Use of STAMP-seq to enable spatial transcriptomics through single-nucleus in situ tagging and exploration of relationship of dynamic plasma cell modules and immunotherapy response in non-small cell lung cancer.

Abstract

e20040 Background: Immunotherapy has revolutionized cancer treatment, particularly in non-small cell lung cancer (NSCLC), yet a large fraction of patients remain refractory to treatment. Understanding treatment response requires characterization of immune cell populations and their spatial organization, as cellular neighborhoods critically influence immune cell recruitment, activation, and function. Current imaging-based and NGS-based methods struggle to achieve clear separation of individual cells in densely packed regions. Hence, we developed STAMP-seq, a novel method tagging single nuclei with cleavable spatial barcodes using high-density DNA sequencing chips, to uncover immunotherapy response-related modules and dynamic changes in NSCLC. Methods: STAMP-seq employs spatially-encoded DNA arrays to label individual nuclei with cleavable barcodes. Following validation in mouse brain tissue, we analyzed surgically resected NSCLC samples from immunotherapy-treated patients, integrating spatial transcriptomics with BCR profiling to characterize treatment response features. Results: STAMP-seq achieves precise spatial barcoding, detecting over 30,000 cells per adult mouse brain section with reliable cell subtype assignment. In immunotherapy-treated NSCLC samples, spatial analysis revealed distinct cellular communities associated with treatment response. Responding patients showed enrichment of tertiary lymphoid structures and specialized plasma cell niches, with tumor cells exhibiting elevated TNF-α signaling and reduced proliferation. These tumor cells showed proximity to ROBO2 + CAFs while remaining distant from ACTA2 + CAFs. BCR clonotype and pseudotime analysis revealed that IGHG1 + plasma cells are recruited from TLS or vasculature by stromal-rich "transit stops" containing antigen-presenting capabilities CAFs through HMGB1 - CXCR4 interaction. Conclusions: STAMP-seq enables high-resolution spatial transcriptomics at single-nucleus level. Our characterization of NSCLC spatial architecture revealed previously unrecognized immune response organization, particularly how specialized stromal-rich "transit stops" coordinate cancer-suppressing plasma cell recruitment. These findings demonstrate how spatial information can advance diagnostic and prognostic module development for human diseases.