189 patients who had 200 groin dissections between July 2002 and February 2008. From this group 177 patients had one RLND and were assessed for predictors of survival.

ID occurred in 93 patients and I-ID in 84 patients. ID had median LN retrieval of 11, inter-quartile range 10–14, ≥8 LN were retrieved in 90% of cases, and 38% of cases had ≤10 LNs. I-ID had median LN retrieval of 21.5, inter-quartile range 17–25, ≥14 LN were retrieved in 90% of cases, and 0 cases had ≤10 LN. There were no disease-free or melanoma-specific survival differences between ID and I-ID groups but the I-ID group had higher lymph node disease burden. The strongest predictors of survival on multivariate analysis were LN ratio, macroscopic LN disease and ulceration of the primary melanoma. Overall 39% of patients having I-ID had positive pelvic LNs but only 9.3% of those having I-ID for positive sentinel nodes. Pelvic failure after earlier inguinal dissection occurred in 7 of the 93 inguinal dissections.

In conclusion, the high rate of involved pelvic LNs and 7.5% pelvic failure in patients having ID suggests I-ID should be considered in all patients, especially those with macroscopic metastases in groin LNs. The authors are seeking expressions of interest for a randomized controlled trial of ID versus I-ID with reliability of pre-operative imaging, disease-free survival and morbidity as endpoints.


Indocyanine green fluorescence-guided sentinel lymph node biopsy in dermato-oncology

I. Stoffels1, H. von der Stück1, C. Boy2, T. Pöppel2, N. Körber1, J. Dissemond1, D. Schadendorf1, J. Klode1

1 Department of Dermatology, Venerology and Allergology, University of Essen-Duisburg, Germany; 2Department of Nuclear Medicine, University of Essen-Duisburg, Germany

Sentinel lymph node biopsy (SLNB) for cutaneous malignancies usually carried out with radioactive nanocolloids (Tc-99m) and/or blue dye. But the use of radioactive nanocolloids involves expensive equipment, radiation protection measures and is not generally available. The SLNE is controversially discussed internationally. This is especially given to the high false-negative rate up to 44%. An alternative could be the fluorescent dye indocyanine green (ICG). Preliminary data from toxicity and feasibility studies suggest that SLN detection guided by ICG fluorescence could emerge as a new method for SLNB. ICG-injection can be performed without allergic reactions or toxic side effects. Fluorescence imaging enables transcutaneous real-time lymphography and intraoperative lymph node detection thus combining the advantages of radiocolloid method and blue dye method. One reason for this could be the large size of the Tc-99-nanocolloid-complex with 6–12 nm. The ICG bound to albumin has a size of about 2.1 nm and can thus stain lymphatic channels and SLN, which escape of a staining by Tc-99m-nanocolloid.

We investigated the advantage of intraoperative fluorescence detection of lymphatic vessels and SLN with a Near-Infrared (NIR) camera in comparison to conventional methods using preoperative lymphoscintigraphy and SPECT/CT in 24 patients with cutaneous malignancy.

A total of 68 SLNs were removed in 24 operative procedures. In 7 SLN (10.3%; 7/68) the histopathological assessment could demonstrate a metastatic involvement. Thirteen additional SLN (19.1%) in 9 patients were only identified using the the fluorescent labeling. Two of these additional SLN (8.3%; 2/24) showed metastatic involvement.

The ICG fluorescence-guided SLNB is an innovative imaging technique for dermato-oncology, reliable and providing additional information in the detection of SLN. Therefore SLNB with fluorescence-dye is an attractive option with intraoperative real-time lymphoscintigraphy to improve the detection of SLN in cutaneous malignancies and potential reduction of the false negative rate in SLN.


Preoperative SPECT/CT: the new gold-standard in the detection of sentinel lymph nodes in patients with cutaneous malignant melanoma?

I. Stoffels1, C. Boy2, T. Pppel2, J. Kuhn1, K. Kltgen1, D. Schadendorf1, J. Dissemond1, J. Klode1

1 Department of Dermatology, Venerology and Allergology, University of Essen-Duisburg; 2Department of Nuclear Medicine, University of Essen-Duisburg

There is some controversy around the value of sentinel lymph node excision (SLNE) because of the high false negative rate, the potential high morbidity and the high operating costs. In order to improve preoperative three-dimensional mapping of sentinel lymph nodes (SLN) by means of hybrid single photon emission computed tomography/computed tomography (SPECT/CT) is gaining significance. But some authors therefore recommend SPECT/CT only for patients with SLN in head and neck region or when no sentinel node is shown in preoperative lymphoscintigraphy. Our study seeks to identify the potential medical and economic advantages of preoperative SPECT/CT in direct comparison to standard SLNE without SPECT/CT in patients with primary malignant melanoma and SLN in the head and neck region, axillar and groin region.

The data of 404 clinically lymph node-negative patients with early stage melanoma, who underwent SLNE with or without preoperative SPECT/CT within 7 year were analyzed.

The SLNE with SPECT/CT-technique significantly reduced operating time. In the SPECT/CT cohort significantly more SLN were detected than in the control cohort (2.36 SLN/patient versus 1.85 SLN/patient; P < 0.001). Additionally the discovered metastatic involvement of SLN was significantly higher in the SPECT/CT cohort (31% (44/140) versus 21% (55/264; P < 0.001). The median follow-up time was 26.9 months and the false negative rate in the SPECT/CT cohort was significantly lower than in the control cohort (8.3% versus 17.9%, P < 0.001). Moreover, SLNE with SPECT/CT-technique was feasible even in head and neck surgery, in patients who were obese or had multiple comorbidites, by using local anesthesia (LA) resulting in significant cost reduction (91.69/SLNE with LA versus 359.00/SLNE with general anesthesia, P < 0.0001).

The SPECT/CT imaging is an innovative and reliable imaging technique, providing additional information in the detection of SLN. Therefore SLNE with SPECT/CT is an attractive option to improve the detection of SLN in patients with malignant melanoma and to reduce false negative SLN results. We recommend the SLNE in SPECT/CT-technique as the new gold-standard in the detection of SLN in patients with cutaneous malignant melanoma.


Lentigo maligna mimicking invasive melanoma in Mohs surgery

T. Tsakok1, N. Sheth1, A. Robson2, C. Gleeson1, R. Mallipeddi1

1 Dermatological Surgery and Laser Unit, St John's Institute of Dermatology, St Thomas’ Hospital, London, UK; 2Department of Dermatopathology, St John's Institute of Dermatology, St Thomas’ Hospital, London, UK

A 76-year old retired sailor presented with a 3 × 2.5 cm unevenly pigmented macule on the left neck. This was at the site of a previous lentigo maligna that had been excised with a standard 5 mm margin. Biopsy showed a lentiginous proliferation of atypical melanocytes, tracking down adnexae. This confirmed recurrence, and the patient was referred for Mohs micrographic surgery (MMS), during which horizontal sections confirmed lentigo maligna in the epidermis. However, one section revealed a nodule of severely atypical melanocytes, apparently lying mid-dermis. This presented an interpretational problem. Lentigo maligna commonly involves hair follicle epithelia; thus, the nodule may reflect either:

1 Complete replacement of follicular epithelial cells still confined by basement membrane (in-situ disease).

2 A focus of invasive melanoma – in which case MMS should be converted to wide local excision.

To clarify this, immunocytochemical staining was performed for laminin-5, a component of basement membrane. This revealed immunopositivity of the basement membrane of follicular epithelium, confirming that the atypical melanocytes were confined to hair follicle. MMS was therefore completed in a 2-stage, 10-section procedure.

Lentigo maligna is a form of melanoma in-situ with high prevalence in Caucasians, typically arising on chronically sun-damaged skin. Following biopsy and exclusion of invasive disease, therapy involves conventional excision, MMS, topical treatment or radiotherapy. Lentigo maligna presents several challenges in diagnosis and management, including its propensity to involve hair follicle adnexae. This may create histological difficulty in distinguishing such foci from invasive melanoma. In this case, a nodule of atypical melanocytes appeared to lie within the dermis, potentially altering treatment and prognosis. The use of laminin-5 provided a means of resolving this diagnostic dilemma, facilitating continuation of MMS to tumour clearance. Laminin-5 staining to facilitate the treatment of lentigo maligna by MMS has not previously been reported and we wish to highlight this fascinating example.


Utilization and costs of healthcare resources in patients with metastatic melanoma in the US

S. Wang, Z. Zhao, B. Barber, S. Gao, V. Wagner

Amgen, Inc., Thousand Oaks, CA, USA

The incidence of metastatic melanoma (MM) has increased during the last three decades and the death rate has risen faster than most other cancers. MM carries severe burdens to patients, their families, and the payers. However, little information exists regarding the economic burden of MM.

Using a large US medical claims database, this study assessed healthcare resource utilization and costs among patients with MM in the real-world practice setting. Patients were identified between 2005 and 2010 using ≥2 melanoma diagnoses (ICD-9-CM: 172.xx, V10.82) and ≥2 diagnoses for metastasis (ICD-9-CM: 197.xx, 198.xx). The index date was the first date of metastasis diagnosis. Patients who had other primary malignant tumors prior to the melanoma diagnosis, were younger than 18 years old at the index, or had a pre-index period of <6 months, were excluded from the analysis. Patients were followed from the index date to death, disenrollment, or end of the study period (6/30/2010), whichever occurred first. Healthcare utilization and costs (adjusted to 2010 dollars) were examined per patient-year for office visits, outpatient visits, outpatient pharmacy, emergency room (ER) visits and inpatient hospitalization.

A total of 2546 patients with metastatic melanoma were identified. Mean (± standard deviation) age at the index date was 60.6 (± 14.0) years and 36.5% were female. Overall, 87.3% of the patients had physician office visits with a mean of 24.0 visits per person-year, 64.7% had ER visits with a mean of 12.9 visits per person-year, 90.6% had outpatient visits with a mean of 8.3 visits per person-year, and 82.2% were hospitalized with a mean of 12.4 hospitalizations per person-year. The mean total costs per patient-year were $127 204, which was driven mainly by inpatient costs ($60 355/patient-year) and outpatient costs ($34 540/patient-year).

Inpatient and outpatient care are key cost drivers in the medical management of patients with advanced melanoma.


Analysis of factors influencing drug treatment for patients with metastatic melanoma in the US

Z. Zhao1, S. Wang1, B. Barber1, M. Hsiao2, S. Gao1, V. Wagner1

1 Amgen, Inc., Thousand Oaks, CA, USA; 2University of Southern California, Los Angeles, CA, USA

Metastatic melanoma (MM) is an aggressive disease. Little research has been conducted to describe the treatment patterns for MM and the factors influencing drug treatment.

Using a large US medical claims database, this study assessed treatment patterns and factors influencing use of drug treatment in patients with MM. Patients were identified between 2005 and 2010 using ≥2 melanoma diagnoses (ICD-9-CM: 172.xx, V10.82) and ≥2 diagnoses for metastasis (ICD-9-CM: 197.xx, 198.xx). The index date was the first date of metastasis diagnosis. Patients were followed from the index date to death, disenrollment, or end of the study period (6/30/2010), whichever occurred first. Logistic regression was used to assess driving factors for use of drug treatment by adjusting for the patients’ demographic and clinical characteristics.

A total of 2546 patients with MM met the study inclusion and exclusion criteria. Mean (±SD) age was 60.6 (±14.0) years old and 36.5% were female. Mean length of follow-up observation period was 322 days. Overall, 38.7% received drug treatment. Among them, 48.7% received temozolomide, followed by paclitaxel (22.3%), carboplatin (19.4%), interleukin-2 (IL-2) (17.6%), and dacarbazine (DTIC) (17.2%), 14.4% interferon alfa-2b (IFN), 9.9% cisplatin, 6.3% vinblastine, and 4.7% granulocyte-macrophage colony-stimulating factor (GM-CSF). Logistic regression revealed that female gender and age were negatively associated with likelihood of receiving drug treatment (P < 0.01 and P < 0.0001, respectively); the presence of brain metastases most strongly increased the odds of receiving drug therapy compared to other visceral metastases (58.7% increase, P < 0.001); patients with liver or lung metastasis were also more likely to receive drug treatment (P < 0.05 and P < 0.05, respectively).

Given the high proportion of patients who do not receive drug treatment, it is important to have a better understanding of the factors influencing drug treatment.


A dose confirmation and signal-generating study of the immunocytokine L19-IL2 in combination with dacarbazine in patients with metastatic melanoma

B. Weide1, T. K. Eigentler1, G. Spitaleri2, F. G. De Braud2, A. Romanini3, R. Gonzalez-Iglesias4, A. Tasciotti4, L. Giovannoni4, K. Schwager5, V. Lovato5, M. Kaspar5, E. Trachsel5, D. Neri6, H. D. Menssen4, C. Garbe1

1 Department of Dermatology, University of Tübingen, Tübingen, Germany; 2European Institute of Oncology, Milan, Italy; 3Santa Chiara University Hospital Pisa, Pisa, Italy; 4Philogen S.p.A., Siena, Italy; 5Philochem AG, Zurich, Switzerland; 6Federal Institute of Technology, Zurich, Switzerland

L19-IL2 is a tumor-targeting immunocytokine composed of an antibody fragment specific to the EDB domain of fibronectin, a marker of tumor angiogenesis, and of human interleukin-2 (IL2). L19-IL2 is able to selectively deliver IL2 to tumor tissues exploiting the expression of EDB on newly formed blood vessels. The recommended monotherapy dose (RD) of 22.5 MioIU IL2 equivalents (day 1, 3, and 5) of L19-IL2 was defined earlier. In this study, we investigated L19-IL2 in combination with dacarbazine in patients with metastatic melanoma, assessing safety, tolerability, and activity. The study consists of a dose confirmation part, during which 10 patients with metastatic melanoma were enrolled, followed by a fixed dose part, during which 22 chemo-naive metastatic melanoma patients were treated at the RD. Patients received intravenous infusions of L19-IL2 (10, 15 and 22.5 MioIU during dose confirmation, and 22.5 MioIU in the fixed dose part of the study) on day 1, 3 and 5 in combination with 1 g/m2 of dacarbazine on day 1, every 21 days. Up to six treatment cycles were given, followed by a maintenance phase during which only L19-IL2 was applied biweekly. Tumor assessment was performed every 6 weeks. Data on safety and efficacy were evaluated using CTC v3.0 and RECIST criteria, respectively. Median age was 55 years (30–83). The RD was confirmed to be 22.5 MioIU. Toxicity was manageable and reversible, with no treatment related deaths. Twenty-nine out of 32 patients were evaluable for efficacy according to RECIST. In a centralized radiology analysis, 8/29 patients achieved RECIST-confirmed objective responses, including a complete response, still ongoing 21 months after start of treatment. The 1-year survival rate was 61.5% and median overall survival was 14.1 months for all patients treated at RD. L19-IL2 combined with dacarbazine, both at RD, can be safely and repeatedly administered to patients with metastatic melanoma, and showed promising signs of clinical activity. To further evaluate this combination therapy for metastatic melanoma, a controlled randomized phase II trial is ongoing.


Neurofibroma-like nodular spindle cell melanoma: CD34 fingerprint and CGH for diagnosis

I. Yeh, S. S. Vemula, T. H. McCalmont

Departments of Pathology and Dermatology, University of California, San Francisco, CA, USA

We present a case which highlights the use of two maneuvers useful in the diagnosis of spindle cell melanoma. A shave biopsy from the cheek of a 58 year old man demonstrated a thin invasive melanoma of 0.3 mm in thickness with a <2 mm wide intraepidermal component. Below this melanocytic lesion but not contiguous with it, there was an S-100 positive, Melan-A negative spindle cell proliferation that was transected. Upon re-excision, no residuum of conventional melanoma was identified, but a residual spindle cell population that was 4 mm in diameter, nodular, well-circumscribed, cytologically bland, and S-100 positive was noted. At our consensus (peer-review) conference, we discussed the differential diagnosis of spindle cell melanoma and neurofibroma, and the group favored neurofibroma but agreed a conclusive diagnosis could not be reached based on histopathologic features alone. To further address the differential diagnosis, we performed CD34 staining which demonstrated lack of a CD34 fingerprint. We also completed array-based comparative genomic hybridization (aCGH), which demonstrated gain of chromosome 6p, loss of 6p and gain of 7. These two methods of analysis both supported a diagnosis of spindle cell melanoma.


β-blockers and survival from melanoma

S. Field, F. Elliott, J. H. Barrett, D. T. Bishop, J. A. Newton-Bishop

Section of Epidemiology and Biostatistics, Leeds Institute of Molecular Medicine, University of Leeds, UK

Beta-adrenergic receptor antagonists (β-blockers) are widely prescribed. There is in vitro and epidemiological evidence that blockade of β-adrenergic receptors (β-AR) reduces tumor progression and metastasis, and melanomas express β2-AR. A recent Italian study in 121 cases suggested that β-blocker usage for >1 year was associated with prolonged melanoma disease-free survival. We report the effects of β-blocker usage on tumor thickness and outcome in a much larger UK cohort.

Thousand three hundred and seventy-three cases of melanoma with Breslow thickness >1 mm were recruited to the Leeds Melanoma Cohort between 2001 and 2011. A survival analysis was carried out as reported by the Italian study using exposure of >1 year as a definition of β-blocker usage. One hundred and thirteen participants were exposed to β-blockers for >1 year.

Breslow thickness was significantly higher (median 2.1 mm, range 1.0–18.0) in the exposed group compared to the unexposed group (median 2.0 mm, range 1.1–9.3) in unadjusted analysis (P = 0.01) but not in age and sex-adjusted analysis (P = 0.25).

In survival analyses adjusted for age and gender β-blockers conferred a non-significant protective effect for relapse, hazard ratio (HR) 0.81 (95%CI 0.54–1.22, P = 0.32) and when further adjusted for Breslow thickness, tumor site, body mass index (BMI) and smoking status HR = 0.81 (95%CI 0.53–1.24), P = 0.33).

For overall survival (OS) β-blocker usage (adjusted for age and gender) conferred a non-significant protective effect HR 0.87 (95%CI 0.56–1.35, P = 0.53) and for melanoma-specific survival (MSS) there was also a non-significant protective effect HR 0.75 (95%CI 0.44–1.26, P = 0.28). Further adjusting the model for tumor thickness, site, BMI and smoking status, β-blockers conferred a similar non-significant protective effect on OS (HR = 0.90, 95%CI 0.56–1.44, P = 0.66) and on MSS (HR 0.71, 95%CI 0.40–1.26, P = 0.24).

This study provides no support for the hypothesis that β-blockers confer a significant protective effect for relapse, or survival, but we cannot exclude a small effect.


A case for performing molecular analysis on challenging melanocytic lesions

M. Moore, R. Gasparini

NeoGenomics Laboratories, 12701 Commonwealth Drive, Suite 5, Fort Myers, FL, USA

The characterization of melanocytic progression to melanoma presents specific challenges derived from the diverse phenotypic nature of the disease. While benign nevi contain few or no genetic alterations, and melanoma are known to possess frequent gross genetic alterations, there continues widespread discussion on the fundamental underlying biology of dysplastic nevi. Central to the argument is the contested proposition of the nature of melanocytic dysplasia; Are a subset of dysplastic nevi premalignant lesions of melanoma, or are all dysplastic nevi fundamentally benign and genetically distinct from melanoma?

Recently, Fluorescent in situ hybridization has been utilized to identify atypical melanocytic lesions with gross chromosomal aberrations, as a surrogate marker for progression to melanoma. Having utilized this assay in a clinical setting for several months, this article will provide a statistical review of the first 200 clinical atypical melanocytic cases submitted for analysis (including a breakdown of samples by morphology and percent positivity by FISH), and highlight a subset of cases (10) where the FISH result was clinically important in rendering a final diagnosis.


Fluorescent in situ hybridization: utility and limitations for the analysis of melanocytic lesions

M. Moore, R. Gasparini

NeoGenomics Laboratories, 12701 Commonwealth Drive, Suite 5, Fort Myers, FL, USA

Fluorescent in situ hybridization is a common diagnostic tool for the determination of malignancy, and subtyping of several solid tumors (i.e., invasive breast carcinoma, prostate, bladder cancer, etc.) and leukemias (i.e., CML, ALL, CLL, myeloma, etc.).

Recently, Fluorescent in situ hybridization (FISH) has been utilized to identify atypical melanocytic lesions with gross chromosomal aberrations, as a surrogate marker for progression to melanoma.

From a technical perspective, performing FISH for the analysis of melanocytic lesions is particularly challenging. Melanocytes are long lived, subject to constant UV light exposure, and undergo periods where they exhibit a high rate of cell division coupled to replicative senescence. They congregate in small nests and morphologically under a DAPI stain, look very similar to several co-localizing cell types. Analysis is technically demanding with the scoring technologist requiring training in both melanocytic histology and FISH analysis. Scoring is compounded by the intrinsic challenges associate with detecting deletions from thin sectioned material, and further complicated by the observation that genetic aberrations are not always found in the most phenotypically worrisome region which necessitates the scanning of the entire section and analysis of several regions of tissue. Additionally, several authors, including this study have identified a disparately high proportion of tetraploid melanocytic cells via melanocytic FISH analysis. While tetraploidy has been correlated with carcinogenesis in several cancer models (bladder cancer is the most recognized example), in melanocytes, tetraploid cell complements have not been linked to poor outcome in multiple studies, which significantly increases the complexity of analysis.

With a number of university and commercial laboratories now offering this assay, this presentation will review the technical challenges of adapting FISH technology for differentiating melanoma from atypical (dysplastic) melanocytic nevi.


Development of Nanolipolee-007: a unique drug targeting the PI3 kinase, STAT and MAP kinase pathways deregulated in melanomas

R. Gowda1,2, S. V. Madhunapantula1,2,3, A. Sharma1,2,3, G. P. Robertson1–7

1 Department of Pharmacology, The Pennsylvania State University College of Medicine, Hershey, PA, USA; 2Penn State Melanoma Center, The Pennsylvania State University College of Medicine, Hershey, PA, USA; 3Penn State Melanoma Therapeutics Program, The Pennsylvania State University College of Medicine, Hershey, PA, USA; 4Department of Surgery, The Pennsylvania State University College of Medicine, Hershey, PA, USA; 5Department of Pathology, The Pennsylvania State University College of Medicine, Hershey, PA, USA; 6The Foreman Foundation for Melanoma Research, The Pennsylvania State University College of Medicine, Hershey, PA, USA; 7Department of Dermatology, The Pennsylvania State University College of Medicine, Hershey, PA, USA;

Malignant melanoma is a difficult cancer to treat due to the rapid development of resistance to drugs. A case in point is the V600EBraf inhibitor PLX4032, clinically known as Vermurafenib, which while initially effective in patients, is not so for long-term treatment since resistant tumors recur. This observation has cause a major shift in the melanoma field to identify pharmacological agents that target multiple key pathways involved in melanoma development and create bioavalable delivery approaches to get them into the tumor cells. To identify a novel drug targeting multiple key pathways in melanoma, a cell-based screen was undertaken of a natural product library comprised of 480 compounds. The screen identified a potent compound capable of inhibiting melanoma development, called Leelamine (Dehydoabietylamine). Leelamine is derived from the bark of pine trees and inexpensive as well as easy to isolate and purify. Using a combination of protein arrays and system biology followed by validation studies, leelamine was found to inhibit the PI3K, STAT and MAPK pathways deregulated in melanoma. To overcome the poor bioavailability of leelamine, a novel nanoliposomal delivery system was developed called Nanolipolee-007, which loads 60% of the compound and was stable at room temperature or at 4°C for >1 year. Furthermore, the nanoliposomal formulation was as effective at killing melanoma cells, as was leelamine in DMSO. As with leelamine, Nanolipolee-007 was found to be more specific at killing melanoma cells than normal cells. Mechanistically, Nanolipolee-007 decreased cellular proliferation and triggered apoptosis mediated through a G0/G1 block resulting in fewer cells in the S-phase of the cell cycle. Nanolipolee-007 significantly retarded melanoma xenograft tumor development without affecting animal weight or organ functions through inhibition of the PI3 kinase, STAT and MAP kinase pathways. Melanoma tumor inhibition was attributed to decreased melanoma cell proliferation, which consequently increased tumor cell apoptosis and decreased tumor vascular development. Collectively, these studies show preclinical development of a novel natural product in a unique nanoliposomal formulation for the treatment of metastatic melanoma.


Desmoplastic melanoma: sentinel node biopsy is indicated in pure and mixed variants

D. Han1, J. S. Zager1, S. S. Marzban1, S. Shah2, B. M. Walls1, N. Rao3, D. Yu4, X. Zhao4, V. K. Sondak1, J. L. Messina1,5

1 H. Lee Moffitt Cancer Center, Cutaneous Oncology Program; 2University of South Florida, College of Medicine; 3H. Lee Moffitt Cancer Center, Radiation Oncology Program; 4H. Lee Moffitt Cancer Center, College of Medicine, Department of Biostatistics, Oncologic Sciences, University of South Florida; 5Department of Pathology, University of South Florida

Desmoplastic melanoma (DM) has been associated in the literature with a lower likelihood of nodal metastases and the utility of sentinel lymph node (SLN) biopsy (SLNB) for DM is debated. We describe a large single institution experience with SLNB in patients with DM to determine clinicopathologic factors that predict nodal metastases. Retrospective review identified 205 patients with DM who underwent SLNB at our institution from 1992 to 2010. Clinicopathologic characteristics were reviewed and correlated with SLN status. Median age was 66 years and 69% of cases were male. Median Breslow thickness was 3.7 mm and the majority (51.2%) of tumors were located on the head and neck (HN). Overall, 28/205 (13.7%) cases had a positive SLN (+SLN). In 128 cases, histologic subtype data was available; 61 (47.7%) were mixed DM (mDM) and 67 (52.3%) were pure DM (pDM). A +SLN was seen in 24.6% of mDM cases but still in 9% of pDM cases (P < 0.05 for mDM versus pDM). After controlling for age, multiple regression analysis demonstrated that mDM correlated with a +SLN (P < 0.05; OR: 2.9, 95% CI: 1.0–8.4 for mDM versus pDM). Completion lymph node dissection (CLND) was performed in 24/28 (86%) +SLN patients. In four cases, an additional positive non-SLN was found; three of these four cases had mDM. After a median follow-up of 6 years, regional recurrence occurred in seven cases. In two of these cases, recurrence was after CLND performed for a +SLN; both cases had mDM. In five of the regional recurrences (4/5 cases HN primary), SLNB was negative and 3/5 of these cases had mDM. Patients with mDM have high rates of nodal metastases and SLNB in these patients is warranted. However, patients with pDM should still be evaluated for SLNB since nodal metastases are seen in a considerable percentage of these cases.


The Melanoma Institute Australia Melanoma Research database: past, present and future

L. E. Haydu1,2, J. F. Thompson1,2, M. Coleman1, H. M. Shaw2, MIA Database Working Group1

1 Melanoma Institute Australia, Sydney, NSW, Australia; 2Discipline of Surgery, Sydney Medical School, The University of Sydney, Sydney, NSW, Australia

The Melanoma Institute Australia (MIA, formerly the Sydney Melanoma Unit) has been collecting clinical research data on melanoma patients since the late 1960s. The platform of the melanoma research database (MRD) has evolved from punch cards to a clinical reports system (CRS) to what now exists on a Microsoft SQL server. The database contains research records for >33 000 patients, 57 800 primary melanoma operations, 6000 sentinel node biopsies, and 9000 regional node dissections. Melanoma Institute Australia made a major contribution of data to the melanoma staging system revision by the American Joint Committee on Cancer in 2001 and again in 2007. We intend to present historical trends in data capture at the MIA at a significant turning point in the timeline of the MRD. To accommodate a growing tumor bank and burgeoning requests for molecular and medical oncology studies, the MRD is currently being restructured. The new structure of the database will include sufficient links to enable tracking of individual melanoma specimens, pathology, and mutation test results. Additionally, algorithms are being implemented to re-stage patients beyond first presentation and to manage multiple pathology records which will increase the efficiency of database queries at the institute. The goal of these initiatives is to continue to produce high-quality, contemporary data items to support the investigation of expanding clinical research questions.


A prospective, population-based study of 40 000 women regarding host factors, UV exposure and sunbed use in relation to risk and anatomic site of cutaneous melanoma

K. Nielsen1,2, A. Måsbäck2,3, H. Olsson2,4, C. Ingvar2,5

1 Division of Dermatology, Department of Clinical Sciences, Lund University, Sweden; 2Lund Melanoma Study Group, Lund University, Sweden; 3Division of Pathology, Department of Clinical Sciences, Lund University, Sweden; 4Division of Oncology and Cancer epidemiology, Department of Clinical Sciences, Lund University, Sweden; 5Division of Surgery, Department of Clinical Sciences, Lund University, Sweden

Prospective cohort studies about cutaneous melanoma (CM) risk are still few. Host factor- and UVR exposure data were collected prospectively by questionnaire in this population-based cohort study including 40000 Swedish born women, aged 25–64 years at enrolment (1990). Risk for CM [Cox regression and Stepwise Cox regression (SCR), hazard ratios (HRs) with 95% Confidence Intervals (CI)] in relation to risk factors, age groups (older or younger than 40 years) and primary site, were analysed.

In 29 520 women with complete follow-up through 2007, 155 invasive and 60 in situ CM were recorded.

In the whole cohort high numbers of nevi (HR, 2.9; 95% CI, 1.7–5.0) as well as heredity (HR, 3.7; 95% CI, 2.0–6.8) were associated with risk for CM. SCR analysis added red hair as a risk factor. Sunbed use >10 times/year was associated with an increased risk for CM in women <40 year (HR, 2.5; 95% CI, 1.0–6.2) and a trend for risk associated with sunbathing vacations (HR, 1.4; 95% CI, 1.0–2.0) was shown for women >40 year. Trunk melanoma showed correlations with high numbers of nevi (HR, 3.0; 95% CI, 1.2–7.3) and heredity (HR, 3.2; 95% CI, 1.1–9.4). Head/neck site was correlated to sunbathing vacations (HR, 2.5; 95% CI, 1.2–5.3) and heredity (HR, 7.6; 95% CI, 1.8–31.8).

Our study supports divergent etiologic pathways to CM, with high numbers of nevi correlated to increased risk for trunk CM. Furthermore, it confirms that high numbers of nevi, red hair and heredity for CM are the most important risk factors and that frequent sunbed use might be an important risk factor for younger women.


Sentinel lymph nodes containing very small (<0.1 mm) deposits of metastatic melanoma cannot be safely considered tumor-negative

R. Murali1,2,3, C. Desilva3, S. W. McCarthy1,3,4, J. F. Thompson3,5, R. A. Scolyer1,3,4

1 Tissue Pathology and Diagnostic Oncology, Royal Prince Alfred Hospital, Sydney, NSW, Australia; 2Human Oncology and Pathogenesis Program, Memorial Sloan-Kettering Cancer Center, New York, NY, USA 3Melanoma Institute Australia, Sydney, NSW, Australia; 4Discipline of Pathology, Sydney Medical School, The University of Sydney, Sydney, NSW, Australia; 5Discipline of Surgery, Sydney Medical School, The University of Sydney, Sydney, New South Wales, Australia

Some authors suggest that patients with very small (<0.1 mm) deposits of metastatic melanoma in sentinel lymph nodes (SLNs) should be considered SLN-negative. We sought to determine whether such patients have benign clinical outcomes, and whether tumor burden estimates following extensive sectioning of SLNs were more accurate in predicting prognosis.

Patients with a single primary melanoma in whom maximum size of the largest tumor deposit (MaxSize) in SLNs was <0.1 mm in the original sections were identified. Five groups of additional sections were cut from SLN tissue blocks at intervals of 250 μm. In each group of five sections, the first and fifth were stained with hematoxylin-eosin, and the second, third and fourth stained immunohistochemically with S-100, HMB-45 and Melan-A, respectively. In each group, features of tumor deposit(s) in SLNs were evaluated: MaxSize; tumor penetrative depth (TPD) (maximum depth of tumor from inner margin of nodal capsule) and intranodal location (subcapsular or parenchymal). Associations with overall survival were assessed for parameters measured in each group of sections.

Of 20 eligible patients (15 females, five males, median age 60 years), four died from melanoma and two patients of unknown causes, after median follow-up of 40 months. Compared to the original sections, all three parameters were upstaged in subsequent groups of sections, but no further upstaging of MaxSize, TPD or location was seen beyond group 3, group 4 and group 2, respectively. Increasing MaxSize was associated with significantly poorer overall survival in groups 1, 2 and 3. However, 3 patients with MaxSize <0.1 mm (even after extensive examination) died of melanoma.

Very small (<0.1 mm) deposits of melanoma in SLNs may be associated with adverse clinical outcomes, partly due to underestimation of SLN tumor burden in initial sections. Therefore, such patients cannot be assumed to have the same prognosis as those who are SLN-negative.


Clinico-pathologic factors associated with distant metastasis and survival in patients with thin primary cutaneous melanoma

R. Murali1,2,3, L. E. Haydu2,4, J. F. Thompson2,4, M. J. Quinn2,4, R. P. M. Saw2,4, K. Shannon2,4, A. J. Spillane2,4, J. R. Stretch2,4, G. V. Long2,5, R. F. Kefford2,5

1 Tissue Pathology and Diagnostic Oncology, Royal Prince Alfred Hospital, Sydney, NSW, Australia; 2Melanoma Institute Australia, Sydney, NSW, Australia; 3Discipline of Pathology, Melanoma Institute Australia, Sydney, NSW, Australia; 4Discipline of Surgery, Melanoma Institute Australia, Sydney, NSW, Australia; 5Discipline of Medicine, Sydney Medical School, The University of Sydney, Sydney, Australia

Approximately 3–5% of patients with thin (≤ 1 mm) primary cutaneous melanomas develope distant metastasis. In this study, we sought to identify clinico-pathologic factors associated with distant metastasis and survival in a large number of patients with thin melanoma treated at a single institution.

We identified patients with a single primary cutaneous melanoma ≤ 1 mm in thickness diagnosed between January 1983 and December 2003 who developed distant metastasis (cases), and matched patients with no recorded recurrence during follow-up (controls). The cases and controls were matched for matched for age, sex and year of primary melamnoma diagnosis. Associations of clinicopathologic parameters with distant metastasis-free survival (DMFS) and melanoma-specific survival (MSS) were analyzed.

One hundred and seventy-eight cases and 178 control patients were identified. Factors associated with development of distant metastasis were: increasing Breslow thickness (P < 0.001), increasing Clark level of invasion (P < 0.001), increasing mitotic rate (P = 0.001), ulceration (P = 0.025), and AJCC T-subcategory (P < 0.001). Multivariable models including Breslow thickness (but not Clark level) showed that factors independently associated with poorer DMFS were increasing age (HR = 1.01, 95%CI 1.00–1.02); increasing Breslow thickness (HR = 3.21, 95%CI 1.73–5.94, and HR = 3.77, 95%CI 2.11–6.74 for 0.51–0.75 mm and 0.76–1.00 mm, respectively, compared with 0.01–0.50 mm); ulceration (HR = 1.87, 95%CI 1.14–3.06) and mitotic rate (HR = 1.13, 95%CI 1.05–1.21). Similar associations with MSS were found.

Clinico-pathologic predictors of distant metastasis and survival identified in this large study of patients with thin primary cutaneous melanomas will enable more accurate stratification of risk of distant metastasis and poor survival in such patients, and assist in formulating clinical management and follow-up regimens based on the level of risk.


Four-year survival rates for patients with metastatic melanoma who received ipilimumab in phase II trials

J. D. Wolchok1, J. S. Weber2, M. Maio3, B. Neyns4, K. Harmankaya5, K. Chin6, V. de Pril7, R. Humphrey8, A. Hoos6, C. Lebbé9

1 Memorial Sloan-Kettering Cancer Center, New York, NY, USA; 2Moffitt Cancer Center, Tampa, FL, USA; 3University Hospital of Siena, Siena, Italy; 4UZ Brussel, Brussels, Belgium; 5Medical University of Vienna, Vienna, Austria; 6Bristol-Myers Squibb, Wallingford, CT, USA; 7Bristol-Myers Squibb, Braine-l’Alleud, Belgium; 8Bristol-Myers Squibb, Lawrenceville, NJ, USA; 9Hôpital Saint-Louis, Paris, France

Ipilimumab is a fully human, monoclonal antibody that blocks cytotoxic T-lymphocyte antigen-4 to potentiate antitumor T cell responses. Ipilimumab, alone or in combination with dacarbazine, has demonstrated a statistically significant improvement in overall survival (OS) in two phase III, randomized controlled trials in metastatic melanoma with long-term survival in a proportion of patients. The most common ipilimumab treatment-related adverse events were inflammatory in nature, which could be severe and life-threatening, but most were reversible using treatment guidelines. To evaluate the long-term survival effect of ipilimumab, we conducted follow-up analyses on 3 phase II trials in pretreated and untreated metastatic melanoma. Ipilimumab was given every 3 weeks × (induction); eligible patients could receive maintenance ipilimumab every 12 weeks from week 24. Primary efficacy and safety data have been previously published.

Median OS, previously reported 1-, 2-, and 3-year survival rates, and current analyses for 4-year survival rates are given in the Table. In these phase II trials, ipilimumab monotherapy is associated with long-term survival exceeding 4 years in a proportion of patients with metastatic melanoma. Further studies are needed to define the survival results between 10 and 3 mg/kg doses.


A standardized approach to isolated limb infusion for in-transit melanoma on the extremities: perioperative data and outcomes

J. Wong1, G. M. Beasley2, D. S. Tyler2, J. S. Zager1

1 Department of Cutaneous Oncology, Moffitt Cancer Center, Tampa, FL, USA; 2Department of Surgical Oncology, Duke University Medical Center, Durham, NC, USA

Isolated limb infusion (ILI) has been demonstrated to be an effective method for treatment of extremity in-transit melanoma metastases. Although frequently performed at select high volume centers, there is lack of a standardized approach.

Patients with extremity in-transit melanoma from Moffitt Cancer Center and Duke University Medical Center underwent ILI using the same standardized approach. Perioperative parameters and outcomes were recorded.

From 1/2008 to 7/2011, 81 patients underwent ILI: 64 lower extremity (LE) and 17 upper extremity (UE). The standardized approach included heating the extremity to ≥37°C before tourniquet occlusion and administering 60 mg papaverine into the ILI circuit. Melphalan, based on limb volume and dose corrected for ideal body weight, and Dactinomycin were then circulated for 30 min. Washout with normal saline occurred before tourniquet release. Median length of stay (LOS) was 6 days (range 3–15). Serologic toxicity, measured by serum creatine kinase (CK) levels, peaked at a median of 535 IU/l on post-operative day 4. Mild to moderate (grade I–II) and severe (grade III–V) regional toxicity occurred in 75% and 25%, respectively. Shorter LOS, lower CK levels, earlier CK peak, and less toxicity, were seen in UE ILIs compared to LE. Three-month follow-up was available for 76 patients (range 3–193 months). Overall response rate at 3 months was 62% (36% CR). Seventeen (27%) patients developed disease progression in the extremity, at a median 6 months (3–16) and 24 (33%) progressed outside the extremity at a median 4.1 months (3–16). Eleven (14%) patients died during follow-up; median overall survival was 35.6 months.

ILI, ideally performed via a standardized approach, is an effective method for controlling in-transit (stage IIIb/c) melanoma metastases. Patients undergoing UE ILIs have less toxicity and shorter LOS than LE. ILI offers an 86% 3-year overall survival with a third having complete response.


Prognostic significance of host immune factors in melanoma

H. Wu, T.-Y. Li, X.-Z. Xu

Departments of Pathology and Biomedical Statistics, Fox Chase Cancer Center, Philadelphia, PA, USA

Patient outcome is not only determined by the intrinsic biologic characteristics of a tumor, but also by the function of the host immune system and the patient's overall physical condition. In the age of personalized medicine, it is important to identify easily available clinical data to further substratify patients into various risk groups. We retrospectively analyzed the prognostic significance of pre-operative peripheral blood counts in melanoma patients (121 invasive and 107 metastatic). In patients with metastasis, absolute lymphocyte count (ALC) and hemoglobin (Hb) level are significantly associated with disease-specific (CSS) as well as overall survival (OS), using both univariate (Kaplan-Meir estimation) and multivariate (Cox proportional hazard model) analyses. Specifically, in multivariate analysis the P values for ALC is 0.00004 and 0.00006 for CSS and OS; Hb 0.00015 and 0.0002 for CSS and OS. In patients with invasive melanoma, univariate analysis shows that an ALC > 1.3 × 109 cells/l (the lower limit of normal in our laboratory) and normal Hb level (male >14 g/dl, female >12 g/dl) are significantly associated with better disease-specific survival (P = 0.04 and 0.0015, respectively). In addition, in multivariate analysis, Hb level as a continuous variable remains a significant predictor for disease-specific and overall survival (P = 0.04 and 0.00022, respectively). Our study relates to similar findings in other cancers, i.e. prognostic significance of ALC in lymphomas, myelodysplastic syndromes and advanced rectal cancer, and Hb level in lymphoma, gastric, rectal, cervical, vulvar and metastatic renal cell carcinomas. Our data raise the possibility of constructing a prognostic score containing both tumor- and host-specific factors to estimate patient outcome in melanoma. The data should be further validated in a prospective study that will also analyze the significance of T-cell subsets and NK cells.


Diphencyprone: effective topical immunotherapy for cutaneous metastatic melanomacase

D. L. Damian1,2, J. F. Thompson1,3

1 Melanoma Institute Australia; 2Dermatology, Sydney Cancer Centre; 3Department of Surgical Oncology, University of Sydney at Royal Prince Alfred Hospital, Sydney, NSW, Australia

Extensive locally recurrent or cutaneously metastatic melanoma is often unsuitable for or resistant to surgery, radiotherapy or regional or systemic chemotherapy. We used DPCP in aqueous cream, applied topically once per week, to treat patients with extensive cutaneous metastatic melanoma. For each patient, a concentration of DPCP was chosen which caused a definite but not excessive contact hypersensitivity response at treated sites. Of 32 patients who have thus far completed at least 3 months of treatment, 50% achieved complete clearance of skin metastases, and four patients were also noted to have regression of metastatically involved lymph nodes. One patient displayed complete regression of all metastases in his skin, draining nodes and lungs. Thirty-one percent of patients showed partial response to DPCP, with either regression of some but not all skin lesions, or substantial slowing of the rate of disease progression. Hence 81% of treated patients showed at least partial response to this inexpensive and relatively non-invasive treatment. Topical immunotherapy with DPCP should be considered for extensive cutaneous metastases unsuitable for other therapies.


Red hair and inherited MC1R variants are associated with better prognosis for melanoma patients

J. N. Bishop1, J. Davies1, M. Harland1, D. T. Bishop1, S. Madhusudan2, R. van Doorn3 on behalf of the BioGenoMEL consortium4

1 The University of Leeds, Leeds, UK; 2The University of Nottingham, Nottingham, UK; 3Leiden University Medical Center, NL;

BioGenoMEL is a new consortium of research groups formed with the aim of identifying inherited genes having an effect on outcome thereby giving important biological insights into host/tumor interaction. The aim of this BioGenoMEL study was to test the hypothesis that inherited variants in the gene coding for the melanocortin 1 receptor (MC1R) are associated with better survival in cutaneous melanoma. Our hypothesis was that MC1R variants might modulate MITF transcription factor signaling; affecting tumor cell proliferation, apoptosis and DNA repair, all of which could have an impact on patient survival.

Participant groups carried out MC1R genotyping and the survival and genotyping data were pooled and analyzed in Leeds. The variants were classified as ‘R’ and ‘r’ variants, according to their reported effect on hair color. A survival analysis in the Leeds Melanoma Cohort was carried out adjusting for all factors known to impact on survival. The results were then compared with data from nine additional, smaller cohorts. The data sets were broadly similar in terms of age and sex distribution. Median Breslow thickness varied. The absence of any wild type MC1R alleles was associated with a significantly lower risk of death in the Leeds set (HR 0.65, 95% CI 0.46–0.90) and overall (10 data sets) (HR 0.79, CI 0.66–0.95) with some support from the nine smaller data sets considered together (HR 0.83, CI 0.66–1.04).

Testing small effects of inherited genetic variation on outcome for cancer requires collaboration between multiple research groups to provide sufficient validation data sets but even then individual studies still need to be large. The data are suggestive of a protective effect for red hair and with prognostic significance for inherited MC1R variants. Studies of inherited variation having an effect on outcome have the potential to identify biological pathways crucial to host/tumor interaction and therefore survival.


Cutaneous leiomyosarcoma: a single institution experience with treatment and outcomes

J. L. Deneve, J. L. Messina, D. Salmasinia, S. S. Marzban, M. Bui, D. G. Letson, D. Cheong, R. J. Gonzalez, V. S. Sondak, J. S. Zager

Moffitt Cancer Center, Tampa, FL, USA

Cutaneous leiomyosarcoma (CLMS) is a rare cutaneous soft tissue sarcoma for which treatment algorithms, including appropriate resection margins and adjuvant therapy have been poorly defined. We report our single institution experience treating CLMS with narrow margin resections. An IRB approved retrospective review of patients with CLMS was performed from 2005–2010. Clinicopathologic and outcomes variables were further analyzed.

Thirty-three patients with CLMS were treated over a 5 year period, 76% were male, 97% were Caucasian with a median age of 63.5 years (range, 20–91). The majority of tumors (67%) were located on the extremities. Pre-operative staging included a diagnostic CXR (39%) or CT scan of the thorax (42%), all of which were negative for distant metastasis. Eighty-eight percent of tumors were low grade with a median tumor size of 1.3 cm (range, 0.5–2.8 cm). Mitoses were evident in 76% of tumors, with a median of 10.8 (range 1–58). Eighty-two percent underwent immunohistochemistry analysis, 100% of tumors tested for smooth muscle actin were positive, 96% tested for S100 were negative. All patients underwent primary surgical resection with a median margin of 1 cm (range 1–2 cm). Three patients (9%) required re-excision with 1 cm margins for a microscopic positive margin. Final resection margin was negative in 32/33 (97%). Adjuvant radiotherapy was used in 15%; for a persistent positive margin (N = 1) or surgeon preference (N = 4). There were no recurrences, and at a median follow-up of 15.5 months (range, 2–49 months) 100% were alive with no evidence of disease.

Cutaneous LMS is a low grade malignancy affecting primarily elderly male Caucasians. When treating CLMS, adequate oncologic control and favorable overall outcome are possible with a 1 cm surgical margin of resection. The benefit of preoperative chest imaging and adjuvant XRT to the primary site (in the case of negative margins) is likely minimal.


Primary malignant melanoma arising in a cystic ovarian teratoma

T. Tsakok1, S. Burbridge2, A. Andi3, A. Jacques3, P. Menon4, M. Harries2

1 Academic Foundation Programme, Guy’s and St Thomas’ NHS Foundation Trust, London, UK; 2Department of Oncology, Guy’s and St Thomas’ NHS Foundation Trust, London, UK; 3Department of Radiology, Guy’s and St Thomas’ NHS Foundation Trust, London, UK; 4Department of Histopathology, Guy’s and St Thomas’ NHS Foundation Trust, London, UK

A 58 year-old lady presented with abdominal pain and swelling. Physical examination yielded a distended abdomen with a palpable mass. CT scan showed a mixed-attenuation pelvic lesion with subcapsular liver deposits. MRI showed characteristics of a mature teratoma, with a large soft tissue component suspicious for malignant transformation. The liver lesions were reported as containing fat – a highly unusual finding. The patient underwent left salpingo-oophorectomy and the mass was removed. Its cystic component contained necrotic material and hair. Its solid fragment exhibited a brown/black cut surface. Histology showed a malignant neoplasm with immunohistochemistry positive for S100 protein/HMB 45. The diagnosis was malignant melanoma arising within a mature cystic teratoma. Post-operatively, PET-CT demonstrated small volume intra-abdominal disease; this was stable on PET-CT at 2 months. After a further 4 months, the patient described worsening abdominal symptoms, and PET-CT revealed significant disease progression. The patient declined systemic treatment, and PET-CT 3 months later confirmed further progression. She was commenced on Dacarbazine chemotherapy, but this was discontinued due to deteriorating performance status. The patient died soon after.

Ovarian teratomas are relatively common in women of reproductive age, but only 0.2–0.4% are thought to undergo malignant transformation. Although the teratoid elements of such tumours harbour a capacity to differentiate into many malignancies, melanoma is the rarest, with only 37 individual cases reported to date. Diagnosis can be hampered by clinical and pathological challenges, including the need to exclude cutaneous or mucosal primaries, as well as morphological diversity of the lesion itself. In this case the T1-intense signal caused by melanin on MRI was initially thought to represent fat, and eventual diagnosis was only made on histopathology. We wish to highlight the utility of sequential PET-CT imaging, and suggest that if malignant transformation of a teratoma is suspected, it should be considered as a component of pre-operative staging.


Electrochemotherapy for treatment of locally advanced superficial cancer: results from a single Institution

E. Pennacchioli1, C. Margarino1, S. Gandini2, F. Verrecchia1, G. Tosti1, G. Spadola1, F. Baldini1, M. Mosconi1, P. Ferrucci1, A. Testori1

1 Melanoma and Sarcoma Unit, European Institute of Oncology, Milan, Italy; 2Epidemiology Division, European Institute of Oncology, Milan, Italy

We analyse the outcome of electrochemotherapy in superficial spreading of different sutaneous and non-cutaneous diseases.

Between May 2006 and August 2011, 179 ECT treatments in 108 evaluable patients. Treated primary diseases (62 pts), are SCC 40% (25/62), BCC 37% (23/62) and melanoma 8% (5/62). Sixty-five percent of patients had at least two treatments. In patients treated for metastatic disease 76% (89/117) are melanoma. The treatment was performed using intravenous Bleomycin in 159 cases (89%), local Bleomycin in 20 cases (11%). For post-operative monitoring of response a recorded clinical evaluation was done in all patient at day 30, 60, 90, 180, 360.

Results at 1 month are: CR 12% of cases, PR 84% of cases; NR 4% of cases; at 3 months CR 14%; PR 47% PRO 26% and 5% DOD; 1 yr after treatment 18% CR; 25% PR 40% PRO and 10% DOD. The mean length of response in melanoma was 3 months (mean 2–8 months). Concerning the histotype, CR was obtained in 57% cases of Kaposi sarcoma, 50% of BCC and 9% of melanoma. PR was obtained in melanoma in 89% of cases at 1 month, 35% at 6 months and 25% at 1 yr. The treatment was never used to obtain surgical operability. Local toxicity consist in mild pain, self-retaining serum effusion for at least couple of weeks and rare ulceration of treated nodules. A possible systemic effect of this treatment is being studied in our Institution, as an immunological effect has been hypotesised.

ECT is a safe procedure. In relation to the histotype it should be considered a palliative local treatment (melanoma, breast cancer) or a definitive curative treatment (Kaposi sarcoma and BCC). The length of response should be used in order to associate a systemic treatment, possibly an immunotherapy.


Role of sentinel node biopsy in thin and thick melanoma

F. Verrecchia1, E. Pennacchioli1, S. Gandini2, G. Tosti1, G. Spadola1, F. Baldini1, M. Mosconi1, P. Ferrucci1, A. Testori1

1 Melanoma and Sarcoma Unit, European Institute of Oncology, Milan, Italy; 2Epidemiology Division, European Institute of Oncology, Milan, Italy

Sentinel node biopsy (SNB) procedure has been accredited as standard of care in patients with melanoma but its role is still under discussion, in particular for thin (≤1 mm) and thick (≥4 mm) melanomas. We retrospectively analyzed data from the European Institute of Oncology (Milan) and Istituto Nazionale dei Tumori (Naples) considering patients with thin (490 cases) or thick (298 cases) melanoma undergone SNB. In thin melanoma subgroup, with a median thickness 0.80 mm, SNB positive rate is 5%. In the Thick melanoma group, with a mean thickness 6.6 mm, positive SNB accounted for 39.3%.

As shown in the analysis of our data, the SNB technique is useful to provide reliable diagnostic and prognostic information also for melanoma patients with Breslow thickness ≤1 mm. The association between OS and SNB positivity with age, gender, Breslow thickness increasing, ulceration has been analyzed too, but none of them is statistically significant.

SNB should be the standard method for melanoma patients ≥4 mm, not only for staging, but also for guiding therapeutic decisions, either for adjuvant therapy decisions or to confirm the indication for a completion lymph node dissection.


Use of tissue marking ink to intraoperatively identify the ‘hot spot’ in sentinel lymph nodes for melanoma

J. K. John, J. M. Pimiento, C. A. Puleo, R. A. Mathew, T. M. McCardle, A. A. Sarnaik, J. S. Zager, V. K. Sondak, J. L. Messina

Moffitt Cancer Center, Tampa, FL, USA

Focused examination of the sentinel lymph node (SLN) may maximize detection of metastatic disease and accuracy of staging. Surgeons frequently observe localization of radioactivity and/or lymphazurin dye to circumscribed regions of SLNs (‘hot spot’), raising the possibility of metastatic disease colocalizing with radiotracer or dye. We determined the feasibility of marking the hot spot with permanent ink to allow the pathologist to identify this area, and evaluated whether metastatic disease preferentially localizes there. Retrospective, single-institution evaluation of operative & pathology reports from patients with primary melanoma >0.76 mm in depth undergoing SLNB in 6-month period. If present, the ‘hot spot’ was marked with gentian violet (Sandel 4 in 1TM Marker, Sandel Medical Industry, Chatsworth, CA) by the surgeon immediately after excision, and re-inked with permanent tissue marking dye (Triangle Biomedical Sciences, Inc.) during gross dissection. SLN were submitted entirely for pathologic evaluation, stained with hematoxylin and eosin, S-100 and Melan-A. Additional levels were cut as necessary until the hot spot was visualized microscopically. Metastatic melanoma was diagnosed if atypical cells positive for S-100 and/or Melan-A were seen. Two hundred and one patients underwent SLNB. Of these, 104 patients had 204 SLN with identified hot spots. Median age = 63 years. Median depth = 1.55 mm. Eighteen SLN from 16 patients (15%) demonstrated metastasis. In 10 SLN the metastasis was beneath the hot spot, and in 8 SLN it was unrelated to the hot spot.

Intraoperative inking of the hot spot is feasible and reproducible. Permanent tissue marking ink can be easily visualized during microscopic examination of SLNs. In this small study, the hot spot does not appear to preferentially harbor metastatic disease. We are expanding this study to include larger numbers of patients and utilize an intraoperative permanent inking kit to evaluate whether it adds sensitivity to evaluation of the SLN.


Whole brain radiotherapy trial and bio-specimen banking

V. Jakrot1, E. Paton1,2,3, B. Shivalingam1,4, R. Scolyer1,4, G. Mann1,2,5, G. Fogarty1, A. Hong1,4, J. Thompson1,2,3,4

1 Melanoma Institute Australia, North Sydney, NSW, Australia; 2University of Sydney, Sydney, NSW, Australia; 3Australia and New Zealand Melanoma Trials Group, Sydney, NSW, Australia; 4Royal Prince Alfred Hospital, Sydney, NSW, Australia; 5Westmead Millennium Institute, Sydney, NSW, Australia

The ANZMTG 01.07 Whole Brain Radiotherapy in Melanoma Trial seeks to evaluate the role of whole brain radiotherapy (WBRT) following local treatment of intracranial melanoma metastases. Previous WBRT trials in other diseases have been slow to accrue. The weekly Melanoma Institute Australia (MIA) Multi-disciplinary Team (MDT) provides a regular forum for the review of complex cases, alert referrals, discuss clinical trials; facilitating excellent working relationships between clinicians. The team consists of surgical, medical and radiation oncologists; neurosurgeons, dermatologists, pathologists, care-coordinators, and representatives from clinical trials, research and bio-specimen bank.

The WBRT trial is open at 15 hospitals world-wide. The primary endpoint is distant intracranial failure. Secondary endpoints include the effect of WBRT on neurocognitive function and overall survival. Bio-specimen banking is crucial to let us see whether a molecular signature within the resected metastases will show which patients might benefit the most from WBRT.

Fifty-eight patients have been accrued to date including 20 patients from MIA, with the highest recruitment rate of any WBRT trial. As interest in the trial grew, we observed a corresponding increase in the number of brain metastases banked in the Melanoma Bio-Specimen Bank, with exponential growth since 2007. A total of 35 metastases have been banked to date and more than 10 in 2011 alone. We noted the numbers of tumours banked and patients entered on trial are essentially identical. The activity of the trial appears to be promoting and enhancing identification of cases of melanoma brain metastases that are suitable for tissue banking.

The MIA Bio-specimen Bank has been an important link between the treating Neurosurgeon and Radiation Oncologist responsible for coordination of the WBRT Mel Trial. The availability of previously scarce brain metastases from melanoma patients will enable translational research into many aspects of melanoma, not only the correlates of response to WBRT.

  • Numbers in parentheses refer to abstract number of proffered submissions accepted for oral presentation

  • Available faculty abstracts provided in order of presentation at the conference. Any faculty abstracts not available by the time of publication will be provided to participants at the Congress.

  • *,†

    These authors contributed equally to this work.