SEARCH

SEARCH BY CITATION

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Sequencing of dermatological tumor genomes: Melanoma
  5. Squamous cell carcinoma and basal cell carcinoma
  6. Outlook
  7. References

Activated intracellular signaling pathways based on mutations in oncogenes and tumor suppressor genes play an important role in a variety of malignant tumors. In dermatology, such mutations have been identified in melanoma, basal cell carcinoma and squamous cell carcinoma. These have partly led to the establishment of new, targeted therapies. Treatment successes have been particularly impressive for melanoma with small molecule inhibitors directed against the mutated BRAF oncogene and in basal cell carcinoma with inhibitors directed against the hedgehog signaling pathway. New sequencing technologies, in particular next generation sequencing, have led to a better and more comprehensive understanding of malignant tumors. This approach confirmed the pathogenic role of BRAF, NRAS and MAP kinase pathways for melanoma. At the same time, a series of further interesting target molecules with oncogenic mutations such as ERBB4, GRIN2A, GRM3, PREX2, RAC1 and TP53 were identified. New aspects have recently been shown for squamous cell carcinoma by detection of mutations in the NOTCH signaling pathway. A better understanding of the pathogenesis of these and other tumors should lead to improved and maybe even individualized treatment. The current developments in dermatological oncogenetics based on the new sequencing technologies are reviewed.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Sequencing of dermatological tumor genomes: Melanoma
  5. Squamous cell carcinoma and basal cell carcinoma
  6. Outlook
  7. References

Insights into the genetic causes of melanoma and basal cell carcinoma have led to new therapeutic approaches [1, 2]. In melanoma, mutations in the BRAF, NRAS and KIT genes are most relevant; in basal cell carcinoma, mutations in the hedgehog signaling pathway. In squamous cell carcinoma one frequently finds mutations in the TP53 tumor suppressor gene and in the RAS oncogenes, which have to date remained without therapeutic implications [3]. Recently, the knowledge about mutations in malignant skin tumors has increased exponentially through the use of novel sequencing methods [4]. The classical Sanger sequencing method for DNA was first improved through automation and then in 2005 through the introduction of next or second generation sequencing (NGS) [5]. The currently available technology platforms allow through what is termed massively parallel sequencing the simultaneous sequencing of millions of sequence fragments of a genome. It is essential that a certain site of the genome be sequenced as deeply as possible, i.e. often parallel, in order to keep the error rate low (deep sequencing). A promising new technology was recently introduced by Oxford Nanopore Technologies. Here, DNA is sequenced directly on the basis of an electrical voltage alteration that develops when DNA is drawn through a small pore [6]. This type of sequencing is termed third generation sequencing, but has of yet not been employed in larger tumor collectives. Together, these technologies form the basis for the wealth of new knowledge in dermatological oncology.

Sequencing of dermatological tumor genomes: Melanoma

  1. Top of page
  2. Summary
  3. Introduction
  4. Sequencing of dermatological tumor genomes: Melanoma
  5. Squamous cell carcinoma and basal cell carcinoma
  6. Outlook
  7. References

Whole-genome sequencing

In the first study of the total melanoma genome of a single melanoma cell line with NGS, SPDEF, a member of the ETS transcription family, and the matrix metalloproteinase MMP28 were identified as new mutated candidate genes [7]. The majority of the mutations found were C>T/G>A- or CC>TT/GG>AA transitions, which indicates a causal role of UV light. This was also confirmed in further studies. In a recent study 25 melanoma metastases were examined using NGS [8]. Besides mutations in the BRAF and NRAS gene, the PREX2 gene was most frequently mutated with 44 % (Figure 1, Table 1). In the end effect a total of 14 % of all cases (including more than 100 cases in the prevalence screen) showed PREX2 mutations at various sites. PREX2 (phosphatidylinositol-3,4,5-trisphosphate-dependent Rac exchange factor 2) interacts, among others, with the tumor suppressor PTEN. PREX2 mutations increased the growth of an NRAS-transformed melanoma cell line after transplantation to immunosuppressed mice [8].

Table 1. Current studies for the identification of genetic variants in skin tumours using automated Sanger sequencing/PCR sequencing and next generation sequencing (NGS)
No.TissueDNA/RNATechniquesGenes with variationsRef.
  • *)

    Abbreviations: Meta, metastases; PM, primary melanomas; PT, primary tumour;

  • **)

    Percentages in brackets represent the mutation rate of genes in the analysed samples sets. No percentages are given when only one sample was analysed or only different samples from one patients were analysed (see ref. nos. [7] and [19]).

 Malignant melanoma Whole-genome sequencing
1.Cell line COLO-829DNANGSBRAF, SPDEF, MMP28, RARB, TOPB2, PTEN[7]
 Meta*) (n = 25)DNANGSBRAF (48 %) **), NRAS (36 %), PREX2 (44 %), MUC4 (76 %), PRG4 (32 %), MST1 (32 %), KIT (4 %)[8]
 PM*), Meta (n = 2)DNANGSSCAF1, WNT1, ASB9, FAT2, PTRF, RHOB, CNDP2, DROSHA, ERCC5, LRRK1, LRRFIP1[9]
2.Transcriptome sequencing
 Short-term cultures and cell lines, inter alia MeWo, 501 Mel (n = 10)RNANGSRB1-ITM2B (10 %), RECK-ALX3 (10 %), A2M (10 %), CAST (10 %), CENTD3 (10 %), FUS (20 %)[10]
3.Whole-exome sequencing
 Meta, cell lines, inter alia 501 Mel (n = 14)DNANGSBRAF (50 %), GRIN2A (42.9 %), TRRAP (14.3 %), DCC (14.3 %), ZNF831 (28.6 %), PLCB4 (28.6 %)[11]
 Cell lines, inter alia LAU149, LAU165 (n = 7)DNANGSBRAF (85.7 %), NRAS (14.2 %), MAP2K1/2 (28.5 %), PTEN (14.2 %), FAT4 (57.1 %), DSC1 (57.1 %), LRP1B (57.1 %)[12]
 Meta, PM, short-term cultures (n = 121)DNANGSBRAF (63 %), NRAS (26 %), PPP6C (9 %), RAC1 (5 %), SNX31 (7 %), TP53 (19 %), TACC1 (7 %), STK19 (4 %), PTEN (12 %), ARID2 (9 %), KIT (3 %)[13]
 Meta, PM, short-term cultures (n = 61)DNANGSBRAF (45.9 %), NRAS (21.3 %), RAC1 (9.8 %), PPP6C (13.1 %), DCC (34.4 %), PTPRK (19.6 %), GRM3 (19.6 %), TNC (18 %), TP53 (14.7 %), KIT (3.4 %)[14]
 Meta (n = 20)DNANGSBRAF amplification (20 %)[15]
4.Targeted exome sequencing (MMPs, receptor tyrosine kinases, G protein-coupled receptors)
 Meta (n = 79)DNAPCRMMP8 (6.3 %), MMP24 (3.8 %), MMP27 (7.6 %), MMP28 (2.5 %)[16]
 Meta (n = 79)DNAPCRERBB4 (19 %), FLT1 (10.1 %), EPHA10 (6.3 %), PDGFRA (5.1 %), PTK2B (10.1 %)[17]
 Meta (n = 80)DNAPCRGRM3 (16.3 %), GPR98 (27.5 %), CHRM3 (10 %), GRM8 (8.8 %), LPHN2 (10.1 %)[18]
5.Combined techniques (Whole-genome/ exome/ transcriptome sequencing)
 Meta (n = 3)DNA, RNANGSHRAS, ELK1, CDKN2C[19]
 PM (n = 77), Meta (n = 53), cell lines (n = 168)DNANGSTERT (33 % in PM; 85 % in Meta; 74 % in cell lines)[20]
6.Squamous cell carcinoma Whole-exome sequencing    
 PT*) (n = 11)DNANGSNOTCH 1 (81.8 %), Notch 2 (45.4 %), TP53 (87.5 %), CDKN2A (25 %)[21, 22]
7.Basal cell carcinoma Whole-exome sequencing
 PT (n = 8)DNANGSNOTCH3 (13 %), MAML1 (13 %)[22]
image

Figure 1. Simplified schematic representation of signaling pathways with activating mutations in malignant melanoma. In current studies using next generation sequencing or automated Sanger sequencing new mutated genes besides BRAF and NRAS were identified, which encode for proteins of intracellular signaling pathways. The figure shows a simplified representation of these pathways. The protein products of often mutated genes (including BRAF and NRAS) are indicated by asterisks. Explanations of abbreviations not given in the main text: FN1, fibronectin 1; ITGA, integrin alpha; NRG1–4, neuregulin 1–4; PAK, p21-activated kinase 1; CaM, calmodulin; PKB, protein kinase B; Erk1/2, extracellular signal-regulated kinase 1/2; JNK, c-Jun N-terminal kinase; VCL, vinculin.

Download figure to PowerPoint

In the comparison of a primary acral melanoma and a simultaneously existing lymph node metastasis, 39 of the identified non-synonymous mutations overlapped [9]. In the primary melanoma a mutation in the SCAF1 gene and in the metastasis a mutation in the WNT1 and SUPT5H gene were found in addition (SUPT5H is a regulator of transcriptional elongation). In both, mutations in PREX2, RHOB and DROSHA were found.

RNA sequencing

The only paper to date that sequenced the transcriptome (total coding RNA) studied 8 short-term cultures of melanoma metastases and 2 melanoma cell lines [10]. Here a series of new gene fusions was identified, among others, two fusions with involvement of the tumor suppressor gene RB1 and of RECK. RECK is a well-known inhibitor of tumor invasion and metastasis. None of the gene fusions were present in more than one cell line and also were not found in 90 further cell lines and short-term cultures.

Whole-exome sequencing

In the first sequencing of the whole exome of 14 metastatic melanomas, GRIN2A demonstrated a mutation rate of 43 %, after validation in a total of 135 samples a mutation rate of 25.2 % [11]. GRIN2A codes for a receptor subunit of an ionotropic glutamate receptor (Figure 1). Based on this knowledge and the further mutation described below, in the metabotropic glutamate receptor GRM3, glutamate receptor antagonists might be of therapeutic interest in melanoma. The glutamate receptor antagonist riluzole is already being employed in ongoing clinical studies. In a further study of 7 melanoma cell lines one of the cell lines originated from a lymph node metastasis appearing 3 years after the primary tumor and one from a distal metastasis appearing after 15 years [12]. Both overlapped in 70 % of the mutations, which according to the opinion of the authors speaks for an early development of mutations in melanoma progression. With inclusion of an additional 127 melanoma samples, in 8 % mutations were found in the MAP2K1 (MEK1) or the MAP2K2 (MEK2) gene. Interestingly, cells with MAP2K2 E207K mutations were more resistant towards the MEK1/2 inhibitor AZD6244 (selumetinib) than cells with a BRAF V600E mutation.

In a recently published sequencing study of 121 melanoma samples attention was focused on regions with an accumulation of mutations in exons without involvement of neighboring introns in order to better identify what is termed driver mutations [13]. Three of the newly identified genes demonstrated recurrent mutations at the same site (RAC1, PPP6C and STK19). The RAC P29S mutation located within a functionally important region was found in 5 % of the cases. RAC1 is a member of the family of RHO-GTPases and is involved in cytoskeleton organization and cell migration (Figure 1).

In a further study with 147 samples (61 samples with detailed sequencing information) RAC1 P29S was the most common recurrent mutation after BRAF and NRAS mutations with 4.7 % [14]. GRM3, PPP6C and PTPRK belonged to the most frequently mutated genes. Mutated RAC1 increased the proliferation and migration of melanoma cells as well as ERK phosphorylation.

In a study of pairs of metastases before treatment with the BRAF inhibitor vemurafenib and after progression under BRAF inhibitor, gene amplifications of the mutated BRAF gene (BRAF V600E) were found in 4 of 20 recurrences [15]. In vitro the resistance towards the BRAF inhibitor could in part be overcome by a combination of BRAF and MEK1/2 inhibitor.

Targeted exome sequencing

Focused investigations using automated PCR sequencing were focused on the exomes of matrix metalloproteinases (MMP), receptor tyrosine kinases and G protein-coupled receptors [16-18]. In the study of 32 metastatic melanomas with respect to MMPs including a prevalence screen of further 47 tumors, MMP mutations were found in 23 % of all melanoma samples [16]. MMP8 and MMP27 were mutated most frequently. Mutated MMP8 promoted lung metastasis in a mouse model.

Of the receptor tyrosine kinases, 86 kinases in 29 metastases were examined. ERBB4 was mutated in 19 % of all 79 samples (including the prevalence screen) (Figure 1). Beyond this, mutations in FLT1 and PDGFRA were found. Mutated ERBB4 had a similarly strong transforming property on NIH-3T3 cells as oncogenic KRAS. The Inhibition of mutated ERBB4 protein with lapatinib (GW2016) reduced the growth of ERBB4-mutated melanoma cells, but not of wild-type cells [17].

In the examination of G protein-coupled receptors, GRM3 and GPR98 were mutated most frequently, GRM3 in 16.3 % and GPR98 in 27.5 % [18]. GRM3 demonstrated a mutation hotspot in 4 of 57 melanoma samples in the prevalence screen. Mutated GRM3 increased migration and the metastatic behavior of melanoma cells. An activation of GRM3 led also to MEK1/2 activation. Correspondingly, the MEK1/2 inhibitor selumetinib inhibited the growth of melanoma cells with mutated GRM3.

Combined techniques

In a study with 4 cutaneous melanoma metastases using a combination of different sequencing methods (exome, transcriptome and whole genome sequencing) an activating HRAS mutation (HRAS Q61L) and a mutation in ELK1 (ELK1 R74C), a member of the ETS transcription factor family was found [19]. As RAS signaling activates both the MAPK as well as the PI3K-mTOR signaling pathway, the patient was viewed as a suitable candidate for a clinical study with PI3K and MEK1/2 inhibitors. In a recently published paper on a genetic association study in combination with an NGS analysis in a family with familial melanoma, a functional mutation in the promoter (Ets binding site) of the TERT gene was identified. In studies on spontaneous melanomas and melanoma metastases the TERT promoter was mutated in 33 % and 85 %, respectively [20].

Squamous cell carcinoma and basal cell carcinoma

  1. Top of page
  2. Summary
  3. Introduction
  4. Sequencing of dermatological tumor genomes: Melanoma
  5. Squamous cell carcinoma and basal cell carcinoma
  6. Outlook
  7. References

Two recent NGS studies address cutaneous squamous cell carcinoma [21, 22]. In the first study, 8 squamous cell carcinomas were examined; afterwards in the second study 3 further tumors were included. Remarkable was the high mutation burden of 1,300 mutations per tumor with 85 % of mutations as C>T transitions, which again indicates the causal role of UV radiation. Newly detected were mutations in NOTCH genes. Including a further tumor and 14 squamous cell carcinoma cell lines, the mutation prevalence for NOTCH1 or NOTCH2 was 75 %. In the first study in 7 of 8 tumors mutations in the TP53 gene, further in two of 8 cases, mutations in the CDKN2A gene (codes for p16 tumor suppressor) and in one of 8 cases in the HRAS gene were found. NOTCH receptors primarily regulate cell differentiation in the hematopoietic system and in neurogenesis. They are often mutated in T- and B-cell leukemias. After receptor binding and modification by the enzyme secretase NOTCH1 translocates into the nucleus and activates nuclear factor-kappa B (NF-κB), VEGF, MMP-9, ERK, Akt, c-Myc, p21, p27 and p53 here [23]. In the second study referred to, it could be shown for individual mutations of NOTCH1 that they abolish the transcriptional activity of NOTCH1 [22]. The exact role of NOTCH1/2 in tumor development is, however, still unknown. It is known from mouse models that NOTCH1 functions as a tumor suppressor in the skin.

In a further paper there was a search for viral sequences in cutaneous squamous carcinomas using RNA sequencing. No viral sequences with oncogenic potential or related sequences could be found, which does not, however, exclude the involvement of as of yet unknown viruses [24].

Up to now there are no NGS papers that have examined basal cell carcinoma in a targeted fashion. In older studies on 42 basal cell carcinomas using Sanger sequencing with respect to mutations in classical oncogenes PATCHED mutations were found in 67 % and TP53 mutations in 40 % of the cases. No mutations were found in GLI1, NRAS, KRAS, HRAS, BRAF or CTNNB1. In the paper mentioned by Wang et al. 5 basal cell carcinomas were included as controls and examined using exome sequencing [22]. One basal cell carcinoma demonstrated a NOTCH3 mutation and one mutation in the MAML1 gene, that codes for a NOTCH interaction partner. Further mutations were not reported.

Outlook

  1. Top of page
  2. Summary
  3. Introduction
  4. Sequencing of dermatological tumor genomes: Melanoma
  5. Squamous cell carcinoma and basal cell carcinoma
  6. Outlook
  7. References

A significant further development in the field of NGS are the recently presented bench top sequencers for a targeted mutation search. An important advantage of this technique is that the results are available within one day. The rapid and uncomplicated analysis of certain mutations or mutation combinations could in the future facilitate treatment decisions [25]. Based on currently available data, in addition to BRAF, NRAS and MEK1/2, particularly PREX2, ERBB4, GRIN2A/GRM3 and RAC1 are promising candidates for such analyses and mono- or combination therapies based on them.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Sequencing of dermatological tumor genomes: Melanoma
  5. Squamous cell carcinoma and basal cell carcinoma
  6. Outlook
  7. References