New models for patient-derived head and neck cancers

Cancer biologists have an arsenal of tools and models that can be used to study different aspects of the disease. These models have distinct strengths and weaknesses. For example, immortal cell lines are useful to tease apart basic cellular processes that impact disease, while mouse models are more appropriate to study how tumor cells interact with each other and the surrounding healthy cells. One tool to study cancer progression is called a patient-derived xenograft. This system involves taking tumor cells from a patient and transplanting them into a mouse. The xenografted cells proliferate in the mice much as they would in patients.  From there, researchers can explore questions about how the tumors behave. Xenograft models are most informative when the cancer is transplanted orthotopically, or into the mouse equivalent of whatever human tissue the cells originated from. For example, human colon cancer should be transplanted into a mouse colon or human brain cancer should be transplanted into mouse brain for the model to be as clinically relevant as possible.

Head and neck squamous cell carcinoma (HNSCC) is a type of cancer that forms solid tumors in the mouth, throat, or other parts of the head and neck. Models to study HNSCC are plentiful; researchers have been culturing immortal HNSCC cell lines since the 1950s, organoids capturing some of the three-dimensional structure of the cancer have been developed, and researchers have even used patient samples to create xenograft models. However, problems persist with each of these. Immortal cell lines can cross contaminate, and culturing them in a dish can lead to genetic drift or clonal expansion that can change how similar the cell line is to a patient’s cancer. Organoid models lack the surrounding tumor environment that is known to impact disease progression. Current patient-derived xenograft models require researchers to grow the cells in a dish before transplanting them to mice. This opens the xenograft model up to the same contamination and clonal selection problem that cell line models have. On top of that, current HNSCC xenograft models only establish orthotopic tumors in mice about 30% of the time. To overcome these limitations, Dr. Peiran Zhou, a T32 research fellow in Head and Neck Surgery at UW, teamed up with Dr. Brittany Barber at UW and Dr. Slobodan Beronja in the Fred Hutchinson Human Biology Division to create a better xenograft model for HNSCC.

Typically, xenografts are established by transplanting a tumor cell suspension into mice, but researchers studying other types of solid cancers have recently discovered that transplanting intact tumor pieces helps the cancer thrive in the mice. Because of this, the team thought that this could be a way to improve the success rate of HNSCC xenografts. However, HNSCC cells are typically transplanted under a mouse’s tongue and transplanting a larger piece of tumor in such a small and delicate area is not technically feasible. Zhou decided to first transplant the piece of tumor outside of its native environment, wait for the tumor to grow at this site, isolate the tumor cells, and finally dissociate them for an orthotopic cell suspension injection under the tongue. Using this approach, the team was able to generate xenografts for eight out of the nine patient samples they tested. This almost 90% engraftment rate was a marked improvement over previous models, highlighting the promise of this approach to efficiently generate xenografts. Still, how closely the genetics of their xenografts matched the genetics of the actual tumor was unclear.

To establish whether their xenografts represented actual human tumors, the team used a lentiviral system to add barcodes to the tumor cells isolated from the heterotypic injection site. When the orthotopic tumors were large enough, the group sequenced the barcodes present before and after the transplant. If the sequencing revealed that the orthotopic tumors had the same distribution of barcodes as the non-transplanted cells, the team could conclude that their xenografts retained the heterogeneity present in the patient tumor. Using mathematical modeling, they found that their orthotopic tumors were very similar to the non-transplanted cells, confirming that this approach creates a clinically relevant xenograft model.

HNSCC tumors can be driven by different combinations of genetic mutations. To ensure that the mutational landscapes of their xenografts accurately represented that of the primary tumor, the group characterized the mutations present in the primary tumor and in their orthotopic xenografts. They found that the xenografts retained most of the genetic characteristics of the primary tumor, indicating that this approach generates a genetically stable model to study HNSCC tumors. In the future, Beronja hopes that other groups can use these models to further their own HNSCC research.

This work was supported by the STTR PDX Annotation Grant to Fred Hutchinson Cancer Center, Dick and Loraine Burger Pilot Funding, and an AAO-HNSF Resident Research Grant.

Zhou P, Mills CB, Dong ZM, Barber BR, Beronja S. 2025. Barcoded Orthotopic Patient-Derived Head & Neck Squamous Cell Carcinoma Model Demonstrating Clonal Stability and Maintenance of Cancer Driver Mutational Landscape. Cancer Med. 14(15):e71137. doi: 10.1002/cam4.71137.