Rogue Cell State Fuels Colorectal Cancer Spread — and a Drug Can Reverse It
A high-MAPK, low-WNT tumor cell state drives CRC metastasis, and blocking mutant KRAS suppresses it in mice.
Summary
Researchers at Genentech discovered that colorectal cancer cells most likely to spread to the lungs and liver adopt a specific molecular identity: high activity in the MAPK signaling pathway combined with low activity in the WNT pathway. Using a mouse model that closely mimics human metastatic colorectal cancer, they showed that these cells carry extra copies of MAPK genes, boosting that pathway while silencing stem cell programs normally driven by WNT. Blocking the mutant KRAS protein — a key MAPK activator — reversed this cell state and reduced metastases. Analysis of human patient data confirmed that tumors displaying this gene signature are linked to worse survival. The findings reveal a surprising switch: while MAPK and WNT cooperate to start tumors, they oppose each other during metastasis.
Detailed Summary
Colorectal cancer (CRC) is one of the leading causes of cancer death worldwide, largely because it frequently spreads to distant organs like the liver and lungs. Most CRC tumors are driven by mutations in two major signaling pathways — WNT and MAPK — which are known to cooperate in initiating tumor growth. But what drives the cancer to spread has remained poorly understood.
Researchers at Genentech used a technique called serial in vivo orthotopic passaging to develop an immunocompetent mouse model that faithfully recapitulates human metastatic CRC. By repeatedly transplanting tumor cells through living mice, they selected for highly metastatic cell populations and then studied what made them different.
The key discovery was a distinct cell state defined by high MAPK pathway activity and low WNT pathway activity. Metastatic cells had acquired chromosomal amplifications in MAPK pathway genes, ramping up that signaling while simultaneously suppressing WNT-driven transcriptional programs — including genes associated with cancer stem cells. This represents a striking reversal of the cooperation these pathways show during tumor initiation.
Critically, pharmacological inhibition of mutant KRAS G12D — a common oncogenic driver — reduced this MAPK-high, WNT-low transcriptional state and significantly decreased both lung and liver metastases in the mouse model. Analysis of human CRC patient datasets confirmed that a gene signature reflecting this cell state correlates with poorer survival outcomes, validating clinical relevance.
These findings reframe how we think about CRC metastasis: rather than a fixed genetic program, it appears to involve dynamic cell state plasticity shaped by the balance between two major oncogenic pathways. Targeting KRAS or the broader MAPK pathway may therefore offer a strategy not just for controlling primary tumors but for preventing or treating metastatic spread. Caveats include the mouse model's limitations and the abstract-only availability of full methodology.
Key Findings
- A high-MAPK, low-WNT cell state specifically drives metastatic spread in colorectal cancer.
- Metastatic CRC cells acquire chromosomal amplifications in MAPK genes, suppressing WNT stem cell programs.
- Inhibiting mutant KRAS G12D reversed the metastatic cell state and reduced lung and liver metastases in mice.
- Human CRC patient data confirmed the MAPK-high, WNT-low gene signature correlates with worse survival.
- MAPK and WNT pathways cooperate in tumor initiation but oppose each other during metastatic dissemination.
Methodology
The team used serial in vivo orthotopic passaging in immunocompetent mice to generate and select highly metastatic CRC cell populations, then characterized them using transcriptomic and genomic profiling. Pharmacological KRAS G12D inhibition was tested to assess reversal of the metastatic cell state. Human CRC patient datasets were analyzed to validate the clinical relevance of the identified gene signature.
Study Limitations
This summary is based on the abstract only, as the full paper is not open access, so detailed methodology and statistical analyses cannot be fully evaluated. The primary experimental model is a mouse system, and translation to human biology requires further validation in clinical settings. All authors are employed by Genentech and hold Roche shares, representing a potential conflict of interest.
Enjoyed this summary?
Get the latest longevity research delivered to your inbox every week.