Lung cancer is the leading cause of cancer death worldwide, with a 5-year survival rate of 14 %. The first-line treatment for lung squamous cell carcinoma (SqCC), the second most common subtype of lung cancer, relies heavily on standard cytotoxic therapies and so there is a need for the development of novel targeted therapies. Typical in vivo models of lung cancer for preclinical drug trials use cell lines engrafted in immunocompromised mice; however, there are vast genomic and phenotypic differences between cancer cell lines and tumour tissue from patients. These differences undoubtedly affect the activity of novel therapies and as such, there is a need for in vivo models that more accurately recapitulate human tumours. To this end, our laboratory has developed a bank of patient-derived xenografts (PDXs) in immunocompromised mice using freshly resected lung tumours from patients undergoing surgery at Melbourne hospitals.
We have successfully transplanted lung squamous cell carcinoma (SqCC, 23/38 patient samples engrafted) to generate PDXs that recapitulate the phenotype of the primary tumour. RNA sequencing, whole exome sequencing and SNP array of three of our SqCC models have shown the range of genetic alterations present in the human tumour is conserved in the resulting mouse xenograft tumours.
We have used our molecularly characterized SqCC PDXs to evaluate the activity of new therapeutic strategies. Fibroblast growth factor receptor 1 (FGFR1) amplification has recently been described in 20% of lung SqCC and can be targeted by FGFR inhibitors such as BGJ398. Treatment of PDXs with BGJ398 identified a sample that responded to therapy and yet was FGFR1 non-amplified. We found this tumour overexpressed FGFR1 mRNA by in situ hybridization, suggesting overexpression of FGFR1 mRNA may be a novel biomarker of FGFR inhibitor response. In addition, we found combination treatment of cisplatin and FGFR inhibitor significantly prolonged survival.
Our results demonstrate that PDXs permit the evaluation of specific therapies based on the molecular characteristics of a patient tumour, providing highly relevant preclinical models to develop improved personalized medicine.