SOX9 links genomic activation to spatial tumor ecosystem organization and immune context across cancers
Abstract Tumors are spatially organized ecosystems in which genomic alterations, cellular plasticity, and immune interactions are tightly coupled, yet the principles linking these layers remain unclear. Here, we identify SOX9 as a cross-cancer molecular feature associated with spatial tumor ecosystem organization. Integrative pan-cancer multi-omics analyses indicated that SOX9 upregulation was associated with copy-number gain and promoter hypomethylation and with adverse clinical outcomes in a cancer-type-dependent context. Single-cell RNA sequencing (scRNA-seq) showed that SOX9 expression was predominantly restricted to epithelial lineages, where higher SOX9 levels were associated with increased epithelial heterogeneity and less differentiated, progenitor-like states. In silico virtual-knockout analysis predicted coordinated changes in a collagen- and extracellular matrix (ECM)-centered transcriptional module rather than demonstrating direct SOX9 regulation. An independent I-SPY2 breast-cancer cohort reproduced the positive association between continuous SOX9 expression and a prespecified ECM/collagen score after adjustment for hormone-receptor and human epidermal growth factor receptor 2 (HER2) status. Cell-cell communication analysis indicated that SOX9-high epithelial states displayed distinct pathway compositions and epithelial-centered signaling architecture, without a significant increase in overall T-cell-directed communication strength. Spatial transcriptomic analyses, supported by permutation-based spatial statistics, showed reproducible segregation between SOX9-high tumor-associated niches and tertiary lymphoid structure (TLS)-rich immune niches across cancers. Complementary immune deconvolution approaches yielded concordant broad-lineage patterns; in immune checkpoint inhibitor (ICI)-treated cohorts, SOX9-high status was associated with inferior survival and an inflammatory yet exhaustion-skewed phenotype. Collectively, these findings support an association-based model in which SOX9 marks genomic, epithelial-state, matrix-associated, spatial, and immune features of tumor ecosystems, while direct mechanistic and clinical utility require prospective experimental validation.
Authors
- Mengjuan Long
- Wendan Shi (ORCID: https://orcid.org/0000-0002-9043-9249)
- Xue Zhang (ORCID: https://orcid.org/0000-0002-3856-5053)
- Wei Xia (ORCID: https://orcid.org/0000-0001-7491-8090)
- Shuye Liu
- Qibin Song
- Nana Yu
- Xiaoting Tong
Institutions
- Wuhan University (CN)
- Renmin Hospital of Wuhan University (CN)
Publication Details
- Journal
- Discover Oncology
- Published
- 2026-09-30
- DOI
- https://doi.org/10.1007/s12672-026-05979-x
- Primary Topic
- Single-cell and spatial transcriptomics
- Type
- article
- Field-Weighted Citation Impact
- 0.00