Gut microbiome regulates immunotherapy response through systemic immune environment in patients with hepatocellular carcinoma

Abstract

Background Immunotherapy including immune checkpoint inhibitors (ICIs) has emerged as a key pillar of anticancer treatment for hepatocellular carcinoma (HCC). However, patients respond to treatment heterogeneously, and currently biomarkers to predict treatment response are limited. Gut microbiota plays an important role on modulating immune systems, which has a potential to serve as the biomarker for immunotherapy response. We hypothesized that gut commensal bacteria can impact on systemic immune that are important for treatment responses to immunotherapy in patients with HCC. To be specific, we aimed to characterize the gut microbiome and systemic immune environment and their prospective roles on outcomes in HCC patients treated with immunotherapy and establish their causal links.

Materials and Methods HCC patients treated with ICIs were eligible. Feces samples were collected for metagenomic sequencing. The baseline peripheral immune subtypes and cytokines were measured using flow cytometry and Luminex technique. Treatment responses were classified as responders (Rs), including complete response (CR), partial response (PR) or stable disease (SD) > 6 months, and non-responders (NRs), i.e., progression disease (PD) within 6 months. Response was evaluated according to the modified Response Evaluation Criteria for Solid Tumors (mRECIST). Rs were further subclassed as long-responders (LRs), CR, PR or SD > 12 months and short-responders (SRs), CR, PR or SD of 6-12 months. The linear discriminant analysis effect size (LEfSe) analysis was used to detect significantly diverse microbial species in Rs and NRs. The causal relationships were verified in the fecal microbiota transplantation (FMT) mice model.

Results A total of 63 patients were enrolled, with 48 Rs and 15 NRs. Clostridiaceae family, Clostridium and Citrobacter genus were significantly enriched in Rs, while Phocaeicola vulgatus was abundant in NRs. The microbial compositions of 33 LRs were significantly different from those of non-LRs (15 NRs and 8 SRs) (p = 0.024), which were characterized with higher abundance of Bifidobacterium adolescentis, Klebsiella pneumoniae and Klebsiella genus. The baseline blood immune profiles were compared between Rs and NRs, and Rs exhibited significantly higher peripheral Th17 cell proportion compared with NRs (p = 0.025), and the LRs showed higher proportions of CD71+ (p = 0.035), ICOS+ (p = 0.020) and HLA-DR+PD-1+ (p = 0.015) CD8+T cells, natural killer T (NKT) cells (p = 0.025), and conventional dendritic cells (cDCs) (p = 0.002) than that of non-LRs. In paired blood and fecal samples, the abundance of Bifidobacterium adolescentis was positively correlated with the proportion of peripheral CD8+T and NKT cells, and the expressions of function markers on CD8+T cell, including CD71, CX3CR1 and ICOS. Furthermore, the FMT from Rs not NRs combined with anti-PD-1 significantly inhibited tumor growth, along with upregulation of intratumoral expressions of IFN-γ, GzmB and Ki-67 on CD8+T cells.

Conclusions The gut microbiota enriched in Bifidobacterium adolescentis was correlated with long-term benefit of ICIs in HCC patients, likely via inducing anticancer immunity like activated CD8+T cells. The gut microbiota profiles of Rs can be transferred by FMT to improve treatment response, suggesting a potential intervention target for immunotherapy.