Early Life Stress Differentially Modulates Distinct Forms of Brain Plasticity in Young and Adult Mice

Background Early life trauma is an important risk factor for many psychiatric and somatic disorders in adulthood. As a growing body of evidence suggests that brain plasticity is disturbed in affective disorders, we examined the short-term and remote effects of early life stress on different forms of brain plasticity. Methodology/Principal Findings Mice were subjected to early deprivation by individually separating pups from their dam in the first two weeks after birth. Distinct forms of brain plasticity were assessed in the hippocampus by longitudinal MR volumetry, immunohistochemistry of neurogenesis, and whole-cell patch-clamp measurements of synaptic plasticity. Depression-related behavior was assessed by the forced swimming test in adult animals. Neuropeptides and their receptors were determined by real-time PCR and immunoassay. Early maternal deprivation caused a loss of hippocampal volume, which returned to normal in adulthood. Adult neurogenesis was unaffected by early life stress. Long-term synaptic potentiation, however, was normal immediately after the end of the stress protocol but was impaired in adult animals. In the forced swimming test, adult animals that had been subjected to early life stress showed increased immobility time. Levels of substance P were increased both in young and adult animals after early deprivation. Conclusion Hippocampal volume was affected by early life stress but recovered in adulthood which corresponded to normal adult neurogenesis. Synaptic plasticity, however, exhibited a delayed impairment. The modulation of synaptic plasticity by early life stress might contribute to affective dysfunction in adulthood.


Introduction
Brain plasticity is highly sensitive to stress. Acute and chronic stressors downregulate both morphological and functional plasticity.
An important form of morphological brain plasticity is adult neurogenesis [1]. It has been shown in several species that the number of proliferating cells in the dentate gyrus diminishes shortly after acute and chronic stress [2,3]. Moreover, other morphological changes in response to stress have been shown in the hippocampus, including retraction of apical dendrites of CA3 pyramidal neurons and loss of astrocytes [4,5]. These findings might explain, at least in part, the hippocampal volume loss that has been observed after stress [6,7]. In humans, there have been repeated reports of decreased hippocampal volume in stressrelated disorders as depression and posttraumatic stress disorder [8,9].
Whereas adult neurogenesis and stress-related brain volume loss are limited to distinct brain regions, plastic alterations of synaptic transmission are found ubiquitously in the brain. Long-term potentiation (LTP) increases synaptic efficacy, whereas long-term depression (LTD) decreases synaptic efficacy. As long-term synaptic plasticity can be measured for weeks or even months in living animals, it has been postulated that it is the molecular correlate of learning and memory [10,11]. Stress has a profound impact on long-term synaptic plasticity. Both acute and chronic stress have been shown to facilitate hippocampal LTD and to impair LTP [12,13].
Most of these findings were obtained directly after the end of the stress protocol, describing immediate effects of stress on brain plasticity. Much less is known about the long-term effects of stress. For example, little is known about how stress during early development modulates brain plasticity in adulthood. Remote effects of early life stress could be highly relevant; adverse childhood experiences that are associated with high levels of prolonged stress have been identified as important risk factors for a multitude of psychiatric and medical disorders, including depression, posttraumatic stress disorder, personality disorders, and cognitive dysfunction, as well as obesity, diabetes, and cardiovascular disease [14,15,16].
We applied a combination of in vivo and ex vivo methods to elucidate the effects of early life stress on different forms of plasticity in the hippocampus. We used the early deprivation (ED) paradigm, in which individual mouse pups were isolated from their dam and littermates repeatedly during the first two weeks of their postnatal development. In rats, ED has been shown to induce reduced body weight of pups and to modify the acoustical features of their ultrasonic vocalizations when separated from their dam. ED has also been found to affect maternal care upon retrieval of pups [17,18]. In adult rats that had undergone ED, decreased emotionality in novel settings, a tendency for increased anxiety, anhedonia-like traits, and reduced social motivation were found [19,20]. In addition, an ED-induced reduction in astroglial density was seen in different brain regions of adult rats [4]. In marmosets, ED has a long-term effect on the hippocampal expression of genes implicated in synaptic function and plasticity but does not affect hippocampal volume [21].
To assess immediate and long-term effects of ED on brain plasticity, we measured neurogenesis, hippocampal volume, longterm synaptic plasticity, stress-related behavior, neuropeptides and their receptors after termination of ED and in early adulthood. We found a clear dissociation between different forms of brain plasticity in response to early life stress.

Results
C57/BL6 mice were individually separated from both their dam and littermates for three hours daily from postnatal day 1 (P1) to P14 and were compared to control animals that had been reared under routine animal facility conditions. Most of the measurements were performed at P15 and P70.

Hippocampal volume
To assess morphological brain plasticity, we measured hippocampal volume longitudinally in vivo using magnetic resonance volumetry in individual animals at three different time points (P15, P30, and P70). We found that, at P15 and P30, hippocampal volumes of early deprived mice were significantly lower than those of control mice; however, at P70, no significant volumetric differences could be seen ( Fig. 1 A and B). To control for an unspecific brain maturation deficit in deprived mice, we measured the cumulative areas of the consecutive whole-brain cross-sections containing the hippocampi and found no significant difference between stressed and non-stressed animals at all three time points (Fig. 1 C).
These results suggests that early deprivation induced specific and reversible morphological alterations in the hippocampus that normalized by adulthood.

Hippocampal neurogenesis
Adult neurogenesis in the hippocampal dentate gyrus was assessed at P70 using immunohistochemistry for doublecortin. This protein is expressed selectively in newborn hippocampal granule cells during the first three weeks after mitosis and can therefore be used as a marker for adult neurogenesis. We found no significant difference in the number of newly generated neurons at P70 between ED and control animals (42.563.1 labeled cells in control animals, n = 11, vs. 48.263.9 cells in ED animals, n = 9, in the 30 mm projection image, n.s.; Fig. 1 D and E). We conclude that ED in early development is not sufficient to impair adult neurogenesis, corresponding with normal hippocampal volumes in adult animals from both experimental groups.

Membrane and EPSP properties
We collected brain slices from deprived and non-deprived animals after the termination of the early stress protocol on day P15 and in adulthood (P70) and performed whole-cell patch-clamp measurements of hippocampal CA1 pyramidal cells. We first determined basic intrinsic membrane and action potential (AP) properties (Fig. 2, Table 1). The membrane input resistance (R M ) was calculated from the current response to a hyperpolarizing voltage pulse (25 mV from a holding potential of 270 mV for 50 ms). There was no significant difference between R M values in deprived or non-deprived mice at any time point (Fig. 2 A). Action potentials in CA1 neurons were evoked by a current injection of 300-800 pA for 1 s. Neither the AP half-width nor the frequency of the AP burst were different between the groups; however, there was a small but significant difference in the AP amplitude between the two experimental groups at P70 which could not be detected at P15 (Fig. 2 B, C). Excitatory postsynaptic potentials (EPSPs) were evoked by extracellular stimulation of the Schaffer collateral pathway. Neither EPSP rise time nor half width were significantly altered ( Fig. 2 D). These findings argue against a functionally significant modulation of basic membrane conductances by the early deprivation protocol.

Long-term synaptic plasticity
Our results so far suggested that ED impaired hippocampal volume in infant and juvenile mice but that these alterations recover by adulthood. We proceeded to examine putative effects of ED on long-term synaptic plasticity.
LTP was evoked using an associative, spike time-dependent, theta burst stimulation (TBS) protocol using a total of 125 EPSPaction potential (AP) pairings [22]. Five EPSPs were paired with five postsynaptic action potentials at a frequency of 50 Hz (delay between onset of EPSP and AP onset, 5 ms). Five of these pairing sequences were repeated at 5 Hz and five of the resulting thetaburst blocks were repeated at a frequency of 0.1 Hz (Fig. 3 A, B). Baseline EPSP amplitudes were not significantly different between the groups. The effect of the stimulation protocol within one experimental group was assessed by comparing baseline and poststimulation EPSP amplitudes by a Wilcoxon signed rank test. Differences in the amount of LTP between the control and ED groups were compared by a Mann-Whitney test at P15 and P70.
In non-stressed animals at P15, the stimulation protocol resulted in a stable potentiation of synaptic transmission to 280.2622.5% of baseline EPSP amplitude (n = 18, p,0.001 vs. baseline; Fig. 3 C). When the same protocol was applied to animals after the termination of ED at P15, the amount of potentiation was similar (273.4629.6% of baseline EPSP amplitude, n = 12, p,0.001; Fig. 3 D). At P15, the amount of LTP was not significantly different between the control and the ED group.
When LTP was induced in non-stressed control animals at P70, an identical TBS protocol caused a potentiation to 216.4%615.5% of baseline EPSP amplitude (n = 9, p,0.001 va. baseline; Fig. 3 E). However, in animals that had been subjected to ED in their postnatal period, tetanic stimulation resulted in a reduced amount of potentiation (161.9%615.8% of baseline EPSP amplitude, n = 11, p,0.001 vs. baseline; Fig. 3 F). At P70, the amount of LTP was significantly lower in the ED group (p,0.05). Taken together, these results show that ED had no acute effect on LTP immediately after the end of the stress protocol. However, we noted a delayed impairment of long-term synaptic potentiation in adult animals after ED.

Forced swimming test
In the forced swimming test (FST), rodents are placed in a small water basin; immobility time (floating) is compared to time spent with active escape movements (struggling). The FST tests the behavioral consequences of a short episode of stress. Many antidepressants selectively reduce floating time in the FST.
The FST was performed in animals at P70. Previous exposure to ED significantly increased floating time (165.0612.3 sec in ED mice vs. 114.9617.0 sec in control animals, n = 9 each, p,0.05; Fig. 4). These findings suggest that postnatal exposure to ED has the opposite effect of antidepressant in the FST in adult mice.

Neuropeptides and their receptors
It is unclear how early life stress causes prolonged or delayed plastic alterations in adult life. Substance P and neurokinin B are two neuropeptides that are increased in different brain regions after exposure of animals to stress [23]. We quantified the cerebral concentrations of substance P and neurokinin B and their respective receptors NK-1 and NK-3 (Table 2). We found that, at P15, early deprivation was associated with significantly increased synthesis of substance P and neurokinin B in the frontal cortex, as measured by enzyme immunoassay. Substance P remained significantly increased until P70 (Fig. 5 A and B). At the same time, the synthesis of NK-1 and NK-3 receptors was downregulated as measured by mRNA real-time PCR; however, the decreased receptor mRNA concentration did not reach significance ( Fig. 5 C and D). These results confirm that ED acts as an effective stressor in newborn mice and activates stress related pathways. Moreover, permanently increased substance P levels  Table 1. Membrane Properties, Action Potential and Synaptic Function in the Hippocampal CA 1 Region of Juvenile and Adult Mice with and without Early Deprivation. might suggest an involvement of tachykinine signaling in adult life alterations in response to early stress.

Discussion
In sum, we found that ED was associated with reduced hippocampal volume in infant and juvenile mice; however, these alterations were corrected by adulthood. Neurogenesis in the dentate gyrus of adult mice after ED was unchanged. Early life stress caused a delayed impairment of LTP and increased immobility in the FST. Moreover, ED was associated with a sustained increased in the synthesis of substance P (Table 3). Taken together, these results suggest a dissociation of the effects of early life stress on different forms of brain plasticity. Whereas ED impairs hippocampal volume as an important measure of morphological plasticity in infant and juvenile mice, it causes delayed impairment of synaptic plasticity in adult animals.  . Increased immobility in the forced swimming test after early deprivation. Mice were placed in a water basin at P70 and behavior was videotaped. Immobility time represents the cumulative time spend with passive floating from minute 2 to 6 after the start of the experiment. Immobility time was significantly increased in adult mice which had been deprived after birth (n = 9 each). doi:10.1371/journal.pone.0046004.g004 Reversible impairments of morphological brain plasticity after early life stress A volumetric approach using MR imaging has been applied in a number of recent animal studies on rodents [24,25,26,27], promoted by the increasing availability of dedicated animal imaging systems. Non-invasive in vivo measurements are especially useful because longitudinal studies offer more reliable measurements when they are repeated on the same animals over a longer period of time [28].
Despite its obvious advantages, volumetric measurements might not be able to detect small effects on hippocampal volume; moreover, it cannot be ruled out by volumetric assessments that early life stress is associated with long-lasting changes in other morphological measures like dendritic arborization, spine density, and spine morphology. Given the modulation of long-term synaptic plasticity by early life stress shown here, it is even highly likely that associated structural changes occur as LTP has been demonstrated to modify the morphology of dendritic spines [29,30].
Hippocampal volumes have not been studied in rodents following ED. Kalish et al. [25] reported reduced hippocampal volume in a rat model of trait anxiety. However, Schubert et al. did not find alterations in hippocampal volume in rats following seven weeks of social isolation after weaning [27].
In humans, MR volumetry has revealed a number of structural abnormalities in a variety of psychiatric disorders, including posttraumatic stress disorder [9], borderline personality disorder [31] and first-episode schizophrenia [32]; however, many of these results have been difficult to replicate. In major depression, the reductions in hippocampal volume were small and heterogeneous but overall significant in a meta-analysis. They correlated with chronicity of depression (i.e., total lifetime number of affective episodes) but not with mood state [33].
The normalization of hippocampal volumes after ED corresponds to our finding of unimpaired neurogenesis in adult mice. Similar results have been reported for maternal separation in rats in which normal numbers of BrdU-labeled cells were observed when measured 2-3 weeks after BrdU injection [34,35]. Adult  hippocampal neurogenesis has been implicated in the pathophysiology of depression and the behavioral effects of antidepressants [36]. Our findings argue against persistent or delayed alterations of hippocampal volume or adult neurogenesis after early life stress.

Modulation of synaptic plasticity by acute, chronic and early life stress
The modulation of long-term synaptic plasticity by stress has been demonstrated in numerous behavioral models. Acute stress protocols have been found to facilitate LTD and to impair the induction of LTP [13,37,38]. These effects were generally dependent on the activation of glucocorticoid receptors [39]. Chronic stress produces comparable effects on synaptic plasticity [12]. In previous work, we have shown that chronic mild stress in young adult rats upregulated subsequent LTD induction, an effect that could be prevented by treatment with an antidepressant [3]. The effects of chronic stress have typically been independent of increased basal glucocorticoid levels. Accordingly, in parallel experiments using the identical ED procedure, the glucocorticoid receptor mRNA levels were not altered at P15 [40]. Acutely increased glucocorticoid levels might cause secondary effects and affections of the HPA axis might be more readily detectable by functional testing [41].
In our experiments, we show long-term, remote effects of early life stress on synaptic plasticity in adult animals. Correspondingly, Brunson et al. found impaired LTP in middle aged rats, but not in young rats, after fragmented maternal care during the early postnatal period. This impaired LTP was accompanied by a deterioration of complex cognitive behavior [42]. A similar earlylife stress protocol led to decreased LTP at P53-57 and deficits in spatial learning. This effect could be overcome by exposure to an enriched environment, which was applied following the stress procedure [43]. Another group described a delayed impairment of the reinforcement of LTP after postnatal maternal separation [44]. Moreover, maternal separation had a delayed effect on hippocampal synaptic development as assessed by synaptophysin immunoreactivity [45]. A putative modulation of LTD by early deprivation is unknown and has not been the focus of this study.

Mechanisms of modulation of synaptic plasticity by early life stress
The cellular mechanisms underlying delayed modulation of synaptic plasticity by early life stress are unclear. Our data argue against a modulation of basic membrane conductances or excitability by the early stress protocol. More relevant mechanisms might include alterations in gene expression or second messenger cascades probably caused by initial glucocorticoid signals. An important candidate that may explain the mediation of hippocampal plasticity by early life stress is CRH [46,47,48]. This hypothesis is supported by the finding that the effects of early life stress on cognitive function in middle-aged rats could be reproduced by administration of CRH into the immature rat brain [49]. Alternatively, a reduction of astrocytes might impair later LTP induction; this phenomenon has been observed after early deprivation in Fischer rats [4,50].
In this study, we found a prolonged increase of frontal Substance P concentrations and a transient increase of NK B concentrations after ED. NK 1 and NK 3 receptor mRNA levels were downregulated accordingly, but these reductions did not reach significance which is most probably due to the underpowered sample size in this set of experiments. Substance P has been shown to be involved in regulation of brain plasticity. In NK1 receptor knockout mice, neurogenesis in the dentate gyrus and BDNF secretion were increased, while immobility time in the FST was reduced [51]. This finding has also been demonstrated in TAC1 mutant mice, which lack substance P, and in wild-type mice after application of a NK1 receptor antagonist [52,53]. An NK1 receptor antagonist could prevent stress-induced reductions in neurogenesis and hippocampal volume [7]. Mechanisms by which substance P could impair LTP might be in an increase in GABAergic inhibition or intracellular IP 3 levels [54,55,56]; but see [57].
Different forms of stress have been shown to affect substance P secretion [23]. Substance P levels were increased after chronic mild stress, an animal model of depression [58]. Acute or chronic stress protocols caused a downregulation of NK1 receptors and increased substance P release in the amygdala [59,60].
In humans, intravenous infusion of substance P lowered mood and memory, whereas increased substance P levels could be measured in major depression and post-traumatic stress disorder. NK1 receptor antagonists have been evaluated as antidepressants [23,61].
Taken together, this evidence suggests that substance P may be another attractive candidate that could explain the modulation of brain plasticity by early stress. However, rather than a simple correlation between substance P levels and plasticity measures, plasticity and behavior might be modulated by a complex interaction between neurokinins and such interrelated signaling systems as BDNF, CRH and monoamines. A putative etiological role for substance P after early stress can only be assessed by interventional studies using NK1 antagonists or with knockout mice, which will be the focus of future studies.

Behavioral impairment in adulthood after early life stress
What are the functional and behavioral consequences of impaired synaptic plasticity after early life stress?
Synaptic long-term plasticity is believed to be the molecular basis of learning and memory. Alterations of synaptic plasticity might explain learning defects in animal models of early stress [42,62]. Similar cognitive dysfunctions have been described after early stress in primates and humans [63,64].
Moreover, it has been proposed that synaptic plasticity is involved in emotional regulation and that synaptic plasticity or its modulation might be disturbed in depressed patients [65,66]. We have observed facilitation of LTD after chronic mild stress, a valid animal model of depression [3]. In adult rodents, the antidepressant fluoxetine facilitated ocular dominance plasticity, a form of plasticity in the visual cortex [67]. In addition, different antidepressants, the mood stabilizer lithium and electroconvulsive therapy have been shown to effectively modulate synaptic plasticity in different brain regions including the hippocampus [68,69,70,71]. In depressed humans, plastic modulation of visually evoked potentials was altered, while chronic application of an antidepressant in healthy humans greatly enhanced plasticity [72]. Different learning paradigms which were used as behavioral correlates of plasticity in humans were altered in major depression [73]. In this set of experiments, we show that an impairment of LTP due to early life stress occurs alongside an increased floating time in the FST [74,75]. The FST is widely used as a screening test for antidepressive properties of pharmacological substances but cannot be regarded as an animal model of depression [76]. Recently, efforts have been made to separate its stress-inducing properties (forced swimming) from its potential utility as a read-out of the behavioral state of the animal (immobility time) [77]. In any case, under our conditions ED had an opposite effect compared to antidepressant in adult animals. However, further research is necessary to clarify how impairments of synaptic plasticity and increased immobility on the FST are related and how they relate to the risk for depression.
Previous findings on affective and fear-related behaviors after early trauma in mice are contradictory: increased immobility time in the FST has been observed after maternal separation, which could be prevented by treatment with an antidepressant or with a glucocorticoid receptor antagonist [78,79]. Anxiety and fearrelated behaviors were differentially modulated in adult mice after maternal separation depending on sex and estrous cycle of the animal [80]. However, Millstein & Holmes [81] did not observe any consistent effect of an ED procedure on different anxiety-and depression-related phenotype in several mouse strains. As we left the pups at room temperature without additional warming during the ED procedure, this additional stressor might explain the positive findings under our conditions.

Pathophysiological implications
In humans early life trauma is one of the most important risk factors for psychiatric disorders, most notably depression. Different early life traumatic events have been shown to increase the risk for depression in adults. These traumatic events include birth complications, physical, sexual, and emotional abuse and severe somatic disease [82]. In post-traumatic stress disorder, severe acute or chronic traumatic events cause deficits in memory functions [9] but also predispose patients to depression. An impairment of synaptic plasticity in adulthood caused by early life stress might predispose susceptible individuals to exhibit increased sensitivity of brain plasticity to stress, which might impair the function of neuronal circuits involved in mood regulation, causing the phenotype of affective disorders.

Materials and Methods
Animals C57/BL6 mice were kept with ad libitum access to food and water. Animals were maintained in a temperature-controlled room with a light/dark cycle of 12:12 h (lights on at 07:00 h). Experiments were performed during the light period of the cycle at 21uC and were conducted in accordance with the principles and procedures of local authorities for the Care and Use of Laboratory Animals (Regierungsprä sidium Freiburg, permit numbers G-04/ 65, G-05/48, G-08/35, T-06/01).

Early deprivation stress
Each litter of newborn mouse pups was randomly assigned to an experimental group at P1. We compared two experimental groups: control (standard animal facility rearing) and ED, where all pups of a litter were simultaneously separated from their dam. All pups were individually isolated in a small plastic container lined with bedding (10615610 cm) for three hours daily in the evening in a separate room within the same animal facility. ED was performed at room temperature. After separation, the pups were returned to their home cages, where they were reunited with the dam. ED was performed from P1 to P14. All animals were weaned at P21. Not more than three animals were used from a single litter for each experimental protocol. With the exception of the longitudinal MR measurements, no individual animal was used for more than one experiment. Both sexes were used. Two time points were chosen for the described analyses: P15 (immediate effect of ED) and P70 (long-term effect of ED). For MRI studies, one additional time point was chosen: P30 (mediumterm effect of ED).

MR volumetry
MRI measurements of hippocampal volume were obtained invivo using a 9.4 Tesla animal scanner (BioSpec 94/20, Bruker BioSpin) in anesthetized mice under spontaneous breathing conditions using isoflurane. The respiration rate was continuously monitored and gating was used to reduce movement artifacts during the scan. The MRI protocol consisted of a localizer sequence to plan the brain imaging; a T2-weighted RARE Sequence (TEeff 60 ms, slice thickness: 0.4 mm, MTX 1966196, FOV 18618 mm 2 ) was used to depict the hippocampus in mouse brain images. Total hippocampal volumes were calculated from sets of contiguous images by summing products of area measurements and slice thickness using MIPAV, a freely available medical image processing software package from the National Institutes of Health (Bethesda, USA). Additionally, the areas of 8 cross-sections from contiguous sets of images of the complete brain of each mouse were measured; starting with the slice where the frontal end of the hippocampus was detected. Data were analyzed with two-way ANOVA using 'age' and 'treatment' as main factors. Next, one-way ANOVA with 'group' as the factor was carried out, followed by the Holm-Sidak post hoc test (Sigma Stat 3.5).

Electrophysiology
Synaptic plasticity was measured at P15 and P70 using wholecell patch clamp experiments on hippocampal brain slices. Mice were anesthetized with isoflurane and were killed by decapitation, in accordance with governmental and institutional guidelines. Transverse brain slices (300 mm) were cut from the hippocampus with a vibratome (DTK-1000, Dosaka). Slices were incubated at 35uC for 20 min and were then stored at room temperature in a saline solution containing (in mM): 125 NaCl, 25 glucose, 25 NaHCO 3 , 1.25 NaH 2 PO 4 , 2.5 KCl, 2 CaCl 2 , 1 MgCl 2 (equilibrated with 95%O 2 /5% CO 2 ). In the recording chamber, slices were superfused at a flow rate of 5 ml/min with saline containing 20 mM picrotoxin to pharmacologically isolate excitatory transmission. Infrared differential contrast video microscopy was used (Axioskop 2 FS plus, Zeiss, IMAGO-VGA, Till Photonics). CA1 pyramidal neurons were identified visually and by their characteristic firing pattern. Patch pipettes were pulled from borosilicate glass tubing (3-6 MV, Hilgenberg). The recording pipette contained a filtered solution with (in mM): 135 K-gluconate, 20 KCl, 2 MgCl 2 , 4 sodium adenosine triphosphate, 0.3 sodium guanosine 59-triphosphate, 10 HEPES, and 0.1 EGTA, equilibrated with KOH to pH 7.3. Whole cell recordings were made with an EPC-9 amplifier (HEKA). Signals were filtered at 5 and 10 kHz and were stored offline for later analysis using Pulse and Pulse Fit software (HEKA). All chemicals were obtained from Sigma-Aldrich, Merck or Tocris.
Schaffer collateral fibres were stimulated with a stimulus isolator (Stimulator 2100, A-M Systems) via a patch pipette that was filled with a HEPES-buffered NaCl solution (1-2 MV). The pipette was placed in stratum radiatum of the CA1 region, 20-50 mm away from the pyramidal cell layer. 100 ms voltage pulses of 10-100 V were applied at a frequency of 0.1 Hz to evoke an initial subthreshold EPSP amplitude of 3-6 mV. Before initiation of the induction protocol, a stable baseline was recorded for 10 min.
The resting membrane potential of most cells was between 268 and 272 mV; holding potential was set to 270 mV. Input and series resistances were continuously monitored and recorded over the course of the experiment by applying a 500 ms current pulse after every 10 th EPSP. The criteria for discarding an experiment were as follows: unstable baseline (.20% difference of EPSP amplitudes between 210 min and start of the stimulation), unstable EPSP amplitude after stimulation (.20% difference between 10 and 20 min after termination of the stimulation protocol), occurrence of sudden EPSP amplitude steps or distinct amplitude oscillations, change of .25% in series resistance during the course of the experiment, and evidence of ictal discharge or spiking. In addition, only experiments in which the LTP stimulation protocol caused a significant increase in EPSP amplitude were included into the final analysis (as calculated with a Wilcoxon signed rank test between the last 60 EPSP amplitudes before the stimulation and the 60 EPSP amplitudes between 20 and 30 min after stimulation). The number of excluded recordings was comparable between the different experimental series. Experiments examining the different groups were interleaved.
To quantify synaptic plasticity, 60 consecutive EPSPs were averaged 10 min before the initiation and 20-30 min after the termination of the induction protocol. Changes of the mean EPSP amplitude pre-and post-induction within an experimental series were assessed by a Wilcoxon signed rank test. Normalized plasticity measurements were compared between separate series using a Mann-Whitney test. GraphPad Prism software was used to average and plot EPSP amplitudes.

Immunohistochemistry and imaging of neurogenesis
For doublecortin immunohistochemistry, 350 mm thick slices (one slice per animal) were used, with the same preparation technique as that of brain slices for patch clamp recording. Slices were fixed in 4% paraformaldehyde and were washed. After incubation overnight at room temperature with the primary antibody goat-anti-doublecortin (C-18, immunoglobulin G, 1:100; together with 5% donkey serum and 0.3% Triton X-100, Santa Cruz Biotechnology), slices were washed three times for 30 min with 0.1 mol/l phosphate buffered saline. The secondary antibody donkey-anti-goat-Alexa 568 (1:200, Invitrogen) was co-applied with 0.3% Triton X-100 for 24 hours at 4uC. After rinsing, slices were embedded in ProLong Antifade (Invitrogen). Immunofluorescence was measured with a confocal laser-scanning microscope (LSM 510, Zeiss). A projection image was constructed from confocal stacks with a total thickness of 30 mm. Stained cells were manually counted in the granule cell layer and in the subgranular zone of the dentate gyrus by two blinded investigators. Values were compared with the Mann-Whitney test (GraphPad Prism).

Forced swimming test
In a calm room, mice were gently placed for 6 min in a cylindrical glass container (volume 3 l, diameter 19 cm) that was filled to a depth of 15 cm with water at a temperature of 25uC. Water was changed after every experiment. Behavior was videotaped and analyzed offline from 2 to 6 min after the start of the test by two independent investigators who were blinded to the experimental group of the animals. Floating was defined as either being stationary, with only enough movements of the tail or the forepaws to keep the head above water, or as calm swimming. Calm swimming did not involve lifting the paws above the water surface, with the body oriented parallel to the sides of the cylinder. Struggling was defined as active pawing of the side of the cylinder, reminiscent of escape movements, with the head of the animal oriented toward the wall and with the body oriented perpendicularly to the side of the cylinder [74,75,83]. Data were compared using Mann-Whitney test. As the FST is neither meaningful nor defined for very immature animals, these experiments could only be performed at P15.

RNA isolation and quantitative real-time PCR
RNA isolation and quantitative real-time PCR was used to quantify the expression of NK1-and NK3-receptor mRNA. Total RNA was extracted from mouse brain (frontal cortex) using the guanidine isothiocyanate method. The quantity of total RNA was determined by spectrophotometric measurement at 260 nm. Total RNA (1 mg) was subjected to reverse transcription using 1 mg of random hexamer oligonucleotides in a 15 ml reaction volume. (sense) and 59-ATCTGCTGGAAGGTGGACAG-39 (antisense) (expected size: 362 bp). ß-actin was used as the endogenous control for normalization of initial RNA levels. Control samples with no added cDNA were also included with each run to detect any possible contamination. Thermal cycling conditions were designed as follows: initial denaturation at 95uC for 600 s followed by 40 cycles of 95uC for 10 s, 60uC for 15 s, and 72uC for 10 s. Fluorescence was monitored during each annealing step. At the end of each PCR run, the data were automatically analyzed by the system and an amplification plot was generated for each cDNA sample.

Enzyme Immunoassay
Enzyme Immunoassay (EIA) was used to quantify the concentration of substance P and neurokinin B in frontal cortex. Peptides/proteins were isolated as described above. Neurokinin B levels were determined using the Neurokinin B EIA kit provided by Phoenix Pharmaceuticals (Burlingame, CA, USA). Substance P levels were measured using the Substance P EIA kit provided by Assay Designs (Ann Arbor, Michigan, USA). Protein levels were related to the total protein content of the lysate determined by a BCA Protein Assay Kit (Thermo Fisher Scientific, Rockford, USA). Molecular biology data were compared using a Mann-Whitney test.