Analyses of 81 Chronic Myelomonocytic Leukemia (CMML) for EZH2, TET2, ASXL1, CBL, KRAS, NRAS, RUNX1, IDH1, IDH2, and NPM1 Revealed Mutations In 86.4% of All Patients with TET2 and EZH2 Being of High Prognostic Relevance

Abstract 296

Chronic myelomonocytic leukemia (CMML) is a clonal hematopoietic malignancy that is characterized by features of both a myeloproliferative neoplasm and a myelodysplastic syndrome. Recently, we investigated 81 CMML cases (45 CMML-1, 36 CMML-2) diagnosed between 10/2005 - 9/2008, which had been characterized by chromosome banding analysis and mutation analysis in 6 candidate genes: Mutations were detected in TET2 (44.4%), CBL (22.2%), NRAS (22.2%), KRAS (12.3%), JAK2 (9.8%), RUNX1 (8.7%) (Kohlmann et al., J Clin Oncol. 2010 Jul 19). We now applied amplicon-based deep-sequencing using the small volume Titanium chemistry assay (454 Life Sciences, Branford, CT) to investigate additional 4 candidate genes: IDH1 (exon 4), IDH2 (exon 4) and NPM1 (exon 12) (at known mutational hotspot regions) and the complete coding region of EZH2. EZH2 encodes a catalytic subunit of the polycomb repressive complex 2, a highly conserved histone H3 lysine 27 methyltransferase that influences stem cell renewal. Mutations in EZH2 were recently described to play a role in MPN/MDS. The sequencing library preparation for IDH1, IDH2, NPM1, and EZH2, in total 22 amplicons, was performed using 48.48 Access Array technology (Fluidigm, South San Francisco, CA) to cope with the number of amplicons. In median, 498 reads per amplicon were obtained, thus yielding sufficient coverage for detection of mutations with high sensitivity. After excluding polymorphisms and silent mutations aberrations were detected in IDH1 (1/81; 1.2%), IDH2 (3/81; 3.7%), NPM1 (1/81; 1.2%), and EZH2 (10/81; 12.3%). Another gene recently described in hematological diseases is ASXL1 (additional sex combs like 1) on chromosome 20q11.1. Therefore, the hotspot region of ASXL1 exon 12 was additionally investigated by Sanger sequencing in those 20 cases, in which no mutation had been observed thus far. Nine of these 20 cases (45%) harbored a mutation in ASXL1, thus only 11 cases (13.6%) remained in this cohort in which no mutation was detected. Summarizing this data, 86.4% of these CMMLs harbored at least one molecular aberration with a median of two genes mutated (range 1–4). In more detail, we observed 11 novel distinct EZH2 mutations in ten patients: 7 missense, 3 frameshifts (2 deletions, 1 insertion), and one splice site mutation. EZH2 mutations were found to be heterogeneous and were spread over several exons, predominantly located in the four conserved regions (6/11 in the conserved SET domain; e.g. H680R, N659S). No case revealed a Tyr641 of EZH2 mutation as described for follicular and diffuse large B-cell lymphomas. In median, the burden of EZH2 mutations was 42.5% of sequencing reads per patient (range 1.4–98%). Similarly, a high mutation burden was detected in RUNX1 (median 46.7%), TET2 (median 44.6%), and CBL (median 42.5%) whereas the burden was low in RAS pathway alterations, i.e. NRAS (median 11.1%), KRAS (median 27%), or JAK2 V617F mutations (median 6.9%). With respect to associations of distinct mutations no specific pattern was observed, i.e. EZH2 mutations were concomitantly detected with TET2 (4/10), RUNX1 (3/10), CBL (3/10), JAK2 (3/10), NRAS (2/10), KRAS (1/10), and IDH2 (1/10), respectively. Further, EZH2 mutations were associated neither with morphologic CMML subtype or dysplastic or myeloproliferative characteristics nor with age, white blood cell count, thrombocytes count, or hemoglobin. However, with respect to clinical data a very poor outcome was observed for patients that carried EZH2 mutations compared to EZH2 wild-type cases (median OS 4.3 vs. 130.4 months; p<0.001). In contrast, a significantly better outcome was seen for patients who carried TET2 mutations compared to TET2 wild-type cases (median OS 130.4 vs. 53.6 months, p=0.013). Subsequently, we performed a survival analysis taking both EZH2 and TET2 mutations into account. Here, the cohort was significantly separated into three distinct prognostic groups, i.e. EZH2-mutated with a poor median OS of 4.3 months, EZH2/TET2 wild-type with a median OS of 90 months and TET2-mutated cases with a median OS of 130.4 months (p<0.001). In conclusion, our study revealed molecular mutations in 86.4% of 81 CMML patients providing new insights into the molecular heterogeneity of this disease. Besides alterations in TET2, CBL, ASXL1, and the RAS pathway, EZH2 is targeted by various types of frameshift and point mutations and is a novel biomarker with unfavorable prognosis and clinical utility.