KLF7 as a biomarker for the pre-metastatic state promotes colorectal cancer liver metastasis via TGFβ autocrine signaling.

Although metastasis-initiating cells drive metastasis, only a certain subpopulation of these cells can successfully disseminate from the primary tumor to colonize metastatic sites. The identification and characterization of this subpopulation remain poorly studied. We designated this specific subpopulation as pre-metastatic state (PMS) cells. Identifying biomarkers and understanding the mechanisms underlying PMS formation are vital for preventing metastasis. In this research, we employed machine learning and bioinformatics techniques, integrating data from human cell lines, bulk RNA sequencing, single-cell sequencing, and spatial transcriptomics, to pinpoint PMS and classify its subtypes (PMS1 and PMS2). We revealed that PMS1 exhibits high epithelial-mesenchymal transition (EMT) characteristics and is associated with poor prognosis, while PMS2 is associated with tumor stemness. Notably, KLF7 was identified as a key biomarker of PMS, highly expressed in PMS1 and colorectal cancer liver metastasis (CRLM) lesions. Through dual-luciferase reporter assays and both in vitro and in vivo experiments, we demonstrated that KLF7 promotes EMT characteristics in CRC via TGFβ autocrine signaling, thereby facilitating CRLM. Moreover, it was determined that the small molecule drug epigallocatechin gallate (EGCG) can inhibit the transcriptional activity of KLF7, thereby suppressing the TGFβ signaling and EMT. This result highlights EGCG as a promising compound for the treatment of colorectal cancer (CRC) with high KLF7 expression. Through this research, we established a systematic framework defining two PMS subtypes and identified KLF7 as a pivotal driver of CRLM through TGFβ autocrine signaling.