Classification Of Breast Cancer By Multiple Layers Of Molecular Features

Breast cancer(BRCA),as one of the most common malignancies in women,is a worldwide public health problem.It is a heterogeneous disease characterized by different histopathology and clinical features with different treatment measures.Therefore,the study on breast cancer heterogeneity is essential for improving diagnosis and optimizing treatment of the disease.In this text,the first chapter mainly summarizes the global impact of BRCA,the research significance and status of DNA methylation(DNAm),hormone receptor,tumor microenvironment(TME),long non-coding RNA(lnc RNA)on BRCA progression and classification(Chapter 1).Hormone receptor status and aberrant DNAm have important roles on the pathogenesis and heterogeneity of BRCA,however,the DNA methylation signatures associated with the pathological development and diagnosis of breast cancer under the context of certain hormone receptor have not yet been fully explored.In the present study,we performed methylation-based BRCA classification using the Illumina Infinium Human Methylation450K(HM450K)dataset,which contained 785 invasive BRCA and 98 normal breast tissue samples from The Cancer Genome Atlas(TCGA)dataset.Interestingly,we identified two methylation-based subgroups in ER-positive BRCA and further determined that the hypomethylated subgroup presents a good survival probability.This finding was further supported by another cohort of ERpositive BRCA containing 30 subjects.Additionally,we identified 977 hypomethylated Cp G loci showing significant associations with good survival probability in ERpositive BRCA.Among them,transcriptional activation of 47 genes by certain hypomethylated loci was also in line with good survival probability of ER-positive BRCA.Functional assay in numerous literatures provided further evidences supporting that some of the loci indeed play roles in activating tumor suppressive mechanisms via regulation gene transcription(e.g.,SFRP1 and WIF1)(Chapter 2).In addition,the tumor microenvironment,which is composed of host stromal cells,infiltrating immune cells and extracellular matrix,plays an important role in tumor growth,invasion and metastasis.Here we inferred the stromal scores(SSs)and immune scores(ISs)of BRCA samples based on the RNA-seq data using ESTIMATE algorithm and found the ISs were positively associated BRCA survival,suggesging the roles of tumor immune microenvironment(TIME)in tumor progression.Furthermore,we identified the differentially expressed genes(DEGs)between the BRCA subgroups with high and low ISs.The DEGs were associated with ISs and enriched in immune-related biological processes and pathways.Specially,we identified 3 TIME-related lnc RNAs.(i.e.,MIAT,LOC100130148 and LOC100128977)that associated with BRCA survival.Receiver operating characteristic(ROC)curve was analyzed for testing the influence on the 3-lnc RNA signature associated with overall survival in BRCA(2-year overall survival ROC was 0.732 or 5-year overall survival ROC was 0.634).(Chapter 3).Based on variable DNA methylation(DNAm),estrogen receptor(ER)-positive breast cancer(BRCA)is composed of two major subtypes,with the hypomethylated subgroup displaying good survival.Evidence indicates that the tumor microenvironment(TME)plays an important role in tumor progression and metastasis;however,its role and biological characteristics in DNAm-based subtypes of ERpositive BRCA remain largely unknown.Transcriptome data and matched clinical information of BRCA were downloaded from the Cancer Genome Atlas.Immune(ISs)and stromal scores(SSs)of BRCA patients were calculated using the ESTIMATE algorithm.The hypomethylated ER-positive BRCA subtype displayed high ISs,echoing the finding that higher ISs are associated with good BRCA survival.In addition,we analyzed the differentially expressed genes between the hypo-high-IS and hyperlow-IS BRCA subtypes in ER-positive patients and identified a co-expressed gene module enriched in immune-related biological processes(e.g.,leukocyte activation involved in immune response).Moreover,four hub genes(i.e.,PLEK,CD53,EVI2 B,and CD4)in this module showed significant association between their expression and ER-positive BRCA survival(Chapter 4).In summary,our study classifies the breast cancer by using the information of DNA methylation,hormone receptor,and tumor microenvironment.In especial,the study highlights the contribution of DNAm and tumor microenvironment to BRCA subtyping in ER-positive patients,and identidies novel biomarkers for BRCA diagnosis,classification,and prognosis.