Molecular Biomarker Of Drug Resistance Developed From Patient-derived Organoids Predicts Survival Of Colorectal Cancer Patients

Background: In the era of precision oncology,implicit molecular characterization of the tumor is essential in defining the best therapeutic plan.Therefore,the establishment of prognostic and predictive molecular biomarkers is increasingly becoming more valuable in cancer treatment.Colorectal cancer(CRC)was the 4th most common diagnosed cancer and second leading cause of cancer related death worldwide.5-fluorouracil(5-Fu)is the critical composition of CRC treatments.Prognostic and predictive molecular biomarkers for CRC patients(CRCpts)treated with 5-Fu based chemotherapy can provide assistance for tailoring treatment approach.Patient-derived tumor organoids(PDTOs)have been proven to recapitulate tumor’s pathological morphology,marker expression,chromosomal stability,genomic characterization and tumor heterogeneity.PDTOs can predict the response of chemotherapy,chemoradiation and targeted therapy,suggesting that PDTOs may represent a companion preclinical tool in precision oncology.Purpose: This study aimed to identify an effective CRCpts stratification method based on validated gene signature by using drug sensitivity data(5-Fu)of colorectal cancer organoids(CRCOs)to predict the survival of CRCpts who received 5-Fu based chemotherapy.Methods and Material: Based on our experience of the establishment of mouse small/large intestinal organoids and CRCOs,CRCOs biobank was established from surgical specimens from CRCpts.The drug sensitivity of CRCOs to 5-Fu was tested by using organoid size change and differentially expressed genes(DEGs)related to 5-Fu resistance were generated by transcriptome sequencing.Gene signature and drug resistant score model(DRSM)were developed in TCGA-CRC cohort by LASSO regression analysis and 5-fold cross validation.The DRSM was validated in four GEO datasets for predicting the survival of CRCpts by using drug-resistant genes(DRGs)associated with 5-Fu resistance.Result: After the successful establishment of mouse small/large intestinal organoids and CRCOs,41 CRCOs cultures were generated from 50 CRC tumor tissues after surgery(82%).Following tests revealed a great diversity in drug sensitivity for 10 μM 5-Fu treatment.Fourteen cases(34.1%)were 5-Fu sensitive and the other twenty-seven(65.9%)were resistant.Then DEGs associated with 5-Fu resistance were outputted by transcriptome sequencing.In particular,DEGs were generated in two comparison groups: 1)5-Fu sensitive and resistant untreated CRCOs;2)CRCOs before 5-Fu treatment and surviving CRCOs after 5-Fu treatment.Some molecules and most of the pathways that have been revealed to be participated in 5-Fu resistance were identified in current research.By using DEGs correlated with 5-Fu resistance and survival of CRCpts,the gene signature and DRSM containing five molecules were established.Multivariate analysis revealed that DRS(drug resistant score)was an independent prognostic factor for OS(overall survival)in CRCpts in TCGA-CRC cohort(P < 0.001).Further validation results from four GSE cohorts elucidated that the DRSM based on five genes related to 5-Fu chemosensitivity and developed from patient-derived organoids can predict survival of CRCpts.Meanwhile,our model could predict the survival of CRCpts in different subgroups.Besides,the difference of molecular pathways,tumor mutational burden(TMB),immune response related pathways,immune score,stromal score and immune cell proportion were dissected between DRS-high and DRS-low patients in TCGA-CRC cohort.Conclusions: The DRSM developed in current study by using 5-Fu resistant genes derived from CRCOs can predict the survival of CRCpts.This model may be useful in tailoring therapeutic regimen and act as a supplement of PDTOs-guided personalized treatment for CRCpts.