Metabolism-focused CRISPR screen in B cell tumor identifies sphingolipid metabolism promotes resistance to CAR-T cell therapy Free

Significance Chimeric antigen receptor (CAR) T cell therapy has revolutionized the management of relapsed/refractory B-cell acute lymphoblastic leukemia. However, intrinsic or acquired resistance to CAR-T cell therapy remains a major obstacle for successful B-ALL treatment, accounting for both initial treatment failures and relapses. To overcome these challenges, it is of vital to understand the underlying mechanisms of resistance to CAR-T cell therapy. Here, we show that targeting sphingomyelin synthase 1 (encoded by SMGS1) enhances B-ALL susceptibility to CAR-T cell therapy via disrupting sphingomyelin-ceramide homeostasis and subsequently amplifying death receptor signaling mediated on-target and off-target tumor destruction.

Results We employed a metabolism-focused CRISPR loss-of-function screen on B cell malignancy to identify candidate metabolic genes that modulate susceptibility to CD19 CAR-T cells. To resemble long-term exposure of tumor cells to CAR-T cells, NALM6 cells with sgRNA were cocultured with CD19 targeted CAR-T cells or UTD T cells for five rounds. Resistant living NALM6 cells were harvested for targeted sequencing with control cells. KEGG pathway analysis revealed various metabolic pathways that could account for resistance to CAR-T elimination, with top enrichment of sphingolipid metabolism. Among the most depleted genes, SGSM1 involved in sphingolipid metabolism ranked as one of the top hits.

To confirm the role of SMGS1 and sphingolipid metabolism in mediating CAR-T resistance, we generated SMGS1 disrupted NALM6 and SEM cells. Depletion of SGMS1 in NALM6 and SEM had no effect on the CD19 expression as well as on cellular proliferation and survival, but significantly increased sensitivity to CD19 CAR-T treatment. We engrafted NSG mice with control or SGMS1^KO NALM6, and then infused with UTD T cells or CD19 CAR-T cells. Consistently, SGMS1 deficiency had no impact on tumor growth in vivo, but significantly delayed tumor relapse and prolonged survival of CAR-T treated recipients.

We next sought to determine the underlying mechanism by which SGMS1 mediates tumor immune evasion to CAR-T cell. We found caspase-8 cleavage was increased in SGMS1^KO tumors after exposure to CD19 CAR-T cells and pan-caspase inhibitor could abolish this pro-apoptotic effect of SGMS1^KO. As the ligands of death receptors can initiate caspase-8 activation, we found both purified FasL and TRAIL could induce higher level of apoptosis in SGMS1^KO tumor cells. Furthermore, SGMS1 depletion fails to induce tumor susceptibility to CAR T cells if the death-adaptor protein FADD is deficient. To verify that death receptor signaling is involved in CAR-T cell cytotoxicity, we generated FasL^KO and TRAIL^KO CAR-T cells using electroporation. Notably, neither FasL^KO nor TRAIL^KO CAR-T cells can elicit greater tumor lysis in SGMS1^KO NALM6.

To elucidate the biochemical consequence of SGMS1 depletion in B cell tumor, we performed targeted lipidomics analysis. The results show a broad increase of various ceramides, and a decrease of most sphingomyelin species in SGMS1^KOtumor cells. To further explore whether directly intervening cellular ceramide amount could sensitize B cell tumors to CAR-T cell therapy, we added exogenous ceramide to NALM6 and SEM cell cultures, and observed enhanced tumor susceptibility to CAR-T cells with increased cleavage of caspase-8.

Death receptor signaling has been shown to play critical role in off-target tumor killing in T-cell immunotherapy, we next sought to determine whether targeting sphingolipid metabolism could be leveraged to prevent antigen-escape after CAR-T treatment. CD19^KO NALM6 cells were generated to present relapsed tumor cells with CD19-loss. To model in vivo therapy with antigen-loss escape, we inoculated heterogeneous (90% WT + 10% CD19^KOor 10% CD19^KOSGMS1^KO NALM6) tumors prior to CAR-T therapy, and found SGMS1 deficiency slowed down the relapse of antigen-negative tumor and prolonged recipient survival.

Conclusion Overall, our study identifies SGMS1 mediated sphingolipid metabolism dampens the sensitivity of B cell tumors to CAR-T cell therapy. Mechanistically, SGMS1 reduces ceramides levels in tumor cells to desensitize death receptor signaling initiated by FasL/TRAIL provided by CAR-T cells. Targeting this metabolic activity could be utilized to enhance efficacy of CAR-T therapy in primary treatment and reducing antigen-loss related relapse.