Poster #36 - Miranda Brown

Visium Spatial Transcriptomics Reveals Localized Host Responses to Bacterial Exposure in Vaginal Epithelium

Miranda Brown, BS¹; Mallory Cunningham, BS²; Chris Oldfield, PhD²; Megan Williams, PhD²; Myrna Serrano, PhD²; Kimberly Jefferson, PhD²; Gregory Buck, PhD²,³,⁴; Katarzyna M. Tyc, PhD¹,⁵ Affiliations: ¹ Department of Biostatistics, School of Public Health, Virginia Commonwealth University, Richmond, VA ² Department of Microbiology and Immunology, School of Medicine, Virginia Commonwealth University, Richmond, VA ³ Center for Microbiome Engineering and Data Analysis, Virginia Commonwealth University, Richmond, VA ⁴ VCU Genomics Core, Virginia Commonwealth University, Richmond, VA ⁵ Massey Biostatistics and Informatics Shared Resource (BISR), Virginia Commonwealth University, Richmond, VA

Spatial transcriptomics offers unprecedented insight into gene expression within intact tissue architecture. While spatial technologies are increasingly applied to host tissue profiling, their use in precisely mapping host-microbe interactions-particularly at barrier surfaces where microbial-driven inflammation contributes to the initiation and progression of epithelial cancers-remains limited. Here, we apply 10X Genomics Visium V1 chemistry to investigate spatial localization and host transcriptomic response to two vaginal microbial species, Sneathia amnii and Gardnerella vaginalis, both implicated in bacterial vaginosis, adverse reproductive outcomes, and cervical cancer risk. Human EpiVaginalTM VLC-100 raft cultures, with and without bacterial exposure, were profiled using 10X Visium. Samples were aligned with Space Ranger, aggregated and then analyzed in R/Seurat. Visium spots were checked for bacterial presence. Our integrated analysis framework included normalization, clustering, and differential gene expression analysis across clusters and conditions. We found increasing expression of innate immune genes (e.g., CXCL8, SAA2, CCL20) from the basal layer closer to Sneathi-positive spots at the superficial layer, and downregulation of structural regulators such as FMNL2, suggesting impaired cytoskeletal function in infected regions. Gardnerella-positive regions were enriched for stress and autophagy-related genes (e.g., ATG9B, CGRRF1, ABCG1). Shared upregulated transcriptional responses included GPRC5A, MALAT1, CEACAM6, TIMP2, and PADI1. This study introduces a spatial framework for mapping host-microbe interactions by integrating bacterial localization with host gene expression. While focused on vaginal epithelium, this analytical approach applies broadly to situations where microbial dysbiosis shapes inflammation, barrier disruption, and even tumor microenvironment. With tissues like the cervix, colon, and stomach, where these dynamics influence cancer initiation and progression, our method provides a generalizable platform for probing the spatial interplay between microbes and host responses in early disease, with implications for cancer risk, progression, and therapeutic targeting driven by spatial patterns.