Project summary This application is being submitted in response to the Notice of Special Interest (NOSI) identified as NOT-CA- 24-029. For colorectal cancer (CRC), the efficacy of current immunotherapies like checkpoint inhibition has been disappointing, and scores of new strategies are being developed to bolster efficacy by targeting the numerous features of the TIME that can impede robust and lasting anti-tumor immune responses. However, we have little understanding of which features dominate the response to any given class of immunomodulators since they all work together in complex ways and we don’t have retrospective patient data to probe since most of these are not yet in clinical trials. Our overall goal here is to understand how various patient- specific features of colorectal cancer (CRC) and its tumor-immune microenvironment (TIME) govern differential responses to different types of immunotherapies using our novel in vitro ‘avatars’ of the tumor- immune circuit (‘TIvitars’) that we developed in our parent grant. To accomplish this, we are partnering with clinician-scientist Prof. Ben Shogan, who runs a CRC biobank and collects corresponding stool samples for microbiome analysis. We will establish and perform immunotherapy screens using human CRC tissues from 20 patients comparing responses of each tumor to a broad range of immunomodulators that can (a) compare efficacy among numerous therapies for each patient’s tumor in a response-predictive way, and (b) correlate these responses with the molecular and cellular characteristics of each tumor (obtained from standard -omics analysis of the samples) as well as with the patient’s microbiome (obtained from sequencing stool samples). We hypothesize that by interrogating multiparameter -omics data and clinical data using immunotherapy responsiveness data collected using TIvitar, we will identify tumor-intrinsic features that contribute to immunotherapy responsiveness in CRC patients. In Aim 1, we will systemically characterize each patient’s tumor, gut microbiota, and systemic immune status. For each patient, we will determine dominant tumor mutations, tumor gene expression patterns, characteristics of the immune infiltrate, stromal features, systemic immune status, and microbiota populations from fecal samples collected prior to surgery. In Aim 2, we will use TIvitar to determine each patient’s response to a screen of different immunomodulators. In Aim 3, we will pair the immune response data with the -omics data collected in Aim 1 to identify shared tumor, immune, or microbiota features that could predict susceptibility to immunotherapy. The resulting data will form a completely unique dataset whereby standard -omics data from cancer patients will be coupled with differential immunotherapy responses for each.