Integrative multi-omics analysis prioritizes DPEP1 as a genetically supported candidate associated with skin cancer risk

Abstract

Background

Skin cancer (SC) is among the most prevalent malignancies, yet causal circulating proteins and tractable therapeutic targets remain poorly defined.

Methods

We performed a proteome-wide Mendelian randomization (MR) analysis of 4972 proteins across large GWAS datasets (discovery: 23 694 cases and 372 016 controls; replication cohorts: > 42 000 cases and > 810 000 controls). Sensitivity analyses, bidirectional MR, Steiger filtering and SMR testing supported robustness. Based on an integrated prioritization framework, DPEP1 was selected for downstream analyses. We then explored DPEP1-associated transcriptional programs using bulk and single-cell transcriptomic datasets and performed virtual screening, docking and molecular dynamics (MD) simulations to prioritize candidate DPEP1-binding ligands for future validation.

Results

MR identified six genetically supported proteins associated with SC risk, and DPEP1 was prioritized for downstream characterization based on an explicit target-prioritization framework. In the murine Dpep1-knockout dataset, Dpep1 deficiency was associated with altered stromal and immune programs, including reduced fibroblast abundance, lower YAP-TEAD/TGF-β pathway activity and increased CD8⁺ T-cell cytotoxicity scores. In human melanoma single-cell analyses, DPEP1 expression was enriched in fibroblasts and inferred ligand–receptor analyses suggested potential communication between DPEP1⁺ ·fibroblasts and immune subsets, including CXCL12–CXCR4 and MIF–CD74/CXCR4 interactions. Structure-based screening identified Balsaminone B as a top-ranked predicted high-affinity ligand, supported by stable docking and MD trajectories, with additional candidates including Ergotamine, Evobioside, MOL012579 and Venetoclax.

Conclusion

Together, these findings prioritize DPEP1 as a genetically supported candidate associated with skin cancer risk and provide supportive transcriptomic and computational evidence for future mechanistic and therapeutic investigation.