Super-Enhancer Profiling Identifies Novel Oncogenes and Therapeutic Targets in Multiple Myeloma

Background: Multiple myeloma (MM) is an aggressive neoplastic plasma cell cancer characterized by diversely cytogenetic abnormalities. MM can be divided into subtypes with immunoglobulin heavy chain (IGH) gene translocations involving CCND1-3, FGFR3/MMSET, MAFs and hyperdiploid myeloma containing trisomies of several odd numbered chromosomes 3, 5, 7, 9, 11, 15, 19, and 21. Although several new drugs have been introduced into clinic, treatment for MM patients remains challenge and refractory/resistant to therapy is often seen. Thus, a better understanding of the molecular pathogenesis of MM can lead to generate new prognostic classification and identify new therapeutic targets. Super-enhancers (SEs) are defined as large clusters of cis-acting enhancers, marked by high level bindings of acetylation of histone H3 lysine 27 (H3K27ac) and mediator complex. SEs have been shown to control genes for maintaining cellular identity and also key tumor drivers in various malignancies.

Methods: H3K27Ac ChIP-seq and RNA-seq were performed on primary MM patient samples, MM cell lines. Normal plasma cells and lymphoma cell lines were served as controls. We systematically compared SEs and their associated genes of normal and cancerous tissue. THZ1, a CDK7 inhibitor, was used to efficiently down-regulate SE-associated genes. Combinatory analysis of THZ1-sensitive and SE-associated gene revealed a number of promising MM oncogenes. CRISPR/Cas9 technology and ectopic expression experiments in conjunction with cellular functional assays were performed to determine the effects of candidate SE-genes on MM cells. Circularized chromatin conformation capture followed by sequencing (4C-seq) was applied to explore the direct contact of SE and promoter.

Results: SE analysis uncovered some cell lineage-specific transcription factors (TFs) and known oncogenes in MM. Several key TFs, including IRF4, PRDM1, MYC and XBP1, were identified in most MM samples, confirming the origin of MM cells. These data reinforce the concept that SE establishment is a key component of MM biology. The acquisition of SEs around oncogene drivers is widely observed during tumorigenesis. ST3GAL6 and ADM were two known oncogenic drivers in myeloma cells, which were associated with super-enhancers in all MM samples but not in normal plasma cell and lymphoma cells. We also found SE constituents for multiple subtype-specific key oncogenes such as CCND1 in t(11;14) cells, C-MAF in t(14;16) cells, and NSD2 and FGFR3 in t(4;14) cells.

Furthermore, THZ1 showed prominent anti-neoplastic effect against MM cells. SE-associated genes were more sensitive to THZ1 compared with those genes associated with typical enhancers (TEs). By overlapping THZ1-sensitve gene with SE-associated genes, we identified a number of novel MM oncogenes, including MAGI2, EDEM3, HJURP, LAMP5, MBD1 and UCK2 as a potential druggable kinase. The expression level of MAGI2 and HJURP confers poor prognosis in several MM datasets. MAGI2 silencing in MM cells decreased cell proliferation and induced apoptosis. qRT-PCR and Western blot analysis confirmed the overexpression of HJURP in t(4;14) cells relative to non-t(4;14) MM cells. Furthermore, 4C-seq analysis revealed the physical interaction between HJURP-SE and promoter and THZ1 treatment diminished this interaction. Motif search at SE constituents revealed a highly significant enrichment of NSD2 recognition. Significant reduction of NSD2 binding at HJURP-SE region was observed in KMS11 infected with NSD2-specific shRNAs. Interestingly, blocking SE sites by CRISPR/Cas9i or silencing HJURP by shRNA led to decreased HJURP expression and cell apoptosis, whereas overexpression of this gene promoted cell growth. Taken together, our data demonstrated that HJURP is a novel SE-associated oncogene in t(4;14) MM.

Conclusions: Our integrative approaches by combing H3K27Ac ChIP-seq, RNA-seq and THZ1-sensitive transcript defined the landscape of SE and identified SE-associated novel oncogenes, as well as lineage-specific TFs in MM. Furthermore, we also revealed subtype-specific SE-driving oncogenic program in MM. Taken together, these results not provide novel insight into the MM pathology, but also offer novel, potential therapeutic targets, such as MAGI2, and HJURP for the treatment of MM patients.