Association of Schwann cells with immune-cold tumor microenvironment in triple-negative breast cancer.

Abstract

e12592 Background: Although intratumoral nerves have recently been recognized as a critical regulator of tumor biology, its roles in breast cancer remain poorly understood, largely due to their rarity and technical challenges in transcriptomic profiling, as neuronal cell bodies are absent within tumors. Schwann cells (SCs) ensheath nerve fibers and constitute the most abundant intact cellular component of peripheral nerve system, thus is an ideal target for transcriptomic detection of peripheral nerves. We therefore hypothesized that a SC score will estimate the amount of nerve fibers and reveal its association with tumor biology in triple-negative breast cancer (TNBC), the most aggressive breast cancer subtype. Methods: Spatial transcriptome data was obtained from Zenodo repository. Bulk transcriptome and clinical data of TNBC patients were analyzed in the TCGA (n = 170), METABRIC (n = 335), and SCANB (n = 174) cohorts, as well as cohorts that underwent neoadjuvant chemotherapy (GSE194040, 25066, 163882, 123845, 20271, 230881, and 41998). Signature genes for SC, myelinating SC (mSC), and non-myelinating SC (nmSC) were obtained from PanglaoDB and Tabula Sapiens, and GSVA scores were calculated. Patients were stratified into high- and low-SC score groups using the median. Results: To determine the most appropriate SC score, we compared the PL-SC, TS-SC, TS-mSC, and TS-nmSC scores in ST analyses, and found that the PL-SC score showed the strongest spatial concordance with pathologically annotated nerve fiber regions and was used for all subsequent analyses. The SC score correlated with neuron score defined by xCell algorithm. Across 3 cohorts, high SC scores were associated with lower MKI67 expression, lower Nottingham histological grade, and reduced proliferation scores, accompanied by negative enrichment of cell proliferation-related Hallmark pathways in GSEA. The SC showed no association with genomic instability or homologous recombination deficiency. Although high-SC tumors showed fewer non-silent mutations, no differences in major gene alteration patterns were observed. Consistently, the SC score was negatively associated with multiple immune cell populations, including B cells, CD8⁺ T cells, Th1/2 cells, DCs, and macrophages, while positively associated with stromal cell populations. Immune-related pathways were also downregulated in high-SC tumors. However, the SC score was not associated with pathological complete response in the 8 chemotherapy cohorts. In survival analysis, a high SC score associated with worse disease-specific and overall survival in TCGA, whereas no significant association was observed in the other 2 cohorts. Conclusions: The SC score is associated with a less proliferative, immune-cold tumor microenvironment in TNBC, however, its association with chemotherapy response and prognosis appears to be cohort-dependent.