Abstract 6224: Exploring the integration of CMS subtyping and other omics in colorectal cancer classification and its reflection in stool miRNome and metagenome for non-invasive precision medicine

Abstract

Abstract Background: The Consensus Molecular Subtypes (CMS) is a subtyping system used for colorectal cancer (CRC) to facilitate traditional tumor classification and clinical translation. Relying on RNA-seq data, the system classifies patients into four molecular subtypes (CMS 1-4). Each subtype is enriched with several mutations, but no genetic aberrations are peculiar to a single CMS. A more in-depth characterization of each subtype's genomic and epigenomic background by integrating different omics may help in refining this classification, including the missing assignment of a subgroup of tumors. The possible translation of this improved classification to surrogate tissues, such as stool, has never been investigated. Aims: We aim to comprehensively characterize the CMS by the integrative analysis of target genomic sequencing, RNA-Seq, small RNA-Seq, and immune infiltrate profiling data of a cohort of CRC patients. These data will be used to identify integrative CMS-specific signatures and to characterize those tumors not assigned to any CMS. In addition, differences in stool miRNome and metagenomic profiles will be explored to identify potential specific signatures characterizing each subtype. Methods: Tumor tissue and adjacent mucosa were collected from 87 CRC patients and preserved in RNA later. mRNA and other non-coding RNA expression levels were evaluated with ribosomal depleted RNA-seq and small RNA-seq. Tumors were assigned to a CMS subtype using the CMScaller algorithm and genomically characterized with the TruSight Oncology 500 (TSO500) cancer panel. Differential expression analyses were performed using the DESeq2 package for the R software to identify CMS-specific expression signatures. Small RNA-seq and shotgun metagenomic analyses were performed on stool samples collected from the same patients. Preliminary Results: Tumors were assigned to CMS1 (n=9), CMS2 (n=24), CMS3 (n=26), and CMS4 (n=22), except for six samples. Differential expression analyses between paired tumor samples and adjacent mucosa within each CMS revealed 1096 differentially expressed genes in CMS1, 4218 in CMS2, 27 in CMS3 and 190 in CMS4. From the genomic point of view, CMS3 cases showed 384 unique missense variants, while CMS1 included almost all samples with microsatellite instability. Small RNA-seq on the same tissue sample identified 380 dysregulated miRNAs (215 up and 165 downregulated) in the tumor compared to adjacent mucosa. Conclusions: A more in-depth CMS subtyping by integrating different omics analyses could help refine this classification system. Moreover, the possible CMS reflection in stool miRNome and metagenomic could shed new light on the host gut-microbiota interactions and allow a non-invasive classification applicable for a routine personalized clinical management of CRC. Citation Format: Amedeo Gagliardi, Giulia Francesca, Giulio Ferrero, Sonia Tarallo, Giulia Piaggeschi, Antonio Francavilla, Carla Di Battista, Enrico Berrino, Caterina Marchiò, Francesca Cordero, Barbara Pardini, Alessio Naccarati. Exploring the integration of CMS subtyping and other omics in colorectal cancer classification and its reflection in stool miRNome and metagenome for non-invasive precision medicine [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 6224.