Ketogenic Diet Enhances CAR T Cell Antitumor Function Via β-Hydroxybutyrate

Introduction. Several mechanisms of resistance have been identified for chimeric antigen receptor (CAR) T cell therapy; however, there are limited data on the impact of lifestyle factors, such as diet, on CAR T cell efficacy. Indeed, diet-derived metabolites can modulate T cell functions through the regulation of metabolic, epigenetic, and transcriptional processes. Here, we investigated whether diet can affect responses to CAR T cell treatment.

Methods and Results. To this end, we implanted diffuse large B-cell lymphoma (DLBCL) tumors (A20) into immunocompetent mice (Balb/c), and fed them one of five representative diets, including ketogenic, high-fiber, high-fat, high-protein, Western (i.e., high cholesterol), and a macro- and micronutrient content-matched control diet. Upon tumor engraftment, we infused murine anti-CD19 CAR T cells (CART19). Mice fed a ketogenic diet showed improved tumor control and overall survival relative to all other screened diets. As expected, the level of β-hydroxybutyrate (BHB), the principal metabolite produced during ketosis, was significantly higher in ketogenic diet-fed mice compared to other groups. Since effector T cells can utilize BHB to fuel the citric acid cycle, we hypothesized that BHB can serve as a more efficient energy source than glucose for active CAR T cells, leading to more pronounced effector responses.

Therefore, we tested the effect of BHB alone on several xenograft cancer models in immunodeficient NOD-SCID gamma chain-deficient (NSG) mice. First, DLBCL tumors were implanted into NSG mice (5x10⁶ CD19⁺ OCI-Ly18 cells, subcutaneous), and BHB was oral gavaged daily. On day 11 post-implantation, a suboptimal dose of CART19 (3x10⁶ CAR⁺ cells) was infused. Mice receiving both oral BHB and CART19 demonstrated potent tumor control (complete response, CR 6/7) compared to mice receiving a vehicle gavage (CR 1/7). BHB+CART19-treated mice also showed higher peripheral CAR T cell expansion and elevated serum effector cytokines. Furthermore, BHB given ad libitum in the drinking water led to improved tumor control of anti-mesothelin CAR-T in AsPC-1 (pancreatic cancer, subcutaneous) and anti-CD19 CAR-T in Nalm6 (B cell acute lymphoblastic leukemia, intravenous) xenografts.

Mechanistically, using carbon-13 (¹³C) isotope labeling, we found that antigen-activated CAR T cells preferentially integrate carbon from BHB into the citric acid cycle components over that of standard glucose. This process fuels CAR-T oxidative phosphorylation and increases their oxygen consumption, as confirmed by the Seahorse Assay (Agilent Seahorse XF). To determine whether BHB modulates epigenetic accessibility in CAR T cells, we performed an assay for transposase-accessible chromatin with sequencing (ATAC-seq) on CAR T cells treated with or without BHB. We found that BHB-treatment exposed crucial effector and memory genes (e.g., FOXO1, TCF7, GZMB) in CAR T cells. This likely reflects increased Acetyl-CoA from BHB entering the CAR-T nucleus and acetylating histones. Finally, we found that deleting BDH1, the key enzyme involved in BHB metabolism to Acetyl-CoA, in CAR-T reduces the functional benefit of BHB. This further confirms that BHB enhances CAR-T functions through metabolic modulation.

To test the translational relevance of these findings, we retrospectively analyzed CART19 patient serum (n=17, all LBCL) taken 7 days post-infusion from the NCT02030834 trial using mass spectrometry. In line with our preclinical results that BHB can support CAR-T proliferation, BHB serum concentration positively correlated with CART19 expansion (P = 0.0164). Lastly, we produced CAR T cells from the leukapheresis products of two CART19 patients. We found that providing BHB in the cell culture media during ex vivo manufacturing boosted patient T cell proliferation, resulting in a ~4X increase in absolute cell count.

Conclusions. Our results demonstrate that ketogenic diet-derived BHB can be provided as a dietary intervention to augment CAR-T function in multiple cancer models. The key mechanism involves the enhancement of the citric acid cycle and histone modifications during CAR T cell activation. The results of this study will be translated into a first-in-human clinical trial of BHB-supplementation during CART19 treatment for relapsed or refractory B cell lymphoma.

*S. Liu and P. Guruprasad contributed equally.

**M. Ruella and M. Levy jointly supervised this work.