Activin A and follistatin are dynamically regulated during human pregnancy

Activin A (βA–βA) and activin B (βB–βB) are related dimeric proteins that regulate numerous cellular activities. Activin activity is bioneutralized by follistatin, a specific and high-affinity binding protein. Recently, our group developed specific and sensitive enzyme-linked immunosorbent activin assays that do not detect either activin isoform when bound to follistatin, therefore, the assays are specific for biologically relevant ligands. Activin A is measurable in the serum of pregnant women (cross-sectional sample collection), while activin B is not detected in maternal serum. However, activin B is measurable in amniotic fluid and cord blood sera. The purpose of this study was to measure serum activin A, activin B, and follistatin prospectively in longitudinally collected samples during pregnancy. This study design offered observations of relative changes in serum hormone concentration with each person serving as an internal reference. Serum samples were collected bimonthly from seven pregnant women beginning within the second month of gestation, and up to, but not including, the onset of labor. Six of the seven women had normal labor and delivery. One patient required pitocin (an oxytocin agonist) for induction of labor which led to delivery. Activin A, activin B, total follistatin, free follistatin, human chorionic gonadotropin, estradiol, progesterone, FSH, and LH were measured in maternal serum samples using specific assays. Serum activin A levels increased in the final month of pregnancy in the six patients who delivered following normal labor (<0·78 ng/ml (first trimester) to 1–6 ng/ml (term)). Activin B was not detected in any serum sample (<0·78 pg/ml). Total serum follistatin (free follistatin, follistatin–activin, and follistatin–inhibin) increased 10- to 45-fold in the final month of pregnancy in four of the women undergoing normal labor (10 ng/ml (first trimester) to 100–450 ng/ml (final month)). Total follistatin was high and variable in two women throughout pregnancy. Total follistatin returned to basal serum concentration in three of the patients during the last 2 weeks of pregnancy. Free follistatin was detected throughout pregnancy (range <2–35 ng/ml). Free follistatin represented a small percentage of the total follistatin throughout the time of pregnancy and did not rise coincident with the rise in total follistatin. Serum activin A and activin B were not detected during the entire course of pregnancy in the one patient who did not have normal labor and total follistatin did not rise in the last trimester of pregnancy. Gonadotropin and steroid hormones were measured in all patients and were within normative ranges for human pregnancy (inclusive of the non-laboring patient). The results suggest that immunodetectable activin A is present in the third trimester of pregnant women who have normal onset labor. The total follistatin assay results suggest that follistatin–activin (or –inhibin) complexes are upregulated during the third trimester of pregnancy. Importantly, activin A production exceeds the binding capacity of circulating follistatin. Because binding protein free activin A is biologically active we conclude that the activin A detected in late pregnancy is biologically relevant. The findings are consistent with our hypothesis that activin A is an endocrine factor during the last trimester of human pregnancy and may be involved in normal labor. Journal of Endocrinology (1997) 152, 167–174