Identification Of Hormone-correlated Genes In Breast As Biomarkers For Menstrual Phases

[Objective]Gene and protein expression in the breast varies relative to menopausal status andmenstrual phase, but cancer risk biomarker research using archived benign breast samples isusually uninformed by menstrual phase data. We hypothesized that gene expression variationrelative to ambient hormones can be exploited to develop markers of menstrual phase thatcan be measured directly in breast tissue.[Methods]Random fine needle aspiration samples from the unaffected breasts of18women wereutilized for Illumina microarray analysis. Gene expression was correlated with simultaneousserum levels of estradiol, progesterone, and follicle stimulating hormone (FSH). Truemenstrual/menopausal states were defined by age and serum hormones. Genes highlycorrelated to serum hormones were used to classify samples into menstrual/menopausalgroups. Findings were validated using Openarray with different sample set.[Results]From35,964genes and12,838undesignated transcripts, we identified genes/transcriptshighly correlated with estradiol (1091genes), progesterone (127genes) or FSH (58genes),generating four panels:21estradiol-stimulated genes,22progesterone-stimulated genes;7FSH-stimulated genes, and10FSH-suppressed genes. By using216independent samples, asubset of genes correlated with estradiol (PDZK1, TNFRSF11A) and progesterone(TNFSF11, DIO2and MYBPC1) were validated. TNFSF11is also found to be highlycorrelated with several proliferation and breast cancer related Genes, including CCND1,PRLR, PGR,TFF1,PARP1,PDZK1and PI3KR1. [Conclusion]True menstrual phase status was reflected in the expression of specific gene sets in thenormal breast, potentially allowing the retrospective menstrual phase classification ofarchived breast samples, and reducing the biological noise in risk biomarker research.