Integrating single-cell and spatial transcriptomics reveals glycolysis heterogeneity and NEK6-mediated progression in colorectal cancer

Background Metabolic reprogramming is a pivotal driver of tumor microenvironment (TME) remodeling and colorectal cancer (CRC) progression. However, the spatial organization of glycolysis heterogeneity and the molecular drivers maintaining the malignant high-glycolytic state remain poorly understood. Methods Single-cell RNA sequencing (scRNA-seq), spatial transcriptomics, and Mendelian randomization (MR) analysis were integrated. Glycolytic malignant subtypes were identified using AUCell and Non-negative Matrix Factorization (NMF). Cell-cell communication networks were inferred via CellChat. Pseudotime trajectory analysis and spatial deconvolution were utilized to resolve the developmental dynamics and spatial architecture of subtypes. Finally, key pathogenic regulator genes were screened by integrating large-scale GWAS data with the TCGA cohort, and candidate gene NEK6 was validated through in vitro functional assays. Results Three distinct malignant subtypes were identified; the Glycolysis-C1 subtype exhibited a high-glycolytic state. Cell-cell communication analysis revealed that Glycolysis-C1 cells function as a dominant signaling hub, secreting MIF ligands that target Macrophages and B cells via the CD74/CD44 receptor complex. Pseudotime analysis traced a lineage trajectory from low-glycolytic C3 to high-glycolytic C1. Spatially, Glycolysis-C1 cells concentrated in tumor domains, functioning as a communication hub at the tumor-stroma interface and remodeling the immune niche via MIF-CD44 signaling. Multi-omics integration and MR analysis identified NEK6 as a candidate gene associated with CRC risk. High NEK6 expression correlated with poor prognosis and an immunosuppressive TME. Functional assays confirmed that NEK6 knockdown significantly suppressed CRC cell proliferation and glycolysis, as well as abrogated M2 macrophage polarization. Conclusion This study reveals the spatial and molecular landscape of glycolysis heterogeneity in CRC and identifies NEK6 as a gene functionally associated with the high-glycolytic phenotype. These findings suggest that NEK6 may represent a potential therapeutic vulnerability for future investigation in CRC.