Phase II trial of panobinostat, an oral pan-deacetylase inhibitor in patients with primary myelofibrosis, post–essential thrombocythaemia, and post–polycythaemia vera myelofibrosis


Correspondence: Dr Daniel J. DeAngelo, Dana-Farber Cancer Institute, 450 Brookline Avenue, Dana D1B30, Boston, MA 02215, USA.



Dr Kapil N. Bhalla, Cockrell Center for Advanced Therapeutics The Methodist Hospital Research Institute, 6670 Bertner Ave., R9-113, Houston, TX 77030, USA.



Myelofibrosis (MF) is a Philadelphia chromosome–negative stem cell myeloproliferative neoplasm (MPN) associated with cytopenias, splenomegaly, constitutional symptoms, and poor prognosis. MF patients commonly express JAK2 V617F mutation and activation of Janus kinase (JAK)/signal transducer and activator of transcription (STAT) signalling. Agents targeting the JAK/STAT pathway have demonstrated efficacy in patients with MF. This study evaluated panobinostat, a pan-deacetylase inhibitor that depletes JAK2 V617F levels and JAK/STAT signalling in MPN cells, in patients with primary MF, post–essential thrombocythaemia MF, and post–polycythaemia vera MF. Patients received panobinostat 40 mg administered three times per week. Dose reductions were permitted for toxicities. The primary endpoint was response rate at 6 months using International Working Group for Myelofibrosis Research and Treatment (IWG-MRT) consensus criteria. Analyses of peripheral blood cells from treated patients revealed that panobinostat inhibited JAK/STAT signalling, decreased inflammatory cytokine levels, and decreased JAK2 V617F allelic burden. However, panobinostat was poorly tolerated at the dose and schedule evaluated, and only 16 of 35 patients completed ≥2 cycles of treatment. One patient (3%) achieved an IWG-MRT response. Common adverse events were thrombocytopenia (71·4%) and diarrhoea (80·0%). Although molecular correlative analyses suggested that panobinostat inhibits key intracellular targets, limited clinical activity was observed because of poor tolerance.

Myelofibrosis (MF) is a clonal stem cell, Philadelphia chromosome–negative, myeloproliferative neoplasm (MPN) (Barbui et al, 2011). Clinically, MF is characterized by splenomegaly, anaemia, bone marrow fibrosis, and extramedullary haematopoiesis (Tefferi & Vardiman, 2008). Patients may present with primary MF (PMF) or progress from other myeloproliferative diseases including polycythaemia vera (post-PV MF) or essential thrombocythaemia (post-ET MF) (Tefferi & Vardiman, 2008; Barbui et al, 2011; Tefferi, 2011). Patients with MF have a poor prognosis with a median survival of 5·8 years (69 months) (Tefferi & Vardiman, 2008; Barbui et al, 2011). Patients with an increased risk, as determined by the presence of one or more poor prognostic factors have a shortened overall survival (Cervantes et al, 2009). Patients classified in the intermediate-2–risk group (two poor prognostic factors) demonstrate a median survival of 48 months, whereas high-risk patients (three or more poor prognostic factors) have a median survival of only 21 months.

Myelofibrosis is associated with dysregulated cytokine and chemokine production, which impacts cell proliferation and maturation, leading to bone marrow fibrosis (Tefferi & Vardiman, 2008; Tefferi, 2011). Up-regulation of several cytokines, including interleukin 1 (IL1) receptor agonist, macrophage inflammatory protein 1β (MIP1β), tumour necrosis factor α (TNFα), and IL6, has been observed in patients with MF treated in clinical trials (Quintas-Cardama et al, 2011). The intracellular tyrosine kinases Janus kinase 1 (JAK1) and JAK2 mediate cytokine signalling and contribute to MF cell survival and growth (Tefferi & Vardiman, 2008; Tefferi, 2011). These observations, together with the discovery that approximately 50% of patients with MF possess a JAK2 V617F somatic mutation, have led to the investigation of agents that target the JAK/signal transducer and activator of transcription (STAT) pathway (Quintas-Cardama et al, 2011). Evaluation of these agents is an area of ongoing and intense research, and the JAK1/2 inhibitor ruxolitinib was recently approved by the US Food and Drug Administration for the treatment of MF (Verstovsek et al, 2010).

The JAK1/2 inhibitors ruxolitinib and CYT387 and the selective JAK2 inhibitor TG101348 have demonstrated encouraging clinical activity, including reduction of splenomegaly and improvement of disease-related symptoms in patients with MF (Verstovsek et al, 2010; Pardanani et al, 2011; Quintas-Cardama et al, 2011). Of note, responses with ruxolitinib were observed in patients with the JAK2 V617F mutation and wild-type disease (Verstovsek et al, 2010). Other available therapies for patients with MF are palliative and include splenectomy, splenic irradiation, steroids, and erythropoiesis-stimulating agents (Tefferi & Vardiman, 2008; Tefferi, 2011). Allogeneic stem cell transplantation remains the only curative option available to a relative minority of patients (Kroger & Mesa, 2008).

Several intracellular signalling pathways that contribute to MPN cell growth and survival, as well as enhanced cytokine production, have been implicated in MF. As a result, novel agents that target these pathways continue to be investigated. Deacetylase inhibitors (DACi) are a class of novel agents that lead to increased acetylation of intracellular proteins implicated in oncogenic pathways (Xu et al, 2007; Lane & Chabner, 2009). Intracellular acetylation targets include histones, transcription factors (TP53, hypoxia-inducible factor 1α, HF1A), α-tubulin (TUBA), and the protein chaperone heat shock protein 90 (HSP90; HSP90AA1) (Xu et al, 2007; Lane & Chabner, 2009). Preclinical studies have shown that treatment with a DACi induces cancer cell cycle arrest, differentiation, and apoptosis (Xu et al, 2007; Lane & Chabner, 2009).

Panobinostat is a potent pan-DACi with nanomolar inhibitor activity against class I, II, and IV DACs (Atadja, 2009). Panobinostat represents a novel approach to targeting JAK signalling in patients with MF. In vitro studies with MPN cells demonstrated that panobinostat depleted the mutant JAK2 V617F protein and inhibited JAK-mediated intracellular signalling (Wang et al, 2009). The down-regulation of the JAK2 V617F protein was shown to occur through inhibition of the association with HSP90, leading to proteasomal degradation of the JAK2 V617F protein (Wang et al, 2009). In a phase I study of panobinostat in patients with haematological malignancies, four of 13 patients with MF achieved clinical improvement, including reduction of spleen size (57–86%) and improvement in disease-related symptoms. Patients were treated at doses ranging from 30 to 60 mg, administered orally three times per week (DeAngelo et al, 2013). Three of the four patients had confirmed JAK2-mutated disease. Of note, three patients demonstrating clinical improvement received therapy long term (>10 months).

Based on intriguing preclinical and preliminary clinical results, this phase II study of panobinostat in patients with PMF, post-PV MF, or post-ET MF was initiated. This study sought to evaluate the safety and efficacy of panobinostat 40 mg administered three times per week. In addition, biomarker data were collected to determine the effect of panobinostat on mRNA, protein, and cytokine expression, as well as JAK2 V617F allelic burden.

Design and methods


Adult patients with intermediate-2–risk and high-risk PMF, post-PV MF, or post-ET MF with symptomatic splenomegaly, splenectomised due to splenomegaly, or anaemia [haemoglobin (Hb) <100 g/l or requiring red blood cell transfusions] with an Eastern Cooperative Oncology Group (ECOG) performance status ≤2 were eligible for the trial. Presence of a JAK2 V617F mutation was not required for eligibility. Exclusion criteria included prior DACi or valproic acid for the treatment of cancer. Patients who required valproic acid treatment for any indication or prior radiotherapy to ≥30% of their bone marrow or had impaired cardiac function or clinically significant cardiac disease were not eligible. The study was designed and conducted in accordance with the International Conference on Harmonisation, Harmonised Tripartite Guidelines for Good Clinical Practice, with applicable local regulations and the ethical principles of the Declaration of Helsinki. The protocol was reviewed and approved by the Institutional Review Board/Independent Ethics Committee/Research Ethics Board at each study site. All patients were required to sign an informed consent.

Study design and treatment

This study was a multicentre, open-label, Simon 2-stage phase II clinical trial of single-agent panobinostat administered orally in patients with MF. Panobinostat was administered once daily three times weekly (Monday, Wednesday, and Friday) at a dose of 40 mg on 28-day cycles. Evaluations were performed weekly for the first 4 weeks, once every 2 weeks for the next two cycles, and then once per cycle thereafter. Patients continued on study until either disease progression or unacceptable toxicity, and dose reductions were permitted for toxicities. The first dose reduction was to an every-other-week schedule at the same dose followed by a reduction to 30 mg on the every-other-week schedule. The study was originally designed to enrol two cohorts, one with the JAK2 V617F mutation and another with JAK2 wild-type patients, with 20 patients to be enrolled into each. Because of the disproportionately high enrolment of patients with the JAK2 mutation, the study was amended to stop enrolment in stage 1 after 20 patients with the JAK2 mutation regardless of the number of wild-type patients. According to the Simon 2-stage design, at least two responses in 20 JAK2-mutant patients were required for the study to proceed to stage 2 and enrol additional patients. The primary endpoint was response rate (complete response, partial response, and clinical improvement) at 6 months using the 2006 International Working Group for Myelofibrosis Research and Treatment (IWG-MRT) consensus criteria (Cervantes et al, 2009). This trial was registered with (NCT00931762).

Assessments of efficacy and safety

Patients were assessed for response once per cycle, from cycle 2 onwards, until the end of treatment (Cervantes et al, 2009). Bone marrow was assessed by aspirate and biopsy at screening and again if a haematological response was observed. As part of the criteria for further enrolment of patients, a response must have been observed within the first six treatment cycles. Adverse events (AEs) and ECOG performance status data were collected throughout the study as appropriate. AEs were graded according to the Common Toxicity Criteria for Adverse Events version 4.0 (

The MF symptom assessment form (MF-SAF) is a patient self-assessment questionnaire on the type and severity of symptoms (Mesa et al, 2009). Patients completed the MF-SAF was at baseline and after cycles 2, 4 and 6.

Reagents and antibodies

All antibodies used in this study were obtained from commercial sources. Anti-JAK2, anti–phosphorylated STAT3 (pSTAT3) (Tyr705), anti-STAT3, anti-pAKT (Ser473), anti-AKT, anti–phosphorylated extracellular signal–regulated kinase 1/2 [pMAPK3/1 (pERK1/2)], anti-MAPK3/1, anti–acetyl lysine, and anti–BCL2-interacting mediator of cell death (BIM) were obtained from Cell Signalling Technology, Inc. (Beverly, MA, USA). Anti-pSTAT5 (Tyr694) for immunoblot analysis and flow cytometry was obtained from BD Biosciences (San Jose, CA, USA). Anti-STAT5, anti-PIM1, anti–myeloid cell leukaemia sequence 1 (MCL1), and anti– BCL2L1 (BCL-xL) were obtained from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA, USA). Anti–β-actin (ACTB) was obtained from Sigma-Aldrich (St. Louis, MO, USA). Anti–acetylated K69 HSP90 was generated by Alpha Diagnostics International Inc. (San Antonio, TX, USA).

Isolation of cells from patients with MF

Whole blood was collected in heparinized tubes. For collection of patient plasma, tubes were centrifuged at 1300 g for 10 min. The plasma was removed with a pipette and stored in aliquots for cytokine enzyme-linked immunosorbent assay (ELISA). The remaining cells were combined with HetaSep (STEMCELL Technologies, Vancouver, BC, Canada) in a 50-ml conical tube and centrifuged at 90 g for 5 min with the brake off. The HetaSep fraction of the cells (top layer) was collected with a pipette, and a portion of the cells were removed for genomic DNA isolation and allelic burden assessment. The mononuclear cells were isolated from the HetaSep fraction by Ficoll-Hypaque density gradient centrifugation. Cells were washed with 1 × phosphate-buffered saline (PBS), and aliquots of cells were separated for pSTAT5 expression and immunoblot analyses.

Real-time quantitative polymerase chain reaction

Whole blood from patients with MF at pretreatment and 6 h after panobinostat treatment was collected in heparinized tubes, and then 2 ml of whole blood were transferred into an RNALater tube (Ambion, Grand Island, NY, USA). Total RNA was isolated using a kit from Ambion. Then 50–1000 ng of total RNA was reverse transcribed with a high-capacity reverse transcription kit (Applied Biosystems, Carlsbad, CA, USA). Resulting cDNAs were used for quantitative polymerase chain reaction (qPCR) with TaqMan probes for JAK2 (exons 8–9) and JAK2 (exons 23–24). Relative expression of JAK2 was normalized to expression of GAPDH and is reported as a percentage of the pretreatment expression levels, as previously described (Wang et al, 2009; Fiskus et al, 2011).

Flow cytometric analysis of JAK2 and pSTAT5

Mononuclear cells from patients with MF were collected at pretreatment and after panobinostat treatment. For determination of JAK2 expression, cells were collected 24 h after panobinostat administration. For determination of pSTAT5 expression, cells were collected at 6 and 24 h after panobinostat administration, as well as before cycles 2, 4 and 6. Cells were fixed with 2% paraformaldehyde for 10 min at room temperature, washed with 1 × PBS, and permeabilized in 100% methanol for 20 min on ice. Cells were washed with 1 × PBS, blocked in 0·5% bovine serum albumin (BSA)/PBS for 10 min and incubated with isotype control, rabbit anti-JAK2, or Alexa Fluor 488–conjugated anti-pSTAT5 antibody for 1 h on ice. At the end of incubation, cells for pSTAT5 expression were washed with 0·5% BSA/PBS and analysed by flow cytometry. For JAK2 expression, cells were incubated with Alexa Fluor 488–conjugated anti-rabbit secondary antibody for 30 min on ice. Cells were washed with 0·5% BSA/PBS and analysed by flow cytometry.

Assessment of JAK2 V617F allelic burden

Genomic DNA was isolated from UKE1 and HL-60 cells with a DNeasy blood and tissue kit from Qiagen (Hilden, Germany). Different amounts of mutant JAK2 genomic DNA (UKE1) were serially diluted with wild-type JAK2 genomic DNA (HL-60) to achieve the desired percentages of mutant JAK2 DNA while the total amount of DNA was constant at 20 ng in the PCR mix. The genomic DNA (20 ng per reaction) was combined with a custom TaqMan primer/probe set using minor groove-binding probes to detect both wild-type and JAK2 V617F. qPCR was performed in triplicate on a StepOne plus real-time PCR system. A cycle threshold (Ct) curve was established from the cell lines (UKE1 and HL-60). For assessment of JAK2 V617F allelic burden in patient specimens, genomic DNA was isolated from HetaSep preparations of cells from patients with MF at pretreatment (n = 15) and after panobinostat treatment [before cycles 2 (n = 15) and 4 (n = 4)]. Twenty nanograms of purified genomic DNA were used for qPCR as described above. The Ct for JAK2 V617F in the pretreatment and post–panobinostat-treated samples was compared with the Ct curve of the serially diluted UKE1 DNA. Post–panobinostat-treated JAK2 V617F levels are reported as a percentage of the pretreatment levels (Hammond et al, 2007).

Sodium dodecyl sulfate polyacrylamide gel electrophoresis and western blot analyses

Seventy-five micrograms of total cell lysate was used for sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE). Western blot analyses of acetyl lysine, acetyl-K69 HSP90, JAK2, p-STAT3 (Tyr705), STAT3, p-STAT5 (Tyr 694), STAT5, p-AKT (Ser473), AKT, pMAPK3/1, MAPK3/1, PIM1, BCL2L1, MCL1, and BIM were performed on total cell lysates using specific antisera or monoclonal antibodies. Blots were washed with 1 × PBS plus Tween 20 (PBST), incubated in IRDye 680 goat anti-mouse or IRDye 800 goat anti-rabbit secondary antibodies (LI-COR Biosciences, Lincoln, NE, USA) for 1 h, washed three times in 1 × PBST, and scanned with an Odyssey infrared imaging system (LI-COR). The expression levels of ACTB were used as a loading control for the immunoblots. Densitometry analysis was performed with ImageQuant 5.2 (GE Healthcare, Pittsburgh, PA, USA).

Cytokine ELISA for IL6, TNFα, and MIP1β

All ELISA assays were obtained from Ray Biotech, Inc. (Norcross, GA, USA). Plasma cytokine levels were analysed according to the manufacturer's protocol. Absolute concentrations of each cytokine in the patients' plasma were compared with a standard curve of concentrations for each cytokine provided by the manufacturer.

Statistical analyses

Significant differences between values obtained in a population of cells isolated from patients with MF at pretreatment and after panobinostat treatment were determined using the Student t-test. P values of <0·05 were assigned significance.


Patient characteristics

A total of 35 patients were enrolled onto the trial from August 31, 2009 to October 25, 2010. Patient demographics and characteristics are listed in Table 1. A majority of the patients were male (71·4%) and white (94·3%), and the median age of patients was 68·3 years (range, 42–83 years). Most patients were positive for the JAK2 V617F mutation (65·7%). A majority of patients (68·6%) had PMF, 17·1% had post-PV MF, and 14·3% had post-ET MF. Most (68·6%) had high-risk disease, whereas 31·4% of patients had intermediate risk at screening. Baseline MF-SAF data demonstrated similar symptoms compared with patients in a recent trial evaluating disease burden in patients with MPN, including 293 with MF (Table SI) (Scherber et al, 2011).

Table 1. Patient demographics
CharacteristicN = 35
  1. IPSS, International Prognostic Scoring System.

Sex, n (%)
Male25 (71·4)
Female10 (28·6)
Race, n (%)
White33 (94·3)
Black1 (2·9)
Asian1 (2·9)
Age, years
Standard deviation10·30
Myelofibrosis onset type, n (%)
Primary myelofibrosis24 (68·6)
Post- essential thrombocythaemia myelofibrosis5 (14·3)
Post- polycythaemia vera myelofibrosis6 (17·1)
JAK2 V617F status, n (%)
Positive23 (65·7)
Negative12 (34·3)
IPSS score at screening, n (%)
IPSS Intermediate-211 (31·4)
IPSS High-324 (68·6)
Spleen length at screening, cm below costal margin
Median (range)15 (7–34)
Haemoglobin at screening, g/l
Median (range)99 (68–161)
White blood cell count at screening, 109/l
Median (range)11·9 (1·2–60·0)
Platelet count at screening, 109/l
Median (range)172 (5–505)
Time since diagnosis, days
Median (range)611 (1–8347)

Patients were permitted to remain on trial until disease progression or unacceptable toxicity or at the discretion of the investigator. Thirty-three patients discontinued from the trial, and two patients continue on treatment in a compassionate-use programme. The reasons for discontinuation are outlined in Table SII. AEs were the most common reason for discontinuation in 11 patients (31·4%). AEs most frequently resulting in discontinuation consisted of thrombocytopenia (n = 4) and neutropenia (n = 2). There was one death on study. This death was secondary to thrombosis in a patient with prior deep vein thrombosis.

The median number of days on study was 57 (range, 8–309 d), which was inclusive of dose holds. The majority of patients (n = 19) discontinued before completing cycle 3. Twenty patients experienced a dose modification (reduction, delay, or change in schedule) at some point in the trial. Of the remaining 15, 12 had a treatment duration of <33 d. Seventeen patients had at least 1 dose reduction, and 20 patients had at least one dose interruption. The median time to the first dose reduction was 29 d (≈1 cycle).


Safety data were available on all 35 patients enrolled on the trial. A list of common AEs occurring in ≥20% of patients, regardless of relationship to study drug, are highlighted in Table 2. Thrombocytopenia was the most common haematological AE of any grade (71·4%) and of grade 3/4 incidence (60·0%). Gastrointestinal-related AEs (diarrhoea, nausea, vomiting, and constipation) were also commonly observed but were generally mild (grade 1/2 incidence). Other common AEs included fatigue (57·1% any grade; 31·4% grade 3/4) and anaemia (31·4% any grade; 31·4% grade 3/4).

Table 2. Incidence of common adverse events (AEs) regardless of relationship to study drug (≥20%, any grade)
AEs, ≥20% any gradeAny grade, n (%)Grade 3/4, n (%)
Thrombocytopenia25 (71·4)21 (60·0)
Diarrhoea28 (80·0)3 (8·6)
Nausea20 (57·1)4 (11·4)
Fatigue20 (57·1)11 (31·4)
Decreased appetite19 (54·3)2 (5·7)
Vomiting14 (40·0)2 (5·7)
Anaemia11 (31·4)11 (31·4)
Dyspnoea9 (25·7)5 (14·3)
Constipation8 (22·9)1 (2·9)
Abdominal pain7 (20·0)1 (2·9)
Pyrexia8 (22·9)1 (2·9)
Neutropenia7 (20·0)5 (14·3)
Dry mouth7 (20·0)0
Dizziness7 (20·0)0

The results of the MF-SAF demonstrated that at the start of cycle 2, no substantial changes in patient-reported symptoms from baseline were observed. Improvement in patient-reported symptoms was observed among patients who remained on study in cycle 6; however, these data were difficult to interpret because of the small number of patients who remained on the trial.


A total of 26 of 35 patients (74·2%) had splenomegaly at the time of enrolment and were evaluable for efficacy based on reduction in spleen size. Only one patient met IWG-MRT criteria for clinical improvement (≥50% reduction in spleen size lasting ≥8 weeks). This patient also had a 20-g/l improvement in Hb; however, the improvement was not maintained. A majority of evaluable patients (69·2%) had a transient decrease in spleen size of ≥25%, and 30·8% had a ≥50% reduction in spleen size (lasting <8 weeks; Table SIII). Among the 18 patients with ≥25% spleen reduction, 11 patients were JAK2 V617F-positive and seven were JAK2-wildtype. Among the eight patients with ≥50% spleen reduction, five patients were JAK2 V617F-positive and three patients were JAK2 wild-type. Additionally, two JAK2 V617F-positive patients and one JAK2 wild-type patient achieved a 20-g/l improvement in Hb; however, these improvements were not maintained for 8 weeks.

Molecular correlative studies

Peripheral blood samples were collected for molecular correlative analysis, as outlined in Materials and Methods, for determination of pharmacodynamic and molecular markers of biological activity, including JAK2 V617F allelic burden and downstream proteins involved in the JAK/STAT pathway. Analysis of peripheral blood mononuclear cells isolated from patients with MF before and 24 h after panobinostat treatment suggested DAC inhibition in patient cells, as shown in the representative immunoblots demonstrating increased acetylation of the intracellular targets TUBA, histone H3, and HSP90 following panobinostat treatment (Fig 1A, C). The effects on JAK/STAT signalling were also investigated. Samples for mRNA expression analysis were available from 19 patients pretreatment and after treatment (6 h) (Fig 1B). Depletion of JAK2 protein was also seen at 24 h after 1 dose of study drug as shown in the representative immunoblot assay and flow cytometric analysis (n = 12) (Fig 1C, D).

Figure 1.

In vivo treatment with panobinostat induces hyperacetylation of histone and nonhistone proteins, including HSP90, as well as significantly depletes JAK2 mRNA and JAK2 protein expression in patients with MF. (A) Representative immunoblot analysis of protein extracted from mononuclear cells from patients with MF (n = 12) before and 24 h after panobinostat administration for anti–acetyl lysine demonstrating expression of acetyl-TUBA and acetyl-histone H3. ACTB expression served as the loading control. (B) qPCR gene expression analysis of JAK2 (exons 8–9) and JAK2 (exons 23–24), normalized to GAPDH from patients with MF (n = 19) before and 6 h after panobinostat administration. (C) Representative immunoblot analysis of protein extracted from mononuclear cells from patients with MF before and 24 h after panobinostat administration demonstrating protein expression of JAK2 and acetylated K69 HSP90. ACTB expression served as the loading control. (D) Flow cytometric analysis of mononuclear cells from patients with MF (n = 12) before and 24 h after panobinostat administration analysing intracellular JAK2 expression levels.

The effects of panobinostat on the inhibition of downstream signalling in mononuclear cells isolated from patients with MF were examined. Decreased phosphorylation of STAT5 was observed by flow cytometry and immunoblot analysis 24 h after panobinostat treatment (Fig 2A, B). Phosphorylation of STAT3, AKT, and MAPK3/1 was also decreased following treatment with panobinostat (Fig 2B). In addition, total levels of STAT5, STAT3, and AKT were depleted by panobinostat. Furthermore, changes in protein expression of BCL2-dependent apoptotic proteins were detected, including down-regulation of the antiapoptotic proteins PIM1 and MCL1 and induction of the proapoptotic protein BIM (Fig 2C).

Figure 2.

In vivo treatment with panobinostat inhibits STAT5 phosphorylation, attenuates downstream signalling, and induces proapoptotic BCL2 family proteins in patients with MF. (A) Representative flow cytometric analysis of intracellular pSTAT5 expression of mononuclear cells from patients with MF collected at pretreatment and at the indicated times after panobinostat administration. Open, black-line histogram indicates pretreatment expression. (B and C) Mononuclear cells from patients with MF-MPN before and 24 h after panobinostat administration were analysed by immunoblot for expression of pSTAT5, STAT5, pSTAT3, STAT3, pAKT, AKT, pMAPK3/1, MAPK3/1, PIM1, BCL2L1, MCL1, and BIM. Representative immunoblots of 12 patients are shown. ACTB expression served as the loading control. Numbers beneath the lanes represent densitometric ratios for phosphorylated STAT5 and STAT3 to unphosphorylated STAT5 and STAT3. Ratio was taken as 1·0 in the pretreatment sample.

The effect of panobinostat on JAK2 V617F allelic burden was also investigated. Peripheral blood samples from patients with heterozygous or homozygous JAK2 V617F mutations before initiation of panobinostat treatment and before treatment at cycle 2 (n = 15) and cycle 4 (n = 4) were assessed for change in allelic burden by qPCR. Fig S1 shows that the median percentage of JAK2 V617F allelic burden decreased from 36·82% prior to treatment to 14·88% at cycle 2 and 0·86% at cycle 4.

Plasma from 12 patients was also analysed for determination of inflammatory cytokine levels before and after panobinostat treatment. Plasma samples before treatment and after one cycle of treatment were analysed by ELISA for expression of IL6, TNFα, and MIP1β. Down-regulation of all three cytokines was observed, and the decrease in MIP1β level reached statistical significance (Fig 3). In summary, panobinostat appeared to affect many biologically relevant pathways within the cells of patients with MF.

Figure 3.

In vivo reduction in inflammatory cytokine levels following treatment with panobinostat in patients with MF. Expression levels of inflammatory cytokines IL6, TNFα, and MIP1β determined by ELISA of whole-blood samples from patients with MF (n = 12) after one cycle of panobinostat treatment compared with pretreatment expression. The bar graph shows the percent expression of each cytokine before and after panobinostat treatment. Values are reported as a percentage of the pretreatment expression.


This phase II study evaluated the efficacy and safety of a novel pan-DACi, panobinostat, in patients with higher-risk MF. Panobinostat was not tolerated at the dose of 40 mg three times weekly, and only 16 of 35 patients completed more than two cycles of treatment. The clinical response rate as assessed by IWG-MRT criteria was low (only one patient achieved clinical improvement). However, correlative studies did demonstrate down-regulation of the JAK/STAT pathway. The low clinical response rate in spite of molecular activity is probably related to the short duration on therapy. The high rate of discontinuation observed in this study was somewhat unexpected. It is possible that the patients in whom panobinostat was discontinued for grade four cytopenias during the first cycle had advanced MF with poor bone marrow reserve.

In a phase I study of patients with various haematological malignancies treated with panobinostat, 13 patients with MF were evaluated at doses ranging from 30 to 60 mg three times weekly (DeAngelo et al, 2013). Of note, seven patients received ≥4 cycles of panobinostat at this dose and schedule, and clinical improvement was noted in three of these patients. Based on these promising results, it was expected that the lower dose of 40 mg thrice weekly evaluated in this study would be tolerable and efficacious. Although early discontinuation was common in the present study, two patients are currently receiving panobinostat 30 mg administered thrice weekly/every other week. Both of these patients have been receiving therapy for more than 2 years, suggesting that lower doses or less frequent schedules of panobinostat are more tolerable.

Another phase I/II study of panobinostat conducted specifically in patients with MF determined that the recommended phase II dose of panobinostat was 25 mg thrice weekly (Mascarenhas et al, 2013). A recent presentation of data from this study reported that five patients received panobinostat for more than 6 months. The mean starting dose in these patients was 20 mg (range, 20–30 mg). Panobinostat was well tolerated, and three of the patients who received panobinostat long-term demonstrated IWG-MRT clinical response (Mascarenhas et al, 2013). These data support the activity of panobinostat in patients with MF and suggest that panobinostat may be more tolerable and may exhibit efficacy at doses less than 40 mg thrice weekly.

The results of the molecular correlative analyses observed in the current study demonstrated that panobinostat affected biologically relevant pathways in patients with MF, including reduction in JAK2 V617F allelic burden and inhibition of JAK/STAT signalling. Panobinostat induced in vivo TUBA and HSP90 acetylation in circulating MPN cells, indicating that biologically effective doses of panobinostat were achieved in patients (Atadja, 2009; Wang et al, 2009). It is noteworthy that, as previously described, HSP90 acetylation led to inhibition of the chaperone function of HSP90, which was associated with depletion of progrowth and prosurvival client proteins including JAK2 V617F, STAT5, AKT, and PIM1 (Hu et al, 2009; Wang et al, 2009; Fiskus et al, 2011; Moulick et al, 2011). Consistent with this observation, panobinostat treatment also inhibited JAK/STAT signalling in MPN cells. Our data from patient samples demonstrated that panobinostat mediated decline in MCL1 and BCL2L1 with up-regulation of BIM levels, which is known to promote apoptosis (Reed, 2008) and could thereby lead to the down-regulation of inflammatory cytokines, thus confirming the preclinical observations of panobinostat in MPN cell lines (Wang et al, 2009).

In the phase I/II study of the JAK1/2 inhibitor ruxolitinib, the down-regulation of cytokine levels, including IL6, TNFα, and MIP1β, was correlated among patients who demonstrated a ≥50% composite symptom score change (Verstovsek et al, 2010). After completion of one cycle of panobinostat treatment, JAK2 V617F allelic burden significantly declined (P = 0·036). The decline was more notable in four of the patients who completed three cycles of therapy (Fig 3B). Of note, the selective JAK2 inhibitor TG101348, which has demonstrated clinical benefit in patients with MF, also led to significant decrease in JAK2 V617F allelic burden (Pardanani et al, 2011). These data suggest that panobinostat may affect overlapping pathways associated with direct inhibition of JAK2 in patients with MF.

The data presented in the current study demonstrated that panobinostat elicits not only intracellular effects associated with direct JAK2 inhibitors, such as TG101348, but also the off-target effects observed with JAK1/2 inhibitors, such as ruxolitinib. Recently, mutations in genes involved in regulating the epigenome in myeloid malignancies (e.g. DNMT3A, TET2, IDH1, IDH2, ASXL1, and EZH2) have been described in MPN (Tefferi & Vainchenker, 2011; Vannucchi & Biamonte, 2011). This raises the possibility that an agent, such as panobinostat, which not only targets the JAK/STAT pathway but can potentially also target the deregulated epigenome in myeloid stem and progenitor cells, may represent a novel approach for the treatment of patients with MPN (Fiskus et al, 2009a,b). Recently, the administration of vorinostat to a JAK2 V617F knock-in mouse model of PV resulted in normalization of the peripheral blood counts, marked reduction in splenomegaly, and a decrease in the mutant JAK2 allele burden. This study provided evidence for the role of pan-DACi in MPNs and suggested that the administration of panobinostat for more than two cycles may be required to obtain more gratifying clinical responses (Akada et al, 2012). Furthermore, the DACi givinostat demonstrated preliminary activity specifically in patients with JAK2 V617F-mutated MPNs along with a reduction in JAK2 V617F allele burden (Rambaldi et al, 2010).

In summary, the results of this study demonstrated that panobinostat inhibited JAK/STAT signalling, decreased inflammatory cytokine levels, and decreased JAK2 V617F allele burden in patients with MF, suggesting that panobinostat is biologically active. However, clinical efficacy was lower than expected, with only one patient demonstrating a clinical response according to IWG-MRT criteria. The lack of clinical activity observed may be due to the poor tolerability at the dose and schedule evaluated in this study (40 mg administered three times per week), which resulted in early discontinuations. As preliminary data have suggested better tolerability at lower doses, future studies will seek to determine the potential role of panobinostat at a reduced dose as well as in combination with other agents used in the treatment of patients with MF.


This study was funded by Novartis Pharmaceuticals Corporation. The authors thank Ghulam Warsi, PhD, for assistance with statistical analysis; Robyn Scherber for analysis of the MF-SAF data; and William Fazzone, PhD, for medical editorial assistance with this manuscript. Financial support for medical editorial assistance was provided by Novartis Pharmaceuticals Corporation.

Authorship and disclosures

KNB, DJD, RAM, MSO, CP, JR, AT, and MW designed research; KB, KNB, DJD, WF, RAM, EKR, DSS, AT, and MW performed research; DJD contributed vital new reagents or analytical tools; KNB, DJD, SG, RAM, CP, EK, DSS collected data; KNB, DJD, WF, SG, RAM, MSO, CP, JR, DSS, and AT analysed and interpreted data. All authors contributed to, reviewed and approved the manuscript. EKR has served on a speaker's bureau for Celgene. KNB has received honoraria and research funding from Novartis. DJD has received consultancy fees and honoraria from Novartis. DSS has received consultancy fees and honoraria from Novartis and has served on a speakers bureau for Novartis. MSO is employed by and has equity ownership in Novartis. CP and JR are employed by Novartis. No potential conflicts of interest were disclosed by the other authors.