Impact of Underlying Mutational Profile, and Changes during Treatment, in MPN Patients Treated with JAK Inhibitors

Background. Most patients (pts) with MPN derive clinical benefits from treatment with JAK inhibitors (JAKi) but predictors of response or loss of response have not been clearly identified.

Aims. To address the impact of underlying mutational profile of driver mutations (JAK2V617F, CALR, MPL), including other genes of JAK/STAT pathway, and "subclonal" (epigenetic, spliceosome) mutations, and their modifications under treatment, on the response to JAKi.

Methods. 100 patients with WHO2008-diagnosis of MPN, 39 PV, 14 ET, 64 MF (37 PMF, 27 PPV/PET-MF), were included. 25 (12PV, 13 MF) had received hydroxyurea (HU), 75 (20 PV, 7 ET, 48 MF) a JAKi (67 ruxolitinib, 8 fedratinib). Response criteria were as by IWG-MRT (Blood 2013; 121:4778; Blood 2013; 122:1395). Mutations in 22 genes (JAK1, JAK2, JAK3, EZH2, ASXL1, TET2, IDH1, IDH2, CBL, SRSF2, DNMT3A, NFE2, SOCS1, SOCS2, SOCS3, SH2B3, STAT1, STAT3, STAT5A, STAT5B, SF3B1, U2AF1) were analyzed in blood DNA at baseline (bl) and at the latest available sample by deep sequencing with Ion Torrent-PGM. CALR mutations were analyzed by capillary electrophoresis; JAK2 V617F allele burden was measured by RT-qPCR assays. The nonparametric Wilcoxon rank-sum test, Kaplan-Meier method and Cox regression, were used for statistical analysis.

Results. A. Impact of baseline mutational profile. PV and ET:At bl, 17/40 (42.5%) pts had subclonal mutations. Spleen volume reduction (SVR) by IWG-MRT criteria was obtained in 100% of pts without subclonal mutations compared to 66% of those with >1 mutations (P=0.04). Presence of NFE2 mutations uniformly predicted for lack of SVR (P=0.03). Loss of SVR was predicted by ASXL1 mut (67% vs 12% in un-mutated; P<0.05). No correlation of mutations with symptomatic improvement, normalization of leukocyte and platelet count, and control of hematocrit to <45% in PV pts was noticed. A trend to lower pruritus responses were seen in PV pts harboring >1 subclonal mutations (P=0.06). JAK2 V617F homozygous pts were more prone to have platelet count control (81%) compared to heterozygous (30%; P<0.04). MF:At bl, 33/64 (51.5%) pts had subclonal mutations. No correlation of mutations with symptomatic improvement, normalization of leukocyte and platelet count was found. SVR was not predicted by a high molecular risk status (Leukemia 2013;27:1861; Blood 2014;123:2157); however, harboring >1 of the 22 mutations was negatively associated with SVR (P=0.02) (HR 1.9, 95% CI 1.0-4.6). No correlation of JAK2 V617F or CALR mutations, or the allele burden, on SVR or loss of SVR was discovered.

B. Changes during treatment. PV and ET:follow-up mutational analysis was performed at median treatment duration of 2.9y HU, 4.7y ruxolitinib, 1.7y fedratinib in 50 pts for driver mutation and 40 pts for subclonal genes. Of HU pts (n=12), 33% each had JAK2 V617F allele burden stable, increased and reduced by >10%. Median reduction of allele burden was -18.3%. 1 pt acquired 2 novel subclonal mutations (ASXL1, EZH2) and 2 showed increased allele burden of ASXL1 and TET2 by >10%. Among JAKi pts (n=30), 20 pts (66.7%) had JAK2 V617F allele burden reduction by a median of -33.3% (-13.6 to -100%), stable in 9. Degree of allele burden reduction was positively correlated with length of JAKi treatment (P=0.04). 1 pt acquired novel ASXL1, 2 pts had increase (ASXL1, EZH2) and 6 pts reduction of >10% of allele burden (4 TET2, 1 DNMT3A, 1 SH2B3). No correlation was seen with clinical response over time. MF: follow-up mutational analysis was performed at median treatment duration of 2.4y HU, 1.9y ruxolitinib, 1.4y fedratinib. HU pts (n=13): of the 7 JAK2 V617F mut, 4 increased and 1 decreased the allele burden, and 1 of 3 CALR mutated reduced allele burden, by >10%. 1 pt acquired CBL mutation, other mutations (n=7) were stable. Among JAKi treated pts, of the 41 JAK2 V617F mutated, 51% showed a median allele burden reduction of -14.8% (-10.5% to -53.9%). Of 10 pts with subclonal mutations at baseline, 4 clones (3 ASXL1 and 1 NFE2) increased and 2 (EZH2, ASXL1) decreased by >10%. Only 1 pt acquired a novel mutation in EZH2. No correlation was seen with clinical response over time.

Discussion. These data suggest that only minimal influence on clinical response is provided by driver mutations and their allele burden, or subclonal mutations, in MPN patients receiving JAKi. The clinical relevance of different clone fluctuations over treatment with conventional therapy and JAKi remains to be addressed.