A multi-omics R-loop-linked risk program highlights CKS2-positive proliferative tumor cells as drivers of glioma growth

  • Int Immunopharmacol. 2026 Apr 15:175:116421. doi: 10.1016/j.intimp.2026.116421.
Weichun Tang  1 ,  Shangshang Hu  2 ,  Xu Tong  3 ,  Qiang Li  4 ,  Xiaoyong Wang  5 ,  Yupo Li  3 ,  Biao Gu  5 ,  Nannan Wang  5
Affiliations
  • 1. The Third People's Hospital of Bengbu Affiliated to Bengbu Medical University, Bengbu 233000, Anhui, China; Anhui Provincial Key Laboratory of Tumor Evolution and Intelligent Diagnosis and Treatment, Bengbu Medical University，Bengbu 233030, Anhui, China.
  • 2. The Third People's Hospital of Bengbu Affiliated to Bengbu Medical University, Bengbu 233000, Anhui, China; School of Medicine, Southeast University, Nanjing 210009, Jiangsu, China. Electronic address: [email protected].
  • 3. The First Affiliated Hospital of Bengbu Medical University, Bengbu 233000, Anhui, China.
  • 4. Anhui Provincial Key Laboratory of Tumor Evolution and Intelligent Diagnosis and Treatment, Bengbu Medical University，Bengbu 233030, Anhui, China.
  • 5. The Third People's Hospital of Bengbu Affiliated to Bengbu Medical University, Bengbu 233000, Anhui, China.
PMID: 41763168 DOI: 10.1016/j.intimp.2026.116421
Abstract

Objective: To develop a generalizable glioma risk signature derived from an R-loop-associated transcriptional program, assess its prognostic and immunotherapy-predictive utility, and define its cellular and regulatory determinants.

Methods: We quantified R-loop activity (ssGSEA) and defined subtypes by consensus clustering across bulk multi-omics glioma cohorts. A prognostic model was trained in TCGA and validated in CGGA/GEO, with immunotherapy relevance, cellular localization, and key drivers/therapeutics assessed through integrative multi-omics analyses and targeted experiments.

Results: R-loop activity was elevated in glioma relative to normal tissue, increased with WHO grade, and consistently predicted poorer survival. Consensus clustering identified an R-loop-high subtype characterized by the worst prognosis and immune-evasive features. The resulting risk score robustly stratified survival across multiple cohorts, suggested a lower probability of immunotherapy benefit, and outperformed 145 published glioma signatures. Single-cell and spatial analyses mapped the high-risk program predominantly to proliferative (cycling) tumor-cell states. Among signature genes, CKS2 was the top contributor: it was upregulated at both mRNA and protein levels, associated with adverse outcomes, and functionally promoted glioma proliferation and clonogenicity while suppressing apoptosis; CKS2 silencing inhibited xenograft growth and reduced Ki-67 staining. Regulon inference and ChIP-qPCR supported MYBL2 as an upstream transcriptional regulator of CKS2. Drug-sensitivity analyses prioritized GSK269962A, which showed greater in vitro activity in CKS2-high cells, and molecular dynamics simulations supported stable binding of GSK269962A to CKS2.

Conclusion: An R-loop-anchored risk signature enables robust prognostic and immunotherapy stratification in glioma and nominates the MYBL2-CKS2 axis and associated vulnerabilities as potential translational targets.

Keywords
CKS2; Glioma; Prognostic signature; R-loop.
Products