by ecancer reporter Clare Sansom

Xenografts of human colorectal cancer (CRC) cells into mouse models have yielded many important insights into the pathogenesis of this disease and its treatment.

However, none of the mouse models of colorectal cancer that are currently available completely replicate the course of the disease; in particular, they are more sensitive to drug therapy and do not reproduce the most commonly observed metastatic pathway, from the gastro-intestinal tract to the liver.

Furthermore, many such models require mice that are immunodeficient, and these cannot be used to study the response of the adaptive immune system to tumour development.

Chemokines are small signalling proteins that are able to induce directed movement, or chemotaxis, in nearby cells by binding to receptors.

Epithelial cells in the colon and small intestine secrete a chemokine, CCL25, which binds to cells that express chemokine receptor 9 (CCR9) on their surfaces and so traffics lymphocytes to these organs.

A large group of researchers led by Xiling Shen of Cornell University, New York and Steven Lipkin of Weill Cornell Medical College, New York, USA have engineered human colorectal cancer cells to express the receptor CCR9.

They used these to generate a number of chemokine-targeted mouse models (CTMMs) that robustly replicate many of the features of the human disease.

Seventeen of these models were produced by injecting the engineered tumour cells into the tail-veins of immunodeficient mice, and another three by injecting the cells into early-stage blastocysts of immunocompetent mice.

Introducing human cells into mice at this early stage of development induces specific immune tolerance so they are not rejected by the mouse immune systems.

Human early-stage CRC cells that endogenously express CCR9 are known to form primary tumours in the intestines of immunodeficient mice following tail-vein injection, attracted there by CCL25 secretion.

The researchers generated a panel of 16 human CCR9 CRC cell lines, each bearing different somatic mutations commonly found in these tumours.

This panel replicated the majority of known CRC types, whether classified by molecular profile or by histology.

In all these cell lines, chemotaxis towards the CCR9 ligand CCL25 increased with increasing CCR9 expression in vitro.

The researchers injected these cells into the tail veins of immunodeficient mice and found that the mice reproducibly developed colorectal tumours within 3 weeks of injection, whereas similar cells without CCR9 expression rarely formed colorectal tumours.

The immunocompetent mouse models were generated by micro-injecting 10-15 cells from each of three CCR9 CRC cell lines into mouse blastocysts to generate human CRC-mouse embryo chimeras.

The blastocysts were implanted into mouse ‘foster mothers’ and 24-40 live pups were obtained for every 80 implantations.

These pups developed locally invasive bowel tumours by 6 months of age, and these tumours were found to consist of human cells and to express CCR9.

The pups were found to tolerate the CRC cells while remaining immunocompetent, and no human CRC cells were detected in any other mouse organs.

In seven of the immunodeficient CTMMs, mice that had originally formed primary colorectal tumours developed liver metastases by 8 weeks, and these metastases had lost CCR9 expression.

This is consistent with a model in which CRC cells lose their CCR9 expression and metastasise to the liver through the portal circulation.

The researchers confirmed this by downregulating CCR9 and showing that this led to more lung metastases, and imaging the cells’ movement through the blood vessels to the liver.

Primary colorectal tumours, liver metastases and tumours generated by subcutaneous xenografts of these cells into the mice were challenged with the chemotherapy drugs oxaliplatin and 5-FU/oxaliplatin.

The liver metastases were found to be significantly more resistant to these drugs than either the primary colorectal or the subcutaneous tumours.

Chemoresistance in the metastases was correlated with up-regulation of the DKK4 and NOTCH signalling pathways.

Taken together, these results suggest that the CTMMs are robust mouse models of human colorectal cancer in which the tumour cells are targeted to the GI tract through CCR9 expression, which can be used to model metastasis and response to drug therapy.

In addition, the immunocompetent models, generated by injecting cancer cells into blastocysts, can be used to model immune responses to tumours and to therapy.

Reference

Chen, H.J., Sun. J,, Huang. Z. and 25 others (2015). Comprehensive models of human primary and metastatic colorectal tumors in immunodeficient and immunocompetent mice by chemokine targeting. Nature Biotechnology, published online ahead of print 25 May 2015.