Persistent SARS-CoV-2 Spike is Associated with Localized Immune Dysregulation in Long COVID Gut Biopsies

Abstract SARS-CoV-2 persistence is a proposed driver of Long COVID (LC), but the in-situ relationship between residual viral antigen and immune dysregulation remains poorly defined. To address this critical gap, we employed a high-resolution, multi-modal approach—combining RNAscope, GeoMx Digital Spatial Profiling (DSP), spatial transcriptomics, and multiplex immunofluorescence—on 25 terminal ileum and left colon biopsies from a clinical cohort of 8 LC participants and 5 healthy controls. We confirmed the persistence of SARS-CoV-2 Spike transcript and protein in the gut tissue of all LC cases and controls tested. Yet, comparison of Spike-positive (Spike+) regions in LC versus healthy control colon tissues revealed a differential, symptomatic state-associated signature, with 57 differentially expressed genes (DEGs) (26 upregulated, 31 downregulated), revealing genes that disrupt the immune response in LC subjects. LC colon Spike+ regions demonstrated increased expression of AQP8 and other absorptive-related genes (SLC26A3, SLC26A2, and CLCA4) which are involved with Chron’s disease along with transcripts involved in tumorigenesis (GUCA2A, S100P, TSPAN1). Simultaneous downregulation of key homeostatic chemokines ( CXCL13, CCL19, CCL21 ), and other transcripts reported to exhibit low expression in colorectal cancers (TMEM88B, NIBAN3, DMBT1), suggesting a paradox of epithelial tissue stress yet dysfunctional immune trafficking. Further analysis comparing Spike+ versus Spike-regions within LC colon tissue demonstrated an active, localized, antigen-driven immune microenvironment, identifying 122 DEGs (82 upregulated, 40 downregulated), including tumorigenesis genes. Cellular deconvolution of Spike+ regions revealed a statistically significant focal enrichment of myeloid-derived cells (macrophages, non-classical/intermediate monocytes), plasma cells, and regulatory T cells, coupled with significant enrichment in T-cell-related pathways, including “Antigen processing and presentation,” and “Th1/Th2/Th17 cell differentiation.” The ileum displayed a similar, though less pronounced, signature, demonstrating these statistically significant findings are specific to the colon of LC subjects. In contrast, corresponding Spike+ vs. Spike-analysis in healthy control colon tissues showed a more modest transcriptional response with 38 DEGs. Our data provide robust evidence that persistent SARS-CoV-2 Spike protein detection in the gut is not immunologically inert. Instead, it is actively associated with distinct, immune cell composition shifts and a dysfunctional pro-inflammatory transcriptional profile, supporting the hypothesis that retained viral antigen drives chronic immune dysregulation in tissue of Long COVID subjects.