Predicting bevacizumab efficacy: the emerging role of ACTL6B in colorectal cancer

Colorectal cancer (CRC) is the third most common malignancy worldwide, and bevacizumab is the backbone antibody against vascular endothelial growth factor (VEGF) for patients with liver metastases. Nevertheless, no clinically applicable biomarker reliably foretells who will benefit, because VEGF expression alone shows limited predictive value. This study aims to discover and functionally validate a molecular signature that can anticipate bevacizumab response and long-term outcome in CRC.

A total of 620 CRC cases with documented heterogeneous bevacizumab exposure were extracted from The Cancer Genome Atlas (TCGA). Multi-omics layers-whole-exome sequencing, RNA-seq, reverse-phase protein array, immune-deconvolution algorithms [Tool for Immune Estimation Resource 2 (TIMER2), QUANTitative Immunogeneic Sequencing (QUANTISEQ), Estimating the Proportions of Immune and Cancer cells (EPIC), Microenvironment Cell Populations (MCP)-counter], microsatellite instability (MSI) and tumor mutational burden (TMB)-were integrated. Pan-cancer enrichment [Gene Ontology (GO), Kyoto Encyclopedia of Genes and Genomes (KEGG)], survival modelling, and nomogram construction were performed, followed by lentiviral over-expression and CRISPR-knockout studies in HT29 and COLO205 cells for proliferation, colony formation, trans-well migration and sphingolipid signaling interrogation.

Actin-like 6B (ACTL6B) emerged as the top predictor, showing inverse correlation with mesenchymal markers and positive association with CD4⁺ cytotoxic infiltration. High ACTL6B transcript consistently predicted better objective response rate (ORR; 67% vs. 31%, P=0.002) and longer median overall survival [OS; hazard ratio (HR) =0.58, 95% confidence interval (CI): 0.41-0.81] and recurrence-free survival (RFS; HR =0.62, 95% CI: 0.44-0.87) in the discovery and two validation sets. Mechanistically, ACTL6B transcriptionally repressed sphingosine-1-phosphate receptor 3 (S1PR3) and activated protein phosphatase 2 regulatory subunit Bbeta (PPP2R2B), thereby dampening pro-migratory sphingosine-1-phosphate signaling. A three-gene logistic model (ACTL6Blow/S1PR3high/PPP2R2Blow) yielded an AUC of 0.84 (95% CI: 0.79-0.89) for progressive disease under bevacizumab.

ACTL6B, alone or combined with S1PR3 and PPP2R2B, constitutes a robust biomarker panel for stratifying CRC patients likely to benefit from bevacizumab, warranting prospective clinical qualification.