Peritoneal metastasis (PM) occurs when colorectal cancer cells escape the colon and seed the peritoneum - the membrane lining the abdominal cavity. It affects 20-30% of colorectal cancer patients and represents the worst prognosis among all metastatic sites. Without systemic therapy, median survival is only 6 months. Even with aggressive chemotherapy, median survival reaches only 16 months, and 5-year survival approaches 0%.
The current best hope for selected patients is cytoreductive surgery (CRS) combined with heated chemotherapy delivered directly into the abdomen (HIPEC). While this approach can extend life for certain patients, it carries significant surgical complications, and 50-90% of patients eventually experience recurrence because microscopic cancer cells survive and grow back. The high failure rate highlights the urgent need to understand the molecular drivers of peritoneal metastasis.
A key molecular characteristic of colorectal cancers that develop peritoneal metastasis is the CMS4 subtype signature - one of four Consensus Molecular Subtypes (CMS) of colorectal cancer. CMS4 tumors are defined by enrichment of stromal and mesenchymal gene programs, increased cellular plasticity (the ability to change states), and the highest rate of therapeutic resistance. Understanding what drives the CMS4 program in peritoneal metastasis could unlock new treatment targets.
Epithelial-to-mesenchymal transition (EMT) is a process by which cancer cells shed their organized, stationary epithelial identity and gain the properties of mesenchymal cells - mobile, invasive, and resistant to death. EMT is considered a critical step in metastasis: cells must undergo EMT to escape the primary tumor, travel through the bloodstream or peritoneal cavity, and colonize distant sites.
TWIST1 is a transcription factor - a protein that binds DNA and controls which genes are turned on or off - that is one of the master regulators of EMT. Prior studies showed TWIST1 is associated with poor outcomes in colorectal cancer, particularly in tumors with lymph node involvement and aggressive invasion front behavior. However, its specific role in peritoneal metastasis had not been investigated.
The researchers' key hypothesis was that TWIST1 is specifically activated in colorectal cancer peritoneal metastasis (compared to other metastatic sites like the liver) and drives the cancer program through a network of downstream target genes and stromal cell interactions. Understanding this TWIST1-driven program could identify new drug targets specific to this particularly lethal form of metastasis.
The researchers used a multi-layered approach combining clinical patient samples, bioinformatic analysis of large datasets, CRISPR-Cas9 gene editing, and mouse models. Patient samples from peritoneal metastases, liver metastases, and primary colon tumors were collected at two major medical centers to enable direct comparisons between metastatic sites.
To identify direct gene targets of TWIST1, the team performed ChIP-Seq (chromatin immunoprecipitation sequencing) - a technique that identifies all genomic regions where TWIST1 physically binds DNA - combined with RNA-Seq to identify which genes changed expression when TWIST1 was removed. Integrating both datasets allowed identification of genes that are both directly bound by TWIST1 and functionally regulated by it.
TWIST1 was knocked out using CRISPR-Cas9 gene editing in three colorectal cancer cell lines representing different CMS subtypes: the CMS4 human line MDST8, and murine lines CT26 and MC38. A peritoneal metastasis mouse model was established by injecting cancer cells directly into the peritoneal cavity of mice, mimicking how colorectal cancer actually spreads to the peritoneum in patients.
Analysis of a published gene expression dataset of colorectal cancer metastases revealed that TWIST1 was significantly upregulated in peritoneal metastases compared to primary tumors (p = 0.031). Importantly, among the canonical EMT transcription factors tested (TWIST1, SNAI1, SNAI2, ZEB1), TWIST1 was the only one specifically elevated in peritoneal metastasis.
When comparing peritoneal metastasis to liver metastasis datasets, TWIST1 was specifically upregulated only in peritoneal metastases and not in liver metastases. Western blot analysis of actual patient tissue samples from both sites confirmed this finding at the protein level. This site-specificity strongly suggests that TWIST1 is not simply a general marker of aggressive CRC but rather a driver of the specific biological program that enables peritoneal seeding.
CRISPR knockout of TWIST1 in CRC cell lines caused significant reductions in migration, invasion, wound closure, and sphere formation - all capabilities essential for metastatic spread. This confirmed that TWIST1 is functionally required, not just coincidentally expressed, in the cancer cells' ability to invade and self-renew in ways consistent with metastatic behavior.
By integrating ChIP-Seq and RNA-Seq data, the team identified 27 genes directly bound and regulated by TWIST1. Of these, SPON2 - a secreted protein - stood out as significantly upregulated in peritoneal metastasis patient samples (p = 0.0084) and strongly correlated with TWIST1 expression (Spearman R = 0.77 in the TCGA colorectal cancer cohort).
SPON2 (Spondin 2) is a protein secreted by cancer cells into their surrounding environment. Analysis of the cancer database TCGA showed that patients with high SPON2 expression had significantly worse disease-free survival, and in Stage 4 colorectal cancer, SPON2 was associated with a hazard ratio of 10.69 - meaning high SPON2 was associated with more than 10 times the risk of disease progression.
TWIST1 knockout cells showed dramatic reductions in SPON2 mRNA and protein levels both inside cells and in the secreted form in culture medium. SPON2 knockout itself also significantly reduced cancer cell migration and invasion, confirming that SPON2 functions downstream of TWIST1 to mediate the metastatic phenotype. This established the TWIST1-SPON2 axis as a functional, clinically relevant signaling pathway in peritoneal metastasis.
Tumors do not exist in isolation - they are embedded in a tumor microenvironment (TME) containing many non-cancer cell types including fibroblasts, immune cells, blood vessel cells, and mesothelial cells (in peritoneal metastases). The researchers' prior work had identified mesothelial cells as major contributors to the stroma of CRC peritoneal metastases and found that these stromal cells highly express SPP1 (Secreted Phosphoprotein 1, also called Osteopontin).
This study demonstrated that SPP1 secreted by stromal mesothelial cells acts as an upstream regulator of the TWIST1-SPON2 axis in tumor cells, signaling through the AKT kinase pathway to activate TWIST1 expression. This creates a bidirectional tumor-stroma crosstalk loop: stromal cells release SPP1, which activates TWIST1 in cancer cells, which produce SPON2, which further shapes the stromal environment.
In SPP1 knockout mice, the in vivo peritoneal metastasis model showed significantly reduced disease burden compared to normal mice, confirming the critical role of stromal SPP1 in enabling peritoneal tumor seeding. This SPP1-TWIST1-SPON2 circuit connecting stromal and tumor cells represents a complete molecular framework for peritoneal metastasis progression.
The SPP1-TWIST1-SPON2 axis identified in this study represents a set of potential biomarkers and therapeutic targets specific to colorectal cancer peritoneal metastasis. Since TWIST1 and SPON2 are specifically elevated in PM but not liver metastasis, they could help identify which patients are at risk of peritoneal spread - enabling earlier intervention or enrollment in PM-specific clinical trials.
Therapeutically, each component of the axis offers a potential target. TWIST1 inhibitors could prevent the mesenchymal, invasive behavior of CRC cells. Blocking SPP1/osteopontin in the stroma could cut off the upstream signal that activates TWIST1. Neutralizing secreted SPON2 could disrupt the downstream effector of the axis. Anti-SPON2 antibodies or small molecules may be particularly attractive because SPON2 is secreted into the extracellular space, making it accessible to therapeutic agents.
The identification of the mesothelial cell stroma as the source of SPP1 that drives this cancer program also suggests a potential treatment approach: targeting the microenvironment itself. Rather than only attacking the cancer cells directly, drugs that reduce mesothelial cell SPP1 secretion could deprive CRC cells of the environmental signal needed to activate the peritoneal metastasis program.