Expression Signature of Myeloma Residual Cells Is Characterized By Genes Associated with Proliferation, Epigenetic Modification, and Stem Cell Maintenance

Abstract Background Despite significant progress in the treatment of multiple myeloma (MM), 30% of patients relapse within 3 years of diagnosis. Recent data indicate that intervention before relapse would benefit most patients with high-risk MM and a significant proportion (38%) of low-risk MM patients. Residual MM cells are responsible for relapse, have been described as drug resistant and as having stem cell-like characteristics including altered cell metabolism, increased drug efflux, ALDH1 activity, and myeloma propagating activity. We sought to molecularly characterize residual myeloma plasma cells (MMPCs) and identify novel biomarkers and therapeutic targets of resistant MM cells. Methods Patient samples. Residual samples were collected at least 3 months after start of therapy (range 4 mths-11yrs), had low tumor burden (0.6-4.4% BM MMPC), 2 were in VGPR or better, 7 had PR, and 3 had not responded. Of the 12 cases analyzed, 2 had relapsed > 1 yr prior to residual sample. RNAseq analysis. Paired MMPCs at diagnosis and during residual disease (n=12) and 4 healthy donors were interrogated by RNAseq. Genes with at least 2-fold difference between groups and at a statistical significance <0.05 were investigated further. String pathway analysis was used to detect pathways and biological functions enriched in resistant myeloma cells. The drug gene interaction database (DGIdb) was used to identify known and potential drug targets. Results RNAseq shows that paired diagnostic and residual samples cluster together. Of genes differentially expressed in residual disease, there was reduced expression of cell adhesion markers (e.g. CXCR1, CXCR2, CEACAM3, CEACAM8, ITGB3, ITGAM), tumor suppressors (e.g. PPARG, TP53I11) and metabolic genes (e.g. HK3, PYGL) and increased expression of genes associated with proliferation (e.g. KIF11, OIP5, MAD2L1, BIRC5, ZWINT, MKI67), chromosomal instability (e.g. TOP2A, KIF4A, PBK, RRM2), epigenetic modification (e.g. EZH2, DNMT3B, HMGN5), high centrosome index (e.g. TYMS, MAD2L1, TOP2A, BIRC5, ZWINT, PLK4), DNA repair (e.g. FANCI, FANCB, RAD51, MND1, BRCA1) and stem cell activity such as Notch signaling, ALDH1 expression, and drug efflux (e.g. DLL4, KIF14, TOP2A, BIRC5). We identified cell surface markers EPHA5, TRPC4, MAGEA1, MAGEA12, GAGE1, NLGN1, SVIL, CCR9, KCNK2, PLK4, CIT, and DMD that will be tested for their ability to identify resistant MMPCs in future studies. Interrogation of the DGIdb for drugs targeting genes upregulated in residual MMPCs revealed 36 targets with known drugs available and an additional 23 targets that are potentially druggable (i.e. cell surface, kinase, transporter, DNA repair etc.). To better understand the mechanisms involved in maintaining residual disease, genes that were exclusively upregulated in residual cells were entered into the STRING bioinformatics platform for pathway analysis. This revealed an interaction network of 186 nodes (genes) and 599 edges (molecular interactions), with the average node degree (the number of edges connected to the node) being 6.44. The clustering coefficient and protein-protein interaction enrichment P-value were 0.374 and 1x10-16, respectively.The genes expressed in residual MMPCs were enriched in cell cycle pathways including, chromosome segregation, nuclear division, mitotic cell cycle, positive regulation of cell cycle, cell division, DNA conformation change, DNA metabolic process, DNA packaging, and chromosome organization (FDR < 1%). Conclusions We have identified a gene expression signature specific to resistant myeloma which may be used for early detection of relapse and development of novel therapeutic targets. Interestingly, many genes associated with high risk in myeloma were expressed in residual samples despite the majority of cases being molecularly-defined low risk at diagnosis (11/12), suggesting that residual myeloma harbors a high risk expression signature without gain of genomic aberrations. Identifying markers of resistance early in the disease course will allow treatments to be modified to prevent relapse and may reduce long-term exposure to chemotherapeutics. Disclosures Epstein: University of Arkansas for Medical Sciences: Employment. Davies:Abbvie: Consultancy, Honoraria; Takeda: Consultancy, Honoraria; Janssen: Consultancy, Honoraria; Amgen: Consultancy, Honoraria; Celgene: Consultancy, Honoraria. Morgan:Takeda: Consultancy, Honoraria; Celgene: Consultancy, Honoraria, Research Funding; Janssen: Research Funding; Bristol-Myers Squibb: Consultancy, Honoraria.