Liang Chen,1,* Xiaomei Ying,2,* Haohao Wang,1,* Jiaheng Xie,3,* Qikai Tang,4 Wen Liu1 

1Department of Hepatobiliary and Pancreatic Surgery, Conversion Therapy Center for Hepatobiliary and Pancreatic Tumors, First Hospital of Jiaxing, Affiliated Hospital of Jiaxing University, Jiaxing, Zhejiang, People’s Republic of China; 2Department of General Surgery, Suzhou Hospital of Anhui Medical University, Suzhou, Anhui, People’s Republic of China; 3Department of Plastic and Cosmetic Surgery, Xiangya Hospital, Central South University, Changsha, Hunan Province, 410008, People’s Republic of China; 4Nanjing Jinling Hospital, Affiliated Hospital of Medical School, Nanjing University, Nanjing, Jiangsu, People’s Republic of China

*These authors contributed equally to this work

Correspondence: Wen Liu, Department of Hepatobiliary and Pancreatic Surgery, Conversion therapy center for Hepatobiliary and Pancreatic Tumors, First Hospital of Jiaxing, Affiliated Hospital of Jiaxing University, Jiaxing, Zhejiang, People’s Republic of China, Email nickwen2006@163.com

Background: The role of cellular senescence in the tumor microenvironment of pancreatic cancer (PC) remains unclear, particularly regarding its impact on prognosis and immunotherapy outcomes.
Methods: We utilized single-cell sequencing datasets (GSE155698 and GSE154778) for pancreatic cancer from the Gene Expression Omnibus (GEO) database and bulk RNA-seq data from the University of California, Santa Cruz (UCSC) and International Cancer Genome Consortium (ICGC) repositories, creating three patient cohorts: The Cancer Genome Atlas (TCGA) cohort, PAAD-AU cohort, and PAAD-CA cohort. Dimensionality reduction cluster analysis processed the single-cell data, while weighted gene co-expression network analysis (WGCNA) and differential expression gene analysis were applied to bulk RNA-seq data. Prognostic models were developed using Cox proportional hazards (COX) and least absolute shrinkage and selection operator (LASSO) regression, with validation through survival analysis, decision curve analysis, and principal component analysis (PCA). Tumor mutation data were analyzed using the “maftools” package, and the immune microenvironment was assessed with TIMER2 data.
Results: We developed a senescence-related (SENR) six-gene prognostic signature for PC, which stratifies patients by risk, with high-risk groups showing poorer prognoses. This model also offers predictive insights into tumor mutations and immune microenvironment characteristics. Caveolin-1 (CAV1) emerged as a significant prognostic biomarker, with functional validation showing its role in promoting cancer cell proliferation and migration, highlighting its potential as a therapeutic target.
Conclusion: This study provides a novel senescence-related prognostic tool for PC, enhancing patient stratification for prognosis and immunotherapy, and identifies CAV1 as a key gene with clinical significance for targeted interventions.

Keywords: pancreatic cancer, cellular senescence, prognosis, biomarker, immunotherapy