Phase II Randomized Double-Blinded Study of Pracinostat in Combination with Azacitidine in Patients with Untreated Higher-Risk Myelodysplastic Syndromes

Abstract

BACKGROUND

The prognosis of patients with higher-risk myelodysplastic syndromes (MDS) remains poor despite available therapies. Histone deacetylase inhibitors have shown activity in MDS and in vitro synergy with azacitidine.

METHODS

We conducted a phase II randomized, placebo controlled clinical trial of azacitidine and pracinostat in patients with International Prognostic Scoring System intermediate-2 or high risk MDS. Primary endpoint was complete response (CR) rate by cycle 6 of therapy.

RESULTS

Of 102 patients randomized, 51 were treated in the pracinostat group and 51 in the placebo group. Median age was 69 years. CR rate by cycle 6 of therapy was 18% and 33% (p=0.07) in the pracinostat and placebo groups, respectively. No significant differences in overall survival (OS) (median 16 vs. 19 months, HR = 1.21, 95% CI 0.66–2.23) or progression-free survival (PFS) (11 vs. 9 months, HR = 0.82, 95% CI 0.546–1.46) were observed between groups. Grade ≥3 adverse events occurred more frequently in the pracinostat group (98% vs. 74%) leading to more treatment discontinuations (20% vs. 10%).

CONCLUSIONS

The combination of azacitidine with pracinostat did not improve outcomes of patients with higher-risk MDS. Higher rates of treatment discontinuation may partially explain these results suggesting alternative dosing and schedules to improve tolerability may be required to determine the potential of the combination.

INTRODUCTION

Myelodysplastic syndromes (MDS) are a group of clonal disorders characterized by defective hematopoiesis leading to cytopenias and risk of transformation to acute myeloid leukemia (AML)¹. Patients with higher-risk disease, based on prognostic scores derived from the International Prognostic Scoring System (IPSS) ¹ and its recently revised IPSS-R ², have a short overall survival and increased risk of transformation to AML. Therapy with hypomethylating agents such as azacitidine or decitabine are the standard of care for this group of patients. Although the use of these agents has shown to improve survival of patients with higher-risk MDS^3–5, only 30–40% of patients achieve significant response and all eventually lose response or succumb to their disease. Hypomethylating agent failure is associated with poor outcomes, with a median survival of only 4 to 6 months ^6–8.

Aberrant promoter region CpG island hypermethylation is associated with transcriptional silencing of tumor suppressor genes leading to leukemogenesis⁹. Deacetylated histones are usually associated with DNA hypermethylation and transcriptional repression ^10,11. Inhibition of histone deacetylation and DNA hypermethylation can induce re-expression of silenced genes in leukemia in a synergistic fashion ¹². Preclinical data suggest the combination of hypomethylating agents, such as azacitidine, with histone deacetylase inhibitors, could be potentially synergistic in vivo^12,13. Several clinical studies have explored these combinations in patients with MDS and AML^14–16. Pracinostat, a histone deacetylase inhibitor, has shown preclinical activity in myeloid malignancies¹⁷. An initial phase II exploratory study of pracinostat in combination with azacitidine in 9 patients with intermediate-2 or high risk IPSS MDS showed overall response rates of 89% (8/9), including 7 complete responses + complete responses with incomplete count recovery, and complete cytogenetic responses in 5 (56%) patients with an acceptable toxicity profile mainly consisting of grade 1 or 2 fatigue and nausea ¹⁸.

On the basis of these data, and with the hypothesis that addition of pracinostat may be able to improve upon the response rates of azacitidine, we designed a randomized phase II study comparing azacitidine versus combination therapy with pracinostat in patients with newly diagnosed higher-risk MDS.

MATERIALS AND METHODS

This study was registered at clinicaltrial.gov as NCT01873703. This study was approved by the institutional review board and ethics committees at each study site.

Patient population

Patients age ≥ 18 years with a diagnosis of MDS by French-American-British or World Health Organization classification, 5–30% bone marrow blasts and classified as intermediate 2 (1.5 to 2 points) or high-risk (≥ 2.5 points) by IPSS ¹ were eligible if they had Eastern Cooperative Oncology Group (ECOG) performance status (PS) of 0 to 2, adequate renal and hepatic function (creatinine < 2mg/dL or creatinine clearance ≥ 60mL/min, total bilirubin < 1.5× ULN or ≤ 2mg/dL, and AST and ALT ≤ 2.5× ULN) and QTcF interval ≤ 470 msec. A negative pregnancy test was required prior to enrollment. Main exclusion criteria were prior treatment with HMAs, prior MDS therapy consisting of chemotherapy, immunotherapy, biological or hormonal therapy within the 21 days prior to start of study therapy, or hydroxyurea within 48 hours prior to start of study therapy. Other exclusion criteria included current unstable arrhythmia requiring treatment, history of symptomatic congestive heart failure (New York Heart Association Classes III or IV), history of myocardial infarction within 6 months of enrollment or current unstable angina; clinical evidence of central nervous system involvement; inability to take oral medication; active infection with human immunodeficiency virus or chronic hepatitis B or C; any life-threatening illness unrelated to cancer and presence of a malignant disease within the last 12 months with the exception of adequately treated in-situ carcinomas, basal or squamous cell carcinoma, or non-melanoma skin cancer.

Study Design and Treatment

This was a randomized, double-blinded, placebo-controlled study of pracinostat or placebo in combination with azacitidine. Patients were randomized in a 1:1 ratio to one of two groups: Group 1 (experimental group) of pracinostat plus azacitidine, and Group 2 (control group) of placebo plus azacitidine, and stratified based on intermediate risk 2 vs. high-risk MDS. The primary objective was complete response (CR) by Cycle 6 of therapy. Assessment of response was performed following revised 2006 IWG criteria¹⁹. Secondary objectives included overall response rate (ORR) defined as combination of CR and partial response (PR), clinical benefit rate defined as the combination of hematological improvement (HI) + partial response (PR) + marrow CR + CR, HI rates, duration of response (the time from the initial determination of response to the time of disease progression, relapse, or death on study), overall survival (the time from the first day of study drug administration to death on study), progression-free survival (the time from the first day of study drug administration (Day 1) to disease progression or relapse as defined by the IWG criteria, or death on study), event-free survival (the time from the first day of study drug administration to disease progression, relapse or failure as defined by the IWG criteria, death on study or discontinuation due to clinical progression or lack of efficacy), and rate of leukemic transformation at 6, 12, 18 and 24 months. Dose adjustments and treatment discontinuation were based on hematological and non-hematological toxicity. Therapy could be held in the event of grade 3 non-hematological toxicity or in patients with baseline absolute neutrophil count (ANC) ≥1 ×10⁹/L or platelets ≥30 ×10⁹/L who experienced a greater than 50% decrease in one of these cell counts at Day 28 and had a decrease in bone marrow cellularity to <15% of baseline cellularity on a bone marrow performed on day 42 if platelets remained <30 ×10⁹/L, ANC remained <1 ×10⁹/L and bone marrow blasts were <5% at that time point, respectively. A sample size of 100 patients was planned to provide a reasonable power to detect a clinically important difference in CR rate between the two groups.

Therapy and Follow up

Treatment consisted of azacitidine at a dose of 75 mg/m² daily intravenously or subcutaneously for 7 days (either on Days 1 to 7 or Days 1 to 5, 8 and 9 of each cycle) in combination with pracinostat 60 mg (experimental group) or placebo 60 mg (control group) administered orally every other day for 3 days a week for the first 3 weeks of each cycle. Treatment cycles were repeated every 28 days, unless delay due to toxicity. Therapy was planned to continue until disease progression, patient request or intolerable toxicity. Patients could receive standard of care supportive care measures such as antiemetics, blood transfusions and antimicrobials. All patients were evaluated at baseline with bone marrow aspiration including conventional cytogenetic assessment. Additional bone marrow biopsy and aspirate were performed at the end of cycles 2 and 6, in case of evidence of disease progression or eligibility for allogeneic stem cell transplantation.

Statistical Considerations

The efficacy and safety analyses were conducted on the randomized and treated patients. The sample size was selected to provide reasonable power to detect a clinically relevant difference in CR rates. All p-values were 2-sided, with significance being p<0.05. The Kaplan-Meier product limit method ²⁰ was used to estimate time to events variables. The effects of covariates were assessed using univariate and multivariate Cox proportional hazards models.

Safety

Safety was assessed through the analysis of the reported incidence of treatment-emergent adverse events (AEs), AEs leading to withdrawal and laboratory toxicities, and graded according to the National Cancer Institute Common Terminology Criteria for Adverse Events (CTCAE) v4.0.

RESULTS

Patient Characteristics

A total of 102 patients were enrolled in 19 US centers, randomized and treated between June 2013 and August 2014: 51 (50%) in the pracinostat group and 51 (50%) in the control group. Baseline demographics and disease characteristics are summarized in Table 1. The groups were generally well balanced. Median age across groups was 69 years (range 26–90) with 69% (70/102) being male. A total of 68 (67%) patients had Intermediate-2 risk and 34 (33%) had High risk MDS. A total of 3 patients had refractory anemia (RA), 1 had refractory cytopenia with multilineage dysplasia with ring sideroblasts (RCMD-RS), 49 had refractory anemia with excess blasts 2 (RAEB-2), 17 had RAEB-1, 3 had RAEB not specified, 11 had refractory anemia with excess blasts in transformation (RAEB-T), 8 had chronic myelomonocytic leukemia (CMML), 6 had refractory cytopenia with multilineage dysplasia (RCMD), 3 had myelodysplastic syndrome/myeloproliferative neoplasm unclassified (overlap syndrome), and 1 had MDS associated with isolated deletion 5q. Two patients (33%) with RCMD and the patient with MDS with isolated del(5q) were considered higher-risk based on the presence of a complex karyotype. Median bone marrow blast percentages of patients with CMML were 15% (range 6–16%) and 11% (range 2–15%) in the pracinostat and placebo groups, respectively. Four patients randomized to the placebo group and not eligible because of a diagnosis of RA or RA with ringed sideroblasts were treated and included in the analysis.

Table 1.

Patient characteristics

Characteristic	Azacitidine + Pracinostat (N=51) n (%) [range]	Azacitidine + Placebo (N=51) n (%) [range]	Total (N=102) n (%) [range]
Age (years)	70 [26–90]	69 [43–83]	69 [26–90]
Male	32 (63)	38 (75)	70 (69)
ECOG
0	16 (31)	18 (35)	34 (33)
1	31 (61)	28 (55)	59 (58)
Not available	4 (8)	5 (20)	9 (9)
MDS Subtype at Diagnosis
RA	0	3 (6)	3 (3)
RCMD-RS	4 (8)	2 (4)	6 (6)
RAEB-1	9 (18))	8 (16)	17 (17)
RAEB-2	21 (42)	28 (55)	49 (48)
RAEB type not defined	3 (6)	0 (0)	3 (3)
CMML	3 (6)	5 (10)	8 (8)
MDS/MPN-U	1 (2)	2 (4)	3 (3)
MDS deletion 5q	1 (2)	0 (0)	1 (1)
RAEB-T	9 (18)	2 (4)	11 (11)
IPSS Risk Group
Intermediate-2	34 (67)	34 (67)	68 (67)
High	17 (33)	17 (33)	34 (33)
Hemoglobin (g/dL)	9.2 [7–13]	9.2 [7–13]
WBC (×109/L)	2.3 [1–15]	2.5 [1–47]
Platelet count (×109/L)	39 [2–1370]	54 [15–701]
Peripheral blasts (%)	0 [0–12]	0 [0–33]
Serum Creatinine (mg/dL)	1 [0–2]	0.9 [1–2]
Total Bilirubin (mg/dL)	0.7 [0–2]	0.7 [0–2]

RA: Refractory anemia. RCMD-RS: Refractory anemia with ringed sideroblasts with dysplasia. RAEB: Refractory anemia with excess blasts. CMML: Chronic myelomonocytic leukemia. WBC: White blood cell count.

Disease Response

The rate of CR by Cycle 6, the study’s primary endpoint, was 18% (9/51) in the pracinostat group compared to 33% (17/51) in the placebo group (p = 0.07). A single patient, in the pracinostat group, achieved a CR after Cycle 6. (Table 2). Of the patients who were transfusion dependent at baseline, 25% (2/8) in the pracinostat group and 28% (2/7) in the placebo group achieved transfusion independency. The cytogenetic response rate in the pracinostat group was 42% (14/33) compared to 55% (17/31) in the placebo group (p = 0.14). Hematologic improvement was 35% (15/43) and 55% (24/44) in the pracinostat and placebo groups, respectively (p = 0.51). The overall clinical benefit rate (i.e., CR + PR + HI + mCR) was 53% (27/51) in the pracinostat group compared to 63% (32/51) in the placebo group (p = 0.14) Table 2 summarizes all response data. Median duration of response for patients achieving a clinical benefit was 12 months and 9 months in the pracinostat and placebo groups, respectively (HR = 0.64, 95% CI 0.23–1.79) (Figure 1).

Table 2.

Summary of Responses in both Treatment Arms.

Response	Azacitidine + Pracinostat (N=51) n (%)	Azacitidine + Placebo(N=51) n(%)
Complete response (CR)	10 (20)	17 (33)
CR by cycle 6	9 (18)	17 (33)
Partial response (PR)	0	0
Marrow CR	14 (28)	11 (22)
Progressive disease (PD)	3 (6)	3 (6)
Hematological improvement (HI)	15/43 (35)	24/44 (55)
HI-E	11/40 (28)	17/38 (45)
HI-P	11/36 (31)	18/34 (53)
HI-N	7/27 (26)	10/26 (39)
Cytogenetic Response	14/33 (42)	17/31 (55)
Clinical benefit rate	27 (53)	32 (63)

Figure 1.

Kaplan-Meier curves for Duration of Response in patients treated on Azacitidine in combination with Pracinostat compared to Azacitidine in combination with Placebo.

Survival

The median duration of follow-up was 15 months. The median overall survival (OS) in the pracinostat and placebo groups was 16 and 19 months, respectively (HR = 1.21, 95% CI, 0.65–2.23) (Figure 2) and the median PFS was 11 and 9 months (HR = 0.93, 95% CI 0.49–1.75), respectively (Figure 3). The median EFS was 9 months in both groups (HR = 0.82, 95% CI 0.46–1.46) and the one-year overall survival was 60% in both groups. A total of 15 (15%) deaths (pracinostat, n=9 [18%]; placebo, n=6 [12%]) occurred while on study. In both treatment groups, the most common reason for death on study was death due to AE including febrile neutropenia and pneumonia, followed by disease progression. Mortality rates after 6 months of study discontinuation were 36.6% (15/40) and 33.3% (17/51) in the pracinostat and placebo groups, respectively.

Figure 2.

Kaplan-Meier curves for Overall survival (OS) of patients treated in the pracinostat arm compared to placebo arm.

Figure 3.

Kaplan-Meier curves for Progression-free survival (PFS) of patients treated with pracinostat arm compared to placebo arm.

Safety

A total of 101 (99%) patients experienced at least one treatment-emergent AE. Drug-related AEs were reported in 49 patients (96%) in the pracinostat group and 45 (88%) in the placebo group. The most frequently reported adverse events (AEs) for both groups are summarized in Table 3. Generally, grade ≥ 3 AEs occurred more frequently in subjects receiving pracinostat compared with subjects receiving placebo (98% vs 74%, p<0.001), with fatigue (24% vs 0%), febrile neutropenia (33% vs 20%), and thrombocytopenia (47% vs 26%) being the most common. There were 12 fatal AEs on study, 7 (14%) on the pracinostat arm and 5 (9%) on the placebo arm. The most common fatal AE in the pracinostat group was febrile neutropenia in 2 subjects while pneumonia occurred in 2 subjects in the placebo group. The 30 and 60-day mortality in the pracinostat and placebo groups were 6% and 8%, and 4% and 6% respectively.

Table 3.

Adverse Events Occurring in ≥10% Patients

Adverse Event	All Grades		Grade ≥3
	Azacitidine + Pracinostat (%)	Azacitidine + Placebo (%)	Azacitidine + Pracinostat (%)	Azacitidine + Placebo (%)
Hematologic	77	63	77	59
Anemia	31	39	20	33
Febrile Neutropenia	33	20	33	20
Neutropenia	45	33	45	33
Thrombocytopenia	49	29	47	26
Non-Hematologic
Nausea	69	57	4	2
Fatigue	55	51	24	0
Constipation	53	55	2	2
Vomiting	47	33	4	2
Dyspnea	43	29	8	2
Diarrhea	39	33	4	2
Peripheral edema	45	24	4	0
Dizziness	37	29	4	2
Pyrexia	26	28	4	0
Cough	24	31	0	0
Decreased appetite	33	22	0	0
Myalgia	20	14	2	2
Dehydration	24	14	8	0
Pneumonia	18	16	16	10
Headache	20	18	0	0
QTc prolongation	2	2	0	0

Treatment related AEs leading to pracinostat or placebo dose-reduction occurred in 10% of patients on the pracinostat group and 6% of patients in the placebo group. Azacitidine dose reductions were required in 18% of patients in the pracinostat group compared to 6% in the placebo group. Discontinuation of pracinostat or placebo due to AEs was required in 20% and 10% of patients respectively, with 2% patients in both groups requiring azacitidine discontinuation. More than twice the number of patients discontinued study drugs in the first courses of therapy in the pracinostat group compared to placebo; 17 (33%) vs. 6 (12%) before the start of Cycle 3 and 32 (63%) vs. 16 (32%) before the start of Cycle 5, respectively. Therefore, two thirds of patients discontinued study drugs before Cycle 5 of therapy compared to one third in the placebo group.

Exploratory Analysis

Because of the high rate of early discontinuations in the pracinostat group, we performed an ad hoc analysis to evaluate the efficacy of the combination in patients who could tolerate at least 4 cycles of. A total of 54 patients, 19 (37%) in the pracinostat arm and 35 (69%) in the placebo arm had started Cycle 5 of therapy. No significant differences in terms of CR after cycle 6 of therapy (42% vs 43%,), ORR (47% vs 43%) or clinical benefit rates (95% vs 80%) were observed between the pracinostat and placebo groups respectively. However, although not statistically significant, a numerically longer duration of response (median 12 months vs 9 months) and OS (median not reached vs 19 months) was observed in the pracinostat group.

DISCUSSION

Although available standard of care therapies such as azacitidine have shown to improve overall survival of patients with higher risk MDS ^3–5, only a subset of patients responds to these treatments, and prognosis after loss of response to therapy remains poor ^6–8. In the search of more effective regimens, and based on the synergistic results in preclinical models and favorable results of an open-label, single center study of azacitidine and pracinostat in previously treated higher risk MDS patients ¹⁸, we conducted this multicenter double-blind, placebo-controlled phase II study to evaluate the efficacy and safety of pracinostat in combination with azacitidine in previously untreated higher risk MDS. In contrast to the observations from the open label study, addition of pracinostat to azacitidine did not improve the rate of complete responses, cytogenetic responses, and overall survival in the current study. Also, there were no benefits with the combination in other secondary survival endpoints such as event-free survival or progression-free survival.

One possible explanation is poor tolerability leading to shorter overall exposure to study therapies. While the number of patients who experienced adverse events (AEs) was similar in both treatment groups, a higher proportion of patients treated with pracinostat and azacitidine experienced grade ≥ 3 AEs, mainly fatigue, gastrointestinal symptoms and myelosuppression, and predominantly within the first cycles of therapy. These initial toxicities, leading to more frequent and earlier drug discontinuations in the pracinostat group, likely limited the overall efficacy of the combination and may explain the negative clinical outcomes. This is in contrast to the earlier open label study of pracinostat and azacitidine in MDS, where toxicities were managed with adapted dose reductions, which allowed patients to remain on therapy for longer periods.

In an attempt to determine whether excess toxicity was responsible for these results we analyzed the outcomes of patients receiving at least 4 cycles of therapy. Although no differences in terms of response were observed between both groups, a tendency to improved duration of response and overall survival was observed in the pracinostat group. The observed differences were not significant since this subgroup analysis does not have the statistical power to demonstrate superiority of the combination. Further studies would therefore be required to confirm these results. Interestingly and, although on a different patient population, a phase II trial of azacitidine and pracinostat in newly diagnosed elderly patients with AML not eligible to receive intensive induction chemotherapy due to comorbidities or unfavorable disease features has shown significant activity of the combination with CR rates of 42% and 1-year overall survival of 60%, and was associated with a lower rate of early discontinuation compared to this study²¹. It is possible that older patients with AML who are debilitated from their disease and for whom alternative therapies are limited, may be more willing to comply with regimens associated with fatigue and gastrointestinal toxicities as compared to patients with MDS who are ambulatory and have more therapeutic options.

Also, the absence of sequencing analysis limits our ability to determine whether unbalanced distribution of patients with high-risk mutational features, such as TP53 mutations^22–24; or TET2 mutations, associated with higher response rates to azacitidine^24–26, may be partially responsible for the observed clinical outcomes. Previous studies suggest achievement of a response to azacitidine is associated with improved survival outcomes in patients with MDS²⁷. However, the tendency to lower CR rates with azacitidine and pracinostat observed in our study, did not correlate with shorter survival outcomes in the combination arm.

Although several preclinical and pilot clinical experiences suggest a potential benefit of adding HDAC inhibitors to hypomethylating agents as combination epigenetic therapy ^12,13, combination trials evaluating such treatment combinations with valproic acid ^14,28, entinostat ²⁹, panobinostat ³⁰ or vorinostat have so far had challenges to significantly improve outcomes mainly due to suboptimal HDAC inhibition potency ³¹, suboptimal dosing or excess toxicity in combination with azacitidine ^16,32. For example, the SWOG S1117 study ³³, a randomized study of azacitidine with or without vorinostat in MDS, showed results similar to those we report here, with lower CR rates (15% vs 24%) and higher treatment discontinuation due to adverse events with the combination therapy as compared to azacitidine monotherapy (24% vs 9%). Although conducted in a different patient population, similar observations have also been reported with azacitidine and vorinostat in patients with AML and MDS with comorbidities, poor performance or organ dysfunction ³⁴ in whom addition of vorinostat did not significantly impact overall survival. Therefore, it can be speculated that the absence of clinical benefit in the present study as well as in previous studies could be partially explained by a not optimal treatment schedules. Also, our study did not include methylation assays or other correlative studies before initiation of therapy and during the course of treatment. We acknowledge this is a limitation.

Pracinostat has been shown to have superior preclinical PK/PD properties when compared to other HDAC inhibitors with favorable best-in-class pharmacokinetic properties^35,36. In fact, as was reported in the initial phase I study of pracinostat in patients with hematological malignancies ¹⁷, appropriate pracinostat exposure and HDAC inhibition is possible with doses as low as 40 mg daily. This suggests that alternative combination regimens with lower pracinostat daily doses or less frequent dosing may be better tolerated and should be investigated in the MDS setting, where chronic tolerability is paramount.

In conclusion, the combination of pracinostat and azacitidine as evaluated in this study did not improve the outcome of patients with MDS compared to azacitidine monotherapy. Excess toxicity in the form of cytopenias, fatigue, and gastrointestinal toxicity resulted in earlier discontinuations and suboptimal long-term exposure to treatment, potentially leading to reduced efficacy. Lower daily doses or alternative dosing schedules aimed at reducing treatment toxicity should be evaluated.