Mesolimbic neuropeptide W coordinates stress responses under novel environments

Significance

The study proposes a physiological role for neuropeptide W (NPW) in modulating mouse behaviors under stress. We found that NPW-producing neurons (which are a small subset of mesolimbic dopaminergic neurons) exclusively innervate the extended central amygdala, where the peptide plays an essential role in stress-induced inhibition of the amygdala neurons. The response of NPW-null mice to either formalin-induced pain stimuli or to a live rat (a potential predator) was abnormal when they were placed in a novel environment: They failed to show the normal self-protective and aversive reactions. In contrast, the behavior of NPW-null mice in a habituated environment was normal. These results demonstrate a critical role of NPW in the gating of stressful stimuli during exposure to novel environments.

Abstract

Neuropeptide B (NPB) and neuropeptide W (NPW) are endogenous neuropeptide ligands for the G protein-coupled receptors NPBWR1 and NPBWR2. Here we report that the majority of NPW neurons in the mesolimbic region possess tyrosine hydroxylase immunoreactivity, indicating that a small subset of dopaminergic neurons coexpress NPW. These NPW-containing neurons densely and exclusively innervate two limbic system nuclei in adult mouse brain: the lateral bed nucleus of the stria terminalis and the lateral part of the central amygdala nucleus (CeAL). In the CeAL of wild-type mice, restraint stress resulted in an inhibition of cellular activity, but this stress-induced inhibition was attenuated in the CeAL neurons of NPW^−/− mice. Moreover, the response of NPW^−/− mice to either formalin-induced pain stimuli or a live rat (i.e., a potential predator) was abnormal only when they were placed in a novel environment: The mice failed to show the normal species-specific self-protective and aversive reactions. In contrast, the behavior of NPW^−/− mice in a habituated environment was indistinguishable from that of wild-type mice. These results indicate that the NPW/NPBWR1 system could play a critical role in the gating of stressful stimuli during exposure to novel environments.

Limbic structures, including the amygdala and bed nucleus of the stria terminalis (BST), where different modalities of environmental stimuli can converge and interact, are critical for the response to stress (1). The basolateral amygdala (BLA) receives sensory information from the thalamus and cortex and transfers that information to the central amygdala (CeA), an output center for the amygdaloid complex. Subsequently, the CeA relays this information through axonal projections to nuclei in diverse areas, such as the hypothalamus, brainstem, and pons (2, 3). These circuits are essential for mediating fear and anxiety, especially for fear conditioning using auditory or visual conditioned stimuli (4, 5). Indeed, in a recent report, optogenetic manipulation of the BLA–CeA pathway was found to modulate the expression of anxiety in the mouse (6).

The lateral BST (BSTL) and CeA (CeAL) are two major components of the central extended amygdala, a continuum of telencephalic structures of the forebrain (7, 8). Both the BSTL and CeAL contain numerous GABAergic as well as peptidergic neurons, such as those producing enkephalin, corticotropin-releasing factor, and somatostatin (9–13); these GABAergic neurons form dense local inhibitory circuits within these nuclei. The GABAergic microcircuit in the CeAL is thought to play pivotal roles in mediating fear and anxiety, as recently dissected at the cellular level by studies that combined optogenetics and electrophysiology (14–17). A subpopulation of GABAergic neurons in the CeAL, which is inhibited by a conditioned stimulus (CS), is differentiated by the expression PKCδ (14). These PKCδ⁺ GABAergic neurons were shown to provide feed-forward disinhibition on output neurons in the medial CeA (CeAM), resulting in a conditioned freezing response following exposure to a CS (14, 15). However, PKCδ⁻ GABAergic neurons, which coexpress somatostatin and make reciprocal inhibitory synapses with PKCδ⁺ GABAergic neurons, do not send any significant intra-amygdaloid inhibitory projections to the CeAM. Instead, they send long-range projections to innervate the periaqueductal gray (PAG) and the paraventricular nucleus of the thalamus, i.e., extra-amygdala areas implicated in defensive behavior, and thus contribute to fear conditioning independent of the CeAM (16, 17).

The G protein-coupled receptors (GPCRs) neuropeptides B and W receptors 1 and 2 (NPBWR1 and NPBWR2) (originally termed “GPR7” and “GPR8”) were identified as closely related orphan GPCRs by a degenerative PCR screen using primers based on sequences from opioid receptors (18). Although humans and other primates express both NPBWR1 and NPBWR2, no functional NPBWR2 ortholog has been demonstrated in rodents (19). A role for NPBWR1 has been characterized by the phenotypical analysis of NPBWR1-null mice, which demonstrate late-onset obesity with a hyperphagic phenotype (20). These mice also show abnormal social interaction and an aberrant autonomic response to physical stress (21). Recently, an SNP of the human NPBWR1 gene has been shown to influence the evaluation of facial expressions (22). We and other groups have isolated neuropeptide B (NPB) and neuropeptide W (NPW) as cognate endogenous ligands for NPBWR1 and NPBWR2 (23–26). NPB and NPW are encoded by two separate genes and together with NPBWR1 (and NPBWR2 in primates) constitute a neuropeptide receptor system (27). NPB consists of 29 amino acids with a unique brominated N-terminal tryptophan moiety (23, 25), whereas NPW, as the mature peptide, exists in two forms: NPW [1–23] and NPW [1–30]. NPW23 and NPW30 bind and activate NPBWR1 and NPBWR2 with similar nanomolar affinities (24, 27).

In a previous report, we described the distribution of NPBWR1, NPB, and NPW mRNA in mouse brain by in situ hybridization (23). In contrast to the relatively widespread distributions of NPBWR1 and NPB mRNA, the neurons expressing NPW mRNA are localized primarily in specific midbrain regions, including the PAG, the ventral tegmental area (VTA), the Edinger–Westphal nucleus, and the dorsal part of the dorsal raphe nucleus (23, 28). This characteristic expression pattern implies a potential role for NPW in the regulation of limbic function. In this regard, we also observed a diffuse expression of NPW mRNA in the CA3 region of adult mouse hippocampus (29). In this study we investigated the potential physiological functions of NPW using histochemical studies and phenotypical analyses of NPW-null mice (Fig. S1), with a particular focus on their behavioral characteristics under environmental stress.

Fig. S1.

Targeted disruption of the murine NPW gene. (A) Strategy for the targeted disruption of the NPW gene. The mouse genomic DNA map, targeting vector, and targeted locus map are shown. Restriction enzymes: B, BamHI; Bs, BssHII; HIII, HindIII; RI, EcoRI; Sm, SmaI; Xb, XbaI. (B) Genomic Southern blot analysis of the targeted ES cell DNA using ³²P-labeled 5′ and 3′ outer probes. Extracted ES cell DNA was digested with EcoRI for the 5′ probe and with BamHI for the 3′ probe. (C) PCR genotyping of mouse tail DNA by using two sets of primer pairs that detect the NPW genomic locus and PGK neo, respectively. (D) In situ hybridization of NPW in the PAG of adult mouse brain using a ³⁵S-labeled probe specific for NPW exon 1, which contains the NPW mature peptide-coding sequence.

Results

NPW Neurons in Midbrain Are Dopaminergic and Project Exclusively to the Central Extended Amygdala.

A systematic immunohistochemical survey of the entire CNS with verified anti-NPW antisera showed localization of immunoreactivity exclusively to the BSTL and CeAL (Fig. 1 A and D). A high-magnification view (Fig. 1 C and F) revealed that specific immunoreactivity surrounded neuronal cell bodies in these two regions, indicating that axonally transported NPW accumulated at the axonal terminals. Consistent with this observation, abundant NPBWR1 mRNA expression has been reported previously in both the BSTL and CeAL (23). No immunoreactive staining was observed in these regions in brain slices from NPW^−/− mice, indicating the fidelity and specificity of the antiserum (Fig. 1 B and E) (see SI Results for the generation of NPW^−/− mice). In summary, these results indicate that NPBWR1-expressing neurons in the mouse BSTL and CeAL are extensively innervated by NPW neurons.

Fig. 1.

Neuroanatomy of NPW neurons and their projection sites. (A–F) Immunostaining of NPW in the BSTL (A–C) and the CeAL (D–F). Arrows indicate NPW immunoreactivity in the BSTL (A) and CeAL (D). (B and E) No immunoreactivity was observed in NPW^−/− mice. (C and F) High-magnification views of A and D, respectively. +/+, wild-type mice; −/−, NPW^−/− mice. (G–J) Distribution of NPW and TH in the midbrain. NPW mRNA in the PAG/VTA (G) and the dorsal part of dorsal raphe nucleus (DRD) (I) and immunostaining of the same regions with anti-TH antibody (H and J). ac, anterior commissure; CPu, caudate putamen; IP, interpeduncular nucleus; opt, optic tract. (Scale bars: 500 µm.)

The dorsocaudal part of the VTA, collectively defined as the A10dc, encompassing the dorsal raphe nucleus and the PAG, contains many NPW neurons (Fig. 1 G and I) as well as dopaminergic neurons (Fig. 1 H and J) (30). Immunohistochemistry using NPW antiserum (Figs. 1D and 2A) and a tyrosine hydroxylase (TH) antibody (Fig. 2B) revealed dense localization of NPW and TH in the medial region of the CeAL. Double immunohistochemistry showed significant colocalization of NPW and TH in axon terminals in the CeAL (Fig. 2 C–E). After colchicine treatment the majority of the NPW-containing cell bodies in the PAG region were found to coexpress TH (Fig. 3 A–C). Essentially the same results were obtained by combining NPW in situ hybridization and TH immunohistochemistry (Fig. 3 D–F). These results demonstrate that the majority of NPW neurons in the midbrain are dopaminergic and that a subset of A10dc dopaminergic neurons expresses NPW.

Fig. 2.

Projection of NPW/TH neurons in the central amygdala. (A and B) Immunostaining of NPW and TH in the CeAL from a preproenkephalin1-EGFP transgenic mouse at low magnification: NPW and TH were stained with Alexa 594 (red). GFP-expressing neurons are enkephalinergic neurons. (C–E) Immunostaining of NPW (green) and TH (red) in the CeAL from a wild-type mouse at high magnification. Yellow arrows indicate the colocalization of NPW and TH in nerve terminals in the CeAL. ic, intercalated cells. (Scale bars: 100 µm in A and B; 50 µm in C–E.)

Fig. 3.

NPW neurons coexpress TH. (A–C) Colocalization of NPW (green) and TH (red) in cell bodies of the PAG. Immunostaining of midbrain sections was performed on colchicine-treated brains. Yellow arrows indicate the colocalization of NPW and TH in the PAG. (D–F) Colocalization of NPW mRNA (maroon) and TH (red) in the PAG. After the colorimetric in situ hybridization of NPW, immunostaining of TH was performed. (Scale bars: 50 µm in A–C; 100 µm in D–F.)

Stress-Induced, NPW-Dependent Modulation of Neuronal Activity in the CeAL.

Using the immediate early gene zif268 as a marker for neuronal activity (31), we noted in wild-type mice that the number of zif268⁺ cells decreases in the CeAL after a 30-min exposure to restraint stress. In contrast, this acute stress-induced decline in the number of zif268⁺ CeAL cells does not occur in NPW^−/− mice (Fig. 4 A and B). This finding suggested that, in the absence of NPW innervation, a subset of CeAL neurons might be disinhibited during exposure to stressful situations. Two populations of CeAL neurons were identified with opposite responses to a CS and shown to make reciprocal inhibitory connections (15). These two populations of neurons are differentiated by the expression of PKCδ. We examined whether such CS-inhibited PKCδ⁺ neurons in the CeAL express NPBWR1. Using NPBWR1 in situ hybridization and PKCδ immunohistochemistry, we noted a clear difference in the localization pattern of these two molecules in the CeAL: NPBWR1 mRNA is expressed preferentially in the medial region of the CeAL, whereas the majority of neurons expressing PKCδ are located in the lateral region of the CeAL (Fig. 4C). From these results, we hypothesized that NPBWR1-expressing neurons do not innervate and disinhibit the output neurons in the CeAM. Rather, these NPBWR1 neurons may modulate CeAL neurons, such as the PKCδ neurons.

Fig. 4.

NPW-dependent suppression of zif268 expression in the CeAL under stress. (A) Immunostaining of the CeAL with the anti-zi268 antibody. (Upper Left) Wild-type mouse, no stress. (Upper Right) Wild-type mouse after 30-min restraint stress. (Lower Left) NPW^−/− mouse, no stress. (Lower Right) NPW^−/− mouse after 30-min restraint stress. Red dotted lines demarcate the CeAL, and the black dotted line delineates the intermediate capsule between the CeAL and BLA. ic, intercalated cells; M, medial part of the central amygdaloid nucleus; opt, optic tract; st, stria terminalis. (B) Quantification of zif268⁺ cells (mean ± SEM) in the CeAL. White bars represent wild-type mice; gray bars represent NPW^−/− mice. **P < 0.05, comparing wild-type mice with and without restraint stress (t test). ***P < 0.05, comparing wild-type and NPW^−/− mice, both after 30-min restraint stress (t test). (C) In situ hybridization of NPBWR1 (Left) and immunohistochemistry using anti-PKCδ antiserum (Right) on adjacent sections from the CeAL of a wild-type mouse. Red lines demarcate the lateral boundary of the CeAL. opt, optic tract; st, stria terminalis. (Scale bars: 100 µm in A; 50 µm in C.)

NPW^−/− Mice Showed an Abnormal Response to Noxious Stimuli When Placed in a Novel Environment.

In a previous study, we found that intracerebroventricular (icv) administration of NPB in the rat induced analgesia in the formalin test (23). Also, as noted above, NPW mRNA is highly expressed in the PAG, an area implicated in the descending analgesic system (23). We therefore submitted NPW^−/− mice to the tail-flick test to investigate whether they exhibited any deficiency in their response to acute pain. Wild-type and NPW^−/− mice were not different in their response in this test (Fig. 5A). We then examined whether NPW^−/− mice demonstrated an altered pain response to formalin injected in a hindpaw. Wild-type and NPW^−/− mice exhibited essentially identical licking behaviors to the formalin injection site after they were returned to the home cage, confirming that NPW^−/− mice have a normal response to pain under baseline conditions (Fig. 5B). In contrast, however, we observed a marked reduction in licking behavior in NPW^−/− mice when they were returned to a novel cage after the injection (Fig. 5C). Wild-type mice exhibited the same licking behavior when returned to either cage. A novel cage environment is known both to be acutely stressful and to stimulate exploratory behavior in mice (32). We thus hypothesized that NPW^−/− mice exhibited reduced pain-induced behavior because of the additional stress and distraction of a novel environment following the formalin injection. Exposure to a new testing cage or laboratory environment has been shown to induce analgesia in rats and mice (33, 34). Here, however, we did not observe any analgesic effect of the novel cage environment in wild-type mice. This absence of an analgesic effect in wild-type mice may reflect the balance between the salience of the pain experienced here versus the novelty of the environment: that is, the pain caused by 20 µL of 5% (vol/vol) formalin in wild-type mice would surpass the novelty effect caused by a different cage. This possibility should be examined in future studies to see if a more balanced condition using a milder pain stimulus combined with a more significant environmental change evokes an analgesic effect in wild-type mice. Most importantly, however, even with such a relatively minor change to environmental conditions in the present study, NPW^−/− mice exhibited a significant reduction in licking after the formalin injection. In view of this result, we thus also considered whether exposure to environmental stimuli might be necessary as an adjunct to reveal further behavioral abnormalities in NPW^−/− mice.

Fig. 5.

Abnormal response of NPW^−/− mice to noxious stimuli in novel environment. (A) The latency in seconds to withdraw the tail in response to a noxious thermal stimulus. White bars represent wild-type mice; gray bars represent NPW^−/− mice. (B and C) Time course of paw licking after the s.c. injection of 20 µL of 5% (vol/vol) formalin into the plantar surface of the right hind paw of each test mouse. Each point represents the amount of time (in seconds) the mouse spent licking the injected paw during a 5-min observation period. (B and C) After the formalin injection, each test mouse was placed into the home cage (B) or into a novel cage (C). **P < 0.05, comparing wild-type (blue) and NPW^−/− (red) mice (t test). Values are displayed as mean ± SEM.

NPW^−/− Mice Showed Abnormal Fear-Oriented Behaviors When Placed in a Novel Environment.

As a way of imposing a compelling stress, we introduced an awake rat, which is a potential predator for a mouse. To verify the efficacy of this stressor, we examined two established stress indicators, c-fos expression in the paraventricular nucleus (PVN) of the hypothalamus and plasma corticosterone levels. As shown in Fig. 6A, a 60-min exposure to a rat induced significant c-fos expression in the PVN region in both wild-type and NPW^−/− mice. Plasma corticosterone levels were also increased to a similar extent in response to exposure to a rat in both genotypes (Fig. 6B). These results showed that the paradigm of rat exposure was effective in inducing the intrinsic stress response independently of NPW. Importantly, these results also demonstrated that NPW^−/− mice had normal awareness of a potentially life-threatening change in their environment. Furthermore, these data showed that NPW^−/− mice had intact pathways for the acute neuroendocrine stress response and produced normal and robust activation of the hypothalamo–pituitary–adrenocortical axis in response to the predator stress.

Fig. 6.

Abnormal predator-fear response of NPW^−/− mice in a novel environment. (A) c-fos immunoreactivity in the PVN induced by the stress of a 60-min exposure to a rat. (Upper) Wild-type mice. (Lower) NPW^−/− mice. (B) Changes in plasma corticosterone level after a 30-min exposure to a rat. **P < 0.05, comparing mice of the same genotype with and without rat stress (t test. (C) Thigmotaxis: the proportion of time spent in the center area of the open field. After a 1-h exposure to a rat in the rat cage, each mouse was tested in the open field, and its activity was recorded for 15 min. **P < 0.05, comparing wild-type mice with and without exposure to rat stress (t test). ***P < 0.05, comparing wild-type and NPW^−/− mice, both after rat stress (t test). (D) The proportion of time spent in the lighted compartment of the light–dark box. After a 1-h exposure to a rat in the rat cage, each mouse was tested in the light–dark box, and its activity was recorded for 12 min. **P < 0.05, comparing mice of the same genotype with and without exposure to rat stress (t test). (E and F) The time that each test mouse spent vigilantly staring at the rat was measured during a 1-h observation period. (E) The test mouse was placed into a large cage where a rat had been singly housed for several days. **P < 0.05, comparing wild-type and NPW^−/− mice (t test). (F) A rat was placed in a large cage where the test mouse had been singly housed for 2–3 d. White bars represent wild-type mice; gray bars represent NPW^−/− mice. Values are displayed as mean ± SEM; n.s., not significant.

We then evaluated the locomotor activity of wild-type and NPW^−/− mice in the open-field test after exposure to a rat. During a 15-min session in the open field, without exposure to a rat, wild-type and NPW^−/− mice showed similar locomotor activity in terms of total distance traveled and center time (Fig. 6C). After a 1-h exposure to a rat, wild-type mice showed a significant decrease in time spent in the center area (Fig. 6C). In contrast, NPW^−/− mice did not show this decrease in time spent in the central area of the open field after rat exposure. Thigmotaxis, or the aversion to open spaces, is a characteristic rodent behavior which is considered a correlate of anxiety (35). The finding that NPW^−/− mice showed less thigmotaxis than wild-type mice suggested that NPW^−/− mice were less anxious after exposure to a rat.

We next assessed this apparent hypoanxiety observed in NPW^−/− mice after rat exposure by using the light–dark box test, in which a less-anxious mouse tends to spend more time in the brightly lit compartment. Under control conditions and after a 1-h exposure to a rat, wild-type and NPW^−/− mice showed similar locomotor activity in terms of total distance traversed. The time spent in the lighted compartment was significantly decreased in wild-type mice after exposure to a rat, as expected. In contrast, NPW^−/− mice exhibited a significantly weakened aversion toward the lit, open compartment (Fig. 6D). These results support the open-field thigmotaxis data and indicate less anxiety in NPW^−/− mice after a 1-h exposure to a rat in a novel environment. However, that NPW^−/− mice were exhibiting less overt anxiety in these tests following exposure to the rat did not necessarily confirm a hypoanxious state. For example, their behavior could reflect increased distraction after rat exposure, as we had postulated as a feasible explanation for the results following the formalin injection.

During the time of exposure to the rat, we frequently observed unusual behaviors in NPW^−/− mice. For a naive wild-type mouse, a rat in the same cage is potentially threatening, and the situation thus is highly stressful for the mouse. Hence, wild-type mice typically demonstrated species-specific, innate fear-oriented behaviors toward the rat, such as marked avoidance, maintaining maximal distance and full vigilance, and orienting themselves toward the rat. In contrast, NPW^−/− mice often appeared indifferent toward the rat: They fell asleep, approached the rat frequently and closely, and in some cases ate and drank in the presence of the rat. These behaviors were quantified by measuring the time that each mouse spent vigilantly directing its gaze toward the rat during the 1-h exposure. As displayed in Fig. 6E, NPW^−/− mice stared vigilantly at the rat significantly less than wild-type mice. Importantly, this difference in behavior was not observed in an alternative paradigm, in which the rat was placed as an intruder into a cage in which a mouse had been singly housed and well acclimated. NPW^−/− mice, like wild-type mice, were highly vigilant toward the intruder rat under these circumstances (Fig. 6F). These data, taken in conjunction with results from the formalin test, indicate that NPW^−/− mice behave inappropriately and do not express normal self-preservation behaviors when they encounter a life-threatening event concurrently with a novel environment. In other words, a novel environment may be overly distracting for these mice and thus may have a major impact on their ongoing behavior. These results support the hypothesis that NPW^−/− mice may have been distracted and thus did not initiate typical self-protective behaviors in the open field and light–dark box tests after exposure to a rat.

SI Results

Generation of NPW^−/− Mice.

We constructed a targeting vector to replace exon 1, containing the NPW mature peptide coding sequence (Fig. S1A), and transfected a 129S6/SvEv mouse ES cell line with the linearized targeting vector. Injection of three correctly recombined ES cell clones into C57BL/6J blastocysts yielded chimeric mice that transmitted the targeted allele through the germ line. In situ hybridization using a riboprobe specific for NPW exon 1 revealed no signal in the homozygote mice (Fig. S1D). NPW^−/− mice were viable and fertile, were identified at the expected frequency at weaning, and showed normal development with no gross anatomical abnormalities.

Analyses of the Metabolic State of NPW^−/− Mice.

No differences were observed in bodyweight between wild-type and NPW^−/− mice of either sex (Fig. S2 A and B). Moreover, wild-type and NPW^−/− male mice showed similar food intake and body composition analysis (Fig. S2 C–E). Analysis of plasma from 5- to 6-mo-old NPW^−/− mice on a regular diet showed normal levels of glucose, free fatty acid, cholesterol, triglyceride, insulin, and leptin (Fig. S2 F–K). On the basis of these findings, we conclude that NPW^−/− mice have a normal metabolic state.

Fig. S2.

Metabolic studies of NPW^−/− mice. (A and B) Bodyweights of male (A) and female (B) mice on a 129S6/SvEv background at the ages of 6 and 12 mo. No significant bodyweight differences between genotypes were detected (one-way ANOVA). (C) Food intake. Food intake (in grams per day) expressed as a proportion of bodyweight. The mean values for seven consecutive days are displayed. (D) Body fat level, as a percentage of bodyweight, measured by NMR. (E) Liver weight, as a percentage of bodyweight, measured immediately after dissection. (F) Blood glucose was measured immediately after cervical dislocation. (G–K) The plasma was frozen, and plasma concentrations of free fatty acid (G), cholesterol (H), triglyceride (I), insulin (J), and leptin (K) were measured immediately after thawing. All values are expressed as mean ± SEM. For all measurements in C–K, 3- to 6-mo-old male mice on a 129S6/SvEv background were used, and no significant differences were detected between wild-type and NPW^−/− mice (t test, P < 0.05). +/+, wild-type mice; −/−, NPW^−/− mice.

Discussion

Contrary to our prediction and to the previously published data obtained from icv injection experiments (36), NPW^−/− mice did not differ from wild-type mice in bodyweight or metabolic state (Fig. S2). We thus hypothesize that it is NPB, and not NPW, that regulates feeding behavior and energy homeostasis in adult mice by acting at NPBWR1. Consistent with this hypothesis, NPB^−/− mice exhibit an obese phenotype (29).

We also have demonstrated that the majority of NPW neurons in the A10dc region coexpress dopamine and that the BSTL and CeAL are the exclusive target regions of NPW neurons in mice. Eliava et al. (37) reported that dopaminergic axons, but not noradrenergic or serotonergic axons, form a dense plexus in the CeAL. A dense dopaminergic innervation of the CeAL as well as the dorsolateral subdivision of the BST has been described also (38). Combined retrograde dye and TH immunostaining have shown that dopaminergic neurons of the A10dc group contain approximately half the total number of retrograde-dye/TH double-positive neurons projecting to the CeA and BSTL (30). Thus, mesolimbic dopaminergic neurons extensively innervate cell bodies in the CeAL and BSTL, and these dopaminergic neurons likely exert a crucial role in these nuclei (39). Together with our findings that the majority of NPW neurons localized in the A10dc region coexpress dopamine and send axons exclusively to the CeAL and BSTL, we now hypothesize that NPW also plays an essential role in modulating limbic system function.

In our previous study, we observed a diffuse expression of NPW mRNA in the CA3 region of adult mouse hippocampus (29). Consistent with this finding, a diffuse expression of NPBWR1 mRNA was detectable in the CA1 region, which is the primary target of the CA3 pyramidal neurons (23). However, we were unable to detect any distinct expression of NPW mRNA in the adult hippocampus using colorimetric in situ hybridization (40). Furthermore, we detected NPW immunoreactivity only in the BSTL and CeAL but not in the target regions of CA3 pyramidal neurons, such as the lateral septal nucleus, CA1, and CA3 regions. Therefore we assume that a subpopulation of CA3 neurons might express NPW mRNA but, if so, at a very low level and that any involvement of the hippocampal NPW/NPBWR1 system in the behaviors examined here would be small.

Burst activity in VTA dopaminergic neurons is elicited by sudden auditory or visual orienting stimuli in the awake cat (41). Restraint stress also increases dopaminergic burst firing in putative VTA neurons in awake rats (42). In these studies, the discharge of dopaminergic neurons in the A10dc region was not specifically investigated, but it is likely that novel sensory stimuli such as those used in these studies would evoke the same pattern of excitation as noted for dopaminergic neurons throughout the VTA. NPW-containing dopaminergic neurons in the A10dc region therefore may also be activated by these stimuli in the same time frame and transduce signals to the CeAL and BSTL. In rat brain particularly high levels of dopamine D2 receptor (D2R) ligand binding were observed in the CeAL (43), and D2R mRNA-positive neurons have been localized to the medial region of the CeAL (37). We have observed the localization of NPW- and TH-immunoreactive fibers, as well as NPBWR1 expression, in the medial region of the CeAL (Figs. 1D and 2 A and B) (21). Although the cellular colocalization of NPBWR1 and D2R remains unconfirmed, considering that both NPBWR1 and D2R are Gi-coupled inhibitory receptors, any activity of NPBWR1⁺/D2R⁺ neurons in the CeAL would be under potent inhibitory regulation by NPW/dopamine neurons. Our histological analysis has shown that the localization patterns of NPBWR1 neurons and PKCδ⁺ neurons in the CeAL are distinct. This result suggests that NPBWR1 neurons are not those that are inhibited by a CS and do not disinhibit the output neurons in the CeAM (14, 15). Instead, NPBWR1 neurons may coexpress somatostatin and send inhibitory axons to PKCδ⁺ neurons in the CeAL and/or send long-range projections to extra-amygdala areas. This interpretation is supported by our previous finding that those CeAL neurons that were hyperpolarized by bath application of NPB or NPW in whole-cell recordings send short axons within the CeAL or relatively long axons outside the CeAL. These axonal projections were shown by neurobiotin injection after recordings were completed (21). Importantly, therefore, a role of NPBWR1 neurons in conveying fear-mediating signals would likely involve such projections.

NPW^−/− mice behaved inappropriately and did not express typical self-preserving behaviors when they abruptly encountered life-threatening events while concurrently placed in a novel environment. Significantly, however, the stress-induced behavior of NPW^−/− mice under habituated conditions was identical to that of wild-type mice. We hypothesize therefore that NPW^−/− mice are easily distracted and are unable to initiate normal species-specific self-protective reactions during and after exposure to novel environmental stimuli. At the cellular level, we have demonstrated that restraint stress inhibited a subset of neurons in the CeAL of wild-type mice. Similar results were previously reported using c-fos expression: Several stressors, including restraint stress, a loud noise, or a novel environment, were found to decrease amphetamine-induced expression of c-fos in the CeAL and BSTL (32, 44). Hence, inhibition of neuronal activity in the CeAL and BSTL appears to be common to many, if not all, stressors. However, this inhibition was not fully exerted in CeAL neurons in NPW^−/− mice, and thus this disinhibition of a group of CeAL neurons may underlie the abnormal behavior of NPW^−/− mice observed here. Further studies to characterize the activity of this specific subset of neurons in the CeAL of NPW^−/− mice in a novel environment may aid in the understanding of the neuronal mechanisms of the stress response and by extension, the pathophysiology of those conditions in which impaired social adaptation is symptomatic.

Our findings have demonstrated that behavioral adaptation in mice to a novel environment and the expression of an appropriate behavioral response are specifically dependent on concurrent stimulus gating. Mesolimbic dopamine projections have long been recognized as playing a role in orienting to, and in gating, significant stimuli (45). Our present results now indicate that a subset of these dopaminergic projections, containing NPW, may play a critical role in the processing of, and adaptation to, novel environmental situations. Elsewhere, we have analyzed and described the behavioral phenotypes of mice with targeted disruption of the NPBWR1 gene (21). NPBWR1-null mice also showed a deficit in social interaction when they encountered a novel conspecific intruder, and they exhibited abnormal autonomic and neuroendocrine responses to physical stress. Taken together, these observations suggest an important role of the amygdala NPW/NPBWR1 system in the adaptation to stress in novel environments.

In conclusion, NPW-containing dopaminergic neurons in the mesolimbic region may be critical for the transduction of stress-related information into the central extended amygdala when novel environmental stimuli are encountered. We have demonstrated that NPW is essential for the appropriate expression of normal species-specific behaviors in response to life-threatening, stressful stimuli under novel environments. Thus, NPW^−/− mice behaved abnormally in response to life-threatening events when the environment was novel and/or distracting. A possible interpretation is that these stimuli are not being gated appropriately in NPW^−/− mice, and stimulus salience is either overwhelming or assigned an incorrect priority, thus causing an inappropriate behavioral response. Determining the functional importance of these findings requires additional studies, but our results have potential implications for psychiatric conditions in which stimulus gating, the orienting response, and/or social adaptation are abnormal.

Materials and Methods

Detailed methods are provided in SI Materials and Methods. All animal procedures were approved by the Institutional Animal Care and Research Advisory Committee of the University of Texas Southwestern Medical Center and were strictly in accordance with National Institutes of Health guidelines. Plasma corticosterone was estimated by the Endocrine Services Laboratory at the Oregon National Primate Research Center (Beaverton, OR) as previously described (46).

SI Materials and Methods

Animals.

Preproenkephalin1-EGFP mice were obtained from GENSAT (MMRRC stock number 000251-MU).

Generation of NPW^−/− Mice.

A mouse (129S6/SvEv) genomic DNA BAC library (Children's Hospital Oakland Research Institute) was screened by using the mouse NPW cDNA fragment (23). Positive genomic DNA fragments were analyzed by restriction-enzyme digestion, Southern blot analysis, and partial sequencing. Construction of the targeting vector (Fig. S1A), electroporation of the targeting vector into ES cells, and selection of the targeted clones were carried out as previously described (47). Briefly, targeting vector construction was based on a universal lacZ-neo-tk (pN-Z-TK2) template plasmid vector that contained nuclear lacZ, neo, and flanking tk cassettes (a generous gift from R. Palmiter, University of Washington, Seattle). NPW exon 1 was replaced with the nlacZ cassette using proximal 8-kb and distal 1.2- kb flanking genomic sequences. The targeting vector was linearized at the unique Xho site. The SM-1 mouse ES cell line (48) was cultured on irradiated Leukemia inhibitory factor-producing STO feeder layers. ES cells were electroporated with the linearized targeting vector and then selected for neomycin as described (48). Neomycin-resistant ES cell colonies were screened by Southern blot analysis using the flanking 5′ and 3′ external probes. Three of these ES cell clones were microinjected into blastocysts to produce germline-transmitting chimeric mice using the previously described methodology (48). PCR genotyping of NPW^−/− progeny used primer pairs for detecting the NPW genomic locus (5′-CTTCTGCTCTTGCTACCGCTG-3′ and 5′-CGGACTCCAGGGTGCATGGAC-3′), with primer pairs encompassing the PGK-neo (5′-GGAACTTCCTGACTAGGGGAG-3′) and the NPW 3′ flanking genomic locus (5′-AGCTATGTGCGCTCTGAAGCTAGA-3′).

In Situ Hybridization.

Sectional in situ hybridization using ³⁵S-labeled or digoxigenin-labeled riboprobes was performed as previously described (49). A 0.45-kb coding region of the mouse NPW cDNA, spanning from the mature peptide to the stop codon, and a 1.0-kb coding region of the mouse NPBWR1 cDNA were subcloned into pBluescript (Stratagene). Antisense riboprobes were generated with T7 phage RNA polymerase using the MAXIscript kit (Ambion) in the presence of ³⁵S-CTP and ³⁵S-UTP or with DIG RNA Labeling Mix (Roche). For dual-label NPW in situ hybridization/TH immunohistochemistry, sections were processed first for in situ hybridization. After the chromogen reaction, sections were rinsed with PBS and processed for TH immunohistochemistry.

Food Intake.

Food intake was measured using singly caged 3- to 6-mo-old male mice (129S6/SvEv background) with an acclimation period of at least 1 wk before measurement. Two food pellets (6.0–6.5 g each) were given and replaced with fresh pellets at the end of the light phase each day for 7 d. Bodyweights were also measured each day for 7 d. The average food consumption over the 7 d was expressed as food consumption⋅bodyweight⁻¹⋅d⁻¹.

Plasma and Body Composition Measurements.

For these studies, male mice (129S6/SvEv background) were used. Blood was drawn from the retro-orbital sinus after cervical dislocation, and the plasma was separated immediately and stored at –70 °C. Blood glucose was measured using the Glucometer Elite (Bayer Corp). Plasma free fatty acids were measured using a NEFA kit (Wako Chemicals USA, Inc.). Plasma insulin and leptin were measured using a rat insulin and leptin RIA kit (Linco Research, Inc.). Cholesterol and triglycerides in plasma were measured as previously described (50). Plasma corticosterone was estimated by the Endocrine Services Laboratory at the Oregon National Primate Research Center as previously described (46). NMR measurement of body fat was performed by using a Lunar PIXImus densitometer (GE Medical Systems Lunar) as previously described (51).

Immunohistochemistry.

Anti-NPW polyclonal antiserum was raised in rabbit by immunization with synthetic [Cys¹]-NPW23, CYKHVASPRYHTVGRASGLLMGL, conjugated to keyhole limpet hemocyanin (Sigma) using m-maleimidobenzoyl-N-hydroxysuccinimide ester (Sigma). Antiserum was affinity purified with a SulfoLink kit (Pierce). Immunostaining was performed as previously described (47). Briefly, brain sections (35 µm in thickness) were incubated in anti-NPW rabbit polyclonal antiserum (1:1,000 dilution), anti-fos rabbit polyclonal antiserum (Ab-5; 1:50,000 dilution; Oncogene Research Products), anti-zif268 rabbit polyclonal antiserum (sc-189; 1:5,000 dilution; Santa Cruz Biotechnology, Inc.), anti-PKCδ goat polyclonal antiserum (sc-937; 1:1,000 dilution; Santa Cruz Biotechnology, Inc.), or anti-TH mouse monoclonal antibody (MAB318; 1:2,500 dilution; Chemicon International) for one to three overnight periods at 4 °C followed by biotinylated anti-rabbit, -goat, or -mouse IgG (1:600 dilution; Vector) for 1 h. For chromogen staining, sections were incubated in HRP streptavidin (Vector) and developed in ImmPACT DAB or with an SG kit (Vector). In the case of fluorescent staining, after the incubation with primary antiserum or antibody, sections were directly incubated with anti-rabbit or anti-mouse IgG conjugated with fluorescent dye Alexa Fluor 488 or 594 (1:200; Invitrogen). In one series of experiments, immunostaining was performed on brain sections from preproenkephalin1-EGFP transgenic mice (GENSAT) (Fig. 2 A and B).

Colchicine Treatment.

Mice were anesthetized with i.p. injections of ketamine (100 mg/kg)/xylazine (10 mg/kg). Colchicine (30 mg/mL in saline) (Sigma) was stereotactically (Kopf) injected unilaterally into the lateral ventricle using the coordinates 0.3 mm caudal to bregma, 0.9 mm lateral to the midline, and 2.0 mm below the brain surface. Manual injection of colchicine (1.5 µL) was performed over a 5-min period, and the injection needle was left in place for at least 5 additional min. Brains were harvested 24–48 h after the injection and processed for immunohistochemistry.

Restraint Stress.

A 3- to 6-mo-old male mouse (C57BL/6J background) was restrained for 30 min in a capped and ventilated plastic 50-mL centrifuge tube, which allowed limited movement. After a 30-min period of restraint, brains were harvested immediately and processed for zif268 immunohistochemistry.

Counting zif268⁺ Cell Numbers.

Zif268 immunopositive cells were counted in the demarcated area in the CeAL (Fig. 4A). The demarcation line was drawn according to the positions of the stria terminalis (st), the intercalated cells (ic), the medial part of the CeA (M), and the BLA. The number of zif268⁺ cells in each mouse was determined as the mean of the numbers of zif268⁺ cells in three consecutive sections of the CeAL.

Tail-Flick Test.

The nociceptive response was evaluated by recording the latency to tail withdrawal in response to a noxious thermal stimulus. Briefly, the tail of the mouse (F2 generation, mixed background) was placed in a groove, which contained a slit under which was located a photoelectric cell (Columbus Instruments). When the heat source was turned on, the tail was exposed to a focused beam of light, and the animal responded by flicking its tail out of the groove. Light then passed through the slit and activated the photocell, and the response latency was measured electronically. Eight response latencies for each mouse were measured in succession, and their mean was taken as the nociceptive response latency. The intensity of the beam was adjusted to produce a mean control reaction time of 3–5 s.

Formalin Test.

The formalin test was performed as previously described (52). A group-housed 3- to 6-mo-old male mouse (F2 generation, mixed background) was separated and placed in a new cage on the day of the test and acclimated to the cage for several hours. The formalin test was performed at the end of the light phase. Twenty microliters of 5% (vol/vol) formalin were injected s.c. in the plantar surface of the right hind paw with a 30-G needle, and the mouse was placed immediately in the acclimated cage or a novel cage. The time spent licking or biting the formalin-injected paw during each 5-min interval was recorded for 50 min.

Rat Stressor.

Sprague–Dawley rats were generally tame and rarely showed aggressive or predatory movements toward the intruder mouse when a mouse was placed into a cage containing a rat. However, a few male rats were highly aggressive toward mice, and we therefore chose only nonpredatory male rats during a preliminary screening for this test. Male 3- to 10-mo-old rats were used for this stress-inducing paradigm, and the stressor rat was housed singly for at least 3 d before the test.

Open Field.

Locomotor activity was measured using a standard open-field apparatus (AccuScan VersaMax System). Each 3- to 6-mo-old male mouse (129S6/SvEv background) was placed in the center of the open field, which was equipped with a 15 × 15 infrared beam array spaced at 2.4-cm intervals. The total distance traveled, horizontal activity, vertical activity, and the time spent in the center area (i.e., center time) were recorded automatically by the associated software for 15 min.

Light–Dark Box.

The light–dark box test was carried out using a standard apparatus (Med-Associates, Inc.): a clear Plexiglas cage (20 × 27 × 27 cm) separated into two compartments of the same size by a partition, which had a small opening (ca. 5.3 × 7 cm) at floor level. The open compartment had transparent walls and was brightly illuminated by the room light; the walls and top of the dark compartment were made from black opaque Plexiglas. A 3- to 6-mo-old male mouse (C57BL/6J background) was individually placed in the center of the open compartment and allowed to explore the apparatus freely for 12 min. The number of light–dark transitions and the time spent in the light compartment were recorded.

Staring Analysis.

A 3- to 6-mo-old male mouse (129S6/SvEv background) was placed into a cage in which a resident male Sprague–Dawley rat had been singly housed for more than 3 d, so that the cage was imbued with the odor of rat. In the opposite paradigm, an intruder stressor rat was placed into a large cage in which a test mouse had been singly housed for 2–3 d. In each paradigm, the behavior of the mouse was video recorded for 1 h. One or two observers, blind as to condition, measured the duration each mouse spent vigilantly staring at the rat during the 1-h observation period.