Dissecting tumor heterogeneity in colorectal cancer: uncovering the role of BCL2L1+ cells through single-cell analysis

Background Colorectal cancer (CRC) ranks among the most prevalent gastrointestinal malignancies with liver metastasis being the primary cause of CRC-related death. Although surgical and chemotherapeutic interventions continue to improve, patients with hepatic metastases frequently experience recurrence and limited treatment benefits. Liver metastasis is driven by tumor heterogeneity and immune evasion. Therefore, defining the cellular composition of CRC liver metastases may help identify new therapeutic targets. Methods Primary CRC and liver metastasis samples were analyzed by single-cell RNA sequencing (scRNA-seq). Seurat was used for quality control, dimensionality reduction, clustering, and cell annotation, and Harmony corrected batch effects. Differential expression with GO, KEGG, and GSEA was performed for enrichment. Copy number variation analysis using inferCNV (v1.1) distinguished malignant from non-malignant cells, with smooth muscle cells (SMCs) and epithelial cells (EPCs) as reference populations. CytoTRACE and Slingshot characterized tumor differentiation trajectories, while CellChat and pySCENIC constructed cell communication and transcriptional regulatory networks. The key factor CEBPG was validated by in vitro functional experiments. Statistical analyses were conducted in R and Python. Results scRNA-seq identified five CRC tumor cell subtypes, among which the C4 BCL2L1 + tumor cells (TCs)subtype was predominantly enriched in liver metastases and displayed enhanced proliferation, metabolic reprogramming, and anti-apoptotic activity. Furthermore, transcription factor analysis suggested that CEBPG might regulate BCL2L1 expression to promote tumor survival and migration. Subsequent CEBPG silencing markedly suppressed CRC cell proliferation and invasion. In addition, a BTRS model derived from the C4 BCL2L1 + TCs subtype effectively stratified patient prognosis, as the high-risk group exhibited elevated expression of immune escape–related genes and impaired immune function. Conclusion This study revealed that the C4 BCL2L1 + TCs subtype might drive CRC progression by promoting metabolic adaptation and immune evasion. In addition, CEBPG functioned as a key regulatory factor that increased tumor malignancy through BCL2L1 -mediated survival pathways to some extent. The BTRS model reflected the molecular and immune heterogeneity of CRC and provided a framework for clinical risk stratification and personalized therapy. In summary, this work provides a comprehensive mechanistic framework linking metabolic adaptation, immune escape, and the progression and metastasis of CRC, and identifies potential therapeutic targets.