Chromatin accessibility and gene expression profiling of primary and metastatic ER+ breast cancer [RNA-seq]

In this project, we generated a resource dataset that includes ATAC-seq and RNA-seq data from ER positive primary breast tumors and matched liver and lung metastases. We collected samples from four patients at diagnosis in addition to samples from eight patients at autopsy following cancer-related death, allowing us to gain insights into the molecular mechanisms driving metastasis. Our research centered on the hypothesis that breast cancer metastasis is driven by tissue-specific enhancer programs, where differential transcription factor motif activity orchestrates distinct gene regulatory networks in metastatic sites. Peaks with divergent TF motif activity between tissues may define key regulatory nodes that control metastatic colonization and tissue adaptation. For ATAC-seq, reads were processed using the PEPATAC pipeline, and DiffBind was applied to identify differentially accessible regions, with ChIPseeker used to annotate metastatic-specific peaks. RNA-seq reads were aligned with STAR and quantified using HTSeq, and DESeq2 was employed to identify differentially expressed genes while controlling for batch effects. We established peak-to-gene correlations within a 0.5 Mbp window between chromatin accessibility and gene expression, assessing significance using a conservative null model. To investigate transcription factor (TF) activity, we applied TOBIAS footprinting, which infers TF occupancy by integrating chromatin accessibility with motif information. Leveraging footprinting scores together with RNA-seq expression, we computed a Total Functional Score of Enhancer Elements to identify the most relevant TFs per tissue. Using this approach, we linked TF motif activity at each peak with gene expression correlations, revealing peaks with tissue-specific TF activity and contrasting motif patterns between Liver and Breast. By connecting ATAC-seq peaks to RNA-seq expression, we identified putative metastasis driver genes, including 99 genes in Liver and 9 in Lung, as well as 23 genes shared across both metastatic sites, such as CCNF, SPINT1, and SLC2A1, which may play critical roles in metastatic progression. Notably, we discovered that the majority of these genes are correlated 3 or more enhancers in the metastatic tissue but not in the primary tissue.