Ectopic NSD2 Remodels H3K36me2 and DNA Methylation to Promote Oncogenic Gene Expression in Multiple Myeloma

Introduction: The t(4;14) translocation causes overexpression of NSD2 in ~12% of patients with multiple myeloma (MM) and is associated with poor clinical outcomes. NSD2 catalyzes dimethylation of histone H3 lysine 36 (H3K36me2), which is dramatically increased in t(4;14) MM. The molecular mechanisms by which altered H3K36me2 potentiates an oncogenic gene expression program are poorly understood. One function of H3K36me2 is the recruitment of DNA methyltransferase enzymes such as DNMT3B, leading to DNA methylation (DNAm). DNAm plays important roles in both plasma cell differentiation and oncogenesis, yet how DNAm is dysregulated in t(4;14) MM is largely unknown. Methods: We analyzed whole genome bisulfite sequencing (WGBS) DNAm data from 415 primary MM samples from the MMRF CoMMpass trial (NCT01454297) which had been integrated with genetic, transcriptional, and clinical information. Differential DNAm analysis between t(4;14) and non-t(4;14) MM samples was performed using the DSS Bioconductor package. KMS11 isogenic cell lines which express either high (NTKO) or low (TKO) levels of NSD2 were profiled for chromatin marks using cleavage under targets and tagmentation (CUT&Tag) and DNAm using WGBS. Three t(4;14) MM cell lines (KMS11, KMS18, H929) were treated with KTX-1031, a catalytic inhibitor of NSD2. Currently, KTX-1001 is a first-in-class catalytic NSD2 inhibitor in a phase 1 clinical trial (NCT05651932). Following 8 days of treatment, chromatin modifications, DNAm, and gene expression were assayed using CUT&Tag, WGBS, and RNAseq. DNMT3B protein levels were assayed by western blot. Results: Analysis of CoMMpass samples showed higher levels of global DNAm in the t(4;14) subtype compared to other subtypes (p = 2.9e-8) and >2M differentially methylated loci (DML) (FDR < 0.01), with 92% of DML having higher DNAm in t(4;14) MM samples. A correlation analysis between DNAm and gene expression determined that ~232K DML correlated with RNA levels of the nearest gene (FDR < 0.01). These RNA-correlated DML were enriched near genes identified by Zhan et al. (Blood, 2006) as defining the t(4;14) gene expression subtype (OR = 55.9, p < 1e-16), suggesting that changes in DNAm at these regions may facilitate the t(4;14) MM gene expression program. Epigenetic profiling of KMS11 NTKO (NSD2-high) and TKO (NSD2-low) cells showed higher global DNAm in NTKO cells (p < 1e-9). Furthermore, regions of elevated H3K36me2 in NTKO cells were enriched for increases in DNAm (OR = 1.85, p < 1e-15). Treating t(4;14) MM cell lines with the NSD2 inhibitor KTX-1031 reversed the chromatin changes observed in t(4;14) patients, resulting in reduced H3K36me2 and increased levels of the EZH2 modification H3K27me3. These epigenetic changes corresponded with downregulation of 539 genes and upregulation of 133 genes (FDR < 0.05) in KMS 11 t(4;14) cells. While the NSD2 inhibitor globally depleted H3K36me2 and elevated H3K27me3, these changes were most pronounced proximal to and in the gene bodies of downregulated genes. NSD2 inhibition for 8 days significantly reduced levels of DNMT3B protein, but only slightly reduced global DNAm (~2%). Importantly, the NSD2 inhibitor reversed much of the t(4;14) transcriptional program described by Zhan et al. (FDR = 5.5e-4). Furthermore, gene set enrichment analysis demonstrated that the NSD2 inhibitor resulted in reduced expression of genes involving NFKB (FDR < 0.001) and KRAS (FDR < 0.05) signaling. Conclusions: Through analyses of patient and cell line data, we provide evidence that NSD2-mediated increases in H3K36me2 directly dysregulate DNAm in t(4;14) MM, leading to DNAm hypermethylation as compared to non-t(4;14) MM. Many DNAm changes are correlated with transcriptional changes and occur primarily near disease-relevant genes. Treatment with catalytic NSD2 inhibitor for 8 days reverses much of the chromatin and gene expression changes associated with t(4;14) MM. NSD2 inhibition also reduces DNMT3B protein levels after 8 days and may disrupt the t(4;14) DNAm hypermethylation program after more prolonged treatment, resulting in further gene expression changes. We will present these results alongside data from extended NSD2 inhibitor treatments. We posit that elevated NSD2 in t(4;14) MM causes excessive H3K36me2 leading to relative DNA hypermethylation and transcriptional dysregulation which can be reversed through extended pharmacologic inhibition of NSD2.