Nuclear SET Domain (NSD) Protein Lysine Methyltransferases (KMT) Family Members Expression in Acute Myeloid Leukemia

The Nuclear SET Domain (NSD) Protein Lysine Methyltransferases (KMT) family is composed of three members: NSD1/KMT3B, NSD2/WHSC1/MMSET and NSD3/WHSC1L1 which regulate gene expression through methylation of lysine 36 of histone H3 (H3K36). NSD2 overexpression was reported in multiple myeloma with t(4;14)/IgH-MMSET. NSDs gene expression profile is unknown in acute leukemias, however NSD1 and NSD3 were described to be fused with the nucleoporin 98 gene (NUP98) in rare AML and myelodysplastic syndrome cases and, both fusion proteins were associated with poor prognosis. The aims of the present study were to characterize the expression of NSD-KMTs in patients with AML and healthy controls, to determine if this expression is associated with specific genetic abnormalities and/or with treatment outcome.A total of four healthy donors and 45 AML patients (27♀, 18♂) at diagnosis were included in the study. Our cohort included 8 patients with acute promyelocytic leukemia (APL), 8 with core binding factor (CBF) leukemias [4 with t(8;21) and 4 with inv(16)], and 29 patients with non-APL non-CBF AML. NSD family gene expression was evaluated by qPCR using the comparative Ct method for analysis. A higher expression of the NSD1 gene was observed in AML cells compared to normal bone marrow (BM) samples {median [range] = 3.202 [0.6804-0.096] vs. 1.003 [:0.7956-1.265], p=0.0243}. Similarly, the expression of NSD3 was higher in AML, but the difference was significant only for the comparison between healthy BM and CBF-AML groups {median [range] = 1.070 [0.6360-1.410] vs. 2.719 [1.238-8.830], p=0.0265}. No significant differences were detected in the analysis of NSD2 expression. Considering the three groups of AML patients, no correlation was found between NSD1, NSD2 or NSD3 expression levels and age, gender, leukocyte counts at diagnosis, karyotype (normal vs. abnormal), frequency of specific genetic abnormalities (t(15;17)/PML-RARA; t(8;21)/RUNX1-RUNX1T1; inv(16)/CBFB-MYH11) or percentage of blasts in bone marrow. NPM1 mutations and FLT3 internal tandem duplications (FLT3-ITD) were detected in 29.6% (13/44) and 21% (9/43) of the patients with AML, respectively. We observed a significant increase in NSD1, NSD2 and NSD3 expression in blasts from patients with FLT3-ITD (p=0.0177), but not in those with NPM1 mutations. These differences remained significant when APL cases were excluded from the analysis. Next, patients were grouped according to NSD1 or NSD2 expression. Patients with NSD1 or NSD2 expression higher or lower than the median value (3.25 and 3.16, respectively), showed no significant differences regarding age distribution, leukocyte counts or percentage of blasts in bone marrow at diagnosis, or presence of genetic abnormalities. Regarding the analysis of treatment outcome, patients with non-APL AML were stratified into high and low NSD1 or NSD2 expression subgroups using the criteria above. The median overall survival of patients in the low NSD2 expression subgroup was of 333,023 days [95% CI:158,541-507,505 days] whereas patients in the high NSD2 expression subgroup was of 817,629 days [95% CI:238,702-1396,555 days] (p=0,633). No significant difference observed between the overall survival of patients in the high and low NSD1 expression subgroups. In order to determine if NSD-KMT levels were associated with changes at histone H3 lysine 4 (H4K4) and H3K36 (known to activate gene transcription), as well at histone H3 lysine 9 (H3K9), H3K27, H3K79 and H4K29, associated to regulatory repression, we ran an experiment using Illumina Infinium Methylation 450k arrays. The comparison between normal and leukemic cells revealed specific histone methylation profiles. There is experimental evidence that histone methylation is a prerequisite for DNA methylation and transcriptional regulation, suggesting interplay between histone and DNA methylation. Our data correlate overexpression levels of NSD-KMT with histone modifications, suggesting that this modification and not only DNA methylation can contribute for epigenomic changes associated to AML pathogenesis.