Enriched pathways in gut microbiome predict response to immune checkpoint inhibitor treatment across demographic regions and various cancer types

Abstract

Understanding the effect of gut microbiota function on immune checkpoint inhibitor (ICI) responses is urgently needed. Here, we integrated 821 fecal metagenomes from 12 datasets to identify differentially abundant genes and construct random forest models to predict ICI response. Gene markers demonstrated excellent predictive performance, with an average area under the curve (AUC) of 0.810. Pathway analyses revealed that quorum sensing (QS), ABC transporters, flagellar assembly, and amino acid biosynthesis pathways were enriched between responders (R) and non-responders (NRs) across 12 datasets. Furthermore, luxS, manA, fliC, and trpB exhibited consistent changes between R and NR across 12 datasets. Follow-up microbiota transplant experiments showed that inter-species signaling by different QS autoinducer-2 (AI-2) molecules (synthesized by luxS) can act on overall community function to promote the colonization of Akkermansia muciniphila, which is associated with superior ICI responses. Together, our data highlight the role of gut microbiota function in modulating the microbiome and antitumor immunity.