Identification of tumor initiating cells and early marker genes in normal colonic epithelium that lead to neoplastic transformation

Background & Aims : Colorectal cancer (CRC) remains a major cause of cancer-related morbidity and mortality worldwide. Although the adenoma-carcinoma sequence and associated genetic alterations are well characterized, the earliest cellular and molecular events that initiate tumorigenesis within histologically normal colonic epithelium remain poorly defined. This study aims to identify tumor-initiating cells (TICs) and early transcriptional markers of neoplastic transformation using single-cell RNA sequencing (scRNA-seq) from paired normal-appearing and transformed human colonic tissues. Methods : Fresh biopsies from histologically normal-appearing colonic mucosa and paired polyps, including tubular adenomas, sessile serrated adenomas, and adenocarcinoma, were collected from 7 human subjects. Single-cell transcriptomes were generated using the 10X Genomics platform and analyzed with Seurat, Monocle 2, CytoTRACE, GSVA, GSEA, RNA velocity, and InferCNV. Tumor-associated states were inferred utilizing clustering, trajectory analysis, pathway enrichment, copy number variation profiling, and validated spatially by RNA-FISH. Results : A total of 51,054 high-quality single-cell transcriptomes were resolved into 33 epithelial and stromal clusters. Tumor-specific stem-like (tSTM, cluster 0) and deep crypt secretory (tDCS, cluster 10) populations were enriched in adenomas, whereas sub-clustering of tSTM identified TICs (subclusters 4 and 6) derived largely from histologically normal mucosa. TICs exhibited strong stemness potential, genomic instability, and early activation of epithelial-mesenchymal transition (EMT) and interferon signaling, accompanied by suppression of oxidative phosphorylation. ETS2 , SLC12A2 , and LEFTY1 were identified as TIC-specific markers, while SOD3 and GPRC5A showed progressive upregulation along the TIC-to-tSTM trajectory. RNA-FISH confirmed spatial expression of candidate genes in adenomatous crypts, and independent validation using the COLON MAP dataset supported the presence and diagnostic performance of TIC-associated markers. Conclusions : This study identifies TICs as the developmental origin of neoplastic stem-like states and delineates early transcriptional and pathway reprogramming events that drive the transition from normal to premalignant colonic epithelium. These findings provide new insight into CRC initiation and nominate biomarkers with translational potential for early detection and therapeutic targeting.