Obstetrics & Gynecology:
Serum Activin A, Inhibin A, and Follistatin Concentrations in Preeclampsia or Small for Gestational Age Pregnancies
Keelan, Jeffrey A. PhD; Taylor, Rennae MHSc; Schellenberg, Jean‐Claude MD, PhD; Groome, Nigel P. PhD; Mitchell, Murray D. DSc; North, Robyn A. MD, PhD
Liggins Institute and Division of Pharmacology and Clinical Pharmacology, and Division of Obstetrics and Gynaecology, University of Auckland Faculty of Medical and Health Sciences, Auckland, New Zealand; Clinique et policlinique d'obstétrique, Maternité de Geneve, Geneva, Switzerland; and Centre for Proteins and Peptides, School of Biological and Molecular Sciences, Oxford Brookes University, Headington, Oxford, United Kingdom.
Address reprint requests to: Jeffrey A. Keelan, University of Auckland, Faculty of Medical and Health Sciences, Liggins Institute and Division of Pharmacology and Clinical Pharmacology, Private Bag 92019, Auckland, New Zealand; E‐mail: firstname.lastname@example.org.
This study was supported by the Auckland Medical Research Foundation, Health Research Council of New Zealand, Lottery Health NZ, and the Auckland University Research Fund.
The authors thank Greg Gamble, Clinical Trials Unit, and Alistair Stewart, Division of Community Health Biostatistics Unit, University of Auckland, for statistical advice.
Received January 16, 2001. Received in revised form September 6, 2001. Accepted September 24, 2001.
OBJECTIVE: To determine whether maternal serum activin A, inhibin A, and follistatin concentrations in idiopathic small for gestational age (SGA) pregnancies are similar to those in normal pregnancies or elevated as in preeclampsia.
METHODS: Maternal serum activin A, inhibin A, and follistatin concentrations were determined in 1) nulliparous women with idiopathic SGA (birth weight <10th percentile; n = 18), preeclampsia (systolic blood pressure ≥140 mmHg or diastolic blood pressure ≥90 mmHg plus proteinuria ≥2+or >0.3 g/24h; n = 22), and normotensive controls, matched for gestational age at sampling (n = 22), and 2) a longitudinal series of samples collected at five intervals throughout pregnancy from nulliparous women with idiopathic SGA (n = 19), preeclampsia (n = 22), preeclampsia plus SGA (n = 15), or who had uncomplicated pregnancies (n = 20).
RESULTS: Serum concentrations of activin A and inhibin A were similar in idiopathic SGA pregnancies to controls. In preeclampsia, activin A and inhibin A levels were markedly increased compared with controls or women with idiopathic SGA (P < .001), particularly in those with early‐onset disease. Follistatin concentrations were only modestly (<twofold) elevated in preeclampsia (P < .001). In the longitudinal study, serum activin A or inhibin A concentrations were increased in women who later developed preeclampsia, whereas in women with idiopathic SGA pregnancy, a small overall increase in activin A levels was observed.
CONCLUSIONS: In contrast to women with preeclampsia, normotensive women with SGA pregnancies do not have markedly elevated circulating levels of activin A and inhibin A. These data support the hypothesis that increased serum activin A concentrations in preeclampsia may be a manifestation of maternal disease rather than just a marker of abnormal placentation.
In preeclampsia and idiopathic small for gestational age (SGA) pregnancies, cytotrophoblast invasion is restricted and remodeling of the spiral arteries is limited, resulting in reduced uteroplacental perfusion.1 Placental villous pathology and reduced fetoplacental blood flow also occur in these conditions.2 Despite similarities in placental pathology, preeclampsia and idiopathic SGA pregnancies show marked differences in maternal patho‐physiology. Preeclampsia involves endothelial cell activation,3 an intravascular inflammatory response with monocyte and neutrophil activation,4 and activated coagulation; resulting in maternal disease. In contrast, in idiopathic SGA, the mother remains normotensive and healthy.
Activin A and inhibin A are related dimeric glycoproteins that belong to the transforming growth factor‐β superfamily.5 They are synthesized and produced by numerous endocrine tissues, including the ovary and placenta, and by such entities as monocytes, fibroblasts, and endothelial cells. In some but not all tissues, inhibin also acts as an activin antagonist.5,6 Maternal circulating concentrations of activin A and inhibin A are increased in preeclampsia, particularly in women with severe early‐onset disease.7–9 The relative amount of free bioactive activin A circulating in preeclampsia has yet to be determined, although a recent study reported that serum follistatin concentrations were not elevated in women with preeclampsia.9
While the marked elevation in activin A and inhibin A in preeclampsia has attracted the attention of investigators seeking a biochemical predictor of preeclampsia,10,11 the pathophysiologic significance of these observations remains uncertain. It has been assumed that the source of activin A and inhibin A in preeclampsia is predominantly placental tissue, reflecting abnormal placentation and trophoblast function.11,12
Activin has diverse biological effects, including regulation of cytotrophoblast differentiation;13 enhanced differentiation of monocytes into macrophages;14 and regulation of prostanoid and cytokine production by monocytes, macrophages,15,16 and placental cells.17 Of note, activin inhibits endothelial cell growth,18 and stimulates endothelin production.19 The biological activity of activin is inhibited by follistatin, a high‐affinity activin‐binding glycoprotein that is produced by a wide range of cells and tissues.20,21 At least three follistatin isoforms are known to exist (288, 303, and 315 kDa in size), all of which can bind and neutralize two activin molecules.5,20,21 A recent study reported that serum follistatin concentrations were not elevated in women with preeclampsia,9 suggesting that relative amounts of free activin A in the maternal circulation may be elevated in preeclamptic pregnancies.
In light of the marked elevation in concentrations of activin found in the circulation of women with early‐onset preeclampsia, we hypothesized that activin A might be a manifestation of maternal disease. Our hypothesis would predict that free activin A and inhibin A levels would be elevated in women with preeclampsia but normal in women with idiopathic SGA pregnancies (ie, without maternal disease). To explore this hypothesis, we measured activin A, inhibin A, and follistatin in the circulation of women with preeclampsia, those with idiopathic SGA, and those with normal pregnancies. We also measured inhibin A and activin A throughout pregnancy in women who subsequently developed preeclampsia (with and without an SGA baby), those with idiopathic SGA pregnancies, and controls with uncomplicated pregnancies.
MATERIALS AND METHODS
We conducted two studies: a cross‐sectional study of women with preeclampsia, idiopathic SGA, and uncomplicated pregnancies, and a longitudinal study of women who had an uncomplicated pregnancy or were destined to develop preeclampsia or idiopathic SGA. The regional health authority's ethics committee approved both studies, and informed consent was obtained from all women.
Activin A, inhibin A, and follistatin 288 were measured in maternal serum from nulliparous women with idiopathic SGA (n = 18) or preeclampsia (n = 22) and normotensive controls who delivered an appropriately grown baby at term (n = 22). Exclusion criteria were multiple pregnancy, underlying medical conditions, and use of aspirin. Gestational age was calculated from the date of the last menstrual period and confirmed by an early ultrasonogram. If there was a discrepancy in the gestation of ±5 days, the gestational age was based on the early ultrasonogram. The gestation of blood sampling in cases and controls was matched to within ±7 days. Blood sampling was performed at a mean (± standard deviation [SD]) of 34.9 ± 3.2 weeks of gestation in the SGA group, 34.8 ± 3.5 weeks in the preeclampsia group, and 35.1 ± 3.5 weeks in controls. Serum samples were stored at −80C until assayed.
A community‐based longitudinal study of 1496 healthy nulliparous women, of whom 71 (4.8%) developed preeclampsia, was conducted between 1993 and 1995.10,22 Exclusion criteria were fetal congenital malformations, multiple pregnancy, essential hypertension, diabetes, renal disease, or use of aspirin. Serial serum samples were obtained at 8–13, 15–19, 21–25, 27–30, and 35–38 weeks of gestation and stored at −80C. Activin A and inhibin A were assayed in maternal serum from women with idiopathic SGA (n = 19), preeclampsia but no SGA baby (n = 22), preeclampsia and an SGA baby (n = 15), or uncomplicated pregnancy (n = 20). From a list of all patients in the study, samples were selected for analysis from the first 19 women with idiopathic SGA pregnancies, 22 with preeclampsia but no SGA baby, and 15 with preeclampsia and an SGA baby who had complete or almost complete sample sets. Follistatin was not measured in light of the data obtained in the cross‐sectional study.
We defined SGA as a sex‐adjusted birth weight less than the 10th percentile.23 The idiopathic SGA group had no clinical evidence of congenital infection, congenital anomalies, or chromosomal abnormalities. Preeclampsia was defined as systolic blood pressure ≥140 mmHg or diastolic blood pressure ≥90 mmHg on two or more occasions after 20 weeks of gestation, but before the onset of labor, plus proteinuria ≥2+ or >0.3 g/24 h.
Total activin A, total inhibin A, and follistatin 288 were measured by using enzyme‐linked immunosorbent assay previously validated for serum studies.24–26 The limit of detection of the activin A, inhibin A, and follistatin assays was 14.0, 2.3, and 39 pg/mL, respectively. Interassay and intraassay precision were 7.1% and 4.7%, 8.6% and 6.2%, and 6.8% and 6.9%, respectively.
Continuous clinical data and the immunoassay data in the cross‐sectional study were analyzed by using analysis of variance to compare findings across all groups, followed by the Tukey test for post hoc comparisons. Categorical clinical data were analyzed by using the Fisher exact test. The longitudinal data were log transformed and analyzed with mixed‐models procedure by using SAS software (SAS Institute, Inc., Cary, NC). The covariance structure for the repeated measures was modeled by using a heterogenous compound symmetry matrix. This method offers an efficient approach to repeated measures and permits maximum likelihood imputation of missing data. Pairwise comparison of groups at each gestational age point was performed by using Student t tests; P < .05 was considered significant.
Clinical characteristics of the women are shown in Table 1. All women with preeclampsia had a normal serum creatinine concentration at the time of blood sampling (mean [±SD] 66 ± 10 μmol/L).
Maternal serum activin A and inhibin A concentrations were elevated in women with preeclampsia compared with controls (Table 2). In contrast, maternal serum activin A and inhibin A levels were not elevated in idiopathic SGA pregnancies compared with gestational age–matched women. In preeclamptic women delivered before 34 weeks, median serum activin A (60.6 ng/mL) and inhibin A (33.9 ng/mL) concentrations were markedly elevated compared with controls (3.6 and 3.5 ng/ mL, respectively; P = .002 for both), with no overlap between the groups; however, the sample was small (n = 7). Activin A and inhibin A were negatively correlated with gestational age at sampling in women with preeclampsia (Spearman's rho [rs] = −.56 and −.57, respectively; P < .01), whereas all three analytes had a weak positive correlation with gestational age at sampling in the control and SGA groups.
When we combined the three patient groups, activin A concentrations correlated with follistatin (rs =.59 P < .001) and inhibin A levels (rs = 0.83, P < .001) levels. The correlation between activin A and follistatin levels remained significant in the preeclamptic (rs = .567, P < .01) and SGA groups (rs = .599, P < .02) when analyzed separately, but not in the control group (rs = −.233, P = .284). Overall, median activin A concentrations were approximately fourfold greater in the preeclampsia group than in controls, compared with the less than twofold difference in follistatin levels. The ratio of activin to follistatin was significantly higher in the preeclamptic group than in controls (mean [± SD], 6.05 ± 5.33 vs 2.08 ± 1.66, P < .001 by analysis of variance).
Table 3 shows the clinical characteristics of the patient groups. In all groups, the concentration of activin A and inhibin A in maternal serum increased with gestational age (P < .001) (Figure 1). An interaction was observed between gestational age at sampling and patient group for both polypeptides (P < .001). Concentrations of activin A in both preeclamptic groups were elevated compared with controls and showed an interaction between group and time (P < .001). Activin A concentrations were significantly elevated in the preeclampsia without SGA group from 8–13 weeks of gestation onward and in the preeclampsia with SGA group from 15–19 weeks onward (Figure 1A). In the idiopathic SGA group, activin A levels differed significantly from those in controls (P = .03), with no interaction between group and time (Figure 1A).
Differences in serum inhibin A concentrations among groups were less marked than those of activin A. Inhibin A levels in the idiopathic SGA pregnancies were not elevated compared with controls (P = .22). Only the preeclampsia without SGA group showed an interaction between group and time compared with controls (P < .001), with significantly increased inhibin A levels at 27–30 weeks and 35–38 weeks of gestation (Figure 1B). In the preeclampsia plus SGA group, the overall difference was of borderline significance (P = .055), reaching statistical significance at 27–30 weeks.
This study represents a comprehensive analysis of maternal serum activin A, inhibin A, and follistatin concentrations in pregnancies with idiopathic SGA compared with those with preeclampsia. Concentrations of activin A and inhibin A in idiopathic SGA pregnancies were similar or slightly elevated than were those in women with normal pregnancies. This finding is in marked contrast to women with preeclampsia, who had significantly elevated levels compared with women who had normal or idiopathic SGA pregnancies. In the cross‐sectional study of women at the time of disease, the degree of elevation of activin A and inhibin A levels in preeclampsia was similar to that reported in other studies, including the observation that women with early‐onset disease have dramatically increased activin A levels compared with normotensive women.12,27 This finding was further demonstrated by the observed negative correlation between gestational age and activin A and inhibin A levels in women with preeclampsia, compared with the positive correlations seen in the control and SGA groups.
Although no differences in serum levels of activin A or inhibin A were detected in women with SGA in the cross‐sectional study, analysis of the longitudinal data suggested that activin A levels in this group may be modestly increased. The lack of observed differences in activin A values between the control and idiopathic SGA groups in the cross‐sectional study may result from heterogeneity of the SGA group; the presence of constitutionally small infants may confound the results. This is unlikely, however, because 67% (12 of 18) of infants in the idiopathic SGA group had a birth weight less than the third percentile, similar to the 53% that were less than the third percentile in the longitudinal SGA group. Power differences may also explain the disparity in the findings between the two studies; the longitudinal study had an 80% power to detect an overall difference of about 15% between the SGA and control groups at α = .05, much higher than the cross‐sectional study. Larger studies using more refined criteria for growth restriction that included a measure of placental pathology, such as abnormal umbilical Doppler ultrasonographic measurements, are required to confirm and expand our findings.
In the longitudinal study, activin A and inhibin A concentrations were elevated weeks before the onset of clinical disease in women with preeclampsia, regardless of whether their baby was SGA. The elevation in activin A level in preeclampsia was less marked in the longitudinal study than in the cross‐sectional study, as expected, because the samples were collected before the onset of clinical disease. The reduced significance observed between activin A and inhibin A concentrations in the preeclampsia plus SGA group compared with controls at 35–38 weeks (Figure 1) probably reflects insufficient power, because half of this group had early‐onset disease and delivered before 34 weeks of gestation. This resulted in a relatively small sample (n = 8), which reduced the power to discern a significant difference. Elevated serum activin A concentrations were recently reported in patients with preeclampsia plus intrauterine growth restriction, similar to those with preeclampsia only.9
Although follistatin levels in preeclampsia were modestly elevated compared with those in controls, activin A concentrations remained considerably higher than follistatin levels. These findings are similar to those of D'Antona et al, who reported normal circulating follistatin levels in women with preeclampsia.9 Although this provides evidence that unbound biologically active activin A circulates in excessive concentrations in the blood of women with preeclampsia, caution must be used in interpreting these data. The follistatin assay that we used preferentially detects follistatin 288,26 but it has not yet been established whether this is the major follistatin isoform circulating in pregnancy. Nevertheless, our findings suggest that preeclampsia is associated with increased activin bioactivity in the maternal circulation.
Our observations are consistent with the hypothesis that activin A and inhibin A may be a manifestation of maternal disease and may play a role in the pathogenesis of the maternal response in preeclampsia. The fact that activin A and inhibin A concentrations in the maternal circulation are normal or minimally elevated in idiopathic SGA pregnancies suggests that abnormal levels of these polypeptides in preeclampsia do not merely reflect placental pathology. It has not yet been determined how much of the increased amounts of activin A and inhibin A in sera of women with preeclampsia is derived from the placenta. Because both endothelial cells and monocytes undergo activation in preeclampsia,3,4 a significant contribution from these sources cannot be discounted. An alternative explanation for the elevated activin A and inhibin A levels in preeclampsia is reduced renal clearance.28 However, this is unlikely to be the sole reason, because all women with preeclampsia had normal serum creatinine levels at the time of sampling and in the longitudinal study, activin A concentrations were elevated before preeclampsia developed.
High circulating concentrations of biologically active activin in preeclampsia could contribute to maternal hypertension through several mechanisms. Production of the potent vasoconstrictor endothelin by vascular endothelial cells has been reported to be stimulated by activin,19 and activin has been shown to increase production of thromboxane and tumor necrosis factor‐α by macrophages.16 More recently, activin was shown to promote vascular smooth muscle–cell proliferation and differentiation toward a contractile phenotype.29
In conclusion, activin A and inhibin A levels in maternal serum are not or are only marginally elevated in women with SGA pregnancies, in contrast to those with preeclampsia. Our findings are consistent with the hypothesis that the increased serum activin A concentrations in preeclampsia may be a manifestation of maternal disease, rather than simply a consequence of altered placental secretion. The possibility that some of the elevated activin A or inhibin A concentrations in the sera of preeclamptic women originate from a nonplacental source and play a role in disease pathogenesis should be explored.
1. Lim KH, Zhou Y, Janatpour M, McMaster M, Bass K, Chun SH, et al. Human cytotrophoblast differentiation/invasion is abnormal in preeclampsia. Am J Pathol 1997; 151:1809–18.
2. Macara L, Kingdom JCP, Kaufmann P, Kohnen G, Hair J, More IAR, et al. Structural analysis of placental terminal villi from growth-restricted pregnancies with abnormal umbilical artery Doppler waveforms. Placenta 1996;17:37–48.
3. Taylor RN, de Groot CJ, Cho YK, Lim KH. Circulating factors as markers and mediators of endothelial cell dysfunction in preeclampsia. Sem Reprod Endocrinol 1998; 16:17–31.
4. Redman CW, Sacks GP, Sargent IL. Preeclampsia: An excessive maternal inflammatory response to pregnancy. Am J Obstet Gynecol 1999;180:499–506.
5. DePaolo LV, Bicsak TA, Erickson GF, Shimasaki S, Ling N. Follistatin and activin: A potential intrinsic regulatory system within diverse tissues. Proc Soc Exp Med Biol 1991;198:500–12.
6. Lewis KA, Gray PC, Blount AL, MacConnell LA, Wiater E, Bilezikjian LM, et al. Betaglycan binds inhibin and can mediate functional antagonism of activin signalling. Nature 2000;404:411–4.
7. Muttukrishna S, Knight PG, Groome NP, Redman CW, Ledger WL. Activin A and inhibin A as possible endocrine markers for preeclampsia. Lancet 1997;349:1285–8.
8. Silver HM, Lambert-Messerlian GM, Star JA, Hogan J, Canick JA. Comparison of maternal serum total activin A and inhibin A in normal, preeclamptic, and nonproteinuric gestationally hypertensive pregnancies. Am J Obstet Gynecol 1999;180:1131–7.
9. D'Antona D, Reis FM, Benedetto C, Evans LW, Groome NP, de Kretser DM, et al. Increased maternal serum activin A but not follistatin levels in pregnant women with hypertensive disorders. J Endocrinol 2000;165:157–62.
10. Muttukrishna S, North R, Morris J, Schellenberg J, Taylor R, Asselin J, et al. Serum inhibin A and activin A are elevated prior to the onset of preeclampsia. Hum Reprod 2000;15:1640–5.
11. Aquilina J, Barnett A, Thompson O, Harrington K. Second-trimester maternal serum inhibin A concentration as an early marker for preeclampsia. Am J Obstet Gynecol 1999;181:131–6.
12. Silver HM, Lambert-Messerlian GM, Star JA, Hogan J, Canick JA. Comparison of maternal serum total activin A and inhibin A in normal, preeclamptic, and nonproteinuric gestationally hypertensive pregnancies. Am J Obstet Gynecol 1999;180:1131–7.
13. Caniggia I, Lye SJ, Cross JC. Activin is a local regulator of human cytotrophoblast cell differentiation. Endocrinology 1997;138:3976–86.
14. Yamada R, Suzuki T, Hashimoto M, Eto Y, Shiokawa K, Muramatsu M. Induction of differentiation of the human promyelocytic cell line HL-60 by activin/EDF. Biochem Biophys Res Comm 1992;187:79–85.
15. Petraglia F, Sacerdote P, Cossarizza A, Angioni S, Genaz-zani AD, Franceschi C, et al. Inhibin and activin modulate human monocyte chemotaxis and human lymphocyte interferon-gamma production. J Clin Endocrinol Metab 1991;72:496–502.
16. Nusing RM, Barsig J. Induction of prostanoid, nitric oxide, and cytokine formation in rat bone marrow derived macrophages by activin A. Br J Pharmacol 1999;127:919–26.
17. Keelan JA, Zhou RL, Mitchell MD. Activin A exerts both pro- and anti-inflammatory effects on human term gestational tissues. Placenta 2000;21:38–43.
18. Kozian DH, Ziche M, Augustin HG. The activin-binding protein follistatin regulates autocrine endothelial cell activity and induces angiogenesis. Lab Invest 1997;76:267–76.
19. Brown MR, Vaughan J, Jimenez LL, Vale W, Baird A. Transforming growth factor-beta: Role in mediating serum-induced endothelin production by vascular endothelial cells. Endocrinology 1991;129:2355–60.
20. Sugino H, Sugino K, Hashimoto O, Shoji H, Nakamura T. Follistatin and its role as an activin-binding protein. J Med Invest 1997;44:1–14.
21. Phillips DJ, de Kretser DM. Follistatin: A multifunctional regulatory protein. Front Neuroendocrinol 1998;19:287–322.
22. North RA, Taylor RS, Schellenberg JC. Evaluation of a definition of preeclampsia. Br J Obstet Gynaecol 1999; 106:767–73.
23. Guaran RL, Wein P, Sheedy M, Walstab J, Beischer NA. Update of growth percentiles for infants born in an Australian population. Aust N Z J Obstet Gynaecol 1994; 34:39–50.
24. Groome NP, Illingworth PJ, O'Brien M, Cooke I, Ganesan TS, Baird DT, et al. Detection of dimeric inhibin throughout the human menstrual cycle by two-site enzyme immunoassay. Clin Endocrinol 1994;40:717–23.
25. Knight PG, Muttukrishna S, Groome NP. Development and application of a two-site enzyme immunoassay for the determination of ‘total’ activin-A concentrations in serum and follicular fluid. J Endocrinol 1996;148:267–79.
26. Evans LW, Muttukrishna S, Groome NP. Development, validation and application of an ultra-sensitive two-site enzyme immunoassay for human follistatin. J Endocrinol 1998;156:275–82.
27. Muttukrishna S, Knight PG, Groome NP, Redman CW, Ledger WL. Activin A and inhibin A as possible endocrine markers for preeclampsia. Lancet 1997;349:1285–8.
28. Harada K, Shintani Y, Sakamoto Y, Wakatsuki M, Shit-sukawa K, Saito S. Serum immunoreactive activin A levels in normal subjects and patients with various diseases. J Clin Endocrinol Metab 1996;81:2125–30.
29. Engelse MA, Neele JM, van Achterberg TA, van Aken BE, van Schaik RH, Pannekoek H, et al. Human activin-A is expressed in the atherosclerotic lesion and promotes the contractile phenotype of smooth muscle cells. Circ Res 1999;85:931–9.
© 2002 The American College of Obstetricians and Gynecologists