Abstract NG07: Association between somatic microsatellite instability, hypermutation status, and specific T cell subsets in colorectal cancer tumors
Claire Elizabeth Thomas,Yasutoshi Takashima,Tomotaka Ugai,Daniel D. Buchanan,Conghui Qu,Li Hsu,Andressa Dias Costa,Stephen Gallinger,Robert C. Grant,Jeroen R. Huyghe,Sushma S. Thomas,Robert S. Steinfelder,Shuji Ogino,Amanda I. Phipps,Jonathan A. Nowak,Ulrike Peters
DOI: https://doi.org/10.1158/1538-7445.am2024-ng07
IF: 11.2
2024-04-07
Cancer Research
Abstract:Background: Colorectal cancer (CRC) is a critical public health concern as the third most commonly diagnosed cancer and second leading cause of cancer death globally. CRC is a heterogeneous disease with many different mechanisms and molecular subtypes. Around 15% of sporadic CRCs are classified as exhibiting high microsatellite instability (MSI-high) due to deficient DNA mismatch repair (dMMR). In the presence of MSI-high, colorectal tumors accumulate somatic mutations, resulting in high neoantigen burden, which, in turn, elicits a strong T cell response. Several previous studies have shown that MSI-high status is associated with overall levels of T cell infiltration in CRC and, specifically, with CD45RO+ T cell density. However, many studies have measured T cell densities using single-plex immunohistochemistry assays of individual T cell markers, which fail to capture the complexity of the T cell response in tumors, and the contributions of different T cell subsets. Using a multiplex immunofluorescence panel, we aimed to characterize the composition of the T cell response in CRC tumors in relation to MSI-high and hypermutation status in CRC. Methods: This study was conducted within a subset of participants from three well-characterized epidemiologic studies included in the Genetics and Epidemiology of Colorectal Cancer Consortium (GECCO): the Ontario Family Colon Cancer Registry (OFCCR), the Nurses' Health Study (NHS), the Health Professionals Follow-up Study (HPFS). In a total of N=637 participants with CRC, we profiled the in-situ T cell landscape of CRC using digital imaging, machine learning, and a multiplexed immunofluorescence panel. Our customized 9-plex panel included antibodies directed against CD3, CD4, CD8, CD45RA, CD45RO, FOXP3, and MKI67, as well as a pan-cytokeratin antibody to identify tumor/epithelial cells and nuclear DAPI (4', 6-diamidino-2-phenylindole) stain. Assays were conducted using tissue microarrays (TMAs) where all slides were sent to Dana Farber Cancer Institute for analysis. In brief, core selection from donor FFPE blocks was guided by pathologist review of H&E-stained slides. For each CRC case, TMA blocks have 2-4 cores (0.6 mm) from tumor areas. T cell immune microenvironment is assessed for each tumor histologically using a multispectral imaging platform (PerkinElmer Vectra 3.0) for multiplexed immunofluorescence (mIF), where TMAs are stained with all markers concurrently and imaged. With pathologist supervision, a machine learning algorithm segmented tissue into epithelial and stromal area regions as well as phenotyped individual cells. The mIF panel has been validated against traditional chromogenic IHC. Based on these markers, we quantified counts and densities of naive, memory, and regulatory subsets of helper (CD3+CD4+) and cytotoxic (CD3+CD8+) T cells, and looked separately at these subsets within epithelial and stromal tissue, resulting in 12 unique subsets. Microsatellite instability and hypermutation status were determined from one of two targeted sequencing panels that included either 205 or 298 genes. MSI status was called using mSINGS. Hypermutation status was defined by plotting point mutations for all samples and observing different peaks, separately by panel. The minimum value between the two peaks was 23 and 26 point mutations per sample, respectively, which were used as cut-points. We used multivariable ordinal logistic regression to estimate odds ratios (ORs) and 95% confidence intervals (CIs) for the association between MSI/hypermutation status with quantiles of specific T cell densities in CRC. As an alternative approach where counts were modeled directly, we used negative binomial count models with an offset term for area. Ordinal logistic regression either used quartiles or tertiles of T cell densities as the outcome, where different subsets were assessed for zero-inflation considering both epithelial and stromal tissue, so the same subset used the same categorization in each tissue. Subsets with less than 25% zero counts were divided into quartiles, subsets with greater than 25% zero counts into tertiles. Study-specific quantile cut points were used to limit potential batch effects across TMAs. Models were adjusted for age, sex, and study, and p-values were adjusted for multiple testing using the false discovery rate method for 12 independent tests. Results: Among 637 CRC tumors with targeted tumor sequencing and T cell immune profile available, 107 were classified as MSI-high, and 106 were classified as hypermutated. 98 tumors were both MSI-high and hypermutated. Compared to tumors with low microsatellite instability (MSI-low) or microsatellite stable (MSS) tumors, MSI-high tumors were strongly associated (ORs around 2 or higher) with higher T cell densities for all epithelial area CD3+CD4+, epithelial area CD3+CD8+, and stromal area CD3+CD8+ T cells (adjusted p-values <0.001). -Abstract Truncated-
oncology