Establishment And Identification Of Ovarian Cancer Human Tumor Xenograft Model And Observation Of The Efficacy Of Platinum Chemotherapy

BackgroundEpithelial ovarian cancer(EOC)ranks first in mortality among all gynecologic malignancies with high malignancy.Most patients of EOC have advanced diseases at the time of diagnosis.Owing to high rates of recurrence and chemotherapy resistance,the five-year overall survival for metastatic cases remains only 29%.Experiments based on animal models can help in screening new anti-tumor drugs and optimizing medication regimen in treatment of EOC.Patient-derived tumor xenograft(PDX)models for ovarian cancer are constructed by direct transplantation of patient materials into immunodeficient mice.This model is supposed to resemble the patient tumor,while some differences between them have been reported.At present,PDX models have been established to study a variety of cancers.For ovarian cancer PDX models,however,the rates of successful engraftments varied in different experiments and there was no further analysis of factors affecting engraftment rates.More evidence was needed before a wide application of PDX models to preclinical cancer research.This study aims to explore risk factors of the engraftment rates in EOC-PDX models and to analyze the resemblance of models to the original tumors.Method1.Establishment of EOC-PDX modelsFresh tissues of EOC were collected and implanted subcutaneously in NOD-Prkdc-il2rg(NPI)mice to generate PDX models.All tumor grafts obtained from models underwent screening of lymphoma by flow cytometry.2.Analyses of the consistency between EOC-PDX models and the original cancerPaired samples of EOC-PDX models and original tumors were compared in the aspect of histomorphology by HE staining.Immunohistochemistry(IHC)of PAX-8,CK7,CK20,P16,P53,WT-1,ER and PR were also used for comparison on protein level in HGSC.Genetic alterations were compared by WES.Besides,RNA sequencing was also performed to analyze the similarity of transcriptome between PDX and clinical tumors.3.Analysis of factors affecting establishment of EOC-PDX modelsThe clinicopathological information of EOC cases was analyzed to indicate correlation between the clinicopathological characteristics and engraftment rates in constructing EOC-PDX models.Whole exon sequencing(WES)was performed to explore factors affecting engraftments in PDX models on gene level.4.Evaluation of platinum response in PDX modelsTo credential PDX models as reliable platform for preclinical tests,platinum response of models was evaluated and compared with that of the corresponding EOC patients.Result1.Establishment of EOC-PDX modelsThis study enrolled 113 EOC patients in Peking Union Medical College Hospital from Jan.2018 to Oct.2019.Fresh tumor tissues were collected during cytoreductive surgery and directly implanted under the flank surface of immunodeficient mice in 12 hours.As a result,69 cases of EOC-PDX models were successfully established.The engraftment rate was 61.1%for the 1st generation.Time for engraftment is mainly during 4-8 months.CD 19 and CD45 markers were detected by flow cytometry and few tumor grafts were screened as lymphoma.The engraftment rate for the 2nd generation was 96.7%with a significantly shorter engraftment time compared to the 1st generation(p<0.001).2.Analyses of the consistency between EOC-PDX models and the original cancer(1)HE staining confirmed that PDX models retain structures of their original tumors.Stroma in tumors tissues tended to lost after engraftment,leading to an enrichment of cancer cells.(2)According to results of IHC in 10 cases of high-grade serous ovarian cancer(HGSC),the expression levels of PAX-8,CK7,CK20,P16,P53,WT-1,ER and PR in graft tumors were in accordance to original tumors.(3)WES was performed for 10 cases of HGSC and their corresponding tumor samples.PDX models recapitulated the diversity of genetic alterations of original cancers.The models preserved the majority of significantly mutated genes,loss of heterozygosity(LOH)modes and 2 signatures associated with endogenous mutational process and failure of DNA double-strand break-repair by homologous recombination in ovarian cancer.In the aspect of copy number variation(CNV),PDX models showed the same tendency with the original tumor.mRNA sequencing of 9 paired samples suggested different patterns of transcription between PDX models and the original tumors.3.Analysis of factors affecting establishment of EOC-PDX modelsMultivariate analysis indicated that neoadjuvant chemotherapy(NACT)was an independent factor(OR=0.315,p=0.006)for engraftment rates,while the age,primary/recurrent type,pathological type,tumor grade,FIFO stage,preoperative CA125 level,TP53 and BRCA1/2 mutation,and the number of transplantation sites may not influence the establishment of PDX models.Further analyses of how NACT effected on engraftment showed that a higher Ki-67 index correlated a higher engraftment rate;neoadjuvant chemotherapy significantly reduced Ki-67 index of ovarian cancer.In addition,the result of WES for 38 cases ovarian cancer indicated no relationship between neoadjuvant chemotherapy and tumor mutation burden(TMB),microsatellite instability(MSI)and transitions of transversions(TiTv).Tumor samples after NACT has a significantly reduced frequency of LOH(p=0.014);reduced LOH was clearly related to a lower engraftment rate(p=0.024).4.Evaluation of platinum response in PDX modelsMice bearing PDXs from 33 patients were treated with chemotherapy based on cisplatin or carboplatin.Platinum response on PDX models was evaluated according to the modified response evaluation criteria in solid tumors(mRECIST).A consistency test indicated that complete remission(CR)after platinum chemotherapy in PDX models was highly related to platinum sensitiveness in patients(Kappa=0.6550,p=0.001).Clinical cases whose PDX models showed CR after platinum chemotherapy had better progression-free survival compared to those cases evaluated as PR,SD or PD in PDX models(p=0.0111).Conclusion1.EOC showed a good engraftment rate for the construction of PDX models,with a relatively long period of time for engraftments.2.EOC-PDX models recapitulated morphology,protein expression and genetic alterations of original tumors,in spite of an altered mode of transcription.3.NACT resulted in a lower engraftment rate by reducing both the Ki-67 index and LOH frequency of cancer tissues.4.Platinum response of PDX models were in accordance with prognosis of ovarian cancer patients.