Single cell and bulk RNA sequencing identifies tumor microenvironment subtypes and chemoresistance-related IGF1+ cancer-associated fibroblast in gastric cancer

Abstract

Background

The tumor microenvironment (TME) significantly influences prognosis and drug resistance in various tumors, yet its heterogeneity and the mechanisms affecting therapeutic response remain unclear in gastric cancer (GC).

Methods

The heterogenous TME were explored with single-cell RNA-sequencing (scRNA-seq) data of 50 primary GC samples. We then identified four GC TME subtypes with nonnegative matrix factorization (NMF) and constructed a pearson nearest-centroid classifier based on subtype-specific upregulated genes. Genomic features and clinical significance of four subtypes were comprehensively evaluated. We reclustered fibroblasts to identify cancer-associated fibroblast (CAF) subtype associated with poor clinical outcomes. RT-qPCR and double immunofluorescence staining were applied to validate the findings. Cellchat analysis elucidated potential molecular mechanisms of the CAF subtype in GC disease progression and chemotherapy resistance.

Findings

The GC TME exhibited high heterogeneity, influencing chemo-sensitivity. Four TME-based subtypes predicting response to immunotherapy and chemotherapy were identified and validated in 1406 GC patients. Among which, ISG1 subtype displayed higher fibroblasts infiltration and heightened oncogenic pathways, and inferior response to chemotherapy with unfavorable prognosis. Microsatellite instability-high (MSI-H) GCs within four TME subtypes showed immunological heterogeneity. We then reported an IGF1-overexpressing CAF was associated with chemo-resistance and GC recurrence. Cell communication analysis revealed IGF1⁺ CAF may induce drug-resistant phenotypes in tumor cells through IGF1-α6β4 integrin ligand-receptor binding and activation of EMT biological process.

Interpretation

We identified four TME-based subtypes with different clinical outcomes and IGF1⁺ CAFs contributing to poor clinical outcomes in GC, which might provide guidance for individualized treatment and facilitate the development of novel therapeutic targets.