Lay summary:
Outcomes for advanced oral cavity squamous cell carcinomas (OCSCC) have remained poor for many years. This highlights a critical need for new treatment methods for OCSCC. Patient derived xenografts (PDXs) can provide a system for testing such new therapies. These xenografts are small pieces of the patient’s tumor that have been successfully implanted and continue to grow as tumors in mice. Here we showed that such xenografts can be successfully created and then used DNA and RNA sequencing to show that they share the same tumor- causing mutations as the primary tumors from which they were derived. This provides a realistic setting for testing the effect of different tumor treatments in a patient-specific manner.

Progress summary:

Outcomes for advanced oral cavity squamous cell carcinomas (OCSCC) have remained stable for many years highlighting a critical need for the integration of newer precision modalities into OCSCC patient management. Patient derived xenograft (PDX) models are a critical component of developing therapeutic platforms but a comprehensive PDX approach for OCSCC has not been described. We have developed a large panel of OCSCC specific PDX models in immunodeficient (NSG) mice. PDXs were developed in two settings (1) patients undergoing standard of care primary resection or (2) as part of an ongoing window clinical trial of the MEK1/2 inhibitor trametinib in OCSCC patients. All xenografts were histologically confirmed squamous cell carcinomas and a subset of these were further passaged into NSG mice. To assess genomic conservation, we performed whole genome sequencing (WGS; n=10), whole exome sequencing (WES; n=15) and transcriptome sequencing (RNAseq; n=15) of matched primary and initial xenografts. The landscape of alterations (including copy-number, loss-of- heterozygosity, mutations, and expression patterns) in primary tumors and xenografts paralleled recently published genomic data and strong conservation of these changes was observed in xenograft samples. Together these data illustrate that OCSCC PDXs can be established with high success rates and retain high fidelity with parental tumors. These data establish a straightforward methodology for OCSCC xenograft generation and deepen our understanding of genomic parallels between primary tumors and PDX models. These models will thus provide an invaluable resource for studying patient-specific responses to established and novel targeted therapies. The initial proposal also described an evaluation of conditionally reprogrammed cells (CRCs) derived from the same primary tumors. These models also represent an attractive model for studying OCSCC. However, our sequence analysis showed that the cell lines derived did not contain any tumor-specific mutations from the primary. This indicates that in fact the published CRC establishment protocol in some cases may culture only normal epithelial cells. Such lines would not be a suitable model for studying tumor progression and testing drug sensitivity. Although negative, this is an important finding and cautionary tale emphasizing the importance of verifying tumor cell content in derived CRCs. CRF funds were

used to fund sequencing experiments including sample management, library construction, sequencing, automated analysis of sequence data, and partial salary support for a graduate student to lead ad hoc analysis. A publication describing these results is in the advanced stages of preparation for submission in early 2017.

Lay summary: 

Outcomes for advanced oral cavity squamous cell carcinomas (OCSCC) have remained poor for many years. This highlights a critical need for new treatment methods for OCSCC. Patient derived xenografts (PDXs) can provide a system for testing such new therapies. These xenografts are small pieces of the patient’s tumor that have been successfully implanted and continue to grow as tumors in mice. Here we showed that such xenografts can be successfully created and then used DNA and RNA sequencing to show that they share the same tumor-causing mutations as the primary tumors from which they were derived. This provides a realistic setting for testing the effect of different tumor treatments in a patient-specific manner. 

Progress summary:  

Outcomes for advanced oral cavity squamous cell carcinomas (OCSCC) have remained stable for many years highlighting a critical need for the integration of newer precision modalities into OCSCC patient management. Patient derived xenograft (PDX) models are a critical component of developing therapeutic platforms but a comprehensive PDX approach for OCSCC has not been described. We have developed a large panel of OCSCC specific PDX models in immunodeficient (NSG) mice. PDXs were developed in two settings (1) patients undergoing standard of care primary resection or (2) as part of an ongoing window clinical trial of the MEK1/2 inhibitor trametinib in OCSCC patients. All xenografts were histologically confirmed squamous cell carcinomas and a subset of these were further passaged into NSG mice. To assess genomic conservation, we performed whole genome sequencing (WGS; n=10), whole exome sequencing (WES; n=15) and transcriptome sequencing (RNAseq; n=15) of matched primary and initial xenografts. The landscape of alterations (including copy-number, loss-of-heterozygosity, mutations, and expression patterns) in primary tumors and xenografts paralleled recently published genomic data and strong conservation of these changes was observed in xenograft samples. Together these data illustrate that OCSCC PDXs can be established with high success rates and retain high fidelity with parental tumors. These data establish a straightforward methodology for OCSCC xenograft generation and deepen our understanding of genomic parallels between primary tumors and PDX models. These models will thus provide an invaluable resource for studying patient-specific responses to established and novel targeted therapies. The initial proposal also described an evaluation of conditionally reprogrammed cells (CRCs) derived from the same primary tumors. These models also represent an attractive model for studying OCSCC. However, our sequence analysis showed that the cell lines derived did not contain any tumor-specific mutations from the primary. This indicates that in fact the published CRC establishment protocol in some cases may culture only normal epithelial cells. Such lines would not be a suitable model for studying tumor progression and testing drug sensitivity. Although negative, this is an important finding and cautionary tale emphasizing the importance of verifying tumor cell content in derived CRCs. CRF funds were used to fund sequencing experiments including sample management, library construction, sequencing, automated analysis of sequence data, and partial salary support for a graduate student to lead ad hoc analysis. A publication describing these results was submitted, received positive reviews and is currently under final revision at Cell Reports. A pre-print copy of this manuscript has been submitted along with this report.