Bioinformatic Screening And Prognostic Analysis Of Platinum Resistance-associated Genes In High-grade Serous Ovarian Cancer

Objective: Ovarian cancer is one of the three most common malignant tumors of the female reproductive system.90% of ovarian cancer patients are epithelial ovarian cancer(EOC).High grade serous ovarian cancer(HGSOC)accounts for 60-80% of all EOC cases.It is also the main histological type that causes the death of ovarian cancer patients.Currently,platinum and taxane chemotherapy are considered standard adjuvant treatment for ovarian cancer after cytoreductive surgery.Although it is effective,the standard adjuvant treatment has imperfections.Many patients suffer platinum resistance.Platinum resistance seriously affects the clinical efficacy and prognosis of patients.It is also one of the main reasons for ovarian cancer recurrence and treatment failure.Therefore,how to predict platinum resistance in patients with ovarian cancer,what are the key factors and targets that lead to platinum resistance in patients,how to improve the sensitivity to platinum-based chemotherapy are the key issues of the current ovarian cancer treatment research.In this study,bioinformatics analysis techniques were used to analyze the differential expression genes in platinum-resistant and sensitive HGSOC samples and to construct a biomarker-based model that could predict the drug resistance and prognosis of HGSOC patients.Methods: Four chips were retrieved from the GEO database containing 66 patients with high-grade platinum-sensitive serous ovarian cancer and 25 patients with platinum-resistant ovarian cancer.RMA package was used for background correction and data normalization.sva package was used to remove the batch difference among the four gene chips.limma package was used to analyze differential expression gene.DAVID was used to perform gene function annotation analysis and pathway enrichment analysis.The univariate Cox regression was performed by using the coxph function in the survival package.Genes with P<0.01 were selected into the Cox multivariate regression.Cox multivariate regression was performed by using the step function in the survival package.According to the formula: Risk Score = h0(t)(ExpmRNA1 * βmRNA1 + ExpmRNA2 * βmRNA2 +...+ ExpmRNAn * βmRNAn),Patients’ risk scores were calculated.Patients were divided into high-risk group and low-risk groupby median risk score.The survivalROC package was used to analyze the predictive power of risk score for recurrence free time in HGSOC patients.It was also used to analyze the predictive power of risk score for recurrence free time in HGSOC patients with different ages,stages,grades,lymph node metastases,residual tumors,therapy outcome and anatomic neoplasm subdivisions.The survival package was used to analyze the impact of risk score on relapse-free survival in HGSOC patients.It was also used to analyze the impact of risk score on recurrence-free time in HGSOC patients with different clinical characteristics.Results: 66 patients with high-grade platinum-sensitive serous ovarian cancer and 25 patients with platinum-resistant ovarian cancer were collected from the GEO database for the analysis of differentially expressed genes.Compared to platinum-sensitive HGSOC samples,a total of 346 genes were differentially expressed in platinum-resistant samples(P <0.05,Fc >1.5).147 of them were down-regulated(platinum-resistant / sensitive)and 199 of them were up-regulated(platinum-resistant/ sensitive).Next,gene functions were annotated and the KEGG pathways the genes involved were enriched by DAVID.We found that these genes are not only involved in multiple drug resistance-related biological processes such as cell signal transduction and cell adhesion,but also participate in the PI3K-Akt signaling pathway and many other signaling pathways associated with drug resistance in ovarian cancer.Finally,Cox regression model was constructed with LAMP3(β =-0.091,HR =0.913,95%CI : 0.815-1.022,P = 0.116),SLAMF7(β =-0.117,HR = 0.889,95%CI : 0.808 – 0.979,P = 0.017),DEPDC7(β = 0.261,HR = 1.298,95%CI :1.101-1.531,P = 0.002)and AKAP12(β = 0.165,HR = 1.180,95%CI : 1.025-1.358,P = 0.021).Among them,DEPDC7 overexpression and AKAP12 overexpression were risk factors for RFS in serous ovarian cancer patients.Both of them could be used as independent risk factors for RFS in serous ovarian cancer patients.High expression of LAMP3 and high expression of were protective factors of RFS in serous ovarian cancer.But only SLAMF7 can be used as an independent protective factor of RFS in patients with serous ovarian cancer.After calculating the risk score for each patient based on the formula,we analyzed the predictive power of risk score for recurrence free time in HGSOC patients.The area under the curve(AUC)was found to be 0.779,indicating that the risk score was accurate for predicting recurrence-free time in patients with serous ovarian cancer.Survival analysis showed that patients with HGSOC in the low-risk group had longer relapse-free survival(RFS)time(P<0.001)compared with those in the high-risk group.And the risk score had a greater prognostic value for serous ovarian cancer patients older than 58 years of age compared with patients younger than 58 years of age.Compared with patients with grade 2 serous ovarian,it had a greater prognostic value for patients with Grade 3-4.Compared with patients who did not suffer lymph node metastases,it had a greater prognostic value for patients with lymphoid metastasis.Conclusions: 1.A total of 346 different expressed genes(P <0.05,Fc >1.5),including 147 down-regulated(drug-resistant / sensitive)and 199 up-regulated(drug-resistant / sensitive)were obtained.The different expressed genes involved in many biological processes related to drug resistance,such as cell signal transduction and cell adhesion.They may also participate in the drug resistance of ovarian cancer through PI3K-Akt signaling pathway and Focal adhesion signaling pathway.2.Cox regression model consisting of four genes(LAMP3,DEPDC7,SLAMF7 and AKAP12)was constructed to predict the RFS time of HGSOC patients.Among them,DEPDC7 overexpression and AKAP12 overexpression were risk factors for RFS in serous ovarian cancer patients.Both of them could be used as independent risk factors.High expression of LAMP3 and high expression of SLAMF7 were protective factors of RFS in serous ovarian cancer.Only SLAMF7 were an independent protective factor.3.The Cox regression model had a greater prognostic value for serous ovarian cancer patients older than 58 years of age compared with patients younger than 58 years of age.Compared with patients with Grade 2 serous ovarian cancer,it had a greater prognostic value for patients with Grade 3-4 serous ovarian cancer.The prognostic value of these regression models was greater for patients with serous ovarian cancer who have developed lymph node metastases compared with patients without lymph node metastases.