Research Supplements to Promote Diversity in Health-Related Research: NIH - R01 Diversity Supplement (GM134307)

This research project aims to investigate how African ancestry and mismatch-repair (MMR) enzyme deficiency/microsatellite instability-high (MSI-H) status associate with immune-related gene expression in colorectal adenocarcinoma (COAD/READ). Motivated by prior reports of ancestry-related differences in immune gene expression and by the clinical importance of MMR/MSI status for immune checkpoint inhibitor eligibility, the study used a two-pronged analytical approach combining public TCGA bulk RNA-sequence data with targeted RT-qPCR validation on an independent, ancestry-balanced set of formalin-fixed paraffin embedded (FFPE) tumor samples enriched for MMR-d/MSI-H cases. In the TCGA-COAD-READ cohort, multivariable linear regression models were applied to identify immune genes associated with African ancestry (AA) versus European ancestry (EA) while simultaneously controlling for MMR-d/MSI-H status, tumor location (colon versus rectum), and disease stage.

These models identified CXCL10, an important chemokine that regulates the tumor immune microenvironment, as differentially associated with ancestry and MMR/MSI status: CXCL10 expression was lower in AA compared to EA groups and higher in MMR-d/MSI-H compared to MMR-proficient (MMR-p)/MSI-Low (MSI-L) plus microsatellite stable (MSS) groups. The TCGA analysis indicated that when ancestry and MMR/MSI status were included in the same model, neither tumor stage nor location had significant effects on CXCL10 expression. However, the authors noted a limitation in the TCGA dataset: the number of AA MMR-d/MSI-H samples was small (n = 9), reducing power to detect interactions or ancestry effects across MMR/MSI strata. To address this limitation, the team performed RT-qPCR assays on a separate set of FFPE COAD samples consisting of 59 AA and 59 EA tumors, deliberately enriched for MMR-d/MSI-H cases. RT-qPCR measurements of CXCL10 mRNA relative to two reference genes (C1ORF43 and RAB7A) confirmed the TCGA finding that CXCL10 expression is higher in MMR-d/MSI-H tumors compared to MMR-p/MSI-L+MSS tumors. However, in this independent cohort, differences in CXCL10 expression between AA and EA groups did not reach statistical significance. Overall, the combined analyses did not provide evidence for a significant effect of African ancestry on CXCL10 expression when accounting for MMR/MSI status.

The work highlights CXCL10 as a gene whose expression is strongly associated with MMR-d/MSI-H status and underscores the challenges of disentangling ancestry-related differences from tumor molecular features when sample sizes for specific combinations (e.g., AA with MMR-d/MSI-H) are limited. Methods included multivariable regression of TCGA RNA-seq data and RT-qPCR validation on well-characterized FFPE samples. The study informs efforts to understand how ancestry and tumor molecular phenotype jointly shape the tumor immune microenvironment, with implications for biomarker development and understanding disparities in immune-related tumor biology.