Indiana University-Purdue University Indianapolis (IUPUI) / Multiple myeloma (MM) is the most frequent cancer to involve the skeleton, with over 80% of myeloma patients developing lytic bone disease (MMBD). Importantly, MM-associated bone lesions rarely heal even when patients are in complete remission. Bone marrow stromal cells (BMSCs) isolated from MM patients have a distinct genetic profile and an impaired osteoblast (OB) differentiation capacity when compared to BMSCs from healthy donors. Utilizing an in vivo model of MMBD and patient samples, we showed that BMSCs from tumor-bearing bones failed to differentiate into OBs weeks after removal of MM cells. Both Runx2 and Osterix, the master transcription factors for OB differentiation, remained suppressed in these BMSCs. However, the molecular mechanisms for MM-induced long-term OB suppression are poorly understood. We characterized both Runx2 and Osterix promoters in murine pre-osteoblast MC4 cells by chromatin immunoprecipitation (ChIP). The transcriptional start sites (TSSs) of Runx2 and Osterix in untreated MC4 cells were co-occupied by transcriptionally active histone 3 lysine 4 tri-methylation (H3K4me3) and transcriptionally repressive histone 3 lysine 27 tri-methylation (H3K27me3), termed the “bivalent domain”. These bivalent domains became transcriptionally silent with increasing H3K27me3 levels when MC4 cells were co-cultured with MM cells or treated with TNF-α, an inflammatory cytokine increased in MM bone marrow microenvironment. The increasing H3K27me3 levels induced by MM cells or TNF-α were associated with the downregulation of the H3K27 demethylase JMJD3 in MC4 cells and murine BMSCs. Knockdown of JMJD3 in MC4 cells was sufficient to inhibit OB differentiation. Further, ectopic overexpression of JMJD3 in MC4 cells partially rescued the suppression of osteoblast differentiation induced by TNFa. We also found that pre-incubation of MC4 cells with the NF-kB inhibitor quinazoline (QNZ) before TNF-a treatment prevented the downregulation of JMJD3. In agreement with our in vitro findings, BMSCs from MM patients had persistently decreased JMJD3 expression compared to healthy BMSCs. Our findings together demonstrate that decreased JMJD3 expression in BMSCs contributes to the long-term OB suppression in MMBD by remodeling histone landscapes at the Runx2 and Osterix TSSs. Thus, developing strategies to restore JMJD3 expression in BMSCs should increase bone formation and possibly decrease tumor burden in MM.