Impact of the Molecular Profile of Malignant Clones on the Response to Interferon Alpha (IFNa) Therapy in JAK2V617F-Negative Essential Thrombocythemia (ET)

Background: The majority of ET patients (pts) without JAK2 or MPL mutations present somatic mutations in the calreticulin gene (CALR). However, a series of mutations in other genes involved in the epigenome, the splicing machinery or leukemic transformation have been described in ET, but contrary to MF, their impact has not been clearly assessed. We have previously shown that IFNa may considerably reduce the JAK2-mutated clones, and we also observed that TET2 mutated clones are resistant to IFNa therapy in JAK2+TET2+ pts with polycythemia vera. As little is known about the IFNa response of clones harboring mutations other than JAK2V617F in ET, we took advantage of a cohort of pts without JAK2 mutation but CALRmutated and treated with IFNa (according to international and local guidelines) to assess the dynamics of the different MPN clones during therapy.

Aims: 1) To determine, using a Next Generation Sequencing (NGS) approach, the molecular pattern of mutations in genes previously shown to have a prognostic impact in MPNs in a series of CALR-mutated ET pts; 2) To study the evolution of these mutational patterns during IFNa therapy.

Methods:JAK2 and MPL-negative ET pts followed in our department fulfilling the following criteria were included in the study: presence of a CALR mutation; availability of at least 3 sequential blood samples including one taken before IFNa; IFNa therapy for at least 3 months; informed consent for molecular analysis. Total DNA was extracted from blood samples (Qiagen blood DNA mini kits) for molecular analyses. Mutations in TET2, ASXL1, EZH2, SRSF2, IDH1, IDH2, SH2B3 and CSF3R were searched through a NGS approach on a MiSeq instrument using a TruSeq custom amplicon approach (Illumina). CALRgene mutation detection was done by direct Sanger sequencing of exon 9, and mutant allele burden (%CALR) was estimated by fragment analysis (GeneMapper software , Life technologies) with a sensitivity of about 1%. Both sequencing and fragment analysis were performed on a 3500xL DX Genetic Analyser (Life technologies). Molecular response was defined as complete (CMR) when CALR mutation was no longer detectable, partial (PMR) when %CALR was decreased by >50%, minor (mMR) when %CALR was decreased by 25 to 49%, and non responders if %CALR was reduced by less than 25%.

Results: Among 238 ET pts without JAK2 or MPL mutations, we identified 52 pts (22%) treated with peg-IFNa-2a, of whom 24 fulfilled the inclusion criteria. Median age was 52 years (range 31-68), 67% were women, median follow up was 12 years (range : 1.5 – 29) and median IFN treatment duration was 30 months (range: 12 – 102). 23/24 (96%) pts achieved complete or partial hematological response to IFNa. CALR mutations (present in all patients, quantifiable in all but 1 because of 1bp difference between mutated and WT) were of type 1 in 10 (42%), type 2 in 8 (33%) and of neither type in 6 (25%) pts, respectively. In addition to CALR, a second mutation was found in 8 pts (33%) by NGS, affecting ASXL1 (n=2), TET2 (n=2), IDH2 (n=2), CSF3R (n=1) and SH2B3 (n=1) genes. Comparison of sequential samples showed that the %CALR decreased from a median of 43% (range: 8-57) to 26% (range: 0-49) in the last sample (p= 0.018). In details, %CALR decreased in 13 (57%) pts, including 1 CMR, 7 PMR and 5 mMR. Interestingly, molecular response to IFNa of CALR+clones appeared poorer in pts with additional mutations (vs. pts with CALR mutation alone): 50% were non responders (vs. 40%), %CALR even increased during follow up in 25% (vs. 7%), and the median decrease in %CALR was 32% (vs. 45%). Dynamics of clones harboring additional mutated genes showed that, in contrast to CALR, mutant allele burden did not vary in 6/8 pts. However, an IDH2 mutation decreased from 9% to 1% in 1 pt, and a TET2 mutation increased from 1% to 21% in another (while %CALR remained unchanged in both pts, at 44% and 19% respectively). Of note, after 12 years of median follow-up, no emergence of new mutation was observed in any patient.

Conclusion: In this cohort of JAK2V617F-negative ET, 96% of pts achieved hematological response with IFNa therapy. However, molecular profiling suggests that the mutational pattern found in malignant clones modulates the molecular response to IFNa: 1) Existence of more than 1 mutation induces poorer molecular response; 2) Clones with CALR mutation alone are sensitive to IFNa; 3) Clones harboring mutations in genes other than CALR seem less responsive to IFNa therapy.