Fetal Hypothalamus-Pituitary-Adrenal Responses to Estradiol Sulfate

Abstract

Estradiol (E₂) is an important modifier of the activity of the fetal hypothalamus-pituitary-adrenal axis. We have reported that estradiol-3-sulfate (E₂SO₄) circulates in fetal blood in far higher concentrations than E₂ and that the fetal brain expresses steroid sulfatase, required for local deconjugation of E₂SO₄. We performed the present study to test the hypothesis that chronic infusion of E₂SO₄ chronically increases ACTH and cortisol secretion and that it shortens gestation. Chronically catheterized fetal sheep were treated with E₂SO₄ intracerebroventricular (n = 5), E₂SO₄ iv (n = 4), or no steroid infusion (control group, n = 5). Fetuses were subjected to arterial blood sampling every other day until spontaneous birth for plasma hormone analysis. Treatment with E₂SO₄ attenuated preparturient increases in ACTH secretion near term without affecting the ontogenetic rise in plasma cortisol. Infusion of E₂SO₄ intracerebroventricularly significantly increased plasma E₂, plasma E₂SO₄, and plasma progesterone and shortened gestation compared with all other groups. These results are consistent with the conclusion that E₂SO₄: 1) interacts with the hypothalamus-pituitary-adrenal axis primarily by stimulating cortisol secretion and inhibiting ACTH and pro-ACTH secretion by negative feedback; and 2) stimulates the secretion of E₂ and E₂SO₄. We conclude that the endocrine response to E₂SO₄ in the fetus is not identical with the response to E₂.

In fetal sheep, the hypothalamus-pituitary-adrenal (HPA) axis is a central component of the fetal stress response and plays a central role in the initiation of parturition (1–4). The activity of the fetal HPA axis is influenced by both physiological (blood gases, blood pressure, etc.) and endocrine variables. We have previously reported that estradiol (E₂) is a potent stimulator of the fetal HPA axis, increasing neuronal activity in brain regions known to control HPA axis function and increasing the expression of genes within the fetal brain that are known to be important for HPA axis activity (5–8). The effect of E₂ on HPA axis activity in the fetus mirrors the effect of E₂ and other estrogens in adult animals. Estrogen has been shown to augment HPA activity in adult rats. Female rats have higher basal and stimulated corticosterone than males (9) and exert maximum responses to stress during proestrus, when plasma estrogen is elevated (10–14). Ovariectomy attenuates adrenal steroid production in response to stress; however, activity of the HPA axis is ameliorated after hormone replacement with E₂ (15, 16).

The major forms of estrogen circulating in fetal plasma are estrone-3-sulfate (E₁SO₄) and estradiol-3-sulfate (E₂SO₄) (17, 18). These sulfoconjugated forms of estrogen are not known to be directly biologically active. Sulfoconjugated steroids are abundant in fetal plasma (18, 19) and are able to bind estrogen receptors within target tissues only after deconjugation by the enzyme steroid sulfatase (STS) (20, 21). Accordingly, we hypothesized and demonstrated the presence of both STS and estrogen sulfotransferase (STF) in the fetal brain, allowing for bidirectional conversion of the conjugated and unconjugated forms of the steroids within an important target tissue (22, 23). We have also demonstrated that short-term infusion of E₂SO₄ into fetal plasma increases fetal HPA axis activity and stimulates cellular responses in the fetal brain that are consistent with increased neuronal activity (18, 24).

Although we have demonstrated that relatively short infusions of E₂SO₄ augment fetal HPA axis activity, we could not conclude that long-term exposure to elevated levels of E₂SO₄ produced persistent endocrine responses. We therefore hypothesized that infusion of E₂SO₄ into either the vascular compartment or the central nervous system of the fetal sheep produce persistent increases in fetal ACTH and cortisol concentrations. The present study was designed to test this hypothesis.

Materials and Methods

Fourteen pregnant ewes with time-dated singleton fetuses (120–125 d gestation) were used in this study. They were divided into three experimental groups: control (n = 5), E₂SO₄ infusion intracerebroventricular (icv; 1 mg/d, n = 5), and E₂SO₄ infusion iv (1 mg/d, n = 4). Plasma hormone data for the fetuses in the control group have been published previously (25). The three groups reported in this study were performed at the same time as a test of estrogen receptor blockade (25). The control group is therefore a time- and conditions-matched control for both studies except for the position of the infusion (into the lateral cerebral ventricle). All experiments were approved by the University of Florida Institutional Animal Care and Use Committee and were performed in accordance with the American Physiological Society's Guiding Principles of the Care and Use of Animals (26). Animals were housed in individual pens and were given access to food and water ad libitum.

Surgical preparation

Ewes were fasted for 24 h before surgery. Ewes were intubated and anesthetized with halothane (0.5–2%) in oxygen during surgery. Before the start of surgery, 750 mg ampicillin (Polyflex; Fort Dodge Laboratories, Fort Dodge, IA) was administered im. The uterus was exposed through a midline incision and the fetal hindlimbs were located by palpation. The hindlimbs were delivered through a small incision in the uterus. Vascular catheters were inserted into fetal tibial arteries and saphenous veins, and the catheter tips were advanced to the femoral arteries and veins, as described previously (25, 27). The catheters were filled with heparin to prevent clotting and closed at the end using a sterile brass nail. An additional catheter (outside diameter 0.09 in.; inside diameter, 0.05 in.) was sewn to the skin of one fetal hindlimb for access to amniotic fluid. Fetuses in the E₂SO₄ iv group received an infusion of E₂SO₄ into one of the venous catheters using an osmotic minipump (size 2mL4; Alza Corp., Palo Alto, CA). Fetuses receiving icv infusions were instrumented with arterial catheters and an extra catheter placed in the lateral cerebral ventricle as previously described (28). E₂SO₄ was infused using an osmotic minipump (size 2mL4; Alza). Antibiotics (750 mg ampicillin) were administered into the amniotic cavity via direct injection.

Postoperatively, ewes were given 1 mg/kg flunixin meglumine (Webster Veterinary, Sterling, MA) for analgesia and returned to their pens in which they were monitored until they could stand on their own. Twice daily during a 5-d recovery period, ewes were given antibiotic (ampicillin, 750 mg, im) and rectal temperatures were monitored for indication of postoperative infection.

Blood collection

After the recovery period, fetal blood samples were drawn from the arterial catheter every other morning (between 0800 and 1000 h) for use in hormone assays. Samples were kept on ice until centrifuged at 3000 × g for 15 min at 4 C to separate red blood cells and plasma. Plasma was stored at −20 C until analysis. Blood gases were measured at the time of blood sampling using an ABL 77 radiometer (Radiometer America Inc., Cleveland, OH) blood gas analyzer.

Plasma hormone assays

ACTH_1–39 was measured using a two-site immunoradiometric assay purchased from DiaSorin (Stillwater, MN; catalog no. 27130). Proopiomelanocortin (POMC) and pro-ACTH (reported herein as POMC) was measured using an ELISA kit from Immunodiagnostic Systems Ltd. (Fountain Hills, AZ; catalog no. AC-71F1) according to the manufacturer's instructions. Cross-reactivity with POMC and pro-ACTH is 100%, as reported by the manufacturer.

Plasma E₂ and E₂SO₄ were measured using the estradiol ELISA kit from Oxford Biomedical Research (Oxford, MI; catalog no. EA70) as previously reported (18). E₂ was extracted from plasma using hexane and ethyl acetate (3:2), and E₂SO₄ was extracted from plasma using ethanol (18). Cross-reactivity with 17β-estradiol, estriol, and estrone in this kit is 100, 0.41, and 0.10%, respectively, as reported by the manufacturer. Cross-reactivity with E₂SO₄ is 100%, as we have reported previously (18). Plasma estrone (E₁) was measured using the estrone EIA kit from Cayman Chemical (Ann Arbor, MI; catalog no. 582301). E₁ was extracted from plasma using ethanol, as described for the estradiol assay. Cross-reactivity with estrone, E₁SO₄, estrone-3-glucuronide, cortisol, and 17β-estradiol are 100, 100, 100, less than 0.01, and less than 0.01%, respectively. Values reported from this assay represent total E₁ (unconjugated plus conjugated E₁).

Plasma cortisol was measured using a cortisol ELISA kit from Oxford Biomedical Research (catalog no. EA65). Cortisol was extracted using ethanol, as described previously (29). Cross-reactivity with cortisol, cortisone, 11-deoxycortisol, and corticosterone in this kit is 100, 15.77, 15, and 4.81%, respectively, as reported by the manufacturer. Dehydroepiandrosterone sulfate (DHAS) was measured with a ¹²⁵I-DHEA-SO₄ Coat-A-Count kit from Diagnostic Products Corp. (Los Angeles, CA; catalog no. TKDS5). Percent cross-reactivity with DHAS, dehydroepiandrosterone, and E₁SO₄ was 100, 0.57 and 0.25%, respectively, as reported by the manufacturer. Progesterone was measured with a ¹²⁵I-Progesterone Coat-A-Count kit from Diagnostic Products (catalog no. TKPG5) according to the manufacturer's instructions.

Statistics and estimation of secretion rates

Two-way ANOVA was used to analyze differences among treatment groups in this study. One-way ANOVA was used to test the effect of time within each treatment group. Pairwise multiple comparisons were performed using the Student-Newman-Keuls multiple-range test. Effect of treatment on day of parturition was analyzed using the modified Wilcoxon method for survival analysis (30). SPSS version 17 (SPSS Corp., Chicago, IL) was used for analyses. A significance level of P < 0.05 was used to reject the null hypothesis. Values are reported as mean ± sem. To estimate endogenous secretion rates for E₂ and E₂SO₄, metabolic clearance rates (MCR) were calculated for each hormone using infusion rates and changes in plasma concentration from this (for E₂SO₄) and one previous (for E₂) study from this laboratory (27) using the following formula: MCR = (infusion rate)/(change in plasma concentration at steady state).

Results

In the control group, we measured an increase in fetal ACTH_1-39, POMC, and cortisol (Fig. 1) before parturition. ACTH_1-39 concentrations were not different among groups (P = NS by ANOVA), whereas POMC was significantly lower in the iv and icv groups compared with the control group (P < 0.001 for main effect of group in ANOVA and P < 0.05 by Student Newman-Keuls test for comparison of groups). Despite the lack of overall significant differences among groups for ACTH_1-39, there was a statistically significant (P < 0.05 by simple effects contrast) increase in plasma ACTH_1-39 concentration in the control group but not in the other experimental groups. Although there was no preparturient increase in plasma concentrations of ACTH_1-39 and POMC in the iv and icv groups near term and although the increase in ACTH_1-39 concentration was limited to the last measurement in utero, fetal plasma cortisol concentrations increased markedly in all groups before parturition (Fig. 1). Fetal plasma cortisol was constant until d −7 when it began to increase until the end of gestation in all experimental groups. There were no significant differences in plasma cortisol among groups. The trend for increased cortisol over time was statistically significant by two-way ANOVA (P < 0.001 for main effect of gestational age), although there was no statistically significant difference between groups (P = NS for main effect of group in ANOVA).

Fig. 1.

Plasma ACTH_1-39 (left panels), POMC (middle panels), and cortisol (right panels) concentrations in chronically catheterized singleton ovine fetuses throughout the last 13 d of gestation. Fetal sheep were treated with no steroid infusion (control, top panels), estradiol sulfate icv (middle panels), estradiol sulfate iv (bottom panels). Values are represented as mean ± sem. Asterisks denote statistically significant change from −14 d.

Plasma concentrations of E₂, E₂SO₄, and total (sulfoconjugated plus unconjugated) E₁, (Fig. 2) did not increase significantly before parturition as analyzed by one- and two-way ANOVA. Infusion of E₂SO₄ by both routes increased fetal plasma concentrations of E₁, E₂, and E₂SO₄ (P < 0.001 for main effect of group for all three hormones; Fig. 2). Plasma concentrations of E₁, E₂, and E₂SO₄ were significantly higher in the icv group than in the iv or control groups (P < 0.05 by Student-Newman-Keuls test), and plasma concentrations of E₂ and E₂SO₄ were significantly higher in the iv group than in the control group (P < 0.05 by Student-Newman-Keuls test). Values of plasma E₂ concentration in the icv group were approximately doubled compared with fetuses in the iv group (197.0 ± 9.9 vs. 82.3 ± 10.5 pg/ml) and increased approximately 6-fold compared with control fetuses (32.3 ± 8.5 pg/ml). Similar to E₂, increases in plasma E₂SO₄ concentrations were highest in the icv group. Infusion of E₂SO₄ significantly reduced the ratio of E₂SO₄ to E₂ (P < 0.001, two way ANOVA). Plasma progesterone (P₄) concentrations were highest in the icv group (P < 0.001for main effect of group in ANOVA; P < 0.05 comparing control with icv for P₄ by Student-Newman-Keuls test). E₂SO₄ infusion icv significantly increased plasma P₄ concentrations beginning on d −15, peaking at d −9 (10.4 ± 1.9 ng/ml), and remained elevated relative to other experimental groups until term. The plasma E₂ to P₄ ratio was significantly different among groups (P < 0.001, two way ANOVA). The E₂ to P₄ ratio increased after the E₂SO₄ treatment iv (0.018 ± 0.0026) and icv (0.029 ± 0.0025) relative to control (0.0071 ± 0.0021) fetuses. The E₂SO₄ treatment significantly decreased the E₂SO₄ to E₂ ratio (9.9 ± 1.2 iv and 5.4 ± 1.1 icv) relative to control fetuses (12.0 ± 1.0).

Fig. 2.

Plasma E₂ (left panels), E₂SO₄ (middle panels), and total (conjugated plus unconjugated) E₁ (right panels) concentrations in chronically catheterized singleton ovine fetuses throughout the last 13 d of gestation. Fetal sheep were treated with no steroid infusion (control, top panels), estradiol sulfate icv (middle panels), estradiol sulfate iv (bottom panels). Values are represented as mean ± sem.

The plasma DHAS concentrations were significantly different among the experimental groups (P < 0.001, the main effect of group in two way ANOVA; Fig. 3). DHAS did not change as a function of gestational age (P = NS for gestational age in ANOVA) but was higher in the E₂SO₄ iv-treated fetuses (0.026 ± 0.003, 0.031 ± 0.003, and 0.058 ± 0.008 ng/ml in the control, icv, and iv groups, respectively; P < 0.05 by Student-Newman-Keuls test).

Fig. 3.

Plasma P₄ (left panels) and DHAS (right panels) concentrations in chronically catheterized singleton ovine fetuses throughout the last 13 d of gestation. Fetal sheep were treated with no steroid infusion (control, top panels), estradiol sulfate icv (middle panels), estradiol sulfate iv (bottom panels). Values are represented as mean ± sem.

Fetal arterial blood gases and pH in all three experimental groups were consistent with known values for chronically catheterized fetuses (Table 1). There were no significant differences in P_aO₂ among groups. P_aCO₂ was decreased in the E₂SO₄ iv and icv groups (P < 0.001 for the main effect of the group in ANOVA; P < 0.05 control > icv > iv using Student-Newman-Keuls test). pH_a was marginally significantly different among the experimental groups (P = 0.05 for main effect of group in ANOVA), but post hoc analysis by Student-Newman-Keuls test did not identify individual differences among groups.

Table 1.

Fetal blood gases and pH^a

Treatment group	PaO2 (mm Hg)	PaCO2 (mm Hg)	pHa
Control (n = 5)	19.0 ± 0.7	59.6 ± 0.6	7.32 ± 0.01
E2SO4 iv (n = 4)	18.4 ± 0.8	55.4 ± 0.7b	7.33 ± 0.01
E2SO4 icv (n = 5)	18.5 ± 0.8	57.9 ± 0.7	7.32 ± 0.01

^aMean values ± sem were calculated using the mean value of arterial oxygen and carbon dioxide partial pressures and pH (P_aO₂, P_aCO₂, and pH_a, respectively).

^bP < 0.001 compared with control.

Treatment of fetal sheep with icv E₂SO₄ advanced the day of spontaneous parturition slightly but significantly (Fig. 4). Survival analysis of gestation length revealed a highly significant difference in parturition day among the three groups (P = 0.001). Lambs delivered on d 144 ± 1 (mean ± sem; control), 141 ± 2 (mean ± sem; E₂SO₄ icv; P < 0.001 compared with control by Wilcoxon test), and 142 ± 2 (E₂SO₄ iv; P = NS compared with control and P = 0.038 compared with E₂SO₄ icv by Wilcoxon test).

Fig. 4.

Survival plot of probability of remaining in utero (probability of nondelivery) in control- (●), E₂SO₄ iv- (○), and E₂SO₄ icv (▾)-treated fetuses.

Multiple linear regression analysis of plasma hormone and blood gas data reveals a statistically significant correlation (r² = 0.59, n = 74, P < 0.001) between plasma concentration of cortisol (nanograms per milliliter) and plasma concentrations of ACTH_1-39 (picograms per milliliter), and POMC (picomoles per liter), and blood concentrations of H+ (nanomoles per liter). The relationship among these variables is shown in the following equation:

Discussion

The results of this study demonstrate the following: 1) E₂SO₄ increases fetal plasma cortisol but not ACTH or POMC; 2) administration of E₂SO₄ centrally and, perhaps iv, increases endogenous E₂ and E₂SO₄ secretion; and 3) E₂SO₄ advances the timing of parturition. We propose that the physiological effects of sulfoconjugated E₂ on the fetal endocrine environment observed in this study are more complicated than for E₂ alone; understanding the complexity of E₂SO₄ action on fetal neuroendocrine control mechanisms will entail understanding the role of local conjugation/deconjugation enzymes, transporters, and receptors in the relevant brain regions.

E₂SO₄ effects on the fetal HPA axis

We measured fetal plasma ACTH_1-39, POMC, and cortisol as a means of investigating HPA activity in this study. Many investigators have shown that HPA axis activity increases near term in this species (31–34), an observation we confirmed in our control fetuses (Fig. 1). Treatment with E₂SO₄, either iv or icv, attenuated the normal preparturient rise in ACTH. E₂SO₄ increased cortisol secretion but tended to not change or reduce POMC and ACTH_1-39 secretion. Interestingly, despite attenuated ACTH secretion in treated fetuses, cortisol output near term increased in the iv and icv groups. These results are consistent with other recent results from our laboratory in which we compared plasma hormone concentrations and gene expression in twin fetuses, one of which received an icv infusion of E₂SO₄ (35). In that study, the infusion increased fetal plasma cortisol concentration and decreased the abundance of POMC mRNA in fetal pituitary as well as the abundance of CRH mRNA in hypothalamus.

Together with the present data, we speculate that there is an increase in fetal adrenal cortisol secretion (caused by increased adrenal sensitivity to the circulating ACTH, adrenal response to a novel adrenocorticotropic stimulus, or a direct action of E₂ or E₂SO₄ at the adrenal cortex) and that hypothalamo-pituitary control of ACTH secretion are inhibited by negative feedback (36). It might also result from secondary influences on adrenal secretion. Multiple regression analysis revealed a statistically significant association between H⁺ and plasma cortisol concentrations. In adult animals, H⁺ stimulates aldosterone biosynthesis (37). Although controversial (38), some investigators have suggested that this direct effect of acid on adrenal steroidogenesis might be mediated by tandem of P domains in a weak inward rectifying K⁺ channel (TWIK)-related, acid-sensitive K⁺ channels, which are known to be expressed in the adrenal cortex of the adult rat (39). Similarly, it is possible that there is a direct E₂ effect at the adrenal cortex (after deconjugation) (40).

With regard to the neuroendocrine regulation of ACTH secretion, the actions of E₂ and E₂SO₄ are somewhat different. Chronic infusions of E₂ stimulate fetal ACTH secretion chronically (8). Previous studies in this laboratory have demonstrated that infusions of E₂SO₄ for 5 d increase fetal ACTH and cortisol secretion (18). The present study demonstrates that the stimulation of fetal ACTH by E₂SO₄ is transient and that because of an apparent stimulatory effect on the fetal adrenal, long-term (>5 d) infusions actually inhibit the preparturient rise in fetal ACTH secretion. Nevertheless, it is possible that E₂SO₄ may be inhibitory to CRH and ACTH through neurosteroid-like interaction with γ-aminobutyric acid (GABA)_A receptors or, perhaps less likely, by acting as an antagonist of estrogen receptor. The paraventricular nucleus (PVN) is known to have GABAergic innervation, with nearly half of all synapses in the medial parvocellular PVN containing GABA receptors (41). Specific GABA_A receptor subunits can be found within hypophysiotrophic CRH neurons (42), and pharmacological studies both in vitro and in vivo have shown GABAergic mechanisms to be involved in the inhibition of CRH secretion (43–45). Activity of the HPA axis is reduced (as measured by ACTH secretion) after icv injection of GABA_A antagonists (46). It is possible that, in the present experiments, E₂SO₄ either acted directly as a neurosteroid or modified endogenous neurosteroid levels in the PVN, consistent with known actions of neurosteroids, which are GABA_A agonists (47–49).

E₂SO₄ effects on E₂ and E₂SO₄ secretion

The data collected in this study allow us to estimate the secretion and clearance of E₂SO₄ and to estimate the changes in E₂ and E₂SO₄ secretion rate in response to infusion of the sulfoconjugated steroid icv. Based on calculated MCR for E₂SO₄ and E₂, we estimate that approximately 26% of the iv infused E₂SO₄ was converted to E₂ and that when E₂SO₄ was infused directly into the brain, there was an increase in the endogenous synthesis of both E₂ and, perhaps to a lesser extent, E₂SO₄. This increase in E₂ synthesis is accompanied by an increase in the plasma P₄ concentration, possibly the result of concomitant increases in the E₂ and P₄ secretion.

E₂SO₄ was approximately 10 times more abundant than E₂ in plasma of control animals. The differences in concentration between circulating conjugated and unconjugated steroids was previously noted for E₂ (18) and E₁ (19). E₂SO₄ treatment, either icv or iv, significantly lowered the E₂SO₄ to E₂ ratio compared with control fetuses, suggesting that infusion of E₂SO₄ may up-regulate STS activity, down-regulate estrogen STF activity, or preferentially stimulate E₂ secretion. This would have the effect of lowering the E₂SO₄ to E₂ ratio by increasing plasma E₂, which is what we saw in both E₂SO₄ treatment groups. Recent data from this laboratory have confirmed that icv infusion of E₂SO₄ does up-regulate STS and down-regulate STF protein in the fetal brain (35). The up-regulation of STS is likely to be quantitatively more important because the abundance of the mRNA for this enzyme is far greater than for STF (35).

The site of biosynthesis of the various estrogens circulating in fetal blood has not been definitively identified. Because the placenta contains a relative abundance of aromatase, it has been assumed that the placenta is the major source of E₁ and E₂ (17, 50). The ovine placenta has a low activity of 17-hydroxylase and 17,20 lyase until late in gestation, suggesting that the placental E₁ and E₂ biosynthesis might depend on a supply of 17α-hydroxylated substrate from other steroidogenic tissues. Mitchell et al. (51) have proposed that androstenedione secretion by the fetal adrenal serves as a precursor for placental estrogen production. In the present study, icv infusion of E₂SO₄ increased fetal plasma P₄ concentration. Because the major source of P₄ in fetal blood is thought to be placenta (52), the data suggest a placental source of the E₂ and E₂SO₄, perhaps from de novo synthesis from cholesterol. Interestingly, the plasma DHAS concentration was increased only in the iv group. We expected increases in DHAS concentrations in fetal plasma to be evident in groups in which steroidogenesis was stimulated (especially in groups in which there was any evidence of E₂ or E₂SO₄ biosynthesis). In the present experiments, the only group in which we observed evidence of increased E₂ synthesis was the icv group. Nevertheless, the nature of the signal from fetal brain to the site of E₂ and E₂SO₄ synthesis is unclear. In a separate study of twin fetuses, one treated with icv infusion of E₂SO₄ and one treated with vehicle, we found that the only difference in pituitary gene expression of known pituitary tropic hormones was a decrease in the abundance of ovine FSH mRNA (35). This might have been the result and not the cause of the increased circulating E₂ concentrations. Several investigators have proposed neural regulation of steroidogenesis in the fetal sheep (53). Such a mechanism could play a role in the present experiment, but we have no evidence either supporting or refuting it.

In summary, we have examined the effects of E₂SO₄ on HPA axis activity and on the timing of parturition in sheep. This sulfoconjugated estrogen has a neuroendocrine action to increase the secretion of cortisol, E₂, and E₂SO₄ and also to advance parturition. E₂SO₄ inhibits fetal ACTH secretion, possibly through interaction as a neurosteroid. The stimulus to cortisol and E₂ secretion, taking into account the reduction in ACTH secretion, is not revealed by the present study. We conclude that E₂SO₄ has effects on the HPA axis that are distinct from E₂ and that this may be an important mechanism for modulating this activity of the axis during gestation.