The Impact of Somatic Mutations upon the Response to Combination Therapy with Ruxolitinib and Interferon in MPN Patients

Introduction: The Philadelphia-negative myeloproliferative neoplasms (MPN) are associated with driver mutations in JAK2, CALR, and MPL genes. Non-driver mutations affect disease progression and treatment response. Combination therapy with pegylated interferon-alpha2 (IFN) and ruxolitinib has recently been shown to induce hematologic and molecular responses in patients (pts) with MPN. We studied 25 pts by targeted next generation sequencing (NGS) of 42 genes and investigated the impact of somatic mutations (mut) on treatment response.

Methods: Twenty-five MPN pts with polycythemia vera (PV) (n=16, all JAK2V617F) and myelofibrosis (MF) (n=9, JAK2V617F (6), CALR (1), MPL (1), triple-negative (1)) participated in the study. NGS was performed on the Illumina platform on DNA from peripheral blood at baseline and after 24 months of combination therapy. Data were analyzed using Biomedical Genomics workbench and VarSeq. Variants with coverage <100x, variant allele frequency (VAF) <1%, introns and germline variants, and SNPs with minor allele frequency >1% were excluded. A mut with VAF <1% in either a pre- or post-treatment sample was retained if VAF was >1% in the paired sample. However, a mut with a number of mutated reads below 7 was considered absent. Only pathogenic mut were reported. Statistical analysis was done using either chi-square test or Wilcoxon signed rank/rank sum test in R 4.0.2. A p-value <0.05 was considered significant. The ELN and IWG-MRT response criteria were used.

Results: At baseline and 24 months, PV-pts had a mean number of mut of 2.3 and 2.4, and MF-pts 2.2 and 3.0, respectively. In all pts, 0 and 1 (0/4%) patient had no mut, 10 and 7 (40/28%) had 1 mut, 10 and 9 (40/36%) had 2 mut, and 5 and 8 (20/32%) had ≥ 3 mut at baseline and 24 months, respectively.

In PV, 1 (6%) achieved CR and 3 (19%) achieved PR and 12 (75%) achieved NR. In MF, 4 (44%) achieved CR and 1 (11%) achieved PR and 4 either NR, PD, SD or CI. There was no association between pts achieving CR/PR or NR and median number of mut at baseline (CR/PR: 1, range 1-3) or (NR: 2, range 1-8), (p=0.28). Excluding driver mut, there was still no significant association.

In all JAK2V617F positive pts, the median JAK2V617F allele burden (%JAK2V617F) decreased from 37% (range: 1.9-95) at baseline to 22.5% (range: 0-85) at 24 months. In pts achieving CR or PR (n=7), the reduction in %JAK2V617F was greater (median: 42% to 18%, p<0.05) compared with pts achieving NR (n=12) (median: 30% to 22%, p<0.005).

Stratified according to molecular response (MR) (n=4) or non-MR (n=13), median % JAK2V617F at baseline was 40 (21-95) and 43 (42-70) in non-MR and MR, respectively, and 31 (11-85) and 13 (1.8-25) in non-MR and MR, respectively during treatment. At baseline, pts achieving MR had no non-driver mut and pts achieving non-MR had a median number of non-driver mut of 1 (range: 0-7) (p<0.02). Eight pts were not evaluable because they were either CALR or MPL positive, triple-negative, or had a baseline JAK2V617F ≤ 20%.

In all pts, there were 30 non-driver mut in 15 genes at baseline and 40 non-driver mut in 19 genes at 24 months of therapy. At baseline and at 24 months, TP53 occurred in 6 and 6 (24%), TET2, ASXL1, RUNX1, or SF3B1 in 2 and 3 (8/12%), and CBL, DNMT3A, or SRSF2 in 2 and 2 (8%) pts, respectively. Eleven new non-driver post-treatment mut occurred in 9 pts and were more prevalent in pts with MF (6/9, 66%) compared to pts with PV (3/16, 19%), p<0.02. However, the VAF was low (median 1.2%, range 1.02% - 3.6%) and none of the new 11 post-treatment mut appeared in the same gene.

Discussion and Conclusions: Earlier studies have shown an association between poor response to IFN and number of mut or high molecular risk mut. In this study, the presence of non-driver mut was associated with a significantly poorer molecular response. However, there was no association between the presence or type of non-driver mut and clinico-hematologic response. At 24 months, there was an evolution of subclones with a significantly higher number of new subclones appearing in pts with MF compared to pts with PV. However, the VAF was low and none of the acquired mut appeared in the same gene contradictory to another study showing an association between treatment-emergent mut in DNMT3A and treatment with IFN. In conclusion, these data show that a clinico-hematologic response is achievable during combination therapy despite the presence of non-driver mutations at baseline.