通过转录、染色质可及性分析及足迹推断的转录因子活性解析乳腺癌转移中的增强子逻辑

Breast cancer metastasis remains a major clinical challenge, highlighting the need to dissect the molecular mechanisms involved. In this project, we generated a multi-omic resource consisting of RNA-seq and ATAC-seq profiles from ER-positive primary breast tumors and matched liver and lung metastases. We collected samples from four patients collected at diagnosis and eight patients collected at autopsy following cancer-related death, enabling a comprehensive assessment of metastatic regulatory programs. We hypothesized that breast cancer metastasis is driven by tissue-specific enhancer networks, in which differential transcription factor (TF) motif activity coordinates distinct gene regulatory programs across metastatic sites. Towards this end, RNA-seq reads were aligned using STAR, quantified with HTSeq, and analyzed with DESeq2 to identify differential gene expression between primary and metastatic tissues. ATAC-seq reads were processed with the PEPATAC pipeline, and differentially accessible chromatin regions were identified using DiffBind. Peak-to-gene correlation analyses integrating ATAC-seq and RNA-seq data revealed putative metastasis-driver genes including 99 liver specific genes and 9 lung specific genes. Notably, there were 23 shared genes across both metastatic sites, including CCNF, SPINT1, and SLC2A1. The majority of these genes were associated with three or more enhancers in metastases but not in primary tumors, suggesting enhancer rewiring as a key mechanism of metastatic progression. To identify TFs that may drive these regulatory changes, we applied TOBIAS footprinting to infer TF occupancy by integrating motif information with chromatin accessibility. By combining footprint scores with gene expression correlations, we uncovered tissue-specific TF activity: KLF5, KLF12, SP3, and KLF10 predominated in lung metastases, whereas SP5, SP1, SP4, KLF14, and KLF15 dominated liver metastases. Most chromatin accessibility peaks exhibited differential TF motif usage between tissues, further supporting distinct enhancer regulatory programs in liver versus lung metastases. Only a minority of peaks shared the same top motifs across tissues, and some peaks contained motifs in one tissue but not the other, underscoring both shared and tissue-selective regulatory mechanisms that likely shape metastatic adaptation. By analyzing TF motif activity utilizing patient-specific depth of coverage across peaks allowed us to refine transcription factor binding predictions in ways that reflect individual tumor biology. Prioritizing TFs based on activity within these regulatory landscapes may ultimately improve the development of targeted therapeutic strategies to combat metastatic breast cancer.