Early Maternal Deprivation Reduces the Expression of BDNF and NMDA

Home , NMDA receptor

Molecular Psychiatry (2002) 7, 609–616  2002 Nature Publishing Group All rights reserved 1359-4184/02 $25.00 www.nature.com/mp ORIGINAL RESEARCH ARTICLE Early maternal deprivation reduces the expression of BDNF and NMDA receptor subunits in rat hippocampus M Roceri1,3, W Hendriks2, G Racagni1,3, BA Ellenbroek2 and MA Riva1

1Center of Neuropharmacology, Department of Pharmacological Sciences and Center of Excellence for Neurodegenerative Disorders, University of Milan, Milan, Italy; 2Department of Psychoneuropharmacology, University of Nijmegen, Nijmegen, The Netherlands; 3IRCCS San Giovanni di Dio-Fatebenefratelli, Brescia, Italy

It is well accepted that events that interfere with the normal program of neuronal differentiation and brain maturation may be relevant for the etiology of psychiatric disorders, setting the stage for synaptic disorganization that becomes functional later in life. In order to investigate molecular determinants for these events, we examined the modulation of the neurotrophin brain-derived neurotrophic factor (BDNF) and the glutamate NMDA receptor following 24 h maternal separation (MD) on postnatal day 9. We found that in adulthood the expression of BDNF as well as of NR-2A and NR-2B, two NMDA receptor forming subunits, were signiﬁcantly reduced in the hippocampus of MD rats whereas, among other structures, a slight reduction of NR-2A and 2B was detected only in prefrontal cortex. These changes were not observed acutely, nor in pre-weaning animals. Furthermore we found that in MD rats the modulation of hippocampal BDNF in response to an acute stress was altered, indicating a persistent functional impairment in its regulation, which may subserve a speciﬁc role for coping with challenging situations. We propose that adverse events taking place during brain maturation can modulate the expression of molecular players of cellular plasticity within selected brain regions, thus contributing to permanent alterations in brain function, which might ultimately lead to an increased vulnerability for psychiatric diseases. Molecular Psychiatry (2002) 7, 609–616. doi:10.1038/sj.mp.4001036 Keywords: animal model; brain-derived neurotrophic factor; glutamate; neurodevelopment; schizophrenia; stress

Introduction such as depression and schizophrenia that, to some extent, can be characterized by altered regulation of the It has been long realized that experiences taking place HPA axis and abnormal responses to stress.10,11 early in life have a profound influence on brain devel- Hence if brain development is a key factor for disease opment, produce persistent effects on its function susceptibility, the capability to reproduce alterations while determining individual differences in the vul- occurring early in life becomes crucial for animal nerability to stress.1,2 Early work by Levine has clearly modeling. Since mother–infant interaction may be a demonstrated that stimulation of neonatal rats affects key factor for brain maturation and disease suscepti- their endocrine and behavioral responses later in life.3 bility, maternal deprivation (MD) represents a useful While brief, repeated manipulations (handling) may be paradigm to study disturbances in brain function that beneficial in terms of stress reactivity and result in a might occur in response to adverse events during more efficient coping with the environment, longer development. separations appear detrimental.4–6 Accordingly adverse Specifically, we investigated short- and long-term life events during the perinatal period produce perma- molecular changes taking place after a single 24 h MD nent alterations at the level of hypothalamic-pituitary- on postnatal day 9 (pnd-9), a paradigm that recapitu- adrenal (HPA) axis, which are accompanied by high lates several features of schizophrenia.12 Although dif- anxiety-like behavior and age-related memory impair- ferent factors may be important for the derangement in ment.7–9 Furthermore early adverse situations may brain development, we reasoned that early adverse life increase the vulnerability to psychiatric disorders, events could produce long-lasting impairment in brain function through changes in important modulators of synaptic plasticity. The present study tested this Correspondence: MA Riva, Center of Neuropharmacology, hypothesis through the analysis of the neurotrophin Department of Pharmacological Sciences, University of Milan, brain derived neurotrophic factor (BDNF) and the glut- Via Balzaretti 9, 20133 Milan, Italy. E-mail: marivaȰmailserver. unimi.it amate N-methyl-D-aspartate (NMDA) receptor. Received 25 July 2001; revised 16 October 2001; accepted 14 Changes in neurotrophin levels as well as in the November 2001 expression of this glutamate receptor subtype have Maternal deprivation downregulates neuroplastic genes in rat hippocampus M Roceri et al 610 been recently documented in psychiatric patients, viously described.24 The brain specimens were frozen pointing to a potential contribution in schizophrenic on dry ice and stored at −80°C for further analysis. pathology and/or etiology.13–16 BDNF can affect neuronal survival and differentiation17 and, together with RNA preparation the NMDA receptor complex, play a role as mediators Brain tissue was homogenized in 4 M guanidinium iso- of activity-dependent brain plasticity as well as in spa- thiocyanate (containing 25 mM sodium citrate pH 7.5, tial learning and memory.18–22 0.5% sarcosyl and 0.1% 2-mercaptoethanol) and total We have therefore analyzed the expression of these RNA was isolated by phenol-chloroform extraction elements in pre-weaning and adult MD rats, as well as according to standard procedures. Quantification was the levels of BDNF in adult animals in response to an carried out by spectrophotometric analysis and RNA acute challenging situation in order to evaluate aliquots were re-precipitated in ethanol for RNase pro- whether the stress-dependent regulation of neurotro- tection assay. phin production could be altered as a consequence of early exposure to MD. RNA probes synthesis and RNase protection assay A transcription kit (MAXI script, Ambion) was used to generate cRNA probes and 32P-CTP was used as a Materials and methods radiolabelled nucleotide. The cRNA probes and the Materials relative protected fragment (p.f.) were the following: General reagents were purchased from Sigma (Milan, BDNF = 800, p.f. = 740; Trk-B = 507, p.f. = 447; NR-1 Italy), whereas molecular biology reagents were = 438, p.f. = 414; NR-2A = 187, p.f. = 171; NR-2B = 310, obtained from Ambion (Austin, TX, USA), New p.f. = 264; GAPDH (pTRI-GAPDH-Rat, Ambion) = 376, England Biolabs (Beverly, MA, USA) and Promega p.f. = 316. The cRNA probes for NMDA receptor sub- (Milan, Italy). units have been previously characterized for their specificity and used for the simultaneous detection of Animals and maternal deprivation their mRNA levels in rat brain regions.25 We used Male and nulliparous female Wistar rats of 3 months 150 000 cpm per sample of the 32P-labeled cRNA to old were put together in standard Macrolon cages BDNF, NR-2B and NR-2A, and 50 000 cpm per sample with sawdust (26 × 42 × 15 cm), in temperature-con- of the 32P-labeled cRNA to Trk-B, NR-1 or GAPDH. trolled rooms (23 ± 1°C). The rats were on a standard The RNase protection assay was performed on a 10 12 h light/dark cycle with light on from 7.00 to 19.00 h, ␮g sample of total RNA as described previously.24 with water and food freely available. Two weeks later Briefly, after ethanol-precipitation, total RNA was re- the males were removed from the cages and females suspended in 20 ␮l of hybridization solution (80% for- were checked twice daily for delivery (08.00 and mamide, 40 mM PIPES pH 6.4, 400 mM sodium acetate 17.00 h). The day of delivery was denoted as pnd 0. pH 6.4 and 1 mM EDTA) containing different 32P-lab- On pnd 1 the litters were weighed and culled to six eled cRNA probes. Samples were hybridized at 45°C males and two females. Maternal deprivation took overnight. At the end of the hybridization, the solution place on pnd 9 according to our standard procedure.12 was diluted with 200 ␮l of RNase digestion buffer con- In brief, the mothers were removed at 10.00 h, after taining a 1/200 dilution of a RNase cocktail (1 mg ml−1 which the pups were weighed and returned to their RNase A and 20 U ml−1 RNase T1) and incubated for homecages. They remained in their homecage at room 30 min at 30°C. Proteinase K (10 ␮g) and SDS (10 ␮l temperature for 24 h. On pnd 10, the pups were of 20% stock solution) were then added to the sample weighed again and the mothers returned to their cages. and the mixture was incubated at 37°C for an The mothers of the control litters were briefly removed additional 15 min. At the end of the incubation the from their homecages and the pups were weighed both sample was extracted with phenol/chloroform and on pnd 9 and pnd 10. Weaning took place at pnd 21, ethanol precipitated. The pellet, containing the where rats were normally housed in groups of three RNA:RNA hybrids, was dried and resuspended in load- males and two females per cage. All the determinations ing buffer, boiled at 95°C for 5 min and separated on were carried out on male rats. a 5% polyacrylamide gel under denaturing conditions (7 M urea). The protected fragments were visualized Stress experiment by autoradiography. The effect of acute stress on BDNF gene expression was assessed in adult sham or MD animals (pnd 72). Rats Analysis of protein levels were randomly assigned to four groups (control, no The hippocampus of control or MD animals was homo- MD/no stress; no MD/stress; MD/no stress; MD/stress). genized in lysis buffer (137 mM NaCl, 20 mM Tris, 1% The animals undergoing restraint stress were placed for NP-40 detergent, 10% glycerol, 1 mM methylphenyl- 2 h in transparent plexiglass containers whereas those sulfonylfluoride, 10 ␮gml−1 aprotinin, 1 ␮gml−1 in the ‘no-stress’ group were left undisturbed in their leupeptin and 0.5 mM sodium orthovanadate). home cages. Rats were killed by decapitation 1 h after Homogenates were centrifuged at 16 000 g for 30 min the end of immobilization. Following death most brain in a cold room and supernatants were used for ELISA regions were rapidly dissected from 2-mm thick slices analysis. The BDNF or NT-3 Emax ImmunoAssay according to the atlas of Paxinos and Watson,23 as pre- system (Promega) was used to measure BDNF or NT-3

Molecular Psychiatry Maternal deprivation downregulates neuroplastic genes in rat hippocampus M Roceri et al 611 protein levels in hippocampus according to the manu- protection assay for the analysis of BDNF mRNA levels facturer’s protocol. in the hippocampus of control or MD adult rats. We observed a significant reduction of the mRNA levels for RNA calculation and statistical analysis BDNF in the hippocampus of MD animals (Figure 1b) The levels of mRNA for BDNF, Trk-B, NMDA-R sub- with no changes in the expression of its high affinity units or GAPDH were calculated using the Quantity receptor Trk-B (Control = 100 ± 9; MD = 103 ± 11). This One software from Biorad. In order to ensure that the effect is rather specific from the anatomical point of autoradiographic bands were in the linear range of view, since no significant changes were detected in intensity, different exposure times were used. GAPDH other brain structures, including cortical regions was employed as internal standard for RNase protec- (prefrontal, frontal, piriform and parietal) as well as tion assay: its expression was not altered by MD or by thalamus and nucleus accumbens (Figure 1b). The stress. Optical densities from experimental groups downregulation of BDNF mRNA was accompanied by were then expressed and presented as mean percent of a significant decrease of its protein levels (−32 ± 6% of control values for graphic clarity. The mean percent controls), whereas the expression of another neurotro- represents the ratio between BDNF, Trk-B or NMDA-R phin, NT-3, was not altered in the hippocampus of MD subunits and GAPDH levels expressed as % of control rats (Figure 2). animals. The mean value of the control group within Since it is known that stress can modulate the a single experiment was set to 100 and all other values expression of BDNF,26 we investigated if the changes were expressed as ‘percent’ of control animals. observed within the hippocampus represent the result Changes produced by MD on BDNF or NMDA-R sub- of the acute deprivation. However the reduction of hip- units in different brain regions, at different ages, were pocampal BDNF appears to be a long-term conse- analyzed by one-way analysis of variance (ANOVA). quence of the early events since, as shown in Figure 3, Significant changes were determined by Dunnett’s test. the mRNA levels for BDNF and Trk-B were not The results of the acute stress experiment were ana- changed 2, 6 or 24 h after the deprivation on pnd-9. lyzed by two-way analysis of variance (ANOVA), with Furthermore the expression of the neurotrophin and its status (sham and MD) and treatment (no stress and receptor was also unchanged on pnd-21 (Figure 3). stress) as independent factors and BDNF levels as As MD during early postnatal life produces persist- dependent variables. When appropriate, further different changes in the stress response,27,28 we decided to ences were analyzed by Single Contrast post-hoc test investigate the expression of BDNF under a challenging (SCPHT). situation in adult animals. For this reason control and MD rats were exposed to a single 2 h restraint stress, a procedure known to reduce BDNF levels in the hippo- Results campus.26 We observed a significant effect of MD on The expression of the neurotrophin BDNF was ana- stress-induced modification of BDNF mRNA levels in lyzed in adult animals that experienced a single 24 h adult animals. In fact the decrease of BDNF mRNA MD on pnd-9. Figure 1a shows a representative RNase expression after immobilization was significant in

Figure 1 Effect of maternal deprivation (24 h on pnd-9) on BDNF mRNA levels in the adult rat brain. (a) RNase protection analysis of BDNF mRNA levels in the hippocampus of control (CTRL) or maternal deprived (MD) adult rats. Arrows indicate the protected fragments for BDNF or GAPDH, used as internal standard. The autoradiographic ﬁlm was exposed at −70°C with an intensifying screen for 5 h (GAPDH) or 18 h (BDNF). (b) The levels of BDNF mRNA in different brain regions are expressed as the percentage of change in MD rats with respect to control animals. Each value is the mean ± SEM of 8–12 independent determinations in the following brain regions: HIP, hippocampus; PFC, prefrontal cortex; FC, frontal cortex; PIR, piriform cortex; PC, parietal cortex; THA, thalamus; NA, nucleus accumbens. *PϽ0.01 vs control animals (one-way ANOVA with Dunnett t-test).

Molecular Psychiatry Maternal deprivation downregulates neuroplastic genes in rat hippocampus M Roceri et al 612

Figure 4 Effects of the acute restraint stress on BDNF mRNA Figure 2 Analysis of BDNF and NT-3 protein levels in the levels in the hippocampus of control or MD rats assessed at hippocampus of adult control or MD rats. The data represent pnd 72. The results, expressed as % of control (no MD/no ± the mean SEM of at least six independent determinations stress) rats, represent the mean ± SEM of 6–8 samples for each Ͻ for each experimental group. *P 0.05 vs control animals experimental group. *PϽ0.01 vs control animals; **PϽ0.01 (one-way ANOVA with Dunnett t-test). vs control and PϾ0.05 vs MD/no stress rats (two-way ANOVA with SCPHT).

get for short-term modification of synaptic function by neurotrophins.21,31 As adults, MD rats display a significant reduction of hippocampal NR-2A and NR-2B subunit expression, while no changes were observed in the mRNA levels for NR-1 (Figure 5a and b). As shown for BDNF, this effect appears to be quite specific for this brain region, although a similar, but not significant reduction in the expression of NR-2A and -2B subunits was observed in prefrontal cortex (but not in other structures). Furthermore the temporal occurrence of the changes in hippocampal NMDA receptor subunits matches those on BDNF, as no modifications of NR-2A and NR-2B mRNA levels were evident in pre- weaning rats (Table 1).

Discussion The present results demonstrate that a single 24 h MD on pnd-9 produces a persistent reduction in the expression of the neurotrophin BDNF as well as of the Figure 3 Time course analysis of the changes in hippocam- glutamate NMDA receptor subunits NR-2A and NR-2B pal mRNA levels for BDNF and its high affinity receptor Trk- in rat hippocampus and, to lesser extent, in cortical B following MD. Animals were killed either 2, 6 or 24 h after areas. Our results point to the fact that variations in removal of the mother (pnd-9/10) or they were killed on pnd- maternal care not only influence the development of 21 after 24 h MD on pnd-9. The results, expressed as % of ± behavior and endocrine responses to stress in the off- change with respect to control rats, represent the mean SEM 1,5,6 of 5–8 samples for each experimental group. spring, but determine alterations in molecular players of cellular plasticity. These effects are specific from the anatomical and temporal point of view and might = = sham (F1,28 17.016, P 0.001; SCPHT) but not in MD represent one of the mechanisms through which early = = rats (F1,28 0.275, P 0.603; SCPHT) (Figure 4), sug- adverse life events determine long-term impairment in gesting that in these animals the stress-dependent regu- brain function and plasticity. lation of BDNF was significantly altered. Several years ago Levine showed that the develop- Glutamate is one of the neurotransmitters that regu- ment of the HPA axis response to stress is modified by lates the expression of BDNF.29,30 We have therefore early environmental events.3 It is well established that investigated, under the same experimental paradigm, neonatal handling or high maternal care determines a the expression of the subunits for the glutamate NMDA reduction of ACTH and corticosterone response to receptor (NR), a system involved in activity-dependent stress,6 whilst adverse life events during early develop- plasticity of the hippocampus, which can also be a tar- ment enhance the HPA axis activity.27 The effects we

Molecular Psychiatry Maternal deprivation downregulates neuroplastic genes in rat hippocampus M Roceri et al 613

Figure 5 Effect of MD (24 h on pnd-9) on the gene expression for NMDA receptor subunits in the adult rat brain. (a) RNase protection analysis of NMDA receptor subunit mRNA levels in the hippocampus of control (CTRL) or maternal deprived (MD) adult rats. The lane marked as probe represents an aliquot of the hybridization solution containing the cRNA probes. Arrows indicate the protected fragments for NR-1, NR-2A, NR-2B or GAPDH, used as internal standard. The autoradiographic ﬁlm was exposed at −70°C with an intensifying screen for 5 h (NR-1, NR-2B and GAPDH) or 18 h (NR-2A). (b) Levels of expression for NMDA receptor subunit mRNA in different brain structures are expressed as the percentage of change in MD rats with respect to control animals. Each value is the mean ± SEM of at least eight independent determinations in the following brain regions: HIP, hippocampus; PFC, prefrontal cortex; FC, frontal cortex; THA, thalamus. *PϽ0.05 vs control animals (one-way ANOVA with Dunnett t-test).

Table 1 Analysis of the mRNA levels for NMDA receptor care, a situation that can be considered the opposite subunits in the hippocampus of control and MD rats, as with respect to MD, favors hippocampal plasticity and assessed at pnd 21. The results are expressed as % of control enhances spatial learning and memory through the animals and represent the mean ± SEM of 6–8 samples for upregulation of BDNF and NMDA-R subunits.35 How- each experimental group ever while the changes of the neurotrophin expression they report are evident on pnd-8, but not in adulthood, Group NR-1 NR-2A NR-2B 24 h MD produces a reduction of BDNF mRNA and (% of control levels) protein only in adult animals. In fact modifications in the neurotrophin biosynthesis were observed neither 2, Control 100 ± 6 100 ± 10 100 ± 6 ± ± ± 6 nor 24 h after acute MD nor in pre-weaning rats (pnd- MD 108 3996 108 5 21), suggesting that the changes in its expression may represent the long-term consequence of developmental abnormalities, rather than being a short-term effect of report display a rather selective anatomical specificity the deprivation. We argue that 24 h MD may interfere for the hippocampus, since less consistent changes in with hippocampal maturation and this may lead to a the expression of BDNF and NR-2A/2B were observed reduction of BDNF levels in adulthood as a conse- only in frontal and prefrontal cortex. The hippocampus quence of changes in neurotransmitters or hormones undergoes an experience-dependent development and involved in the control of the neurotrophin production. its maturation extends into the first 3 weeks of post- The modulation of BDNF levels is somewhat differ- natal life, suggesting that the time of deprivation ent from what we have previously observed using coincides with an active synaptogenesis at this level. another paradigm, namely the hippocampal lesion Moreover high levels of glucocorticoid receptors (GR) with ibotenic acid on pnd-7. This animal model, simi- are present in the hippocampus, where they mediate larly to 24 h MD, reproduces schizophrenia-like dys- feedback control on the HPA axis as well as contribute function that, at least in part, is ameliorated by to the vulnerability of this region to stress-related exposure to antipsychotic drugs.36 We found that, in events.32 A balanced supply of corticosteroids appears adulthood, animals lesioned with ibotenic acid have to be important for hippocampal integrity since reduced basal levels of BDNF mRNA in cingulate cor- chronic elevation of corticosteroid as well as adrena- tex as well as alterations in its acute modulation within lectomy can lead to degeneration of specific cellular the prefrontal cortex.24 This suggests that damage of the phenotypes within the hippocampal formation.33,34 hippocampus during the late stage of maturation may Liu et al have recently reported that high maternal affect regions that are targets of hippocampal output,

Molecular Psychiatry Maternal deprivation downregulates neuroplastic genes in rat hippocampus M Roceri et al 614 whereas stress-related events during early postnatal modulation) may be reduced. However an impaired life have a stronger and more selective impact over the feedback control on the HPA axis will lead to a pro- hippocampus itself. A common element in these two longed elevation of circulating corticosterone that, in paradigms is represented by the emergence of BDNF turn, might determine a more protracted reduction of abnormalities after puberty, which is consistent with a BDNF expression. In this respect Biagini et al have major feature of schizophrenia37 and corresponds to shown that repeated MD from pnd 2 to 6 leads to the post-pubertal development of other schizophrenia- higher corticosterone release and lower hippocampal related abnormalities in both experimental models.12,38 GR immunoreactivity after novelty.49 Accordingly we This suggests that a latency does occur between the have recently found that short-term exposure to adverse events and the full manifestation of impair- corticosterone during late gestation, decreases BDNF ment in cellular plasticity. biosynthesis and produces a less prominent but more It is known that BDNF exerts an important role in protracted reduction of its expression after an acute synaptic transmission and plasticity within the hippo- restraint stress (Molteni and Riva, unpublished campus.18,21,22,39 Its expression is regulated by neuronal observations). Although ‘speculative’, these data sug- activity40,41 and the neurotrophin can modulate gest that a ‘worsening’ in the modulation of the neuro- synaptic function at different levels.18 Application of trophin might occur at later time points, particularly exogenous BDNF facilitates hippocampal long-term under enduring exposure to stress. This will ultimately potentiation (LTP), whereas animals with impaired lead to an increased vulnerability, which might expression of BDNF display defects in LTP and learn- recapitulate what is taking place during the course of 42 ing. This functional abnormality may depend upon psychiatric illnesses. The impairment in the acute impairment in neurotransmitter release, since hetero- modulation of BDNF may also be detrimental for the + zygous ( /-) BDNF mice display a significant reduction performance of specific functional tasks, which are of the vesicles docked at presynaptic active zones, as regulated by the neurotrophin. A rapid and selective well as a reduction of synaptobrevin and synaptophy- induction of BDNF does indeed occur in the hippo- sin, two proteins involved in vesicle docking and 50 43 campus during contextual learning in rats as well as fusion. Interestingly it has been reported that the in the inferior temporal cortex of monkey during upregulation of BDNF associated with high maternal declarative memory formation.51 This suggests that a care is paralleled to a significant increment in the reciprocal modulation between BDNF and synaptic expression of synaptophisin.35 It may therefore be function exists and that the proper control of neurotro- inferred that a reduction in the expression of neurotrophin production may be important for activity-depen- phin in MD animals could lead to an impairment in dent plasticity. neurotransmitter secretion and a functional reduction The reduction of hippocampal BDNF expression in activity-dependent processes. after MD may, at least in part, be related to the gluta- The reduction in the expression of hippocampal mate NMDA receptor, which can be a target for short- BDNF after MD may also increase the vulnerability of term effects of BDNF on synaptic plasticity,21,31 but can this region under challenging situations. BDNF is also operate in the modulation of BDNF expression.29,30 highly protective for different neuronal phenotypes and can prevent neuronal cell death in different experi- Indeed there is a close resemblance in the temporal and mental paradigms in vivo as well as in vitro.44,45 anatomical pattern of the changes for NR-2A and NR- Endogenous production of BDNF is essential for the 2B subunits, which occur in the hippocampus, and to survival and maturation of cortical neurons in vitro46 lesser extent in prefrontal cortex, of adult MD rats. The and can also support cell sprouting after striatal NMDA receptor is a heteromeric complex consisting of 52 injury.47 Conversely, environmental enrichment pro- subunit 1 (NR-1) and 2 (NR-2 A, B, C and D). Genetic duces an upregulation of hippocampal BDNF (NR-1 knock-down mice) and pharmacological expression, which is protective against spontaneous (phencyclidine) manipulations of the glutamate NMDA apoptotic cell death and kainate-induced degener- receptor reproduce some of the deficits observed in 53,54 ation.48 All together these data point to an important schizophrenia and abnormalities in the expression role for endogenously expressed BDNF and its acute of this glutamate receptor subtype have been reported 15,16 modulation, in setting the vulnerability threshold of in the brain of schizophrenic subjects. In line with specific cellular phenotypes. A reduction of its the well-established role of glutamate NMDA receptors biosynthesis could therefore be of relevance in situ- in learning and memory,55 high maternal care ations, such as stress, which imply the elaboration of upregulates the expression of different NMDA receptor effective coping strategies. subunits in the hippocampus of developing and adult Furthermore our results indicate that MD induced a animals, while determining an improvement in cogni- persistent change in the mechanisms involved in the tive activity.35 Hence a reduction of NR-2A and NR-2B acute modulation of BDNF. In fact, the downregulation subunits, in conjunction with BDNF, might not only of neurotrophin expression in control animals follow- reduce brain plasticity but also determine a decline in ing an acute restraint stress was not observed in rats cognitive function, which represents a major feature of that have been exposed to MD on pnd-9. Since MD schizophrenia-related disorders. Accordingly, Oitlz reduces GR expression in the hippocampus,27 the and coworkers have shown that 24 h MD on pnd-3 pro- initial response to stress (in terms of BDNF duces cognitive impairment throughout life,9 whereas

Molecular Psychiatry Maternal deprivation downregulates neuroplastic genes in rat hippocampus M Roceri et al 615 prenatal stress has a stronger impact on the cognitive 16 Akbarian S, Sucher NJ, Bradley D, Tafazzoli A, Trinh D, Hetrick performance of aged rats.7 WP et al. Selective alterations in gene expression for NMDA receptor subunits in prefrontal cortex of schizophrenics. J Neurosci 1996; Taken together, our data suggest that a reduction of 16:19–30. BDNF and NMDA-R subunit expression can contribute 17 Lewin GR, Barde YA. Physiology of the neurotrophins. Ann Rev to long-term impairment in brain function following Neurosci 1996; 19:289–317. early life stress. This provides a rationale for 18 Poo MM. Neurotrophins as synaptic modulators [Review]. Nature Rev Neurosci 2001; 2:24–32. developing novel therapeutic strategies aimed to re- 19 Mizuno M, Yamada K, Olariu A, Nawa H, Nabeshima T. Involve- establish the function of these modulators for treating ment of brain-derived neurotrophic factor in spatial memory forma- disorders, such as depression and schizophrenia, tion and maintenance in a radial arm maze test in rats. J Neurosci which might be characterized by a reduction in cellu- 2000; 20:7116–7121. lar plasticity. 20 Lu B, Gottschalk W. Modulation of hippocampal synaptic transmission and plasticity by neurotrophins. Prog Brain Res 2000; 128: 231–241. Acknowledgements 21 Black IB. Trophic regulation of synaptic plasticity. J Neurobiol 1999; 41:108–118. We thank Dr G Yankopoulos for providing cDNA 22 Thoenen H. Neurotrophins and neuronal plasticity. Science 1995; probes to BDNF and Trk-B. We also wish to thank Dr 270:593–598. 23 Paxinos G, Watson C. The Rat Brain in Stereotaxic Coordinates. Raffaella Molteni for helpful discussion. In part sup- Academic Press: New York, 1996. ported by a NARSAD Young Investigator Award to 24 Molteni R, Lipska BK, Weinberger DR, Racagni G, Riva MA. Devel- MAR. opmental and stress-related changes of neurotrophic factor gene expression in an animal model of schizophrenia. Mol Psychiatry 2001; 6:285–292. References 25 Riva MA, Tascedda F, Molteni R, Racagni G. Regulation of NMDA receptor subunit mRNA expression in the rat brain during postnatal 1 Francis DD, Meaney MJ. Maternal care and the development of development. Mol Brain Res 1994; 25: 209–216. stress responses. Curr Opin Neurobiol 1999; 9:128–134. 26 Smith MA, Makino S, Kvetnansky R, Post RM. Stress and glucocort- 2 Francis DD, Champagne FA, Liu D, Meaney MJ. Maternal care, gene icoids affect the expression of brain-derived neurotrophic factor expression, and the development of individual differences in stress and neurotrophin-3 mRNAs in the hippocampus. J Neurosci 1995; reactivity. Ann N Y Acad Sci USA 1999; 896:66–84. 15: 1768–1777. 3 Levine S. Infantile experience and resistence to physiological 27 van Oers HJ, de Kloet ER, Whelan T, Levine S. Maternal depri- stress. Science 1957; 126:405–406. vation effect on the infant’s neural stress markers is reversed by 4 Meaney MJ, Aitken DH, van Berkel C, Bhatnagar S, Sapolsky RM. tactile stimulation and feeding but not by suppressing corticos- Effect of neonatal handling on age-related impairments associated terone. J Neurosci 1998; 18: 10171–10179. with the hippocampus. Science 1988; 239: 766–768. 28 van Oers HJ, de Kloet ER, Levine S. Early vs late maternal depri- 5 van Oers HJ, de Kloet ER, Li C, Levine S. The ontogeny of glucocorticoid negative feedback: influence of maternal deprivation. Endo- vation differentially alters the endocrine and hypothalamic crinology 1998; 139: 2838–2846. responses to stress. Dev Brain Res 1998; 111:245–252. 6 Liu D, Diorio J, Tannenbaum B, Caldji C, Francis D, Freedman A 29 Bhave SV, Ghoda L, Hoffman PL. Brain-derived neurotrophic factor et al. Maternal care, hippocampal glucocorticoid receptors, and mediates the anti-apoptotic effect of NMDA in cerebellar granule hypothalamic-pituitary-adrenal responses to stress. Science 1997; neurons: signal transduction cascades and site of ethanol action. J 277: 1659–1662. Neurosci 1999; 19: 3277–3286. 7 Vallee M, MacCari S, Dellu F, Simon H, Le Moal M, Mayo W. Long- 30 Marini AM, Rabin SJ, Lipsky RH, Mocchetti I. Activity-dependent term effects of prenatal stress and postnatal handling on age-related release of brain-derived neurotrophic factor underlies the neuro- glucocorticoid secretion and cognitive performance: a longitudinal protective effect of N-methyl-D-aspartate. J Biol Chem 1998; 273: study in the rat. Eur J Neurosci 1999; 11: 2906–2916. 29394–29399. 8 Lemaire V, Koehl M, Le Moal M, Abrous DN. Prenatal stress pro- 31 Song DK, Choe B, Bae JH, Park WK, Han IS, Ho WK et al. Brain- duces learning deficits associated with an inhibition of neurogen- derived neurotrophic factor rapidly potentiates synaptic trans- esis in the hippocampus. Proc Natl Acad Sci USA 2000; 97: mission through NMDA, but suppresses it through non-NMDA 11032–11037. receptors in rat hippocampal neuron. Brain Res 1998; 799: 176– 9 Oitzl MS, Workel JO, Fluttert M, Frosch F, De Kloet ER. Maternal 179. deprivation affects behaviour from youth to senescence: amplifi- 32 De Kloet ER, Vreugdenhil E, Oitzl MS, Joels M. Brain corticosteroid cation of individual differences in spatial learning and memory in receptor balance in health and disease. Endocr Rev 1998; 19: senescent Brown Norway rats. Eur J Neurosci 2000; 12: 3771–3780. 269–301. 10 McEwen BS. The neurobiology of stress: from serendipity to clini- 33 Magarinos AM, Verdugo JM, McEwen BS. Chronic stress alters syn- cal relevance. Brain Res 2000; 886:172–189. aptic terminal structure in hippocampus. Proc Natl Acad Sci USA 11 McEwen BS. Allostasis and allostatic load: implications for neuro- 1997; 94: 14002–14008. psychopharmacology. Neuropsychopharmacology 2000; 22: 108– 34 McEwen BS. Protective and damaging effects of stress mediators. 124. N Engl J Med 1998; 338:171–179. 12 Ellenbroek BA, van den Kroonenberg PT, Cools AR. The effects of 35 Liu D, Diorio J, Day JC, Francis DD, Meaney MJ. Maternal care, an early stressful life event on sensorimotor gating in adult rats. hippocampal synaptogenesis and cognitive development in rats. Schizophrenia Res 1998; 30:251–260. Nature Neurosci 2000; 3:799–806. 13 Nawa H, Takahashi M, Patterson PH. Cytokine and growth factor 36 Lipska BK, Weinberger DR. Subchronic treatment with haloperidol involvement in schizophrenia—support for the developmental and clozapine in rats with neonatal excitotoxic hippocampal dam- model. Mol Psychiatry 2000; 5: 594–603. age. Neuropsychopharmacology 1994; 10:199–205. 14 Takahashi M, Shirakawa O, Toyooka K, Kitamura N, Hashimoto T, 37 Duncan GE, Sheitman BB, Lieberman JA. An integrated view of Maeda K et al. Abnormal expression of brain-derived neurotrophic pathophysiological models of schizophrenia. Brain Res Brain Res factor and its receptor in the corticolimbic system of schizophrenic Rev 1999; 29: 250–264. patients. Mol Psychiatry 2000; 5:293–300. 38 Lipska BK, Swerdlow NR, Geyer MA, Jaskiw GE, Braff DL, Wein- 15 Gao XM, Sakai K, Roberts RC, Conley RR, Dean B, Tamminga CA. berger DR. Neonatal excitotoxic hippocampal damage in rats causes Ionotropic glutamate receptors and expression of N-methyl-D- post-pubertal changes in prepulse inhibition of startle and its dis- aspartate receptor subunits in subregions of human hippocampus: ruption by apomorphine. Psychopharmacology 1995; 122:35–43. effects of schizophrenia. Am J Psychiatry 2000; 157: 1141–1149. 39 Kang H, Schuman EM. Long-lasting neurotrophin-induced

Molecular Psychiatry Maternal deprivation downregulates neuroplastic genes in rat hippocampus M Roceri et al 616 enhancement of synaptic transmission in the adult hippocampus. 48 Young D, Lawlor PA, Leone P, Dragunow M, During MJ. Environ- Science 1995; 267: 1658–1662. mental enrichment inhibits spontaneous apoptosis, prevents seiz- 40 Thoenen H. Neurotrophins and activity-dependent plasticity. Prog ures and is neuroprotective. Nature Med 1999; 5:448–453. Brain Res 2000; 128: 183–191. 49 Biagini G, Pich EM, Carani C, Marrama P, Agnati LF. Postnatal 41 Shieh PB, Ghosh A. Molecular mechanisms underlying activity- maternal separation during the stress hyporesponsive period dependent regulation of BDNF expression. [Review] J Neurobiol enhances the adrenocortical response to novelty in adult rats by 1999; 41:127–134. affecting feedback regulation in the CA1 hippocampal ﬁeld. Int J 42 Linnarsson S, Bjorklund A, Ernfors P. Learning deﬁcit in BDNF Develop Neurosci 1998; 16: 187–197. mutant mice. Eur J Neurosci 1997; 9: 2581–2587. 50 Hall J, Thomas KL, Everitt BJ. Rapid and selective induction of 43 Pozzo-Miller LD, Gottschalk W, Zhang L, McDermott K, Du J, Gopa- BDNF expression in the hippocampus during contextual learning. lakrishnan R et al. Impairments in high-frequency transmission, Nature Neurosci 2000; 3:533–535. synaptic vesicle docking, and synaptic protein distribution in the 51 Tokuyama W, Okuno H, Hashimoto T, Xin Li Y, Miyashita Y. hippocampus of BDNF knockout mice. J Neurosci 1999; 19: BDNF upregulation during declarative memory formation in mon- 4972–4983. key inferior temporal cortex. Nature Neurosci 2000; 3:1134–1142. 44 Barde YA. Neurotrophic factors: an evolutionary perspective 52 Seeburg PH. The TINS/TiPS Lecture. The molecular biology of [Review]. J Neurobiol 1994; 25: 1329–1333. mammalian glutamate receptor channels. Trends Neurosci 1993; 45 Siegel GJ, Chauhan NB. Neurotrophic factors in Alzheimer’sand 16: 359–365. Parkinson’s disease brain. Brain Res Brain Res Rev 2000; 33: 53 Mohn AR, Gainetdinov RR, Caron MG, Koller BH. Mice with 199–227. reduced NMDA receptor expression display behaviors related to 46 Ghosh A, Carnahan J, Greenberg ME. Requirement for BDNF in schizophrenia. Cell 1999; 98:427–436. activity-dependent survival of cortical neurons. Science 1994; 263: 54 Jentsch JD, Roth RH. The neuropsychopharmacology of phencycli- 1618–1623. dine: from NMDA receptor hypofunction to the dopamine hypoth- 47 Batchelor PE, Liberatore GT, Porritt MJ, Donnan GA, Howells DW. esis of schizophrenia. Neuropsychopharmacology 1999; 20: 201– Inhibition of brain-derived neurotrophic factor and glial cell line- 225. derived neurotrophic factor expression reduces dopaminergic 55 Bliss TV, Collingridge GL. A synaptic model of memory: long-term sprouting in the injured striatum. Eur J Neurosci 2000; 12: 3462– potentiation in the hippocampus. Nature 1993; 361:31–39. 3468.

Molecular Psychiatry