Prenatal stress alters glutamatergic system responsiveness in adult rat prefrontal cortex

Authors

  • Fabio Fumagalli,

    1. Center of Neuropharmacology, Department of Pharmacological Sciences, University of Milan, Milan, Italy
    Search for more papers by this author
  • Matteo Pasini,

    1. Center of Neuropharmacology, Department of Pharmacological Sciences, University of Milan, Milan, Italy
    2. Department of Experimental and Clinical Pharmacology, School of Medicine, University of Catania, Catania, Italy
    Search for more papers by this author
  • Angelisa Frasca,

    1. Center of Neuropharmacology, Department of Pharmacological Sciences, University of Milan, Milan, Italy
    Search for more papers by this author
  • Filippo Drago,

    1. Department of Experimental and Clinical Pharmacology, School of Medicine, University of Catania, Catania, Italy
    Search for more papers by this author
  • Giorgio Racagni,

    1. Center of Neuropharmacology, Department of Pharmacological Sciences, University of Milan, Milan, Italy
    2. I.R.C.C.S. San Giovanni di Dio-Fatebenefratelli, Brescia, Italy
    Search for more papers by this author
  • Marco Andrea Riva

    1. Center of Neuropharmacology, Department of Pharmacological Sciences, University of Milan, Milan, Italy
    Search for more papers by this author

Address correspondence and reprint requests to Marco Andrea Riva, Center of Neuropharmacology, Department of Pharmacological Sciences, Via Balzaretti 9, 20133 Milan, Italy. E-mail: M.Riva@unimi.it

Abstract

Exposure to stress during gestation alters brain development resulting in permanent alterations that may increase susceptibility to subsequent cognitive or neuropsychiatric disorders. In this manuscript we examined the effects of prenatal stress on critical determinants of the glutamatergic synapse under basal conditions as well as in response to acute stress. The main finding of this work is that gestational stress altered the responsiveness of the glutamatergic system following a challenge at adulthood. In fact, while in control animals acute swim stress enhanced the phosphorylation levels of the NMDA receptor subunits NR-1(Ser896) and NR-2B(Ser1303) as well as the phosphorylation levels of α calcium/calmodulin-dependent protein kinase II (Thr286), a crucial sensor of calcium fluctuations, prenatal stress prevented or attenuated such activation. This dynamic modulation is restricted to prefrontal cortex since no changes were observed in the hippocampus, in line with the different maturational profile of these brain regions. Changes were also observed in the phosphorylation of the α-amino-3-hydroxy-5-methylisoxazole-4-propionate subunit GluR-1(Ser831) which, however, relied on the acute stress exposure and were independent of gestational stress. These effects point to a unique interference of chronic prenatal stress with the responsiveness of specific determinants of the glutamatergic synapse at adulthood in a region specific manner. The inability to mount an homeostatic glutamatergic response to subsequent stress at adulthood may impair the normal responses of the cell to challenging situations.

Abbreviations used
αCaMKII

α calcium/calmodulin-dependent protein kinase II

AMPA

α-amino-3-hydroxy-5-methylisoxazole-4-propionate

GluR1

subunit 1 of α-amino-3-hydroxy-5-methyl-4-isoxazolepropionate receptor (AMPA) glutamate receptor

NR

NMDA receptor

SCPHT

single contrast post-hoc test

There is growing evidence that adverse events during pregnancy have a negative impact on brain development and may increase the risk for neuropsychiatric disorders and related cognitive deficit at adulthood (Fride and Weinstock 1988; Weinstock et al. 1988; Koehl et al. 2001; Spauwen et al. 2004). It has been demonstrated that stress responsiveness, as measured by circulating levels of corticosterone and AdrenoCorticoTropic Hormone, is altered in rats exposed to prenatal stress (Maccari et al. 1995). This is associated with changes of proteins and mechanisms related to neuronal plasticity, with altered expression of neurotrophic factors such as Brain Derived Neurotrophic Factor and basic Fibroblast Growth Factor (Fumagalli et al. 2004, 2005) as well as changes in neurogenesis (Lemaire et al. 2000, 2006).

Recent work by Son and colleagues has demonstrated that prenatal stress produces an impairment of NMDA receptor-dependent synaptic plasticity (Son et al. 2006). Although changes in the steady-state level of certain proteins do represent a hallmark of prenatal stress, the functionally relevant effect produced by a stressful prenatal environment on brain development can be fully appreciated when the animal is exposed to a further stimulus which can challenge the system; this option is critical particularly in view of the evidence that prenatal stress can enhance vulnerability to psychiatric disorders (Weinstock 2001). To this end, it has been shown that prenatally-stressed rats, without apparent behavioral dysfunctions at basal level, exhibited exaggerated locomotor response to a novel environment (Deminiere et al. 1992; Kippin et al. 2008), enhanced responsiveness to cocaine (Kippin et al. 2008) and amphetamine (Henry et al. 1995); additionally, longer-lasting increases in plasma corticosterone (Louvart et al. 2009) and altered responsivity of trophic factor expression (Fumagalli et al. 2004, 2005) were observed after further stress at adulthood. Taken together, these data suggest that prenatal stress indeed produces neuroadaptive modifications which alter the normal response to challenging events at adulthood.

Along this line of reasoning and given the close relationship between the glutamate system and manifestation of psychiatric disorders (Olney and Farber 1995), we decided to incorporate a paradigm of acute adult stress to investigate the functional responsiveness of this system in animals that have been exposed to stress during pregnancy in order to establish the impact of the prenatal manipulation on glutamatergic dynamics, focusing our analysis on two brain regions which are closely associated with the development of psychiatric disorders such as hippocampus and prefrontal cortex (Koenig et al. 2005; Riva et al. 2005; Buckley et al. 2007; Russo et al. 2009).

Materials and methods

Materials

General reagents were purchased from Sigma (Milan, Italy), and molecular biology reagents were obtained from Cellbio (Pero, Milan, Italy) and Promega (Milan, Italy).

Animals and prenatal stress paradigm

Timed-pregnant Sprague-Dawley rats, purchased from a commercial breeder (Charles River, Calco, Italy), were individually housed in 40 × 25 × 20 cm Plexiglas boxes, with a metal top and sawdust as bedding and were maintained in an air-conditioned room (temperature 21 ± 1°C, relative humidity 60 ± 10%) with lights on from 08:00 to 20:00, and with standard pellet chow and tap water freely available. Restraint stress was performed from embryonic day 14 (E14) until delivery as previously described (Fumagalli et al. 2004).

Pregnant females were individually placed in plastic transparent restrainers for 45 min, three times a day (09:00, 12:00 and 17:00), under bright light. Control pregnant females were left undisturbed in their home cages. Pregnant dams delivered normal size litters. Birth was designated as pnd (postnatal day) 0. On pnd 1 pups were culled to six males and two females to ensure the presence of both sexes in the litters. All pups in the litter were left undisturbed (although some clean sawdust was added from the top of the cage) with their birth mothers until the time of weaning (pnd21), at which point the pups were group-housed in the same type of cages.

For each set of experiments, a maximum of two male or female siblings was taken from each litter in order to avoid ‘litter effects’ (Chapman & Stern 1979), so that the number of litters represented in each group was 5–8. Experimental procedures were exactly the same for males and females.

Acute swim stress in adult rats

Acute swim stress was performed on adult control- and prenatally-stressed rats to evaluate whether in utero stress could modify molecular determinants of the glutamatergic synapse in response to subsequent adult challenges. In the acute stress paradigm, rats (80 days of age) were subjected to a forced swim session for 5 min, and killed by decapitation 15 min after the end of the swim session. Stressed rats swam in plexiglass cylinders filled with 30 cm water at 25 ± 1°C. After the stress session, rats were towel dried and housed into single cages until kill. Control rats were left undisturbed in their home cages until the time of kill.

Brain regions were immediately dissected out, frozen on dry ice and stored at −80°C. Dissections were performed according to the atlas of Paxinos and Watson (1996). In details, the prefrontal cortex was dissected from 2-mm thick slices (prefrontal cortex defined as Cg1, Cg3, and IL subregions corresponding to the plates 6–9 (approximately weight 8 mg), whereas hippocampus (including both ventral and dorsal parts) was dissected from the whole brain. All animal handling and experimental procedures were performed in accordance with the EC guidelines (EEC Council Directive, 1987) and with the Italian legislation on animal experimentation (Decreto Legislativo 116/92).

Preparation of protein extracts

Different subcellular fractions were prepared as described previously (Fumagalli et al. 2007a). Tissues (hippocampus and prefrontal cortex) were homogenized in a glass-glass potter in cold 0.32 M sucrose buffer pH 7.4 containing 1 mM HEPES, 0.1 mM EGTA and 0.1 mM phenylmethylsulfonyl fluoride, in presence of commercial cocktails of protease (Roche, Monza, Italy) and phosphatase (Sigma-Aldrich) inhibitors. The homogenate was clarified at 1000 g for 10 min obtaining a pellet (P1) corresponding to the nuclear fraction, which was resuspended in a buffer (20 mM HEPES, 0.1 mM dithiothreitol, 0.1 mM EGTA) supplemented with protease and phosphatase inhibitors. The supernatant (S1) was then centrifuged at 9000 g for 15 min to obtain a clarified fraction of cytosolic proteins (S2) and a pellet (P2) corresponding to the crude synaptosomal fraction which was resuspended in the same buffer used for the nuclear fraction. Total protein content was measured according to the Bradford Protein Assay procedure (Bio-Rad, Milan, Italy), using bovine serum albumin as calibration standard.

Western blot analysis

Protein analysis was performed on P2 (crude synaptosomes) fraction. The same amounts of total protein for all samples (10 μg) were run on an sodium dodecyl sulfate-10% polyacrylamide gel under reducing conditions and then electrophoretically transferred onto nitrocellulose membranes (Bio-Rad). Blots were blocked with 10% non-fat dry milk, incubated with antibodies against the phosphorylated forms of the proteins and then stripped and reprobed with antibodies against total proteins of the same type. The following antibodies were used: anti phospho-α calcium/calmodulin-dependent protein kinase II (αCaMKII) Thr286 (1 : 2500, Affinity Bioreagents, Golden, CO, USA), anti phospho-NR1Ser831 (1 : 1000, Affinity Bioreagents), anti phospho-NR2BSer1303 (1 : 1000, Upstate, Lake Placid, NY, USA), anti phospho-subunit 1 of α-amino-3-hydroxy-5-methyl-4-isoxazolepropionate receptor (AMPA) glutamate receptor (GluR1)(Ser831) (1 : 500, Affinity Bioreagents), anti-total αCaMKII (1 : 5000, Millipore, Billerica, MA, USA), anti-total NR1 (1 : 1000, Zymed Laboratories, South San Francisco, CA, USA), anti-total NR2B (1 : 1000, Santa Cruz Biotechnology, Santa Cruz, CA, USA), anti-total GluR1 (1 : 2000, Upstate), anti-total GluR2 (1 : 2000, NeuroMab, Davis, CA, USA) and anti-total NR2A (1 : 1000, Zymed Laboratories). Immunocomplexes were visualized by chemiluminescence utilizing the ECL Western Blotting kit (GE Healthcare, Milan, Italy) according to the manufacturer’s instructions.

Results were standardized to β-actin control protein, which was detected by evaluating the band density at 43 kDa after probing with a polyclonal antibody at dilution of 1 : 10 000 (Sigma-Aldrich).

Statistical analysis

Data are presented as means and standard errors, with each individual group comprising 5–10 samples. Changes produced by treatments were analyzed in the different brain regions, using a two-way analysis of variance (anova), with prenatal stress and swim stress as independent variables, and protein levels as dependent variables. When appropriate, further differences between groups were analyzed by single contrast post-hoc test (SCPHT). However, where there was no interaction between prenatal stress and swim, only the main effect was reported, rather than proceeding inappropriately to subtesting of each individual treatment combination. Statistical significance was assumed at < 0.05.

Results

The present study has investigated the influence of prenatal stress on the responsiveness of the glutamatergic synapse to an acute swim stress at adulthood and the potential gender susceptibility to such manipulation. To reach this goal, we performed protein analyses in the crude synaptosomal fraction, i.e. a fraction enriched in components of plasmatic membrane, of rat prefrontal cortex and hippocampus, focusing our attention on crucial determinants of glutamatergic synapse.

Prefrontal cortex

Modulation of NMDA receptor subunits

We first analyzed the phosphorylation of NR1 subunit of the NMDA receptor in male rat prefrontal cortex. Two-way anova showed a significant effect of prenatal stress on P-NR1 levels (F1,27 = 9.7565, = 0.004), a significant effect of acute swim stress (F1,27 = 7.8416, = 0.01) and a significant prenatal stress × acute stress interaction (F3,27 = 4.433, = 0.046) (Fig. 1). In light of the significant interaction between the two types of stress, we made intergroup comparisons, which showed that swim stress significantly enhanced P-NR1 levels in sham animals (F1,16 = 13.718, = 0.002, two-way anova with SCPHT) but not in prenatally-stressed rats (F1,12 = 0.215, = 0.647, two-way anova with SCPHT); additionally, post-hoc test revealed that prenatal stress significantly increased P-NR1 levels (F1,13 = 12.845, = 0.003, two-way anova with SCPHT) as shown in Fig. 1. As indicated in Table 1, two-way anova showed no significant effect of prenatal stress on the levels of the obligatory subunit NR1 and no effect of acute swim stress.

Figure 1.

 Effect of prenatal stress on the phosphorylation of glutamate NMDA receptor NR1(Ser896) subunit and modulation by an acute swim stress at adulthood, assessed in the prefrontal cortex of male (left panel) and female (right panel) rats. The results, expressed as % of control/unstressed rats, represent the mean ± SEM of 5–9 independent determinations. anova across both prenatal and postnatal treatments, when significant, appears at the bottom of each panel. PS, prenatal stress; SS, acute swim stress; PS × SS, prenatal stress × acute stress interaction. *< 0.01 vs. control/unstressed [two-way anova with single contrast post-hoc test (SCHPT)]; #< 0.01 vs. control/unstressed (two-way anova with SCPHT).

Table 1.   Levels of glutamate NMDA receptors (NR-1, NR-2A and NR-2B) and AMPA receptors (GluR1 and GluR2) as well as αCaMKII following prenatal stress and their modulation by an acute stress at adulthood in the rat prefrontal cortex of male and female rats
 CTRL/SHAMCTRL/SWIMPS/SHAMPS/SWIM2-way anova
  1. The results represent the mean ± SEM of at least eight independent determinations, for males (upper panel) and females (lower panel). The significant effects are shown in bold.

  2. CTRL/SHAM = animals that were neither exposed to prenatal stress nor to acute stress at adulthood; CTRL/SWIM = animals that were not exposed to prenatal stress but that were exposed to acute stress at adulthood; PS/SHAM = prenatally-stressed animals that were not exposed to acute stress at adulthood; PS/SWIM = prenatally-stressed animals that were exposed to acute stress at adulthood.

Prefrontal cortex
 Males
 NR1100 ± 10163 ± 18156 ± 21156 ± 19Prenatal stress, = 0.155
Acute swim stress, = 0.072
Prenatal stress × acute swim stress, = 0.068
 NR2B100 ± 10106 ± 16129 ± 10132 ± 10Prenatal stress, = 0.044
Acute swim stress, = 0.764
Prenatal stress × acute swim stress, = 0.909
 NR2A100 ± 6122 ± 9111 ± 9139 ± 14Prenatal stress, = 0.152
Acute swim stress, = 0.016
Prenatal stress × acute swim stress, = 0.068
 GluR1100 ± 2107 ± 10110 ± 6100 ± 11Prenatal stress, = 0.877
Acute swim stress, = 0.830
Prenatal stress × acute swim stress, = 0.272
 GluR2100 ± 1196 ± 5115 ± 397 ± 6Prenatal stress, = 0.357
Acute swim stress, = 0.187
Prenatal stress × acute swim stress, = 0.378
 αCaMKII100 ± 794 ± 8104 ± 7116 ± 12Prenatal stress, = 0.160
Acute swim stress, = 0.695
Prenatal stress × acute swim stress, = 0.310
 Females
 NR1100 ± 10106 ± 16101 ± 13128 ± 7Prenatal stress, = 0.361
Acute swim stress, = 0.218
Prenatal stress × acute swim stress, = 0.416
 NR2B100 ± 992 ± 1196 ± 6109 ± 13Prenatal stress, = 0.490
Acute swim stress, = 0.785
Prenatal stress × acute swim stress, = 0.265
 NR2A100 ± 11100 ± 9115 ± 10107 ± 6Prenatal stress, = 0.261
Acute swim stress, = 0.698
Prenatal stress × acute swim stress, = 0.699
 GluR1100 ± 9100 ± 1498 ± 1195 ± 14Prenatal stress, = 0.780
Acute swim stress, = 0.890
Prenatal stress × acute swim stress, = 0.917
 GluR2100 ± 8102 ± 7116 ± 6112 ± 5Prenatal stress, = 0.069
Acute swim stress, = 0.904
Prenatal stress × acute swim stress, = 0.671
 αCaMKII100 ± 7116 ± 18104 ± 7106 ± 7Prenatal stress, = 0.816
Acute swim stress, = 0.435
Prenatal stress × acute swim stress, = 0.540

Analysis of P-NR1 in prefrontal cortex of females revealed that neither prenatal stress (F1,22 = 0.403, = 0.533), nor acute swim stress (F1,24 = 0.536, = 0.473) were able to modulate the phosphorylation of this subunit (Fig. 1). Similarly to male animals, no significant effects of prenatal stress or acute swim stress were detected on total levels of the receptor subunit (Table 1).

When examining the phosphorylation of the regulatory subunit NR2B in males, no effect of the antenatal stress was observed (F1,27 = 1.099, = 0.305, two-way anova), but a significant effect of acute swim stress was found (F1,27 = 8.709, = 0.007) (Fig. 2). The acute swim stress increased P-NR2B levels in both groups, although the magnitude was higher in control animals. Moreover, as summarized in Table 1, the total levels of NR2B were also significantly increased (F1,27 = 8.012, = 0.044) by the prenatal manipulation, while no effect of acute swim stress was observed. Similar changes were also found in female rats. Indeed, P-NR2B levels were not altered by prenatal stress (F1,24 = 0.999, = 0.329, two-way anova), but were modulated by the acute swim stress (F1,24 = 6.367, = 0.020) (Fig. 2). Two-way anova indicated no effect of prenatal stress as well as acute swim stress on the expression of the NR2B subunit (Table 1).

Figure 2.

 Effect of prenatal stress on the phosphorylation of glutamate NMDA receptor NR2B(Ser1303) subunit and modulation by an acute swim stress at adulthood, assessed in the prefrontal cortex of male (left panel) and female (right panel) rats. The results, expressed as % of control/unstressed rats, represent the mean ± SEM of 6–8 independent determinations. anova across both prenatal and postnatal treatments, when significant, appears at the bottom of each panel. PS, prenatal stress; SS, acute swim stress; PS × SS, prenatal stress × acute stress interaction.

Analysis of total NR2A levels in males revealed a significant increase induced by acute swim stress (F1,31 = 6.644, = 0.016) with no effect of prenatal stress (F1,31 = 2.173, = 0.152) (Table 1) whereas in females we found no significant changes in the levels of NR2A subunit (Table 1).

Modulation of αCaMKII

As the major target of calcium influx through NMDA receptors is represented by αCaMKII, which also plays an important regulatory role in the glutamatergic neurotransmission (Colbran and Brown 2004), we evaluated if the changes produced by stressful conditions on NMDA receptor subunits could influence the activation state and expression of this kinase. The levels of P-αCaMKII were not significantly altered by prenatal stress whereas acute swim stress at adulthood reduced P-αCaMKII levels in male animals (F1,33 = 20.265, = 0.0001). Moreover, the response to the acute swim stress was affected by prenatal stress (prenatal stress × acute stress interaction, F3,33 = 9.5257, = 0.004) (Fig. 3). In fact, acute swim stress reduced αCaMKII phosphorylation in prenatally unstressed (F1,19 = 33.877, = 0.0001, two-way anova with SCPHT) but not in prenatally stressed animals (F1,14 = 0.870, = 0.716, two-way anova with SCPHT) (Fig. 3). As summarized in Table 1, total αCaMKII levels were not affected by prenatal stress and acute swim stress, in male as well as female rats. Although P-αCaMKII levels were also modulated by acute swim stress in female rats (F1,25 = 4.739, = 0.041) (Fig. 3), no effect of prenatal stress was detected (F1,25 = 0.417, = 0.525).

Figure 3.

 Effect of prenatal stress on the phosphorylation of αCaMKII(Thr286) and modulation by an acute swim stress at adulthood, assessed in the prefrontal cortex of male (left panel) and female (right panel) rats. The results, expressed as % of control/unstressed rats, represent the mean ± SEM of 6–9 independent determinations. anova across both prenatal and postnatal treatments, when significant, appears at the bottom of each panel. PS, prenatal stress; SS, acute swim stress; PS × SS, prenatal stress × acute stress interaction. *<0.001 vs. control/unstressed rats (two-way anova with SCPHT).

Modulation of α-amino-3-hydroxy-5-methylisoxazole-4-propionate receptor subunits

When examining the phosphorylation of the α-amino-3-hydroxy-5-methylisoxazole-4-propionate (AMPA) receptor subunit GluR1(Ser831), we found no significant effect of either prenatal stress (F1,32 = 3.102, = 0.089) or acute swim stress (F1,32 = 0.030, = 0.863) in males (Fig. 4). Similarly, no effects were observed in the total levels of this subunit (Table 1).

Figure 4.

 Effect of prenatal stress on the phosphorylation of AMPA receptor subunit GluR1(Ser831) subunit and modulation by an acute swim stress at adulthood, assessed in the prefrontal cortex of male (left panel) and female (right panel) rats. The results, expressed as % of control/unstressed rats, represent the mean ± SEM of 5–9 independent determinations. anova across both prenatal and postnatal treatments, when significant, appears at the bottom of each panel. PS, prenatal stress; SS, acute swim stress; PS × SS, prenatal stress × acute stress interaction.

A similar trend was observed in females with no effect on the phosphorylation levels of the GluR1 subunit by either prenatal stress (F1,21 = 0.235, = 0.634) or acute swim stress (F1,21 = 0.791, = 0.386). Additionally, no changes on the total levels of GluR1 protein were found in animals exposed to prenatal manipulation or acute swim stress (Table 1).

Analysis of the constitutive subunit of AMPA receptor GluR2 revealed no significant changes in the levels of the protein in males as well as in females, as indicated in Table 1.

Hippocampus

Modulation of NMDA receptor subunits

Phospho-NR1 levels in males were not affected by prenatal stress (F1,33 = 0.027, = 0.869), or by the acute swim stress (F1,33 = 3.017, = 0.092) (Fig. 5). Moreover, as shown in Table 2, total levels of NR1 subunit were not altered by both manipulations as well. Similar to male rats, NR1 phosphorylation was not altered in the hippocampus of females, neither by prenatal manipulation (F1,33 = 2.823, = 0.112, two-way anova) nor by acute swim stress at adulthood (F1,20 = 0.189, = 0.670) (Fig. 5). Table 2 shows that we found no effect of prenatal stress and acute swim stress on total levels of the subunit.

Figure 5.

 Effect of prenatal stress on the phosphorylation of glutamate NMDA receptor NR1(Ser896) subunit and modulation by an acute swim stress at adulthood, assessed in the hippocampus of male (left panel) and female (right panel) rats. The results, expressed as % of control/unstressed rats, represent the mean ± SEM of 5–9 independent determinations The results, expressed as % of control/unstressed rats, represent the mean ± SEM of 5–9 independent determinations. PS, prenatal stress; SS, acute swim stress; PS × SS, prenatal stress × acute stress interaction.

Table 2.   Levels of glutamate NMDA receptors (NR-1, NR-2A and NR-2B) and AMPA receptors (GluR1 and GluR2) as well as αCaMKII following prenatal stress and their modulation by an acute stress at adulthood in the rat hippocampus of male and female rats
 CTRL/SHAMCTRL/SWIMPS/SHAMPS/SWIM2-way anova
  1. The results represent the mean ± SEM of at least eight independent determinations, for males (upper panel) and females (lower panel). The significant effects are shown in bold.

  2. CTRL/SHAM = animals that were neither exposed to prenatal stress nor to acute stress at adulthood; CTRL/SWIM = animals that were not exposed to prenatal stress but that were exposed to acute stress at adulthood; PS/SHAM = prenatally-stressed animals that were not exposed to acute stress at adulthood; PS/SWIM = prenatally-stressed animals that were exposed to acute stress at adulthood.

Hippocampus
 Males
 NR1100 ± 12103 ± 1396 ± 9112 ± 9Prenatal stress, = 0.857
Acute swim stress, = 0.461
Prenatal stress × acute swim stress, = 0.602
 NR2B100 ± 489 ± 8104 ± 2104 ± 10Prenatal stress, = 0.156
Acute swim stress, = 0.363
Prenatal stress × acute swim stress, = 0.441
 NR2A100 ± 691 ± 1293 ± 13114 ± 9Prenatal stress, = 0.407
Acute swim stress, = 0.583
Prenatal stress × acute swim stress, = 0.145
 GluR1100 ± 698 ± 7111 ± 6115 ± 6Prenatal stress, = 0.036
Acute swim stress, = 0.870
Prenatal stress × acute swim stress, = 0.659
 GluR2100 ± 896 ± 6103 ± 798 ± 7Prenatal stress, = 0.781
Acute swim stress, = 0.545
Prenatal stress × acute swim stress, = 0.911
 αCaMKII100 ± 689 ± 3100 ± 5100 ± 4Prenatal stress, = 0.271
Acute swim stress, = 0.273
Prenatal stress × acute swim stress, = 0.299
 Females
 NR1100 ± 7101 ± 11102 ± 8107 ± 5Prenatal stress, = 0.666
Acute swim stress, = 0.736
Prenatal stress × acute swim stress, = 0.821
 NR2B100 ± 1293 ± 1987 ± 13100 ± 14Prenatal stress, = 0.841
Acute swim stress, = 0.854
Prenatal stress × acute swim stress, = 0.541
 NR2A100 ± 690 ± 897 ± 487 ± 12Prenatal stress, = 0.692
Acute swim stress, = 0.223
Prenatal stress × acute swim stress, = 0.987
 GluR1100 ± 8124 ± 6111 ± 16129 ± 8Prenatal stress, = 0.404
Acute swim stress, = 0.040
Prenatal stress × acute swim stress, = 0.756
 GluR2100 ± 14101 ± 12102 ± 599 ± 13Prenatal stress, = 0.997
Acute swim stress, = 0.943
Prenatal stress × acute swim stress, = 0.887
 αCaMKII100 ± 9102 ± 7107 ± 6107 ± 8Prenatal stress, = 0.439
Acute swim stress, = 0.901
Prenatal stress × acute swim stress, = 0.870

The analysis of NR2B phosphorylation in male rats revealed no effect of prenatal stress (F1,34 = 1.223, = 0.277) or acute swim stress (F1,34 = 0.130, = 0.720) (Fig. 6). Similarly, in females, P-NR2B levels were unaffected by prenatal stress (F1,24 = 1.396, = 0.251) or acute swim stress (F1,24 = 0.886, = 0.358). Analysis of the total levels of the protein revealed that prenatal stress or acute swim stress were unable to alter total NR2B levels in male as well as female rats (Table 2).

Figure 6.

 Effect of prenatal stress on the phosphorylation of glutamate NMDA receptor NR2B(Ser1303) subunit and modulation by an acute swim stress at adulthood, assessed in the hippocampus of male (left panel) and female (right panel) rats. The results, expressed as % of control/unstressed rats, represent the mean ± SEM of 6–8 independent determinations. PS, prenatal stress; SS, acute swim stress; PS × SS, prenatal stress × acute stress interaction.

Analysis of total levels of NR2A revealed no effect of either prenatal stress or acute swim stress, both in males and females (Table 2).

Modulation of αCaMKII

Prenatal stress did not alter the phosphorylation of αCaMKII (F1,33 = 2.364, = 0.135) in male rats (Fig. 7). However, its levels were significantly reduced by acute swim stress (F1,33 = 26.210, = 0.0001) (Fig. 7). Conversely, total levels of αCaMKII were not modulated by prenatal stress or acute swim stress (Table 2). When we focused on αCaMKII in female rats, we found that its phosphorylation levels were not significantly modulated either by prenatal stress (F1,25 = 2.287, = 0.145) or acute swim stress (F1,25 = 3.113, = 0.092) (Fig. 7). Table 2 shows the total levels of this kinase that were not changed by prenatal stress or acute swim stress.

Figure 7.

 Effect of prenatal stress on the phosphorylation of αCaMKII(Thr286) and modulation by an acute swim stress at adulthood, assessed in the prefrontal cortex of male (left panel) and female (right panel) rats. The results, expressed as % of control/unstressed rats, represent the mean ± SEM of 5–9 independent determinations. anova across both prenatal and postnatal treatments, when significant, appears at the bottom of each panel. PS, prenatal stress; SS, acute swim stress; PS × SS, prenatal stress × acute stress interaction.

Modulation of AMPA receptor subunits

Analysis of GluR1 receptor phosphorylation in hippocampus showed that P-GluR1(Ser831) levels were significantly enhanced by acute swim stress (F1,33 = 6.77, = 0.014), whereas prenatal stress had no effect (F1,33 = 2.218, = 0.147) (Fig. 8). Total levels of GluR1 showed a significant increase in prenatally-stressed groups (F1,34 = 4.841, = 0.036), with no effect of acute swim stress (F1,34 = 0.027, = 0.870) (Table 2). In females, neither prenatal stress (F1,24 = 0.638, = 0.434) nor acute swim stress (F1,24 = 1.764, = 0.199) showed any significant modulation of the phosphorylation of this AMPA subunit (Fig. 8). As shown in Table 2, acute swim stress increased the total levels of GluR1 in both sham and antenatally-stressed animals (F1,22 = 4.915, = 0.040), without any effect exerted by prenatal manipulation (F1,22 = 0.730, = 0.404).

Figure 8.

 Effect of prenatal stress on the phosphorylation of AMPA receptor subunit GluR1(Ser831) subunit and modulation by an acute swim stress at adulthood, assessed in the prefrontal cortex of male (left panel) and female (right panel) rats. The results, expressed as % of control/unstressed rats, represent the mean ± SEM of 5–10 independent determinations. anova across both prenatal and postnatal treatments, when significant, appears at the bottom of each panel. PS, prenatal stress; SS, acute swim stress; PS × SS, prenatal stress × acute stress interaction.

Table 2 shows that neither prenatal stress nor acute swim stress were able to modulate the levels of GluR2 subunit, in both males and females.

Discussion

The results of the present study demonstrate that prenatal stress influences glutamatergic responsiveness to a challenge at adulthood, without altering basal glutamate receptor expression. These changes show gender and anatomical specificity since they affect primarily the function of prefrontal cortex in male rats.

While an acute swim stress increases the phosphorylation of the NMDA receptor subunits NR-1(Ser896) and NR-2B(Ser1303) in the prefrontal cortex of control male rats, such effect was not observed for NR-1(Ser896) and was attenuated for NR-2B(Ser1303) in the same brain region of prenatally-stressed animals. It is feasible to hypothesize that increased phosphorylation of NR-1(Ser896) and NR-2B(Ser1303) subunits represents the rapid and coordinated response of the system to the challenge which, via increased calcium influx, may lead to coping responses: the ability to mount such an homeostatic strategy appears to be at least reduced in animals exposed to stress during gestation.

Interestingly, the acute challenge reduced p-αCaMKII(Thr286) levels in the prefrontal cortex of control/sham animals but not in antenatally-stressed rats whereas, at hippocampal level, both groups showed a similar reduction of p-αCaMKII(Thr286) levels. αCaMKII represents the most sensitive sensor of intracellular calcium levels and the phosphorylation in the threonine residue in position 286 confers autonomous kinase activity, i.e. Ca++-calmodulin-independent activity (Colbran and Brown 2004). Thus, whereas increased αCaMKII phosphorylation locks αCaMKII in a Ca++-independent state, which might render the cell unable to cope with a subsequent stimulation, reduced αCaMKII phosphorylation might represent an efficient mechanism to finely adjust the response to further calcium influx: modulation of p-αCaMKII(Thr286) levels might thus represent a tunable gate whose function is compromised in the prefrontal cortex, but still operating in the hippocampus, of animals stressed during gestation. However, although we show that acute swim stress reduces αCaMKII phosphorylation, Suenaga and associates have found increased phosphorylation of the kinase following restraint stress (Suenaga et al. 2004), suggesting that the type and duration of stress might influence its activity. Reduced phosphorylation of αCaMKII might depend on increased activity of protein phosphatases, in line with previous evidence showing that stress enhances the activity of protein phosphatases 2A or 2B (Takahashi et al. 2001; Morinobu et al. 2003). It is difficult to establish the mechanisms responsible for the changes of glutamatergic receptors following prenatal stress. They might be associated with enhanced activation of the hypothalamo-pituitary axis responses (Maccari and Morley-Fletcher 2007) since, given the relationships between glucocorticoids and glutamate receptors (Jing et al. 2008), glucocorticoid hyperactivation might exhaust the pathway of glutamatergic receptor activation, which no longer responds to the acute challenging stress at adulthood.

Whereas increased P-NR1(Ser896) phosphorylation is regulated by Protein Kinase C (Szabo et al. 2009), it is well known that αCaMKII and the NR-2B subunit of the NMDA receptor interact via an enzyme/substrate mechanism (Strack et al. 2000). Under our experimental conditions, acute stress reduced P-αCaMKII(Thr286) levels while increasing NR2B(Ser1303) phosphorylation; this apparent discrepancy might be explained taking into account that autophosphorylation in Thr286 is crucial for such interaction whereas the phosphorylation of NR2B NMDA subunit may occur even in the absence of autophosphorylation (Fong et al. 1989).

Previous work has shown that basal expression of glutamatergic receptors is altered by prenatal stress (Son et al. 2006). These authors had demonstrated that, at the hippocampal level, the expression of NR1 and NR2B is reduced in the post-synaptic density of prenatally-stressed animals thus pointing to a steady-state, basal hypofunction of this system in adult animals that were exposed to stress during gestation. These results are in contrast with those shown in the present report since we did not find any basal change in glutamatergic receptor expression in hippocampus following gestational stress: such discrepancy could be however because of specific experimental conditions such as length (from embryonic day 8.5–18.5 in Son’s paper vs. embryonic day 14–21 in our report) and severity (6 h a day in Son’s paper versus three times a day for 45 min each time in our paper) of prenatal stress as well as to differences in the subcellular compartment examined (i.e. post-synaptic density in Son’s paper versus crude synaptosomal fraction in our manuscript) that may contribute to regional specificities of glutamatergic receptor expression.

There are three noteworthy features of the evidence presented in this report. First, prenatal stress seems to selectively affect the modulation of NMDA receptor subunits driven by swim stress, with no effect on the expression of AMPA receptor subunits thus highlighting a different sensitivity of the glutamatergic receptors to the adaptive mechanisms set in motion by prenatal stress. Second, prenatal stress influences glutamatergic response in rat prefrontal cortex with no effect in the hippocampus; such area-selectivity is in line with our previous observations that prenatal stress affects the response of the neurotrophic factors Brain Derived Neurotrophic Factor and fibroblast growth factor-2 to stressful stimuli at adulthood in cortical regions (Fumagalli et al. 2004, 2005). The only effect observed in hippocampus is the increased phosphorylation of the AMPA subunit GluR1(Ser831) which is, however, triggered by acute swim stress independently from exposure to prenatal stress. Third, the effects were far more substantial in males than in females. The sex-selectivity is referred to both the response to the acute swim stress in prenatally-unstressed animals as well as to the overall influence of prenatal stress on glutamatergic responsiveness. Whereas gender differences are known to exist in animals exposed to gestational stress (Weinstock 2007) (Zuena et al. 2008), the lack of response of females to acute swim stress is quite unexpected given that females are more prone to stress-related disorders. This issue deserves further attention with specific experiments aimed at examining the adult female stress response at different time points of the reproductive cycle.

To sum up, we found regionally-selective changes in modulatory responses of the glutamatergic system as a result of the prenatal insult in male adult rats. Since it is well established that stressful life events occurring early in life have a substantial causal association with neuropsychiatric disorders (Fumagalli et al. 2007b) the altered responsiveness of the glutamatergic synapse herein reported may well contribute to lifelong susceptibility to adverse environmental conditions and the manifestation of psychopathological symptoms.

Acknowledgement

We thank Dr. Giuseppe Giannotti for participating to part of this study. M.P. is funded by the PhD Program in Neuropharmacology, University of Catania, School of Medicine. This work was supported by a grant to M.A.R. from the Ministry of University and Research (PRIN# 2005059982).

Ancillary