Single-cell and spatial transcriptomics profile the interaction of SPP1+ macrophages and FAP+ fibroblasts in non-small cell lung cancer.

Abstract

Background: Non-small cell lung cancer (NSCLC) remains one of the most prevalent malignancies. A series of differentially expressed genes (DEGs) have been identified in tumor samples and play critical roles in modulating the characteristics of tumor cells. However, some DEGs are specifically expressed in the tumor microenvironment (TME) cells. The underlying mechanisms of the functional DEGs warrant comprehensive investigation to elucidate their contributions to tumor biology of NSCLC. Therefore, the primary goal of our study is to systematically investigate TME-related DEGs using NSCLC as a model.

Methods: DEG analysis was performed by comparing bulk transcriptomes of adjacent and tumor samples across 7 independent NSCLC cohorts. Expression pattern of these DEGs were annotated to specific cell types using a single-cell RNA sequencing (scRNA-seq) dataset from 13 NSCLC studies. Myeloid and stromal cells were re-clustered to achieve a detailed characterization of cell-cell interactions within the TME. Spatial co-localization of distinct subpopulations was validated by immunofluorescence staining and spatial transcriptomics (ST). Finally, functional relevance of these interactions was evaluated using a conditional knockout mouse model.

Results: A total of 82 overlapping DEGs were screened out using bulk transcriptomes across 7 NSCLC cohorts. After clustering the integrated 547,360 cells from 217 adjacent/tumor NSCLC samples with available scRNA-seq data, we observed that most of these DEGs were specifically expressed in epithelial, myeloid, and stromal cells. Notably, SPP1 ranks the top DEG and is specifically expressed in myeloid cells. The elevated SPP1 expression in bulk transcriptome was attributed to both enriched proportion of SPP1 ⁺ macrophages and increased expression level of SPP1 in such myeloid subcluster in NSCLC tumor, which was potentially regulated by several transcriptional factors (e.g., HIF1A, ATF5, and STAT1). Cell-cell communication analysis indicated FAP ⁺ fibroblasts exhibited the strongest interaction with SPP1 ⁺ macrophages within TME. Multiple fluorescent staining demonstrated significantly increased SPP1 ⁺ macrophage and FAP ⁺ fibroblast interactions in tumor compared to adjacent tissues. Nichnet analysis showed VCAN-ITGB1 ligand-receptor axis mediating this cellular interaction. ST analysis revealed that this interaction established an immune-excluding barrier surrounding the tumor core. In addition, immune cells (CD4⁺/CD8⁺ T and NK cell) were recruited to tumor edge. Consistently, macrophage-specific Spp1 knockout in a mouse model resulted in increased CD3⁺ and CD8⁺ T cell infiltration and reduced subcutaneous tumor size. Clinically, patients with FAP ^High SPP1 ^High expression in tumor exhibited significantly poor prognosis for immunotherapy treatment.

Conclusions: SPP1 ranks as the top DEG in NSCLC tumor attributed to its specific expression in SPP1 ⁺ macrophages, proportion of which increases in tumor and interact with FAP ⁺ fibroblasts. Their interactions contribute to the formation of a tumor barrier that inhibits immune infiltration, reducing the immunotherapy efficacy in NSCLC.