Changes in peripheral serum levels of total activin A during the human menstrual cycle and pregnancy.

The main objective of this study was to determine whether activin A concentrations in peripheral blood fluctuate during the normal human menstrual cycle and pregnancy. Blood samples were collected longitudinally from five regularly cycling volunteers (22-30 yr) throughout a spontaneous menstrual cycle and cross-sectionally from normal pregnant women attending the antenatal clinic (8-38 weeks gestation: 3-20 subjects/time point). Total (i.e. bound plus free) activin A concentrations were measured using a recently developed two-site enzyme immunoassay that employs an analyte denaturation/oxidation step to eliminate interference due to endogenous activin-binding proteins. During the menstrual cycle, mean serum activin A levels varied in a biphasic manner (by ANOVA, P = 0.02), with highest levels around midcycle (approximately 220 pg/mL) and the late luteal/early follicular phase (approximately 310 pg/mL) and nadirs in both midfollicular (approximately 125 pg/mL) and midluteal (approximately 120 pg/mL) phases. Between the mid- to late luteal phase, the activin A level increased progressively (approximately 2.5-fold; P < 0.05), whereas inhibin A, estradiol, and progesterone all decreased progressively (approximately 10-fold; P < 0.001). During pregnancy, serum activin A levels were much higher than those in nonpregnant subjects, with a value of 2.12 +/- 0.31 ng/mL recorded in week 8. Levels remained at approximately 2 ng/mL between weeks 8-24, but increased thereafter to reach 25.5 +/- 6 ng/mL by week 38, a value approximately 100 times greater than that during the normal menstrual cycle. Serum activin A levels during pregnancy were significantly correlated with inhibin A (r = 0.69; P < 0.001), estradiol (r = 0.55; P < 0.001), and progesterone (r = 0.74; P < 0.001) values. Gel permeation chromatography indicated that all of the detectable activin A in human follicular fluid, pregnancy serum, and term placental extract eluted with an apparent molecular mass between 70-200 kDa, indicating that little, if any, free activin (molecular mass, 25 kDa) is present in these samples. Although these results support a possible endocrine role for circulating activin A during the human menstrual cycle and pregnancy, the observation that all detectable activin A is associated with binding protein(s) raises questions about its relative bioavailability for action on peripheral target cells.