| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Laboratory of Neuroendocrinology, School of Biomedical and Clinical Laboratory Sciences, University of Edinburgh (A.J.D., P.J.B., J.A.R.), Edinburgh, United Kingdom EH8 9XD; and Institute of Zoology, University of Regensburg (O.J.B., I.D.N.), 93040 Regensburg, Germany
Address all correspondence and requests for reprints to: Dr. Alison J. Douglas, Division of Biomedical Sciences, School of Biomedical and Clinical Laboratory Sciences, University of Edinburgh, Hugh Robson Building, George Square, Edinburgh, United Kingdom EH8 9XD. E-mail: alison.j.douglas{at}ed.ac.uk.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
In the rat ACTH and corticosterone secretory responses in pregnancy and lactation to a variety of stressors are attenuated (4, 5, 6, 7, 8), and this has been proposed as a mechanism for minimizing exposure of the fetus and neonate to glucocorticoids in the peripartum period (4, 9). The hyporesponsiveness can be explained by reduced responsiveness of the CRH and vasopressin neurons in the parvocellular paraventricular nucleus (PVN) because they show reduced CRH and vasopressin synthesis under basal (5) and stress conditions (10), with reduced activation of immediate early gene mRNA expression (c-fos and NGFI-B) (11, 12). Furthermore, during undisturbed parturition, ACTH and corticosterone secretion are not increased (13), and the HPA secretory responses to stress are severely blunted (14). In addition to HPA axis hyporesponsiveness, the stress-induced secretion of the neurohypophysial hormone oxytocin is attenuated in late pregnancy and lactation in rats (4, 6, 15), but in lactating mice stress reduces oxytocin secretion, evidently through opioid inhibitory mechanisms (16).
Exposure to stressors (such as white noise or restraint) in the mouse during early pregnancy has adverse effects on the continuation of gestation (17, 18, 19), and later in pregnancy, stress delays the onset of birth (20). In addition, the protective behavior of the dam postpartum is impaired (21). Thus, in mice, gestation, parturition, and maternal behavior are sensitive to the adverse effects of stress. Although early studies reported greater plasma corticosterone concentration in the second half of gestation after acute stress compared with controls (1), a detailed characterization of basal or stress-induced ACTH, corticosterone, and oxytocin secretion and their regulation by central mechanisms in the mouse in the peripartum period is lacking.
We have now investigated the HPA axis and neurohypophysial hormone secretory responses to stress exposure during pregnancy, parturition, and lactation compared with those in female virgin mice. Plasma corticosterone-binding globulin (CBG) capacity was also measured to evaluate whether the availability of free plasma corticosterone depends upon reproductive state. Furthermore, we have assessed the responsiveness of PVN neurons under stressed and nonstressed conditions in virgin, pregnant, and lactating mice by measuring the expression of the immediate early gene nur77 mRNA (NGFI-B), a putative regulator of CRH gene expression, and its expression in PVN neurons rapidly increases after stress (22).
Mechanisms that may underlie the reduced HPA axis responses to stress in the pregnant rat include endogenous opioid inhibition (23, 24). Central oxytocin also inhibits HPA axis responses in rats (25), and in the mouse brain oxytocin has anxiolytic effects (26, 27), including mediating stress-induced analgesia (28, 29). Therefore, we additionally investigated the contributions of opioids and central oxytocin to the pregnancy-related stress hyporesponsiveness in the mouse.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Effect of stress on ACTH, corticosterone, oxytocin, and vasopressin secretion
Experiments were performed 1 h after lights on. Virgin [body weight (BW), 29.2–33.2 g], pregnant (d 17.5–18.5, i.e. day before expected parturition; BW, 44.6–57.4 g), and lactating (2–5 d postpartum; BW, 32.6–44.3 g) mice were left in their home cages or were stressed either by exposure to a novel environment (a clean glass jar; 8 x 5.5 x 12 cm, length by width by height) for 10 min or by forced swimming (in a cylinder 18 cm diameter containing water to a depth of 15 cm at 17 C) for 10 min. The mice were then quickly removed to an adjacent room and were immediately decapitated (<15 s after removal from home cage or stress exposure) using scissors. Trunk blood (
0.7 ml) was collected into tubes containing ice-cold EDTA (anticoagulant, 5%, 100 µl/tube) and aprotinin (a protease inhibitor, Sigma-Aldrich Corp., Poole, UK; 0.039 trypsin inhibitor units in 10 µl/sample), taking care not to collect from the head. The blood was centrifuged at 2500 x g for 5 min, and plasma was aliquoted and stored at -20 C until RIA for ACTH, corticosterone, oxytocin, and vasopressin.
Plasma CBG capacity in virgin, pregnant, and lactating mice
Plasma CBG capacity was also assayed (see below) in the same plasma samples from the unstressed (home cage) virgin, pregnant, and lactating mice.
Effect of stress on the expression of nur77 mRNA in the PVN in virgin, pregnant, and lactating mice
The brains from the same mice in the virgin, pregnant, and lactating female home cage and novel environment groups were collected, rapidly frozen in crushed dry ice, and stored at -70 C before in situ hybridization for nur77 mRNA.
Effect of stress during parturition on ACTH secretion
As parturition in mice generally occurs during the dark phase, for this experiment mice were kept on a reverse light cycle from 1 wk before mating (lights off, 0700 h) and on the day of expected parturition (d 18.5–19.5) were observed under red light for pup births. Parturient mice (n = 5) were decapitated immediately after the birth of pup 2, and trunk blood (0.8–0.9 ml) was collected into ice-cold tubes containing EDTA (150 µl/sample) and aprotinin, as mentioned above. Controls were time-matched to the killing of a parturient mouse, which generally occurred between 1200–2000 h (i.e. 5–13 h after lights off). The controls were pregnant mice on the day of expected parturition (n = 5; before signs of imminent labor) and virgin female mice (n = 9). Additionally, some parturient mice were exposed to stress by placing them in a novel environment (n = 3; glass jar as described above) immediately after the birth of pup 2 for 15 min and then decapitated. Such a stress completely prevents pup birth (30). As oxytocin is secreted in large peaks at the time of birth of a pup, and we aimed to monitor the effects of stress rather than birth, we calculated the time when its concentration would be likely to be reduced to basal, based on its half-life in plasma of 2 min. Thus, an exposure time of 15 min in the novel environment was used. Further virgin controls for this experiment were either left in their home cage (n = 6) or stressed for the same time (15 min) in the novel environment (n = 6) and immediately killed (at 7 h after lights off). Plasma was separated from the blood samples and stored as described above before performing the RIA for ACTH. Plasma oxytocin concentrations in the same mice have been previously reported (30).
Effect of central oxytocin on stress responses in pregnant mice
Pregnant (10–11 d) and virgin mice were anesthetized with Avertin (2% tribromoethanol, 8% ethanol, 1.2% tetraamylalcohol in isotonic saline; 500 µl/100 g BW, ip) and given 1 mg analgesic (Carprofen, Zenecarp, C-Vet Veterinary Products, Leyland, Lancashire, UK; in 20 µl, sc). An intracerebroventricular (icv) guide cannula (26 gauge, 2.2 mm deep from the top of the skull; Plastics One, Inc., from Bilaney Consultants Ltd., Sevenoaks, UK) was implanted stereotaxically projecting toward a lateral cerebral ventricle (coordinates from bregma: 0.5 mm caudal, 1.0 mm lateral, with the skull level) and secured to the skull via a screw and dental cement, and a protective cap was fitted; each mouse was then given 0.25 ml isotonic saline, ip. The mice were kept in single cages and handled daily. On the day of experiment, i.e. 7 d after surgery, an injection cannula (33 gauge) was inserted into the guide cannula of the virgin and 17- to 18-d pregnant mice (1000 h); the injection cannula was attached by fine polythene tubing (30 cm long) to a 10-µl Hamilton microsyringe (Sigma-Aldrich Corp.) and contained filtered Ringer solution (pH 7.4) or oxytocin antagonist (des-Gly d(CH2)5[Tyr(Me)2,Thr4]OVT; Prof. M. Manning, Toledo, OH). Mice were left undisturbed for at least 90 min. Then each mouse received an icv injection, using the remote syringe, of either Ringer solution (1 µl; virgin, n = 8; pregnant, n = 8) or oxytocin antagonist [1 µl 0.15 µg/µl; same concentration used previously in rats (25); virgin, n = 9, pregnant, n = 8]. From the time of injection, mice remained in their home cages for 10 min before being stressed by exposure to the novel environment (as above) for 10 min, immediately after which the mice were decapitated, and trunk blood was collected as before. Plasma was separated by centrifugation and assayed for ACTH, corticosterone, and oxytocin. To confirm penetration of the cannula into the lateral ventricle 1 µl dye (5% Alcian Blue) was injected icv, and its presence was sought in the fourth ventricle. Only data from mice in which the cannula was correctly positioned were included in the results. The mice were laparotomized to check pregnancy status postmortem.
Effect of endogenous opioids on stress responses in pregnant mice
On the morning of experiment (1000 h) virgin female (29.6 ± 1.2 g BW) mice and 17- to 18-d pregnant (50.7 ± 1.2 g BW) mice were injected sc with naloxone (5 mg/kg, 100 µl/50 g BW; Sigma-Aldrich Corp.; n = 5 and 9, respectively) or vehicle (isotonic saline; n = 5 and 12, respectively) and returned to their home cage for 10 min before being stressed by exposure to the novel environment (as above) for 10 min. Immediately after stressor exposure, the mice were decapitated, and trunk blood was collected as described above. Plasma was separated and assayed for ACTH, corticosterone, and oxytocin.
Hormone assays
Plasma ACTH and corticosterone concentrations were determined using commercially available immunoradiometric and RIA kits (Eurodiagnostica, from IDS, Tyne and Wear, UK; and ICN Biomedicals, Inc., Costa Mesa, CA), respectively. The sensitivities of the assays were 5 pg/ml for ACTH and 10 ng/ml for corticosterone. Oxytocin and vasopressin concentrations were measured in lyophilized plasma samples (35 µl for each) after extraction, using highly sensitive and selective RIAs, previously validated for measuring the hormone concentrations in the mouse (2, 31, 32). The assay sensitivities were 0.3 pg/sample, and the intraassay coefficient of variation was 7%. All samples from a single experiment were performed within a single assay.
Plasma CBG capacity was assayed according to Harris et al. (33). Briefly, plasma was stripped of endogenous steroids using dextran-coated charcoal, and then aliquots were incubated with [3H]corticosterone (Amersham; 1 pmol/100 µl) in the presence (nonspecific binding) or absence (total binding) of excess cold corticosterone. Unbound corticosterone was removed using dextran-coated charcoal, and after centrifugation bound [3H]corticosterone was counted in the supernatant. Corticosterone binding capacity was estimated from the specifically bound [3H]corticosterone and expressed as picomoles bound per milligram of protein. Protein was quantified using the Bradford assay; plasma was diluted 1:40 in Krebs buffer, and the assay sensitivity was 0.1 mg/ml.
In situ hybridization histochemistry for nur77 mRNA
Procedures.
Sections of brain containing the PVN were cut on a cryostat (15 µm), mounted on gelatin-coated slides, and stored at -70 C until processing by in situ hybridization for the immediate early gene nur77 mRNA (syn. NGFI-B). A 36-mer oligonucleotide probe, complementary to bases 201–236 of mouse nur77 mRNA (MWG Biotech AG, Ebersberg, Germany) (34), was 3' end-labeled with [35S]deoxy-ATP [NEN Life Science Products, PerkinElmer Life Sciences (UK) Ltd., Cambridge, UK] using terminal deoxynucleotidyl transferase (Roche, Mannheim, Germany; 25 U/20 pmol probe). Labeled probe was purified using spin columns (QIAquick nucleotide removal kit, Qiagen Ltd., Crawley, UK), and specific radioactivity was estimated by scintillation counting (1330 Ci/mM). Hybridization was performed as follows. Briefly, slides containing PVN sections were fixed in 4% paraformaldehyde, washed in PBS, acetylated in triethanolamine (1.5%)/acetic anhydride (0.25%), dehydrated in ethanols, delipidated in chloroform, partially rehydrated in 95% ethanol, and air-dried. The sections were hybridized in humidified chambers at 37 C using hybridization buffer containing the labeled oligonucleotide probe (100,000 cpm/section; 60 µl/three sections covered with a Parafilm coverslip). After 20 h the sections were washed in 1x standard saline citrate (briefly three times at room temperature, then four times for 15 min each time at 60 C and twice for 30 min each time at room temperature), with a final brief rinse in double-distilled autoclaved water, and then air-dried. The slides (along with 14C-labeled polymer microscale standard strips, Amersham Pharmacia Biotech, Little Chalfont, UK) were apposed to Hyperfilm ßmax (Amersham Pharmacia Biotech) for 28 d at room temperature, after which the films were developed using Kodak D19 developer (100 g/liter; Eastman Kodak Co., Rochester, NY) and fixed using Hypam rapid fixer (diluted 1:4 with water; Ilford Imaging, Paramus, NJ). The slides were then dipped in liquid photographic emulsion (NTB3, Kodak, from Anachem, Luton, UK) and exposed for an additional 12 wk before developing as described above and lightly counterstaining with hematoxylin and eosin, clearing in xylene, and coverslipping with DPX (Merck-BDH, Lutterworth, UK).
Quantification.
In the emulsion-dipped, counterstained sections, the number of positive cells was counted in the medial parvocellular PVN (6880–7200 µm anterior of the interaural line) (35) in 6 PVN profiles/mouse. A positive cell was defined as having more overlying silver grains than 3 SD greater the mean background (from 10 cells in a more lateral, nonhypothalamic, nonlimbic brain region adjacent to the PVN). To confirm appropriate exposure time, grain density per cell was measured in approximately 5–6 selected cells/PVN section by computerized image analysis (NIH Image) to ensure that it was on the linear part of the radioactive standard curve, plotted using measurements taken from the 14C-labeled polymer strip standards.
Statistical analysis
Data are presented as the group mean ± SEM. Statistical analysis was performed using SigmaStat software (Jandel Scientific, San Raphael, CA). One- or two-way ANOVA (followed by Student-Newman-Keuls post hoc test where appropriate) were used for comparisons between groups. P < 0.05 was considered statistically significant. Although some group sample sizes are small, they are adequate to detect the differences demonstrated, as calculated with SigmaStat sample size for ANOVA statistics, using a power level of 0.8, an 
level of 0.05, and the respective SD of the group data.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
). In virgin mice stressed by exposure to a novel environment or forced swimming, the ACTH concentration was 3.7- and 5.1-fold greater, respectively, than that in nonstressed virgin mice, and forced swimming had a significantly different effect from novel environment. However, in late pregnant mice exposed to either stressor, plasma ACTH was significantly less than in virgins and was not significantly greater than that in nonstressed pregnant mice (Fig. 1A) regardless of stressor. In lactating mice stressed by exposure to the novel environment, the ACTH concentration was not different from that in the stressed virgin group and was significantly more than that in stressed pregnant mice, whereas after forced swimming, although the ACTH concentration was significantly elevated compared with that in unstressed lactating mice, it was significantly less than that in stressed virgin mice (Fig. 1A; by two-way ANOVA, factors reproductive group x treatment: F4,76 = 5.35; P < 0.001).
|
). However, corticosterone concentrations were significantly greater in the unstressed late pregnant mice (537 ± 51 ng/ml) compared with all other groups (19-fold greater than in virgin mice). In the stressed virgin and lactating mice, plasma corticosterone was significantly higher compared with that in their respective nonstressed controls regardless of stressor, and there were no differences between the virgin and lactating mice; the effect of forced swimming was not different from that of novel environment. Novel environment had no effect on the plasma corticosterone concentration in pregnant mice, whereas forced swimming significantly decreased the corticosterone concentration compared with novel environment (Fig. 1B; by two-way ANOVA, factors reproductive group x treatment: F4,74 = 4.71; P < 0.002).
Plasma CBG capacity in virgin, pregnant, and lactating mice
In the same unstressed mice as those described above, the plasma CBG capacity was 2.5 ± 0.6 pmol/mg protein (equivalent to 28.1 ± 6.6 ng/ml) in virgin female mice. In late pregnant mice, the binding capacity was 6.2-fold greater (equivalent to 178.0 ± 16.2 ng/ml) than that in virgin controls and significantly greater than in all other groups. Binding capacity was not significantly different in lactating mice (equivalent to 51.7 ± 6.9 ng/ml) compared with the virgin female group (by one-way ANOVA: F2,20 = 55; P < 0.0001; Fig. 1C).
Effect of stress on the expression of nur77 mRNA in the PVN in virgin, pregnant, and lactating mice
Hybridization signal for nur77 mRNA in the PVN was weak in unstressed control mice and strongly increased after 10-min exposure to the novel environment in virgin and lactating, but not in pregnant, mice (Fig. 2A; same unstressed and stressed mice as above). Quantification of hybridization in the emulsion-dipped sections revealed that the number of parvocellular PVN neurons expressing nur77 mRNA was not significantly different between the virgin females and the pregnant or lactating groups under nonstress conditions; there were significantly fewer positive cells in the lactating compared with the late pregnant mice (P < 0.05). Confinement in the glass jar for 10 min induced a significant increase in the number of parvocellular PVN cells expressing nur77 mRNA in the virgin and lactating mice to a similar extent, but not in the pregnant mice (by two-way ANOVA, group x treatment: F2,50 = 12.32; P < 0.01; Fig. 2B). There was no significant difference in nur77 mRNA expression between the lactating and virgin mice exposed to the novel environment (Fig. 2B).
|
) compared with virgin and pregnant mice. Exposure to the novel environment or forced swimming had no significant effect on plasma oxytocin concentration in any group (Fig. 1D; by two-way ANOVA, factor group: F2,69 = 8.55; P < 0.001).
Vasopressin.
The basal plasma vasopressin concentration was 6.6 ± 1.1 pg/ml in virgin mice and was significantly less in all other reproductive groups (Fig. 1E). In virgin mice exposed to a novel environment, plasma vasopressin was significantly lower than in the control group (P < 0.05, post hoctest), but there was no significant effect in the other groups. The vasopressin concentration was also significantly less after forced swimming in virgin mice, but not in pregnant or lactating mice, compared with their respective unstressed groups, and there was no difference between these two groups (Fig. 1E; by two-way ANOVA, factors reproductive group x treatment: F4,70 = 3.99; P < 0.01).
Effect of stress during parturition on ACTH secretion
Parturition (at the birth of pup 2) had no effect on ACTH secretion compared with that in time-matched virgin and pregnant controls (by one-way ANOVA: F2,15 = 2.63, P = 0.10; Fig. 3A). The plasma ACTH concentration in the parturient mice exposed to the novel environment (glass jar) for 15 min after delivery of pup 2 was not significantly different from that in unstressed, parturient mice (Fig. 3B). This lack of ACTH response to a novel environment in parturition is in contrast to the significantly greater ACTH concentration in virgin mice stressed for 15 min compared with the nonstressed virgin group (by two-way ANOVA: F1,16 = 4.65; P < 0.05; Fig. 3B). Exposure to the novel environment prevented any births while the mice were in the jar for 15 min, compared with the expected delivery of about three pups within this time period in unstressed mice (30).
|
(by two-way ANOVA, factor group: F1,29 = 14.38; P < 0.001; Table 1). Centrally administered oxytocin antagonist had no effect on plasma ACTH in either virgin or pregnant mice compared with vehicle treatment (Table 1).
|
), again confirming the above data (Fig. 1B). Oxytocin antagonist had no effect on corticosterone concentrations in either virgin or pregnant mice compared with vehicle treatment.
There was no difference in plasma oxytocin concentration between the vehicle-treated stressed pregnant or virgin mice, similar to the findings shown in Fig. 1D. The plasma oxytocin concentration was significantly greater after icv oxytocin antagonist administration compared with vehicle administration in the stressed virgin mice (by two-way ANOVA: F1,19 = 8.33; P < 0.01). In contrast, this effect was not seen in pregnancy (Table 1).
Effect of endogenous opioids on stress responses in pregnant mice
All mice were stressed by exposure to a novel environment (glass jar) for 10 min. After vehicle treatment, the plasma ACTH concentration in stressed pregnant mice was significantly less than that in stressed virgin mice (by two-way ANOVA, factor group: F1,27 = 17.2; P < 0.001; Table 2), as shown above (Fig. 1A). After naloxone treatment, the plasma ACTH concentration was significantly elevated in stressed virgins, but had no effect in pregnant mice (Table 2).
|
; by two-way ANOVA, factor group: F1,27 = 26.2; P < 0.0001; Table 2). However, naloxone had no effect on the corticosterone concentration in the stressed virgin or pregnant mice (Table 2).
The plasma oxytocin concentration in stressed pregnant mice was significantly less than that in stressed virgin mice (by two-way ANOVA, factor group: F1,27 = 18.7; P < 0.001). Plasma oxytocin was not affected in either group by treatment with naloxone (Table 2).
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Basal and stress-induced activity of the HPA axis in pregnancy, parturition, and lactation
Pregnancy.
The late pregnant mice had greatly elevated basal plasma total corticosterone concentrations compared with all other groups studied, and this has also been reported as early as d 12 in pregnancy (1, 37). We now show that at the end of pregnancy, this does not reflect an increased basal plasma ACTH concentration, indicating a dissociation between basal ACTH and corticosterone and enhanced adrenal sensitivity to circulating ACTH (38), as found in other species at the end of pregnancy [rat (4, 39), human (40), and sheep (41)].
Biologically available glucocorticoid is reduced by increased CBG level and, as far as materno-fetal exchange is concerned, by placental mechanisms. CBG capacity was increased in late pregnancy, as previously reported (42). Although the free corticosterone concentration was not measured in this study, we can estimate from the binding capacity data that approximately 100% of circulating corticosterone could be bound in virgin mice (corticosterone, 28.1 ± 6.6 ng/ml; CBG capacity for binding corticosterone, 28.5 ± 8.1 ng/ml, 101.4%). However, in late pregnant mice, only approximately 33% of the total corticosterone could be bound (corticosterone, 537.0 ± 51.0 ng/ml; CBG binding capacity, 178.0 ± 16.2 ng/ml), so the increase in binding capacity would not be sufficient to compensate for the greatly elevated plasma corticosterone. This percentage seems to be considerably less than in midpregnancy according to Barlow et al. (1, 43) and may relate to changing fetal glucocorticoid production and/or the placental mechanisms. Corticosteroids cross the placenta, and stress exposure in d 14 pregnant mice can lead to fetal abnormalities (43, 44). The activity of the glucocorticoid-deactivating enzyme 11ß-hydroxysteroid dehydrogenase II, which regulates placental transport of glucocorticoids (45), increases at midpregnancy in the mouse (46) and would limit corticosterone transport, as in the rat (47). However, the synthesis and activity of 11ß-hydroxysteroid dehydrogenase II then decline progressively until they are undetectable at term in the mouse. Consequently, placental glucocorticoid transport is not limited near term. It is possible that the source of the elevated maternal plasma corticosterone concentrations is the fetuses themselves; however, this is unlikely because the fetal serum corticosterone concentration evidently reflects the maternal concentration (48). Thus, despite the adverse effects of exogenous glucocorticoid administration to the mother on the offspring, it seems that an increased basal maternal glucocorticoid concentration is normal near the end of pregnancy in the mouse and may be important in preparing the mammary gland for lactation (49) and in fetal maturation (50).
In response to stress in pregnant mice, ACTH secretion was attenuated compared with that in virgin mice, and there was no further increase in the corticosterone concentration, indicating no enhanced adrenocortical response. Early reports suggested that in midpregnancy in the mouse, HPA hormone concentrations were greatly increased 1 h after restraint or surgical stress compared with basal levels and levels in virgin mice (1). The present findings of HPA axis hyporesponsiveness in pregnancy in the mouse thus do not support the previous findings. Instead, the reported prolonged high corticosterone levels may reflect delayed termination of the response, as seen previously in rats (5, 23). On the other hand, the evident lack of corticosterone response in pregnancy in the present study could be explained if the adrenal cortex was unable to produce additional corticosterone (ceiling effect). However, the data indicate that any possible harmful effects of a further increase in circulating corticosterone above the already increased basal levels (relative to nonpregnant mice) are prevented by HPA axis hyporesponsiveness to stress.
Parturition.
During parturition, the plasma ACTH concentration was unchanged compared with that in time-matched pregnant or virgin mice, as reported for rats (13). Thus, parturition is not necessarily perceived as stressful in rodents. During parturition, the maternal plasma ACTH concentration also remained unchanged 15 min after exposing the mouse to a novel environment compared with unstressed parturient controls. Similarly in parturient rats, exposure to airpuff stress in the home cage does not trigger HPA axis secretory responses (14), so there may be a lack of transmission of signals regulating or strong inhibition of HPA axis activity in parturient mice and rats. Although in the present study, transfer to a glass jar from the nesting cage during parturition did not activate ACTH secretion the same stressor halted the birth of pups, such that no additional pups were born during the 15 min (this study) or 60 min (30) in the jar, with births resuming after return to the nest. We have previously investigated stress-induced suspension of parturition in the mouse and shown that it involves reduced oxytocin secretion (30), as in the rat in similar circumstances (51). This is not reversed by opioid antagonist, unlike in the rat (30, 51). Instead, it appears that the rapid response to stress during parturition in mice is a result of sympathetic-adrenomedullary activation, because the delay in births is prevented by the administration of a ß-receptor antagonist, propranolol (30). Consequently, it is plain that the perception of this stressor is intact, so the lack of an HPA axis response is due to either selective inhibition of pathways to the HPA axis or hyporesponsiveness of the CRH neurons.
Lactation.
A key point arising from our current studies is that HPA axis responses in lactating mice were largely similar to those in virgin mice. However, ACTH (but not corticosterone) responses in lactating mice were stressor dependent, showing hyporesponsiveness (reduced by 35%) after forced swimming, but not after novel environment, in contrast with the findings in late pregnant or parturient mice. Importantly, there was no attenuation of the nur77 mRNA response in the PVN in lactating mice exposed to the novel environment, indicating activation of parvocellular CRH/vasopressin neurons as in virgin mice, and contrasting with the lack of response in pregnant mice. The weakly attenuated HPA axis activation in lactating mice, only seen after forced swimming, is in contrast to other mammals that have strongly attenuated HPA axis secretory responses during lactation [human (52) and rat (4, 53, 54, 55)]. Additionally, in lactating mice, baseline levels of plasma corticosterone and CBG capacity were restored to those of the virgin mice, evidently with adrenal gland sensitivity to ACTH like that in virgin mice, unlike the greatly reduced sensitivity in the lactating rat (8, 39, 54, 56). Thus, HPA axis responsiveness is almost regained in early lactation in mice, whereas in rats it is only regained post weaning (55).
The neurohypophysial system
In female mice we have shown that neither novel environment nor forced swimming induced an oxytocin secretory response, whereas in male mice as well as male and female rats (4, 15) the oxytocin concentration increases after stress (2). The vasopressin concentration decreased in the female mice after stressor exposure, whereas in male mice forced swimming has no effect (2), and in rats vasopressin secretion remains relatively unchanged after stress (4, 15, 57). Our preliminary data in male mice exposed to a novel environment confirms that stress increases oxytocin secretion and has little or no effect on vasopressin secretion (data not shown). In pregnancy and lactation oxytocin and vasopressin responses were similar to those of the virgin mice, although oxytocin was previously reported to decrease in lactating mice (16). It is remarkable to note that the oxytocin concentration in lactating, suckled unstressed mice was lower than that in pregnant and virgin mice and may reflect minimal oxytocin secretion between milk ejection bursts during suckling (58) and depletion of the posterior pituitary store (59), as in rats.
Mechanisms underlying HPA axis adaptations during pregnancy
In these experiments the ACTH concentrations after stress were reduced in pregnant compared with virgin mice, supporting the evidence that the hypothalamo-pituitary stress axis is hyporesponsive in pregnancy. The data do not support the hypothesis that increased inhibition by central oxytocin or endogenous opioids of the HPA axis explains ACTH secretory hyporesponsiveness during pregnancy in the mouse.
Oxytocin released in the brain has antistress effects in virgin and male rats (25) and is anxiolytic in pregnant and lactating rats and mice (26, 60). However, the oxytocin antagonist did not alter novel environment-induced ACTH secretion in either virgin or pregnant mice, suggesting the lack of an intracerebral effect of oxytocin on parvocellular PVN neurons. Nonetheless, the oxytocin antagonist increased oxytocin secretion, indicating an autoinhibition of magnocellular oxytocin neurons during stress in mice, as seen in rats (60, 61). Thus, the lack of effect of oxytocin antagonist on ACTH secretion in the same mice is evidently not due to inappropriate dosage or lack of affinity of the antagonist for the mouse oxytocin receptor. The autoinhibition of oxytocin responses to stress indicates that oxytocin is released intracerebrally during stress exposure and thus there is dissociation between intracerebral and peripheral oxytocin release in virgin mice. This autoinhibition was lost in pregnancy and yet peripheral secretion remained unchanged, so a factor other than oxytocin or opioids (see below) is presumably inhibiting oxytocin neurons at this time.
Naloxone increased ACTH secretory responses to the novel environment in virgin mice, indicating a weak restraining role for endogenous opioids. In contrast, in virgin rats ACTH secretion in response to forced swimming is stimulated by endogenous opioids (23). In mice, inhibition of the ACTH response by endogenous opioids was not seen in pregnancy and was not up-regulated. Thus, inhibition of the HPA axis by opioids cannot explain the stress hyporesponsiveness of pregnancy; this is again in contrast to the situation in rats where endogenous opioid inhibition of HPA axis responsiveness to stress emerges in pregnancy (23) and is strongly inhibitory at parturition (13).
An alternative apparent explanation is that the greatly elevated glucocorticoid concentration in late pregnant mice mediates a strong negative feedback to the HPA axis that may inhibit PVN neuron and thus anterior pituitary ACTH secretory responsiveness. Although in vitro studies indicate enhanced glucocorticoid negative feedback in the pregnant rat, there is no firm evidence in vivo (5), and this remains to be comprehensively investigated in the mouse.
An important outcome of the present study is that there are striking differences, and some similarities, in the adjustments of HPA axis function in pregnancy and lactation between mice and rats. In both species HPA axis responses to stressors are reduced in late pregnancy and are absent in parturition despite a greatly elevated basal plasma corticosterone level in mice, but not rats. In lactation, rats continue to show strongly attenuated HPA axis responses to stress, but mice do not. A common mechanism for reduced HPA axis responsiveness in pregnancy has yet to be found.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Abbreviations: BW, Body weight; CBG, corticosterone-binding globulin; HPA, hypothalamo-pituitary-adrenal; icv, intracerebroventricular; PVN, paraventricular nucleus.
Received April 21, 2003.
Accepted for publication August 25, 2003.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  S. J. Spencer, A. Mouihate, M. A. Galic, and Q. J. Pittman Central and peripheral neuroimmune responses: hyporesponsiveness during pregnancy J. Physiol., January 15, 2008; 586(2): 399 - 406. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. A. Slattery and I. D. Neumann No stress please! Mechanisms of stress hyporesponsiveness of the maternal brain J. Physiol., January 15, 2008; 586(2): 377 - 385. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. J. Brunton, M. Sausbier, G. Wietzorrek, U. Sausbier, H.-G. Knaus, J. A. Russell, P. Ruth, and M. J. Shipston Hypothalamic-Pituitary-Adrenal Axis Hyporesponsiveness to Restraint Stress in Mice Deficient for Large-Conductance Calcium- and Voltage-Activated Potassium (BK) Channels Endocrinology, November 1, 2007; 148(11): 5496 - 5506. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. K. Pincus-Knackstedt, R. A. Joachim, S. M. Blois, A. J. Douglas, A. S. Orsal, B. F. Klapp, U. Wahn, E. Hamelmann, and P. C. Arck Prenatal Stress Enhances Susceptibility of Murine Adult Offspring toward Airway Inflammation J. Immunol., December 15, 2006; 177(12): 8484 - 8492. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Mouihate, S. Ellis, E.-M. Harre, and Q. J. Pittman Fever suppression in near-term pregnant rats is dissociated from LPS-activated signaling pathways Am J Physiol Regulatory Integrative Comp Physiol, November 1, 2005; 289(5): R1265 - R1272. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Di, R. Malcher-Lopes, V. L. Marcheselli, N. G. Bazan, and J. G. Tasker Rapid Glucocorticoid-Mediated Endocannabinoid Release and Opposing Regulation of Glutamate and {gamma}-Aminobutyric Acid Inputs to Hypothalamic Magnocellular Neurons Endocrinology, October 1, 2005; 146(10): 4292 - 4301. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Ma, M. J. Shipston, D. Morilak, and J. A. Russell Reduced Hypothalamic Vasopressin Secretion Underlies Attenuated Adrenocorticotropin Stress Responses in Pregnant Rats Endocrinology, March 1, 2005; 146(3): 1626 - 1637. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
; n = 9–11) or were stressed for 10 min by confinement in a glass jar (novel environment; 
; n = 4–11). Data are the mean ± SEM plasma ACTH concentration (A), plasma corticosterone concentration (B), plasma CBG capacity (C), plasma oxytocin concentration (D), and plasma vasopressin concentration (E). A, By two-way ANOVA, P < 0.001 for interaction between group and treatment. B, By two-way ANOVA, P < 0.002 for interaction between group and treatment. C, By one-way ANOVA, P < 0.0001. D, By two-way ANOVA, P < 0.001 for factor group. E, By two-way ANOVA, P < 0.01 for interaction between group and treatment. *, P < 0.05 vs. respective unstressed control; +, P < 0.05 vs. respective virgin group (by post hoc tests).
, P < 0.05 vs. virgin and pregnant unstressed control (by post hoc tests).