Introduction:Multiple myeloma (MM) is a disease of malignant plasma cells, characterized by high CD38 expression. Although the CD38-targeting monoclonal antibodies are highly effective, resistance invariably arises. Tumor CD38 levels decrease after anti-CD38 therapy, but the expression is rarely permanently silenced. This suggests that CD38 expression may offer a tumor cell survival advantage, but the direct impact of CD38 loss on tumor dynamics has not been extensively characterized.Methods:CD38 knockout (KO) cell lines were generated by CRISPR-Cas9. Immunocompetent Balb/c and immunodeficient NSG mice were injected subcutaneously with either non-targeting (NT) or CD38 KO J558 cells. Stromal adhesion was compared using labeled NT and KO cells, with OP-9 murine stroma cells. Cellular NAD content was quantified using the Promega Glo Assay. Mitochondria were isolated with the Mitochondria Isolation Kit (Thermo Scientific). Oxygen consumption rate (OCR) and extracellular acidification rate (ECAR) were quantified using the Seahorse Assay. Response to hypoxia was evaluated using a modular hypoxic chamber. Cell cycle was quantified using propidium iodine staining.Results:To examine the role of CD38 in murine models, we utilized the CD38-expressing, murine plasmacytoma cell line J558. Strikingly, CD38 KO cells injected into Balb/c mice demonstrated significantly decreased tumor volume compared to NT (113 mm 3 (KO) vs. 1293 mm 3 (NT) at day 25, p <0.001). In contrast, in vitro cell proliferation and colony formation between KO and NT J558 cells were nearly identical, suggesting that the effects of CD38-loss were highly context dependent. Since tumoral CD38 expression may negatively modulate the immune response, we next compared CD38 KO and NT cells injected into immunodeficient NSG mice. CD38 KOs demonstrated an approximately 2.2-fold decreased tumor volume compared to the NT (708 mm 3 (KO) vs. 1592 mm 3 (NT), p=0.07). Further examination of the role of CD38 on the immune microenvironment are ongoing. Considering that some tumor growth impairment was maintained in immunodeficient mice, we next interrogated the effect of CD38 loss on other aspects of cell proliferation using J558 as well as human MM cell lines RPMI-8226 and NCI-H929. Daratumumab induced CD38 internalization has been shown to reduce stromal adhesion of MM cells. Similarly, CD38 KO cells demonstrated reduced stromal adhesion (2.5-fold decrease for J558, p<0.005 and 2-fold decrease for H929, p<0.005). Although stroma is a known promoter of cell survival and proliferation, we further questioned whether the NAD-metabolizing activity of CD38 modulates tumor growth. CD38 overexpression can drive down intracellular NAD and impair mitochondrial biogenesis. Accordingly, we found significantly higher NAD levels in the KO J558 tumor cells compared to NT (2-fold change, p <0.05). Additionally, CD38 KO cells demonstrated significantly higher levels of mitochondrial protein compared with the NTs (5-fold in J558 and 2-fold in H929). CD38 KO cell lines also showed markedly increased metabolic activity, with nearly 2-fold increase in basal OCR and ECAR, as well as in spare respiratory and glycolytic capacity. Given the contrast between in vivo and in vitro growth capacity, we questioned whether changes in mitochondrial content and metabolic function could confer an advantage for CD38-expressing cells under conditions of hypoxia, which is an important characteristic of the tumor microenvironment. Strikingly, under hypoxia, but not normoxia, CD38 KO MM cells demonstrated significantly more cell cycle arrest, defined by G0/G1 blockage (p=0.003 for H929 and p=0.004 for RPMI).Conclusion:We have shown that CD38 KO cells demonstrate decreased tumor growth in vivo but not in vitro. While the immune modulatory potential of CD38 is recognized, some of the growth impairment we observed may be explained by non-immune mediated mechanisms such as reduced stroma adherence as well as changes in cell metabolism. Loss of CD38 was associated with increased mitochondrial respiration, but also elevated ECAR and glycolytic rate. Higher reliance on mitochondrial respiration could explain impaired CD38 KO proliferation rates under hypoxia, possibly as a result of increased generation of reactive oxygen species. DisclosuresGhiaur: Menarini Richerche: Research Funding; Syros Pharmaceuticals: Consultancy.