Molecular Response Patterns in Hydroxyurea (HU)-Resistant or Intolerant Polycythemia Vera (PV) during Treatment with Idasanutlin: Results of an Open-Label, Single-Arm Phase 2 Study

Background

Activating mutations in JAK2 are the molecular hallmark of PV. However, achieving molecular remission is not a validated endpoint in PV, as its correlation with patient-relevant endpoints, such as thrombotic events or disease progression, is not well established. The expression of MDM2, a negative regulator of wild-type p53, is significantly higher in CD34⁺ cells from patients with PV vs those from healthy controls. Therefore, MDM2 antagonists such as idasanutlin (idasa) may have selective and potentially disease-modifying effects in PV. This exploratory analysis reports baseline mutational profiling and serial measurements of JAK2 V617F variant allele frequency (VAF) from patients treated with idasa in the Phase II NP39761 study (NCT03287245).

Methods

Patients with PV, as defined by the 2016 World Health Organization criteria, were eligible for inclusion if their disease was HU resistant/intolerant according to European LeukemiaNet criteria and they were phlebotomy dependent (≥ 1 phlebotomy in the 16-wk period prior to screening). Prior PV therapy with ruxolitinib and/or interferon was allowed. Idasa was given orally once daily on days 1-5 of 28-day cycles for up to 24 mo. The starting dose was 150 mg, but dose increase (to 200 mg) or decrease (to 100 mg) was allowed per protocol-defined criteria. The primary efficacy endpoints were hematocrit control (HTC), defined as maintaining acceptable hematocrit levels with ≤ 1 phlebotomy between cycle 1, day 1 (C1D1) and wk 8 and no phlebotomy after wk 8, and complete hematologic remission (CHR), defined as HTC together with platelets < 400 × 10⁹/L and white blood cell count < 10⁹/L on the day of response assessment (or most recent test). JAK2 V617F VAF was analyzed at screening and at C3D28, C5D28 and wk 32. The JAK2 V617F VAF was determined using quantitative polymerase chain reaction. Baseline blood samples were sequenced using the Foundation Medicine FoundationOne Heme next-generation sequencing panel, which interrogates > 400 DNA-sequenced genes, to explore the genetic landscape of PV beyond JAK2 V617F.

Results

All 27 patients enrolled in the study harbored JAK2 V617F. Baseline and ≥ 1 later JAK2 VAF measurements were available in 23 patients, allowing for analysis. The median age was 56 y (range, 34-74), and 7 patients (26%) had been previously treated with ruxolitinib. Median treatment duration was 257 days (range, 5-677), and the median number of treatment cycles was 8 (range, 1-22). JAK2 V617F VAFs at baseline were highly variable, ranging from 7% to 96% (median, 66%).

A reduction in JAK2 V617F VAF was observed as early as after 3 cycles (median reduction, 39%; n = 19) and was sustained in patients receiving treatment, with a median reduction of 58% at C5D28 (n = 17) and of 76% at wk 32 (n = 13).

The median reductions in JAK2 V617F VAF were significantly higher for patients with CHR and HTC than in nonresponders on C3D28 (CHR, P < 0.01; HTC, P = 0.04) and C5D28 (CHR, P < 0.01; HTC, P = 0.03), with a similar trend also at wk 32 (CHR, P = 0.06; HTC, P = 0.19) (Figures).

Baseline targeted sequencing (n = 16) identified several additional subclonal mutations. However, no TP53 mutations were detected. Variants in TET2 and ASXL1 were detected in 37% (6 of 16) and 20% (3 of 16) of patients, with VAFs of 2%-51% and 38%-50%, respectively. Notably, 68% of the patients had variants detected in 1 to 3 DNA repair genes, including ATM (VAF, 47%-52%), ATR (VAF, 49%), BRCA2 (VAF, 47%-51%), PARP1 (VAF, 52%), PARP3 (VAF, 48%), FANCM (VAF, 47%-50%), BLM (VAF, 48%) and MSH3 (VAF, 49%). In the absence of sequencing of patient-matched non-hematologic tissue and the VAFs being ≈50%, it is possible some of these variants are germline. Whether such variants predispose to PV is unknown.

Conclusions

Idasa has relevant clinical activity in patients with PV who are resistant to or intolerant of HU and showed rapid JAK2 V617F VAF reduction in a subset of patients as early as after 3 treatment cycles. The median JAK2 VAF reduction was significantly greater in patients who achieved CHR and HTC after 3 and 5 cycles of idasa treatment, suggesting that early molecular response was associated with higher likelihood of clinical response. Baseline mutations were seen in several genes commonly mutated in PV. A high frequency of variants in DNA repair genes has not previously been reported in PV and may contribute to genomic instability in this disease.