Objective: To investigate the diurnal rhythm of estrogens in normally cyclic women during reproductive life. Design: Multiple saliva sampling in normally cyclic healthy women during reproductive life at different phases of their menstrual cycles was carried out. Methods: Salivary estradiol was measured by radioimmunoassay in samples collected every 2 h for 24 h from 15 normally cyclic healthy women during reproductive life during the menstrual phase, the late follicular/peri-ovulation phase, the early to mid luteal phase and the late luteal phase, respectively, of their menstrual cycles. The levels of salivary estradiol were analyzed by means of periodic regression. Results: A daily biological rhythm of free estradiol was found after quantification with a nonlinear periodic regression model. The observed diurnal free estradiol rhythm consists of two major components: an asymmetrically peaked diurnal cycle and ultradian harmonics in the range of 6 to 12 h. The diurnal and ultradian rhythms were remarkably consistent throughout the menstrual cycle in terms of mesor (24 h mean level), peak width and amplitude. There was a tendency for the 24-h rhythm acrophases to converge in the early morning, while the acrophase of the menstrual phase occurred significantly later than in the late follicular/peri-ovulation phase. Conclusions: The diurnal rhythm of estradiol has a similar complex temporal organization for different menstrual phases. The menstrual cycle mainly modulates the acrophase of the diurnal rhythm.

We investigated the possibility of a direct action of androgens on the expression of the human corticotropin-releasing hormone (CRH), which plays a central role in the hypothalamicpituitary-adrenal (HPA)-axis. Colocalization of CRH and nuclear/cytoplasmic androgen receptor (AR) was found in neurons of the paraventricular nucleus (PVN) in the human hypothalamus. A potential androgen-responsive element (ARE) in the human CRH promoter was subsequently analyzed with bandshifts and cotransfections in neuroblastoma cells. In the presence of testosterone, recombinant human AR bound specifically to the CRH-ARE. Expression of AR in combination with testosterone repressed CRH promoter activity through the ARE. We conclude that androgens may directly affect CRH neurons in the human PVN via AR binding to the CRH-ARE, which may have consequences for sex-specific pathogenesis of mood disorders.