Abstract Background Chronic obstructive pulmonary disease (COPD) is a major risk factor for non–small cell lung cancer (NSCLC), yet the cellular and molecular mechanisms linking chronic inflammation to tumor progression remain poorly understood. Single-cell RNA sequencing (scRNA-seq) has identified COPD-associated transcriptional programs, including exhausted CD8+ T cells, LAMP3+ dendritic cells, CCL18+ macrophages, and malignant CD74+ epithelial–immune hybrid cells. However, their spatial organization and interactions within the tumor microenvironment (TME) remain unexplored. Methods scRNA-seq data from COPD-associated lung squamous cell carcinoma (GSE194070) were reanalyzed to characterize epithelial-immune cross-talk, highlighting CD74+ tumor cells and their macrophage migration inhibitory factor (MIF)-CD74-STAT3-PD-L1 signaling axis. These transcriptional signatures were then spatially mapped using Xenium 5K spatial transcriptomics data (10X Genomics). Integration was performed via Seurat v5.2 using anchor-based label transfer, and spatial localization was validated using canonical markers (EPCAM, EGFR, CD74, CCR7, LAG3). Results Integrated analysis identified CD74+ epithelial clusters colocalized with TREM2+ macrophages and LAG3+ T-cell niches, defining immunosuppressive microenvironments characteristic of COPD-associated tumors. Among the 226 shared DEGs between modalities, CD74 and STAT3-related genes were consistently upregulated, and MIF expression in TAMs highlighted a conserved immune-evasive signaling axis across datasets. An additional 154 genes were unique to Xenium, providing spatially resolved markers that refine COPD-specific tumor microenvironments. Xenium uniquely revealed spatial co-enrichment of CD74+ tumor and immune cells in tumor margins, whereas scRNA-seq provided higher-resolution detection of transcriptional heterogeneity (4,604 unique DEGs). Together, these data illustrate a spatially coordinated CD74-driven axis of immune suppression and epithelial remodeling. Limitations and Future Directions Public datasets differ in sequencing depth, cell-type representation, and clinical metadata, limiting COPD-specific inference. The Xenium dataset represents only NSCLC without COPD and a limited gene panel, constraining coverage of low-abundance transcripts. Future work will extend spatial mapping to COPD-tumor pairs, expand gene coverage, and experimentally validate predicted CD74-MIF-PD-L1 signaling interactions. Conclusions By integrating scRNA-seq and spatial transcriptomics, this study provides a spatially resolved atlas of COPD-associated lung cancer, revealing CD74+ epithelial-immune interactions that may contribute to immune evasion and tumor progression. These data suggest that CD74 and LAG3 are spatial biomarkers, and MIF is a key signaling mediator that may inform therapeutic strategies. This abstract is funded by: None
Khatun et al. (Fri,) studied this question.