A core stemness-associated module reveals PLK1, NUF2, KIF23, CDCA8, TOP2A, CENPF, AURKA, and ASPM as key genes in...

Rectal cancer (RC) presents significant clinical challenges due to a high risk of recurrence and therapy resistance, which are largely driven by cancer stem cells (CSCs). To address the critical need for RC-specific CSC biomarkers and targets, this study aimed to leverage single-cell transcriptomics and high-dimensional co-expression network analysis (hdWGCNA) to define a core stemness-associated signature and identify potential therapeutic vulnerabilities.

We built a single-cell atlas (GSE199726) and applied hdWGCNA to the CSC-like compartment, then validated findings across TCGA-READ and GEO GSE90627 cohorts. Pathway/immune features were assessed by clusterProfiler, GSEA, CIBERSORT, and ESTIMATE. Drug candidates were nominated by Enrichr/DSigDB and molecular docking. Wet-lab assays included qRT-PCR (SW620/Caco-2 vs HIEC-6), PLK1 knockdown (si-PLK1#1/#2), wound healing, transwell, and CCK-8 at different time courses (0/24/48/72 h).

Our analysis identified a CSC-enriched subpopulation from the single-cell atlas of RC, from which hdWGCNA derived four co-expression modules. This yielded an eight-gene panel (PLK1, NUF2, KIF23, CDCA8, TOP2A, CENPF, AURKA, ASPM) that was consistently upregulated in tumors across TCGA and GEO cohorts, correlated with G2/M checkpoint and E2F/MYC pathways, and associated with an immune-altered microenvironment. Crucially, this yielded an eight-gene panel that demonstrated exceptional diagnostic accuracy (AUC > 0.8) and significant prognostic value. In vitro, all eight genes were elevated in cancer cells; si-PLK1 efficiently reduced PLK1 expression and curtailed migration, invasion, and proliferation across time.

By integrating single-cell atlasing with hdWGCNA, we identified an eight-gene rectal CSC signature that captures a G2/M-stemness program, which generalized across cohorts and was functionally substantiated through PLK1 knockdown in vitro. This tractable vulnerability, anchored by kinases PLK1/AURKA, provides a compact biomarker readout for risk stratification and preclinical therapeutic screening, warranting further in vivo and clinical validation.

An eight-gene CSC signature captures a stemness-linked G2/M program that generalizes across cohorts, relates to the microenvironment, and is therapeutically tractable via kinase targeting, providing a compact readout for risk stratification and preclinical screening.