A machine learning-derived immune-related prognostic model identifies PLXNA3 as a functional risk gene in colorectal
Hanzhang Lyu1,2, Song Wang3, Guodong Guo1,2
1Shengli Clinical Medical College of Fujian Medical University, Fujian Provincial Hospital, Fuzhou, China.
Introduction:
Colorectal Cancer (CRC) remains a leading cause of cancer-related mortality, characterized by substantial interpatient heterogeneity and limited effective prognostic biomarkers.
Methods:
To address this gap, we constructed a robust prognostic model by integrating over 100 machine learning algorithms-such as LASSO, CoxBoost, and StepCox-based on transcriptomic and clinical data from The Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GEO) cohorts.
Results:
Plexin-A3 (PLXNA3) emerged as a top risk gene within the ensemble model, which achieved strong predictive performance, surpassing conventional clinical indicators. Multi-omics validation confirmed PLXNA3's prognostic relevance. Spatial and single-cell transcriptomics demonstrated their enrichment in malignant epithelial regions and negative association with immune cell infiltration, particularly CD8+ T cells and plasma cells. Transcription factor (TF) and microRNA (miRNA) correlation analyses revealed potential upstream regulators of PLXNA3 linked to tumor stemness and immune suppression. Functional enrichment indicated its association with cell cycle, DNA damage repair, and interferon signaling pathways. Immunohistochemistry (IHC) confirmed PLXNA3 overexpression in tumor tissues and its correlation with nodal metastasis. Moreover, drug sensitivity profiling and Connectivity Map (CMap) analysis identified potential compounds, including imatinib, MS-275 and fasudil, capable of reversing PLXNA3-driven transcriptional programs.
Discussion:
This study identifies PLXNA3 as a novel immune-related biomarker in colorectal cancer and elucidates its multifaceted role in tumor progression, immune evasion, and therapeutic resistance. These findings provide a foundation for incorporating PLXNA3 into precision oncology frameworks for gastrointestinal malignancies.
