SPATIAL MULTI-OMICS ANALYSIS OF HOST–MICROBIOME INTERACTIONS IN CHRONIC DISEASES
Keywords:
Spatial Multi-omics, Host–Microbiome Interactions, Chronic Diseases, Spatial Transcriptomic, Biomarker DiscoveryAbstract
Chronic diseases arise when multiple cells, immune mediators, metabolites, and resident microbes interact within a complex, spatially organised tissue microenvironment. Bulk omics has shed light on many molecular changes (genes and microbial taxa) associated with disease, but it can't tell you whether the signals are in the same tissue compartments. This limitation has hindered the identification of local ‘biomarkers’ that account for progression across inflammatory and malignant diseases. This original synthetic research study validated the feasibility of a spatial multi-omics approach to map host–microbiome interactions in chronic disease tissues. In total, 64 tissue samples (eight each of colorectal cancer, inflammatory bowel disease and rheumatoid arthritis) were profiled using spatial transcriptomics, 16S rRNA sequencing, proteomics, metabolomics, histology and computational integration. Tissue niches were subdivided into epithelial, stromal, lymphoid, myeloid and damaged barrier niches. To identify spatial biomarkers associated with high disease progression, differential expression, host–microbiome correlation, network analysis and machine learning were used. The integrated model identified co-localisation of these inflammatory niches with Fusobacterium, Escherichia–Shigella and Prevotella, and with expression of CXCL8, IL1B, LCN2, SPP1, CXCL13 and MMP9. Integrating spatial, molecular and microbial features into the XGBoost model yielded an area under the receiver operating characteristic curve of 0.91 for classifying specimens with high progression, and improved performance compared to single-omics models. Candidate biomarkers included an axis of CXCL8–Fusobacterium, a barrier-injury niche of LCN2–Escherichia–Shigella and a lymphoid-inflammation signature of CXCL13–Prevotella. Our findings provide a foundation for using spatial multi-omics as a discovery platform to stratify chronic diseases and target therapeutic interventions

