A Study Focused On DNA Damage Repair Pathway With Predictive Ability For Therapy Based On Single Cell And Tissue RNA-seq Data In Pan-cancer

The DNA Damage repair(DDR)pathway is essential for maintaining the normal function of cells and genome integrity,and abnormalities in this pathway is associated with abnormal proliferation of cells,increased genomic instability,increased risk of cancer,malignant progression of tumors,and sensitivity to treatment.Based on clinical sample tissue and single-cell transcriptome sequencing data,big data of multi-omics provides a basis for exploring the relationship between abnormal DNA damage repair pathway and the molecular mechanism of cancers occurring and developing.However,the specific role and related mechanisms of abnormal expression of DDR-related genes in development of cancer and prediction of treatment effect have not been fully revealed.Therefore,based on the multiomic sequencing data of 6166 cancer tissues in the TCGA database,the single-cell RNA sequencing data of 227397 in 4 cancers,the data of 431 patients in the transcripto of immunotherapy for 5 solid tumors,and the information of 153 drugs of 1016 cancer cell lines obtained by combining CCLE and GDSC database information,the present study analyzed the multiomic changes of DDR pathway in pan-cancer,and defined TR-DDR score to reveal its correlation with cancer-related pathways and its ability to predict drug treatment effect,using Gene set variation analysis(GSVA),Gene set enrichment analysis(GSEA),CIBERSORT deconvolution algorithm,tumor infiltrating lymphocyte(TIL)score calculation,single-cell type identification and characteristic gene analysis,difference analysis,correlation analysis and other technologies and methods.Finally,based on the transcriptome data of four cancer single-cell cells,the evolutionary trajectory of TR-DDR score in the process of cancer development under epithelial cell subclusters was explored.The specific research content and results are summarized below:(1)Based on TCGA data from 11 cancers,we sorted out somatic single nucleotide variants,copy number variants,and m RNA expression levels of DDR genes in pan-cancer,and found that DDR genes were of abnormalities prevalently in cancer tissues,including high SNV mutation frequency,significant CNV amplification,and significant overexpression of m RNA.We screened 20 tumor-related core DDR genes to calculate TR-DDR scores to represent DDR activity in tumor tissues,and found that this score has a considerable predictive capacity of occurrence of tumor in pan-cancer.In addition,we found that compared with the high TR-DDR score group,the low group had a higher proportion of stage I patients,T1 stage patients,N0 stage patients,and M0 stage patients in most cancers(BRCA,KIRC,KIRP,LIHC,LUSC,THCA)in the large sample data.At the same time,the proportion of patients with a higher complete response to drug therapy was higher in the high group in a variety of cancers(BRCA,KIRP,LIHC,LUSC).Further results showed that TR-DDR score showed strong positive correlations with DNA replication,tumor proliferation,G2 M checkpoints,stem cell characteristics and MYC targets in pan-cancer,but negatively correlated with ECM degradation,angiogenesis,TGF-β signaling,apoptosis and p53 pathway.The results of GO analysis showed that DEGs in the high TR-DDR score group were enriched in stem cell pathways and cell development pathways.In the analysis results of CCLE data,TR-DDR score was positively correlated with tumor cell proliferation capacity.(2)In the CCLE database,the TR-DDR score of cancer cell lines is related to their drug sensitivity.It was found that in 128 of 153 anticancer drugs(83.7%),the sensitive group of cell lines showed relatively high TR-DDR scores.Therefore,it was further verified that the TR-DDR score was a predictive indicator of chemotherapy or immunotherapy responses.The predicted AUC value of TR-DDR was 0.432-0.7(average AUC=0.557)in the chemotherapy data set,while in the immunotherapy data set,the predicted AUC value was 0.454-0.680(average AUC=0.594).In order to improve the predictive ability of this indicator,the TR-DDR score is combined with the TIL score and the G2 M checkpoint score.The results showed that the AUC value range of the combination score in the chemotherapy data set was0.564-0.800(average AUC=0.662),and the AUC value range in the immunotherapy data set was 0.679-0.844(average AUC=0.744).Subsequently,the correlation between TR-DDR scores and other independent clinical response predictors was explored.It was observed that in 11 cancers,high TR-DDR groupings were associated with higher levels of genomic instability,including multiple factors such as neoantigen insertion index,MSI sensor scores,mutation counts,and fragment genomic changes.In general,tumor tissue with a high TR-DDR score has more gene mutations and genomic rearrangements,and the tumor tissue also carries more new antigens accordingly.(3)Tumor microenvironments with different TR-DDR scores also vary greatly.The CIBERSORT deconvolution algorithm was used to obtain the proportion of immune cell infiltration in each sample,and the relationship between TR-DDR scores and different immune cell ratios was explored.Based on data of 11 cancers in TCGA database,it was found that activated CD4+ memory T cells,macrophage M1 type cells,follicular helper T cells,CD8+ T cells,regulatory T cells Infiltration levels of resting NK cells and activated dendritic cells were positively correlated with high TR-DDR scores in at least 5 cancer types,while infiltration of resting CD4+ memory T cells,macrophage M2 type cells,resting mast cells,and monocytes was negatively correlated with high TR-DDR scores in more than 5 cancer types.It was also found that high TR-DDR score group exhibited more overexpression of genes related to stimulation of immune system activation,such as chemokines(IFNG,IL-1A,etc.),tumor necrosis factor receptor superfamily genes(TNF),and IFN responses,as well as lower VEGF expression,compared to the low TR-DDR group.(4)Finally,single-cell transcriptome data of four cancers(breast cancer,colorectal cancer,lung cancer,and ovarian cancer)were used to identify multiple cell types,and the results showed that there was a higher TR-DDR score in epithelial cell clusters of all cancers.At the same time,it was found that the TR-DDR score was correlated with the CNV score,which was used to evaluate the function of malignant cells.In epithelial cell clusters,the correlation between TR-DDR score and various cancer characteristic enrichment scores was consistent with the results obtained from TCGA transcriptome data,for example,it was positively correlated with DNA replication,tumor proliferation,G2 M checkpoint,stem cell characteristics,and MYC targets,while it is negatively correlated with inflammatory response,IL-10 anti-inflammatory pathway,and TGF-β Signal transduction,apoptosis,and p53 pathways.For different TR-DDR+/-cells,it has been found in quasi chronological analysis that as cancer cells evolve,DDR+ cells were transformed from DDR-cells gradually during the progression,and were concentrated at the end of the quasi chronological analysis.However,in some cancers,there will still be a cluster of DDR cells that maintain the DDRform along with the evolution process,accompanied by stronger intercellular transformation and migration capabilities.This study indicates that the TR-DDR score is important for defining the degree of malignancy of tumor cells,and can predict the differentiation process of different malignant cell subsets during the evolution and development of cancer.In summary,based on multi-omics and single-cell data,this study systematically analyzed the important pathways and tumor microenvironment related to DNA damage repair pathways in the process of cancer development and progression using bioinformatic analysis methods,and proposed a joint score index(TR-DDR score combined with TIL score and G2 M checkpoint score)that can be used to predict the efficacy of chemotherapy and immunotherapy.This paper adds new content to the molecular mechanism of DNA damage repair pathway in the process of cancer occurrence,development and metastasis,and provides new research perspectives for the prediction of cancer treatment effect.