NSD1 and NSD2 Transcriptional Levels Might Predict Clinical Outcome in AML Patients

Abstract

Aims: The active methylation of histones plays an important role in regulation of gene expression. Nuclear SET Domain (NSD) Histone Lysine Methyltransferases family (KMT) is composed of three members: NSD1, 2 and 3 which regulates gene expression through methylation of H3K36. NSD2 overexpression, initially reported in multiple myeloma, was recently associate with EZH2 mRNA levels through interaction with H3K27me3 which suppresses the expression of miR-203, miR-26a and miR-31 leading to the up regulation of NSD2 mRNA levels. This correlation has been observed in various types of cancer, although there is no data in myeloid malignancies. Aim: To evaluate the gene expression profile of the NSD family in three de novo AML cohorts: one Brazilian cohort and two public databases addressing its applicability in prediction of treatment outcomes, retrospectively. Methods: As a learning set, samples from 146 subjects (age, 18-87y) were analyzed. Seventy (48%) patients were treated in a University Hospital that serves as reference for northeast of Brazil, while 76 patients were treated in a center of similar characteristics in the southeast of the country. The baseline features distribution was similar between centers. The treatment protocol was described elsewhere (Lima AS et al., Blood, 2015), but was based on anthracycline and cytarabine for induction and intermediate to high doses of cytarabine as consolidation. As controls, sixteen samples of CD34+ cells isolated from total bone marrow (BM) of healthy volunteers (age, 18-60y) were analyzed. The external validation cohort composed by 173 patients (age, 18-88y) were obtained from TCGA AML study available online on CBioPortal for Cancer Genomics (Gao et al. 2013). Patients were dichotomized into "low" and "high" NSD1, 2 and 3 expression groups based on survival receiver operating characteristic (ROC) curve and the C index analyses. To validate our data, NSD transcript levels from an independent cohort was used (525 patients from Amazonia! - NSD1 Probe #219084_at; NSD2 Probe #209052_s_at; NSD3 Probe #218173_s_at - and five normal CD34+ samples included in the same databank). The following parameters were used to evaluate treatment outcome: complete hematological remission (CHR); 5-year Disease-Free Survival (DFS) and 5-year Overall Survival (OS) rates. Results: The median age of the learning set was 46y with 70 males (48%). NSD1, 2 and 3 expression levels are lower in de novo AML samples compared to CD34+ cells (P<.001). These results were not validated in an Amazonia! cohort, which patients with AML had a higher-than-normal expression of NSD1 and 3 (P=.01). Expression of all NSD was closely correlated in the learning set cohort (r>0.96) albeit only NSD2 and 3 had a correlation in the validation cohort (r=0.6). Baseline characteristics (sex, age, leukocyte count, cytogenetic risk, FLT3 and NPM1 mutations)were similar between the cohort of patients with NSDs low and high expression, except for higher blast count in BM, but not in peripheral blood, in high expression group for NSD1 and 2 (P=.021 and P=.033, respectively). Overall, 78/146 (57%) patients achieved CHR. Patients with a high NSD1 expression had a lower CHR rate (44%) compared with those a low expression (56%) (OR: 0.5; 95%CI: 0.2-0.9; P=.04). With a median follow up of 21 months (1-61 months) the estimated 5y DFS rate was 49% (95%CI: 30-65%). The group with high NSD1 expression presented a lower 5y DFS rate (13%; 95%CI: 1-42%) than those with low NSD1 expression (72%; 95%CI: 47-87%) (P=.003). This result was consistent with the multivariable proportional hazards analysis (HR: 7.3; 95%CI: 1.3-38; P=.02). The estimated 5y OS rate was 24% (95%CI: 14-35%) which patients with high NSD1 expression exhibiting lower 5y OS rate (13%; 95%CI: 4-29%). The median age of the external validation cohort was 58y with 107 males (62%). Only NSD2 levels had a association with higher diagnosis age (P=.008). With a median follow up of 40 months (1-118 months) the NSD expression may predict only a higher 5y OS rate in patients with high NSD2 expression (30%; 95%CI: 26-49%) retained the result in a multivariable model (HR: 0.53; 95%CI: 0.3-0.9; P=.02). Conclusions: In summary, NSD1 and 2 transcript levels were independent factors associated with treatment outcomes in two AML cohorts treated with standard protocols based on anthracyclines and cytarabine. Disclosures No relevant conflicts of interest to declare.

NSD2 as a Promising Target in Hematological Disorders

“…No evidence has been published to date, and preliminary results do not assign poorer outcomes to AML patients from different cohorts stratified according to NSD1 and NSD2 expression levels [ 105 ]. However, alterations in NSD1 and NSD2 function have been linked to erythroleukemia development, a subtype of acute myeloid leukemia; indeed, it has been reported that NSD1 is a critical regulator of erythroid differentiation, since its knockdown impaired proper erythroblast maturation and induced erythroleukemia in mice [ 106 ].…”

Drug Discovery Targeting Nuclear Receptor Binding SET Domain Protein 2 (NSD2)

“…Since NSD family members share similar structural domains and functions, it is reasonable to believe that NSD2 also serves critical roles in AML. So far, the definitive relationship between NSD2 and AML remains unclear and needs further elucidation . Functional alterations of NSD1 and NSD2 were found to be associated with the development of erythroleukemia, a subtype of AML.…”

“…So far, the definitive relationship between NSD2 and AML remains unclear and needs further elucidation. 165 Functional alterations of NSD1 and NSD2 were found to be associated with the development of erythroleukemia, a subtype of AML. Indeed, NSD1 has been reported to serve as a critical regulator in erythroid differentiation, and its knockout significantly suppressed the proper erythroblast maturation and promoted erythroleukemia in mice.…”