Genome-wide loss of lineage fidelity is a hallmark of aging breast epithelia and reflects a biology convergent with susceptibility to cancer initiation

Purpose: Leveraging genome-wide lineage-specific transcriptional profiling of human mammary luminal epithelial cells (LEPs) and myoepithelial cells (MEPs) to elucidate the molecular mechanisms underlying aging-associated breast cancer susceptibility. Methods: Human mammary luminal epithelial and myoepithelial cells were isolated from finite lifespan, non-immortalized human mammary epithelial cells (HMECs) derived from primary breast tissue of women across different age cohorts and breast cancer risk profiles. Transcriptional profiles of LEP and MEP cells were generated via RNA-sequencing performed on Illumina HiSeq 2500. RNA-Seq reads were trimmed using Trimmomatic software, and the processed reads were mapped back to the human genome (hg19) using TOPHAT2. Count matrices were generated using HTSeq. Results: Genome-wide loss of lineage fidelity is a hallmark of aging breast epithelia. Age-dependent differential expression occurred almost exclusively in luminal cells characterized by luminal epithelial cells of older women expressing markers normally expressed in myoepithelial cells. Luminal epithelial cells from histologically normal breast tissue from younger women who carry germline mutations in BRCA1, BRCA2, or PALB2 genes and who are considered to be clinically high risk for breast cancer also exhibited these hallmarks of accelerated aging. Furthermore, accelerated aging of these genetically high risk luminal epithelial cells could be predicted using a biological clock trained on gene expression and DNA methylation profiles of the luminal-specific ELF5 transcription factor. Conclusions: Our study shows that lineage-specific analysis is critical to understanding the molecular mechanisms underlying aging-associated cancer susceptibility. Our results suggest that loss of lineage fidelity is a general manifestation of epithelia that are susceptible to cancer initiation. Moreover, breast aging hallmarks identified in our study reflect a convergent biology of cancer susceptibility, regardless of the specific underlying genetic or age-dependent risk. Luminal epithelial (LEPs) and myoepithelial (MEPs) cells enriched from finite lifespan, non-immortalized human mammary epithelial cells (HMECs) serve as our experimental model system. HMECs are derived from organoids isolated from collected breast tissue samples and grown to 4th passage [LaBarge et al., 2013, Garbe et al., 2009]. We examined genome-wide transcription in 108 enriched LEP and MEP samples from 46 women across a wide spectrum of ages and cancer risk profiles. We examined two age cohorts who were considered to have average risk for breast cancer: young ≤30y women considered to be premenopausal (age range 16-29y, n=32 LEP and MEP samples, m=11 subjects) and older ≥55y women considered to be postmenopausal (age range 56-72y, n=22 LEP and MEP samples, m=8 subjects). Samples from middle-aged women >30y <55y (n=6 LEP and MEP samples, m=3 subjects) were also included. Breast tissue samples from these subjects were collected from reduction mammoplasties (RM). We also examined women considered to be at higher risk to develop breast cancer based on accepted genetic or clinical risk factors (age range 25-72y, n=48 LEP and MEP samples, m=24 subjects). These women are either known carriers of high risk germline mutations (BRCA1, BRCA2, PALB2, PALB2/APC), and/or have personal history and/or known family history of breast cancer. Breast tissue samples from these subjects were collected from prophylactic mastectomies (PM) and normal contralateral-to-tumor (CLTT) or normal peripheral-to-tumor (PTT) tissue samples. Expression data from 4th passage HMEC were validated in primary organoids from young ≤30y and older ≥55y women with average risk for breast cancer (age range 16-66y, n=12 LEP and MEP samples, m=7 subjects.