Abstract 4435: A causal modeling platform for testing lifestyle interventions on the microbiome and response to immunotherapy

Abstract

Abstract Rational manipulation of the gut microbiome by diet or other interventions is a promising approach to improving the efficacy and safety of cancer immunotherapy. However, linking lifestyle variables to specific microbial populations and the host immune response to cancer therapies is challenging due to their complexity. We established an experimental system to examine such complexity by linking a human dietary intervention and fecal biospecimens to mouse models to test the response to immunotherapy. Longitudinal human microbiome samples collected before and after a specific dietary or other intervention allow us to test the hypothesis that an intervention alters the response to immunotherapy via impact on the host microbiome. One set of samples was taken from the BEWELL study, which examined the impact of 2 × 80 mL black raspberry (BRB) drink boxes per day for 4 weeks in people at high risk for lung cancer. Participant samples were chosen based on enrichment of specific taxa. Pre- and post-BRB intervention samples were gavaged into mice. Mouse mc38 cells were injected subcutaneously and treated with anti-PD1 Ab or isotype control. Tumor size was monitored, and at the end of the study, tumor immune cell composition data were collected. Tumor growth over time was modeled using a linear mixed-effects model with tumor volume as the outcome variable and the predictor variables of time and treatment (PD1 vs. IgG) interacting with gavage and quadratic time to accommodate non-linear tumor growth and including random effects by mouse. For sample 68, enriched for Roseburia CAG 309, the interaction between time, gavage post-BRB, and Anti-PD1 treatment significantly affected tumor volume (p < 0.05) relative control. The same significant effect (p < 0.05) was also seen for sample 79, enriched for Lachnospira pectinoschiza, and samples 84 and 85, enriched for Blautia obeum. To determine the mechanism by which this might occur, we analyzed the tumor's immune cell composition. Post-BRB mice treated with Anti-PD1 showed a significant increase in tumor-infiltrating CD8+ immune cells relative control, as compared to mice gavaged with pre-BRB stool. Modeling results showed samples whose response was improved slightly after the BRB dietary intervention. These samples were associated with the enrichment of the taxon Roseburia CAG 309. A black-raspberry dietary intervention in humans modified the microbiomes of several participants in a way that is hypothesized to improve response to PD1 treatment. Future directions include supplementing individual microbes into pre-intervention gavages to confirm which taxa improve response. These results suggest that this modeling platform is an effective system for testing microbiome modification on tumor growth and assessing mechanisms by which this might occur and is extendable to other lifestyle-based interventions where pre- and post-intervention specimens are collected. Citation Format: Aaditya Pallerla, Bailey Conrad, Amna Bibi, Nyelia Williams, Caroline Wheeler, Rebecca Hoyd, Shankar Suman, Joseph Amann, Yangyang Liu, Marisa Bittoni, Shiqi Zhang, Madison Grogan, Alvin Anand, Najma Afrah, Carolyn Presley, Fred K. Tabung, Lang Li, Yael Vodovotz, Jiangjiang Zhu, David P. Carbone, Steven K. Clinton, Daniel Spakowicz. A causal modeling platform for testing lifestyle interventions on the microbiome and response to immunotherapy. [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 4435.