Abstract 850P
Background
Patient-derived xenograft (PDX) can overcome several key limitations from cancer cell line xenograft model. However, the success rate of generating a PDX is relatively low. It remains unknown about the genetic factors contributing to the success rate of PDX, especially in epithelial ovarian cancer.
Methods
We collected 38 primary tumor samples from high-grade serous ovarian cancer (HGSOC) patients and the PDX success rate was evaluated. The patients were classified into NCF (neoadjuvant chemotherapy-free, 21) and NC (neoadjuvant chemotherapy, 17) groups. All of the tumors, as well as corresponding blood samples, were subjected to whole exome sequencing (WES). Mutation calling was performed according to the criteria established by The Cancer Genome Atlas (TCGA). Significantly mutated genes (SMG), mutational signatures, tumor mutation burden (TMB), microsatellite instability (MSI), chromosome-wide loss of heterogeneity (LOH) were calculated.
Results
Significant difference of PDX success rate was noted between NCF (17 out of 21, 81%) and NC (8 out of 17, 47%) (p<0.02). Based on WES, we identify somatic driver candidates, including significantly mutated genes, TP53, RPTN and FRG1. However, there is no SMG gene enriched in NCF or NC, nor success or failed samples. We find mutations in MT-CO3 (cytochrome c oxidase III, 3 missense and 1 nonsense) significantly enriched in NC (p = 0.032). As for TMB/MSI, there is no significant difference between NCF and NC. Mutational signature analysis indicates that NCF has significantly lower Signature 26 (etiology: associated with defective DNA mismatch repair; p = 0.026), which is likely due to the chemotherapy exposure. Meanwhile, in NCF, the weight of Signature 3 (etiology: associated with failure of DNA double-strand break-repair by homologous recombination) is significantly lower in success than fail samples (p = 0.049). Compared with NC, NCF has significant higher genome-wide LOH (p = 0.014).
Conclusions
In this study, we provide evidence that HGSOC samples from patients receiving neoadjuvant chemotherapy is less likely to generate successful PDX models, which is mainly due to the defective DNA mismatch repair caused by chemotherapy drug exposure.
Clinical trial identification
Editorial acknowledgement
Legal entity responsible for the study
The authors.
Funding
Chinese Academy of Medical Science.
Disclosure
All authors have declared no conflicts of interest.