Results of a Clinical Trial of H3B-8800, a Splicing Modulator, in Patients with Myelodysplastic Syndromes (MDS), Acute Myeloid Leukemia (AML) or Chronic Myelomonocytic Leukemia (CMML)

Background:

Heterozygous somatic mutations in the genes encoding RNA splicing factors SF3B1, U2AF1, SRSF2 or ZRSR2 induce aberrant splicing in cancer cells and are among the most common mutations in patients with MDS, AML or CMML. H3B-8800 is an orally available small molecule that binds to the SF3b complex and induces alternative splicing changes in cells. Because splicing factor mutant cells depend on residual wild-type function of splicing factors for survival, we hypothesized that H3B-8800 would induce preferential cell killing of mutant cells by further perturbing splicing to synthetic lethality. In pre-clinical models, H3B-8800 preferentially kills spliceosome mutant cells and induces antitumor activity in xenograft leukemia models with core spliceosome mutations (Seiler M et al Nature Medicine 2018). Therefore, we conducted a phase I clinical trial (NCT02841540) of H3B-8800 in patients with MDS, AML or CMML. Here we describe the safety profile and clinical outcomes of the dose escalation cohorts in this phase I trial.

Methods:

This phase I trial explored the safety, pharmacokinetics (PK), and pharmacodynamics (PD) of H3B-8800 in patients with myeloid cancers. Dose escalation cohorts, employing a standard 3+3 design, examined 2 different once daily dosing regimens (schedule I: 5 days on/9 days off; schedule II: 21 days on/7 days off) in a 28-day cycle, with stratification based on lower-risk (LR) versus higher-risk (HR) myeloid neoplasms.

Results:

As of June 16th, 2019, 84 patients were enrolled at 24 centers in the US and Europe. Dose ranged from 1-40 mg among 65 patients on schedule I, while 19 patients were enrolled with dose ranging from 7-20 mg on schedule II. The patient population included AML (n=38), CMML (n=4), HR-MDS (n=20), LR-MDS (n=21) and 1 MDS with unknown risk level. Most patients (88%) had spliceosome mutations of interest. The most common mutations were in SRSF2 (p.P95H in 17, p.P95L in 9, p.P95_R102Del in 4), SF3B1 (p.K700E in 11, p.R625C in 4), and U2AF1 (p.Q157P in 6, p.S34F in 4).

Patients remained on treatment from 7 to 819 days; 25 patients (30%) had time on treatment greater than 180 days, 20% more than 1 year and 2% over 2 years. The median therapy duration for LR-CMML/MDS, HR-CMML/MDS, and AML patients were 216, 62, and 47 days respectively. Most observed treatment related treatment-emergent adverse events (TEAEs) were Grade 1 or 2. The most common treatment-related (as judged by the investigator, >10% frequency) TEAEs in the patients treated on schedule I were diarrhea (75%), nausea (37%), fatigue (28%) and vomiting (27%). The most common treatment-related TEAEs in the patients treated on schedule II were diarrhea (68%), vomiting (42%), QTc prolongation (21%), nausea (16%), and fatigue (16%). The most common dose limiting toxicity was prolongation of the QTcF interval >500 msec (n=2, 40 mg on schedule I and n=1, 20 mg on schedule II, all ≥Grade 3) and bradycardia without other arrhythmias (n=1, 14 mg on schedule II, ≥Grade 3). No ophthalmic AEs were observed; 1 patient (LR-MDS) experienced durable marrow aplasia. The maximum tolerated dose (MTD) has not been confirmed for either Schedule I or Schedule II.

PK analysis indicates that H3B-8800 is rapidly absorbed and exhibits dose-proportional increase in plasma exposure. Consistent dose-dependent target engagement (i.e., alteration in mature mRNA transcripts) was observed in blood mononuclear cells from patients enrolled in the 2 mg up to 40 mg dose cohorts on both schedules. Despite this splicing modulation, no objective complete responses (CR) or partial responses (PR) meeting International Working Group criteria were observed. One patient with CMML had a durable platelet response that began in Cycle 1 and persisted through Cycle 13. Nine red blood cell (RBC) transfusion-dependent patients with MDS or CMML and 2 patients with AML did not require RBC transfusions for ≥8 weeks (up to 28 weeks) while on study. One platelet transfusion-dependent patient with LR-MDS did not require platelet transfusions for ≥8 weeks.

Conclusion:

Results from this first-in-human multicenter prospective clinical trial of a splicing modulator in myeloid neoplasms demonstrate dose-dependent target engagement and predictable PK profile of H3B-8800, and safety even with prolonged dosing. Although no objective CR or PR were achieved, decreased RBC or platelet transfusion requirements were observed in 12 (14%) of enrolled patients.