Transcriptomic and epigenetic dynamics of breast cancer progression in the MMTV-PyMT mouse model

Abstract

Malignant breast cancer remains a major health threat to women of all ages worldwide. To improve diagnosis and drug development, it is critical to understand the molecular mechanisms as well as to identify genes and pathways implicated in tumor progression, especially during malignant transition. Gene expression alterations and epigenetic variations of DNA methylation have been widely reported in cancers of different types in both humans and animal models. Global profiles of gene expression and DNA methylation have been suggested to help understand the molecular mechanisms and identify potential biomarkers.;In Chapter 1, I briefly reviewed the current understanding of breast cancer, especially the applications of expression and DNA methylation profiles in breast cancer studies. I reviewed some concepts of DNA methylation and histone modification as well as techniques to measure them. Then I discussed the rationale of using the MMTV-PyMT mouse model in my study and reviewed other mouse models used in breast cancer studies. Finally, I reviewed applications of some biomarkers identified by genomic studies. My studies are data driven: I intended to investigate the dynamics of gene expression and DNA methylation changes without prior hypothesis of responsible genes.;In Chapter 2, I analyzed RNA-sequencing data to study transcriptional alterations. Using the MMTV-PyMT mouse model, which has been widely used to study breast cancer, I analyzed gene expression from four distinct stages of tumor progression (hyperplasia, adenoma/MIN, early carcinoma and late carcinoma) during which malignant transition occurs. I found remarkable expression similarities among the four stages, meaning genes altered in the later stages showed traces in the beginning of tumor progression. I identified a large number of differentially expressed genes in PyMT tumors compared with age-matched normal mammary gland controls from FVB mice. Those genes were enriched in cancer-related pathways as well as E2F targets. Using co-expression networks, I found panels of genes as signature modules with some hub genes that predict metastatic risk. Using text mining on the PubMed literature, I proposed some new breast cancer candidate genes for further study.;In Chapter 3, I profiled DNA methylation with ERRBS (Enhanced Reduced Representation Bisulfite Sequencing) on the same samples. I identified a great number of differentially methylated cytosines (DMCs) in PyMT mice. The premalignant stages were quite similar yet distinct from the malignant ones on DNA methylation patterns. Many differentially methylated loci were preserved from the first to the last stage. Interestingly, genes with promoter hypermethylation were enriched in PRC2 (Polycomb repressive complex 2) targets, and those targets showed reduced expression.;In summary, I discovered that global gene expression and DNA methylation disruptions in the PyMT breast tumor start at early stages. Thus, in this study, our gene expression and DNA methylation profiling of the MMTV-PyMT mouse model sheds new light on the molecular dynamics during breast cancer malignant progression and provides potential mechanisms that regulate this biological process.