Bomedemstat (IMG-7289), an LSD1 Inhibitor, Manages the Signs and Symptoms of Essential Thrombocythemia (ET) While Reducing the Burden of Cells Homozygous for Driver Mutations

Abstract

The ultimate goal in the treatment of ET is to eliminate mutant cells with the capacity to both proliferate and self-renew. Lysine-specific demethylase-1, LSD1, is an enzyme critical for regulating the proliferation of hematopoietic stem cells and the maturation of progenitors (Sprussel et al. 2012). Bomedemstat is an orally active LSD1 inhibitor that reduces platelets, inflammatory cytokines, and the mutant cell burden in mouse models of myeloproliferative neoplasms (Jutzi et. al., 2019). All current treatments - hydroxyurea, interferon, or anagrelide - have their limitations and do not substantially affect the natural history of ET. IMG-7289-CTP-201 was a global, open-label, Phase 2b study of bomedemstat (IMG-7289, MK-3543) taken once daily for 24+ weeks in patients with ET resistant to or intolerant of at least one standard treatment (NCT04254978). The primary endpoint was reduction of platelets to <400 x 10 9/L in the absence of new thromboembolic events. Exploratory endpoints included changes in frequencies of mutant alleles (VAF) assessed by deep sequencing (median exonic depth =1784 reads) of 261 genes of germline and somatic DNA. Homozygosity of mutant alleles was imputed when VAF >50% and/or loss-of-heterozygosity (LOH) was evident based on the difference between the minor allele frequency of flanking single nucleotide polymorphisms (“MAF method”) in germline versus granulocyteDNA. Single cell genotypes (scDNA) (Tapestri ®) were determined in stem/progenitor (CD34+) and monocytes (CD14+) cells in patients with LOH (using the MAF method) showing a decrease in LOH in whom follow-up samples were available. For patients treated ≥24 wks, 95% (61/64) had reduced platelet counts to ≤400 x 10 9/L without new thromboembolic events, in a median time of 10 wks. In the first 48 weeks, four patients experienced a thromboembolic event ( per PI) (1 PE, 2 TIAs, and 1 transient global amnesia). Of the 14/73 (19%) patients with Day 1 WBC count >10 x 10 9/L, 100% treated for ≥24 wks had reduced the WBC count to ≤10 x 10 9/L; mean Hb remained stable. In patients with baseline MPN-SAF TSS >20 (28/73), at Wk 24, 78% (18/23) experienced an improvement with 52% (12/23) improving ≥10 points. The median baseline score for worst fatigue was 4.0 (N=73); by Wk 24, the score had improved to 3.0 (N=62). Among all patients (N=73), the most common AEs regardless of causality were dysgeusia (55%), constipation (38%), thrombocytopenia (34%), and arthralgia (27%). Twenty-four patients reported serious AEs (SAE), with related events ( per PI) in 4 patients. Twenty-one patients discontinued treatment, 11 due to AEs (3 for dysgeusia), 7 subject's decisions, 1 PI decision, 1 disease progression, and 1 unrelated death due to aspiration pneumonia. There have been no safety signals or deaths related to drug in patients treated for up to 979 days. At Wk 24, 86% patients remained on study and 75% at Wk 48. At ~Wk 24, 39 of 46 (85%) of evaluable patients had a decrease in driver VAF (Fig. 1). The greatest changes in VAF occurred in patients with cells homozygous for a driver mutation (Fig. 1). Nine of ten patients with homozygous driver clones had follow-up data: by bulk sequencing of germline+granulocyte DNA, 7 showed a reduction in the proportion of homozygous cells. In all cases tested, scDNA of CD14+ (monocytes) and CD34+ (stem/progenitor) cells were qualitatively similar to results obtained with bulk granulocyte sequencing (Fig.2). By bulk sequencing, the proportion of homozygous cells dropped by a mean of 66% (-4% to -100% (-100% = homozygosity not detectable)). Decreases in LOH by bulk and sc sequencing during treatment with bomedemstat were identified in patients with homozygosity at JAK2, CALR, and MPL. This study demonstrates that treatment with bomedemstat results in a consistent hematologic response in the majority of patients with ET. Further, deep bulk sequencing of germline+granulocyte DNA reveals LOH ( i.e., homozygosity) events with high accuracy that enables determination of the distribution of homozygote, heterozygote and wild-type genotypes. Furthermore, this study shows that sequencing germline+granulocyte DNA provides a reliable qualitative view of the impact of treatment on the genotypic distribution among CD34+ cells as revealed by scDNA genotyping. Finally, bomedemstat showed activity reducing stem/progenitor cells homozygous for mutations in JAK2, CALR, and MPL.