Interaction of neuropeptidase activities in cortico-limbic regions after acute restraint stress

Brain enkephalin, vasopressin and oxytocin are anxiolytic agents involved in the stress response. Acute restraint stress influences certain neuropeptidase activities, such as some enkephalin-degrading peptidases and vasopressinase/oxytocinase, in the medial prefrontal cortex (mPFC), amygdala (AM) or hippocampus (HC), which are involved in this response. Because these regions form a unified circuit and cooperate in their response to stress, it is important to analyze the profile of the regional distribution of these activities as well as their inter-regional model of interaction in this circuit. Regarding the regional study, although most activities showed a marked predominance of the AM over the HC and mPFC, both in control and stressed animals, enkephalin-degrading activity, assayed as membrane-bound alanyl aminopeptidase activity, showed a change after stress, increasing in the HC and decreasing in the AM. The correlational study in controls indicated essentially a positive interaction between the mPFC and AM. In marked contrast, there was a highly significant change in the functional status of this circuit after stress, showing mainly a positive correlation between the mPFC and HC and between the AM and HC. The existence of correlations does not demonstrate a direct relationship between regions. However, reasons for such strong associations after restraint stress should be examined. The present study may indicate a connection between neuropeptidase activities and their corresponding neuropeptidergic substrates due to significant changes in the functional status of the cortico-limbic circuit after restraint stress.


Introduction
Neuropeptidases are proteolytic enzymes that partially regulate neuropeptides at the synaptic level. Therefore, the analysis of function, and it offers us the possibility of pharmacologically controlling the involved systems using activators or inhibitors of their activities [1]. Enkephalins, vasopressin and oxytocin are anxiolytic neuropeptides involved in the brain response to stress. They are partially regulated by the hydrolytic action of some proteolytic enzymes among which, several APs (enkephalin-degrading APs and oxytocinase/vasopressinase) play a key role [3].
Aminopeptidases have a broad substrate specificity and, unfortunately, a confuse terminology: the same AP has been identified under several names and analyzed with more than one substrate leading to considerable confusion on the subject.
We previously reported the influence of acute restraint stress on the levels of some Sol and MB enkephalin-degrading APs (AlaAP and LeuAP) and vasopressinase/oxytocinase (Placental-LeuAP; P-LeuAP) activities of the medial prefrontal cortex (mPFC), amygdala (AM) and hippocampus (HC) [4]. However, stress not only induces morphologic and functional changes in these individual corticolimbic regions but also in the setup of interactions between them, forming a unified circuit that cooperates in response to stress [5]. Study of the regional distribution and the inter-regional interaction of this circuit should be considered. The activity of this circuit is particularly susceptible to the influence of stress: it modifies its functional interaction and consequently alters the functions the brain regions involved [6]. However, how the network responsible for such interactions is established as well as the neurochemical factors involved in its response to stress are not entirely understood. Therefore, to more appropriately understand the role of these neuropeptidase activities in response to acute restraint stress, it is mandatory to analyze (1) the possible changes in the enzyme activities in the selected areas [4], (2) the possible changes in their regional distribution and finally (3) the interaction between the various cortico-limbic regions. These studies constitute the content of the present report.

Materials and methods
Ten adult male Wistar rats weighing 200-250 g were housed under a controlled temperature (25 • C) with a regular 12-h light:12 h dark cycle (light on from 7:00 h) and received food and water ad libitum. For the restraint stress study, five animals were randomly selected and placed individually inside plastic cylinders (21 cm in length and 6 cm in diameter) for 1 h. Both ends of the cylinders were closed with ventilated sliding doors. Five other rats were maintained in their original home cages as controls. The animals were euthanized immediately after the end of the immobilization period [4].
Rat brains were perfused with saline during light conditions (between 9.00 a.m. and 12.00 noon) through the left cardiac ventricle under equithensin anesthesia (2 mL/kg body weight) (equithensin contained: 42.5 g/L chloral hydrate dissolved in 19.76 mL ethanol, 9.72 g/L Nembutal©, 0.396 L/L propylene glycol and 21.3 g/L magnesium sulfate in distilled water), quickly removed (less than 60 s) and cooled on dry ice. To collect the selected tissue samples, the brains were individually defrosted, and after they reached the appropriate consistency, they were sliced by hand with a blade. The selected areas were identified in the slices, and their coordinates were verified in the stereotaxic atlas of Paxinos and Watson [7]. The medial prefrontal cortex was dissected between 12.70 mm and 11.20 anterior to the interaural line (AIL), the hippocampus was dissected between 7.12 and 5.40 mm AIL, and the amygdala was dissected between 7.12 and 5.40 mm AIL.
The preparation of tissue samples and enzymatic assays was performed as previously described [1]. Briefly, tissue samples were homogenized in 10 volumes of 10 mM Tris-HCl buffer (pH 7.4) and ultracentrifuged at 100,000 × g for 30 min (4 • C) to obtain the Sol fraction. The soluble activities and protein content were assayed in triplicate in the resulting supernatants. The pellets were rehomogenized in Tris-HCl buffer (pH 7.4) plus 1% Triton X-100 to solubilize the membrane proteins. After centrifugation (100,000 × g, 30 min, 4 • C), the MB activity and proteins were measured in triplicate in the supernatants. To ensure complete recovery of activity, the detergent was removed from the medium by adding adsorbent polymeric Biobeads SM-2 (100 mg/mL) and shaking the samples for 2 h at 4 • C. AlaAP, P-LeuAP and LeuAP were measured fluorometrically using the aminoacyl-␤-naphthylamides (aaNNap), AlaNNap and LeuNNap as substrates. Each supernatant (10 L) was incubated for 30 min at 25 • C with 1 mL of the substrate solution (2.14 mg/100 mL AlaNNap or 2.92 mg/100 mL LeuNNap, 10 mg/100 mL bovine serum albumin (BSA), and 10 mg/100 mL dithiothreitol (DTT) in 50 mM of phosphate buffer, pH 7.4). For P-LeuAP determination, 20 mM of l-methionine was added to the corresponding substrate solution [8]. Adding 1 mL of 0.1 mol/L of acetate buffer, pH 4.2, stopped all of the reactions. The amount of ␤-naphthylamine, the product of the enzymatic activity, was measured fluorometrically at wavelengths of 412 and 345 nm for emission and excitation, respectively. The sensitivity of the method allows measurements of APs in the picomolar range. The proteins were quantified in triplicate by the Bradford method [9] using BSA as a standard. The specific Sol and MB AP activities were expressed as pmol of AlaNNap or LeuNNap hydrolyzed per min per mg of protein. Fluorogenic assays were linear with respect to the time of hydrolysis and protein content. The experimental procedures for animal use and care were in accordance with the European Communities Council Directive 86/609/EEC.

Statistical analysis
The differences between groups were evaluated using two-way analysis of variance. Post hoc comparisons were performed with LSD tests. p-values below 0.05 were considered significant. To study the association between APs of the mPFC, HC and AM, Pearson's coefficient of correlation was computed. Computations were performed using SPSS 13.0 and STATA 9.0. p-values below 0.05 were considered significant.

Results
The results are presented in Figs. 1-3 and Table 1. Considering the regional distribution of AP activities ( Fig. 1), Sol AlaAP in control and stressed animals as well as MB AlaAP in control animals exhibited the same regional profile with the highest levels in the AM and the lowest levels in the mPFC, demonstrating highly significant differences (p < 0.001) among the three regions. In contrast, after acute restraint stress, MB AlaAP differed markedly in its regional profile distribution, showing the highest levels in the HC and AM. The mPFC differed significantly from the HC (p < 0.001) and AM (p < 0.01) but no differences were observed between the HC and AM.
Soluble LeuAP in control and stressed animals also showed its highest levels (p < 0.001) in the AM compared to the mPFC and HC, but differed in that the lowest levels were observed in the HC. MB LeuAP, both in control and stressed animals, demonstrated its highest levels (p < 0.001) in the AM and the lowest levels in the mPFC and HC without differences between them.
Soluble P-LeuAP also presented a clear predominance (p < 0.001) in the AM compared to the mPFC and HC, with similar levels among the latter. However, MB P-LeuAP, both in control and stressed animals, demonstrated its highest levels in the AM and mPFC and the lowest levels in the HC with highly significant differences compared to the other regions (p < 0.001) (Fig. 1). . The values represent the mean ± SEM of specific AlaAP, LeuAP, and P-LeuAP activities expressed as pmol of alanyl-or leucyl-␤-naphthylamide hydrolyzed per min per mg of protein. *Represents differences between regions: * p < 0.05, ** p < 0.01, *** p < 0.001; +represents differences between control vs stressed animals: + p < 0.05, ++ p < 0.01 (modified from [4]).
The inter-and intra-regional significant correlations between the different enzymatic activities analyzed in control and stressed animals are presented in Table 1 and illustrated in Figs. 2 and 3.
In controls, the region that demonstrated more intra-regional significant correlations between enzymatic activities was the AM compared to the mPFC and HC, and most of the correlations were positive. When the regions were compared, most of the significant correlations (mostly positive) were observed between the mPFC and AM (red line). Some of these correlations showed highly significant positive values such as Sol AlaAP of the mPFC versus MB Fig. 2. Positive (+) and negative (−) correlations between the medial prefrontal cortex (mPFC, white circles), hippocampus (HC, blue circles) and amygdala (AM, red circles) in control and stressed animals according to the number of inter-regional correlations indicated in Table 1. The line thickness is proportional to the number of correlations. As represented in Table 1, arrows indicating inter regional correlations are highlighted in color: mPFC vs AM (red), mPFC vs HC (blue), and HC vs AM (fuchsia).
AlaAP of the AM (r = +0.974; p < 0.001) or Sol P-LeuAP of the mPFC versus MB AlaAP of the AM (r = +0.960; p < 0.002). However, only one negative correlation was observed between the mPFC and HC (blue line), and no negative correlations were observed between the AM and HC (fuchsia line).
After acute restraint stress, fewer significant intra-regional correlations were observed than in control animals. Considering the inter-regional comparisons, while the number of significant correlations between the mPFC and AM was decreased, a high number of correlations were observed between the HC and AM and between the mPFC and HC. The comparison between Sol AlaAP in the mPFC and Sol P-LeuAP in the HC was particularly significant (r = +0.993; p < 0.0001). As in the controls, most of the significant correlations after stress were positive.
To better perceive the differences between both groups, it is interesting to contrast the significant inter-regional comparisons observed in the control or stressed animals with the same enzymatic comparisons of the contrasting group. Therefore, in Fig. 3, we selected the highest levels of significance of the comparisons between regions (mPFC vs AM, mPFC vs HC and AM vs HC) ( Table 1) and then calculated the level of correlation of the same comparison in the contrasting group. For comparisons between the mPFC and AM, the highest level of significance was found between the mPFC AlaAP Sol and AM AlaAP MB (r = +0.974, p = 0.001) of controls, which clearly contrasts with the same comparison in the stress group (r = +0.612, p = 0.196). For the comparisons between the mPFC and HC, the highest level of significance was found between the mPFC AlaAP Sol and HC P-LeuAP Sol (r = +0.993, p < 0.0001) of stressed
Vasopressinase/oxytocinase (EC 3.4.11.3) activity can be determined as cystinyl aminopeptidase (CysAP) or P-LeuAP, using respectively CysNNap or LeuNNap in the presence of 20 mM of l-methionine, which inhibits LeuAP but not CysAP. This assay is more sensitive than that using CysNNap as substrate [15]. Therefore, the terms P-LeuAP, CysAP, vasopressinase and oxytocinase designate the same enzyme [16]. In addition, the insulin-regulated aminopeptidase (IRAP) is the rat homologue of CysAP and the AT 4 receptor for angiotensin IV has been also identified as the MB form of this enzyme [17].
A previous study reported reductions of Sol AlaAP in the HC, MB AlaAP in the AM and Sol P-LeuAP in the AM and an increase of MB AlaAP in the HC after acute restraint stress [4].
However, the analysis of these modifications did not allow to evaluate possible changes in the regional profile of the corticolimbic areas under restraint stress, and therefore to suggest a possible prevalence of certain regions under stress conditions, as reported in the present study (Fig. 1). In addition, we describe here an inter-regional model of interactions between neuropeptidase activities within this cortico-limbic circuit as a possible image of its functional status as well as its efficient relationship between susceptible neuropeptides in basal conditions and under restraint stress (Figs. 2 and 3and Table 1).
The regional enzymatic study demonstrated that most of the neuropeptidase activities in both control and stressed rats showed a predominance of the AM over the HC and mPFC. This indicates a high level of metabolism of the susceptible substrates at this location, and considering neuropeptidergic functions, suggests the relative importance of the AM in this cortico-limbic circuit. In contrast, enkephalin-degrading activity, assayed as MB AlaAP, showed a significant change after stress with a predominance in both the HC and AM over the mPFC. In addition, vasopressinase/oxytocinase activity, assayed as MB P-LeuAP, also exhibited a change after stress, resulting in this case in a predominance of the mPFC and AM over the HC. Considering the possible roles in which the susceptible neuropeptides are involved, these results may indicate a variation in the functions exerted by this circuit after restraint stress in these regions.
Although the existence of significant correlations does not necessarily imply a relationship of the enzymatic activities between two regions, it does imply some causal connection that is accountable for this association. Therefore, this correlation study may suggest that those neuropeptidase activities, and consequently Table 1 Neuropeptidase correlations between and intra cortico-limbic areas (For interpretation of the color information in this figure legend, the reader is referred to the web version of the article.). Significant correlations between the medial prefrontal cortex (mPFC), amygdala (AM) and hippocampus (HC) with respect to neuropeptidase activities including alanyl aminopeptidase (AlaAP), leucyl aminopeptidase (LeuAP) and placental-leucyl aminopeptidase (P-LeuAP), which were in their soluble (Sol) and membrane-bound (MB) forms in control and stressed conditions. Pearson's correlation coefficient (r) and p-values are indicated. Inverse correlations are indicated in italics. Inter-regional correlations are highlighted in color: mPFC vs AM (red), mPFC vs HC (blue), and HC vs AM (fuchsia). Intra-regional correlations are not in color.

NEUROPEPTIDASE CORRELATIONS BETWEEN AND
the functions in which the susceptible neuropeptides are involved, are related to the functional status of this cortico-limbic circuit, which changes significantly after restraint stress. In controls, this connection implied essentially a positive interaction between the mPFC and AM. However, after acute restraint stress, there was a highly significant change in the pattern of interactions within this circuit; a positive correlation emerged between the mPFC and HC and between the AM and HC, whereas the interaction of the mPFC and AM observed in controls was diminished. Such interactions may be established between these regions by positive feedback processes involving paracrine mechanisms and/or by bidirectional axonal transport of neuropeptidases [18]. Neuropeptides such as enkephalins or oxytocin and vasopressin coexist with classical neurotransmitters in these cortico-limbic regions [19]. The presence of neuropeptides in these regions may be due to their synthesis "in situ" or through distant dendritic release into the extracellular space, resulting not only in local action but also in diffusion through the brain to reach distant targets. Furthermore, smaller parvocellular neurons in the paraventricular nucleus also produce oxytocin and vasopressin and project directly to other regions in the brain such as the amygdala or hippocampus [20].
Understanding the functional significance of these results remains difficult. However, because the proteolytic activities analyzed in this study reflect the functional status of their endogenous substrates, these results reveal the possible activation of a neuropeptidergic functional connection between the mPFC, HC and AM after stress, which could be involved in some of the functions performed by this circuit.
Acute restraint is an unavoidable stress situation that causes important autonomic and behavioral changes in rats [21]. A possible modification of the interactions between relevant brain areas has been suggested after exposure to a stressful experience [22]. The HC, AM and mPFC could be important regions that mediate some aspects of the response to stress, especially contributing to the interplay between emotions and memory formation [22]. The HC appears to interact with the AM bidirectionally through direct and indirect connections during the encoding of emotional memories after a stressful event [23]. It has been suggested that the activation of these interactions might enhance memory consolidation [24]. Furthermore, the mPFC is a key component of the neural circuit that mediates responses to stressful situations. It is known that the mPFC modulates neuroendocrine function during stress [25] through its connection to the AM [26] and HC [27]. However, although substantial evidence suggests that acute stress alters different forms of cognitive function by modifying specific brain circuitry involving the prefrontal and limbic regions [28], the exact mechanism by which these regions interact remains to be elucidated, and the present results may be relevant to more accurately understanding this response. Our data suggest a role for neuropeptidases and their corresponding neuropeptidergic substrates such as enkephalins, vasopressin or oxytocin in the response of the cortico-limbic circuitry to stressful conditions that may influence cognitive processes such as memory. Vasopressin [29], oxytocin [30], oxytocinase/vasopressinase activity [31], enkephalin [32] and enkephalin-degrading activity [33] are related to memory processes, and therefore, neuropeptidases and their substrates may constitute a target to develop new therapeutic strategies for the treatment of stress consequences.
In conclusion, considering neuropeptidase activities, acute restraint stress induces a highly significant change in the functional status of the circuit formed by the mPFC, AM and HC. This change consists of a reinforcement of the interactions between the regions studied, but more importantly, it selectively activates a positive correlation between the mPFC and HC and between the AM and HC.
The activation of these interactions involving neuropeptidase activities might be related to changes in cognitive processes such as an enhancement of memory consolidation after acute restraint stress. However, to properly understand the functional significance of these results, further research linked to behavioral studies is required. The use of specific inhibitors or activators of those neuropeptidases would help to better understand the mechanism of the response to stress in activating and inhibiting the functional status of their particular endogenous substrates. This approach could be helpful for developing anxiolytic agents.