The supraoptic (SON) and the paraventricular (PVN) hypothalamic nuclei constitute major neuronal substrates underlying nitric oxide (NO) effects on autonomic and neuroendocrine control. Within these nuclei, constitutively produced NO restrains the firing activity of magnocellular neurosecretory and preautonomic neurons, actions thought to be mediated by a cGMP-dependent enhancement of GABAergic inhibitory transmission. In the present study, we expanded on this knowledge by performing a detailed anatomical characterization of constitutive NO-receptive, cGMP-producing neurons within the PVN. To this end, we combined tract-tracing techniques and immunohistochemistry to visualize cGMP immunoreactivity within functionally, neurochemically, and topographically discrete PVN neuronal populations in Wistar rats. Basal cGMP immunoreactivity was readily observed in the PVN, both in neuronal and vascular profiles. The incidence of cGMP immunoreactivity was significantly higher in magnocellular (69%) compared with preautonomic (∼10%) neuronal populations (P < 0.01). No differences were observed between oxytocin (OT) and vasopressin (VP) magnocellular neurons. In preautonomic neurons, the incidence of cGMP was independent of their subnuclei distribution, innervated target (i.e., intermediolateral cell column, nucleus tractus solitarii, or rostral ventrolateral medulla) or their neurochemical phenotype (i.e., OT or VP). Finally, high levels of cGMP immunoreactivity were observed in GABAergic somata and terminals within the PVN of eGFP-GAD67 transgenic mice. Altogether, these data support a highly heterogeneous distribution of basal cGMP levels within the PVN and further support the notion that constitutive NO actions in the PVN involve intricate cell-cell interactions, as well as heterogeneous signaling modalities.
- nitric oxide
the sympathetic nervous system is critical to the regulation of physiological homeostasis under basal conditions (30, 44). Substantial data indicate that forebrain, brain stem, and spinal networks are involved in this process (13). Located either side of the third ventricle, the paraventricular hypothalamus (PVN) has been established as a convergence point for many regions involved with maintaining homeostasis, such as fluid regulation, metabolism, immunological responses, and thermoregulation (5), as well as being pivotal in the maintenance of cardiovascular function (10, 18).
The PVN is a complex nucleus made up of three functionally distinct subsets of neurons: magnocellular neurosecretory neurons (MNSs), parvocellular neuroendocrine neurons, and parvocellular preautonomic neurons (64). The MNSs produce oxytocin (OT) and vasopressin (VP) peptides and innervate the posterior pituitary gland. The parvocellular neuroendocrine neurons project to the median eminence and regulate the release of hormones from the anterior pituitary gland. Parvocellular preautonomic neurons send long descending projections to brain stem regions, such as the nucleus tractus solitarii (NTS), rostral ventrolateral medulla (RVLM), and sympathetic preganglionic neurons in the intermediolateral (IML) column of the spinal cord (2, 13, 48, 50, 52, 55, 59, 63).
A large bulk of evidence supports a major role of preautonomic PVN neurons in the control of tonic and reflex autonomic function, including modulation of renal sympathetic nerve activity (7, 27, 38, 75), as well as the baroreflex (45, 77) and volume expansion reflexes (21, 51). Autonomic outflow from the PVN is regulated by the action of numerous neurotransmitters, including glutamate, GABA, and ANG II (8, 36, 65, 76). Of particular interest is nitric oxide (NO), a ubiquitous modulatory molecule that acts as a nonconventional neurotransmitter in the central nervous system (54). In general, NO is believed to act as an inhibitory molecule within central regions involved in autonomic and neuroendocrine control (for reviews, see Refs. 3 and 25). For example, incremental increases in NO levels within the PVN result in diminished neurohumoral secretion, decreased sympathetic nerve activity, and concomitant decreases in blood pressure and heart rate (22, 36, 75). Importantly, NO is constitutively produced within the PVN and supraoptic nucleus (SON) (60), and basal NO levels restrain neurohumoral output from these regions. This is supported by recent studies indicating that inhibitors of nitric oxide synthase (NOS) activity within the PVN results in sympathoexcitation and increased blood pressure (76). Moreover, diminished basal NO availability within the PVN has been shown to contribute to elevated neurohumoral drive in pathological conditions such as hypertension, heart failure and diabetes (16, 73, 78).
Accumulating evidence indicates that NO effects on neuroendocrine and autonomic outputs from the SON and PVN are mediated by inhibition of the electrical activity of both magnocellular (3, 37, 59, 62) and preautonomic neurons (34, 35), respectively, and that these actions are mediated through an enhancement of GABAergic transmission (33, 35, 71). One of the classical mechanisms underlying NO signaling in neurons involves activation of soluble guanylyl cyclase (31). In this sense, recent studies indicate that NO facilitation of GABAergic function and inhibitory effects in PVN preautonomic neurons involves activation of a cGMP-dependent pathway (33, 71). Altogether, the above-summarized evidence supports the NO-cGMP signaling cascade as a key mechanism controlling ongoing neuronal activity and neurohumoral output from the PVN.
In the present study, we aim to expand on this knowledge by performing a detailed anatomical characterization of constitutive NO-receptive, cGMP-producing neuronal populations within the PVN. To this end, we combined tract-tracing techniques and immunohistochemistry to visualize cGMP immunoreactivity within functionally, neurochemically and topographically discrete PVN neuronal populations. Overall, our results indicate that basal cGMP immunoreactive levels within the PVN vary in a cell type-dependent manner, and further support the general notion that NO within the PVN acts both in an autocrine and paracrine modulatory manners.
MATERIALS AND METHODS
All experiments were carried out with the approval of the Animal Ethics committees of the respective institutions. Data were obtained from 12 male Wistar rats (aged 7–8 wk old, 250 g) and 4 male transgenic mice (FVBTgN GAD67GFP 45704Swn), aged 6–9 wk old. Animals were housed in a temperature-controlled environment with ad libitum food and water and a 12:12-h light-dark cycle, both presurgery and postsurgery.
Anesthesia was induced using a ketamine/xylazine combination (for spinal cord injections: 40 mg/kg and 5 mg/kg im, respectively and for RVLM and NTS injections, 90 mg/kg and 50 mg/kg ip, respectively). For spinal cord injections, anesthesia was maintained using isoflurane (Isorane, 2–5%, 1 l/min, Veterinary Companies of Australia, Kings Park, NSW, Australia;). Spinal cords were exposed through a midline incision in the upper back, between T1 and T4. Animals were then placed in a stereotaxic frame and a glass micropipette (tip diameter 50–70 μm; Socorex, Ecublens, Switzerland) was filled with fluorescein- or rhodamine-tagged retrograde tracer (Lumaflor, Naples, FL). Following the methods of Cham et al. (6), two injections (200 nl each) were placed 200 μm apart on both the left and right sides of the spinal cord, 0.6 mm from the midline to a depth of 0.9 mm below the dorsal surface, running in a rostral to caudal fashion (total of four injections). To minimize leakage of tracer from the injection site, administration was undertaken over a 5-min period (6, 28). The localization and extension of the tracer in the IML were verified histologically at the end of the experiment (data not shown) and were in agreement with that previously demonstrated by Cham et al. (6). For brain stem injections, the head of the animals was placed in a stereotaxic frame, and the same fluorescent tracers as above were pressure injected (200 nl) Coordinates for RVLM injections were as follows: 12 mm caudal to bregma, 2 mm lateral to midline, and 8 mm below the dorsal surface. Coordinates for NTS injections were as follows: at the level of the obex: 1 mm lateral to the midline and 0.8 mm below the dorsal surface. Injections were performed unilaterally, either in the RVLM or in the NTS. After the injection, muscles were sutured together, and the wound was closed. The location of the tracer was verified histologically, as previously reported (35, 59). Individual animals had retrograde tracer injected in one site only, being either the RVLM (n = 4), NTS (n = 4), or IML (n = 4).
Seven days postsurgery for retrograde labeling, animals were anesthetized with pentobarbital sodium (100 mg/kg ip). Animals were perfused transcardially with 100 ml of ice-cold preperfusate of heparinized saline (0.2% heparin and 0.9% NaCl2, Mayne Pharma, Melbourne, Australia), followed by 500 ml of ice-cold fixative containing 4% formaldehyde in 0.1 M phosphate buffer (PB; pH 7.4). The perfusion was carried out over 25 min with a perfusion pump (rate 20 ml/min). Brains and spinal cord were then removed, and tissue was postfixed for 4 h at 4°C in 4% formaldehyde. The tissue was then immersed in 30% sucrose for a minimum of 24 h, mounted in Tissue Tek OCT compound embedding medium (Sakura Finetek, Zoeterwoude, The Netherlands), and 25-μm-thick transverse sections cut using a freezing cryostat.
To determine the relationship between OT or VP with cGMP in cell populations differentially identified by retrograde labeling, sections were prepared for fluorescence immunohistochemistry. Sections were washed 2 × 15 min in 0.01M PB, then blocked for 1 h in blocking solution of a 0.01M PB saline (PBS, pH 7.4) containing 10 mM Tris (TPBS) with 10% donkey serum (Chemicon, Temecula, CA), 0.01% Triton X-100 + 0.1% NaN3. Sections were then incubated with either a primary antibody cGMP/VP, cGMP/OT, or cGMP/nNOS double-labeling combination (antibody dilutions and source details are provided in Table 1). For cGMP/VP and cGMP/nNOS double labeling, primary antibodies were incubated simultaneously free floating for 24 h at room temperature. When double labeling for cGMP and OT, primary antibodies were incubated consecutively, each for 24 h. Antibodies were diluted in blocking solution. Sections were then washed 3 × 30 min and incubated with species-specific secondary antibodies (Table 1) for 4 h at room temperature, diluted in TPBS + 0.3% Tx + 0.1% NaN3 with 1% donkey serum. Sections were again washed (3 × 30 min 0.01M PBS), mounted on slides and coverslipped in Vector Shield (Vector Laboratories, Burlingame CA). All immunohistochemistry was done with minimal light saturation to prevent fading of the retrograde label. Negative controls were processed in parallel in the absence of primary antibody. Previous studies and manufacturers product information have documented the specificity of antibody reactions for the cGMP (49), OT (69), and VP (41) antibodies using preabsorption studies.
Quantification and analysis.
Sections were examined using a confocal microscope under fluorescence microscopy, and images were acquired of the PVN regions ranging from bregma −1.40 through bregma −2.10. Cells were assessed and counted for staining if they contained single, double, or triple staining, using Image Pro Plus software (Media Cybernetics, Silver Spring, MD). Images were opened as a stack, and the following regions of the PVN were examined for labeling: lateral magnocellular (LM), medial magnocellular (MM), dorsal cap (DC), ventromedial parvocellular (VM), and posterior parvocellular (PaPo) subnucleus (2, 59, 63). Cells were considered labeled if they contained retrograde label and/or OT/VP, and if the nucleus were clearly visible. Only complete cells were included in the count. In all cases and to avoid any bias related to the degree of cGMP staining, the sampled neurons were always observed and traced first on the chromophor channel containing the signal for their phenotypic marker (e.g,. OT, VP, nNOS immunoreactive, and/or retrograde labeling). This was performed using a freehand tool within the Image Pro Plus program. The channel was then switched to the one containing the cGMP immunoreactivity. The mean intensity of cGMP signal within traced neurons was calculated and expressed as fluorescence arbitrary units, ranging from 0 (absolute black) to 255 (absolute white). Background fluorescence was subtracted from all images. Neurons were arbitrarily considered to be cGMP positive when the fluorescence intensity within the neurons was 30% greater than background (see also Ref. 62). The percentage of total magnocellular neurosecretory neurons (determined by their location and OT/VP immunoreactivity), preautonomic neurons (determined by the presence of retrograde label) and nNOS neurons containing cGMP was determined and statistically compared.
To determine differences in the incidence of cGMP immunoreactivity among the various neuronal populations sampled, a Chi square test was used. Values were compared with a χ2 crit values table, obtained from Pagano (43).
General cGMP immunoreactivity within the PVN.
The majority of PVN neurons retrogradely labeled from the three different targets (IML, RVLM, and NTS) were concentrated in mid to caudal regions of the PVN (bregma −1.4 through −2.1). Neurons were clustered within the DC, VM, and PaPo parvocellular PVN subnuclei. Conversely, OT and VP magnocellular neurosecretory neurons were concentrated in the LM and in the MM subnuclei (see Fig. 1A). In general, cGMP immunoreactivity was diffusely observed throughout the PVN. Within neuronal profiles, a characteristic clustered pattern of cGMP was observed, whereas a more diffuse pattern was observed within PVN capillaries and arterioles. Representative examples of cGMP immunoreactivity at different rostrocaudal PVN levels are shown in Fig. 1A2–C2. As we recently reported, in the supraoptic nucleus (62), preincubation of living hypothalamic slices containing the PVN in the presence of NOS antagonists (1 mM l-NAME) or the NO scavenger c-PTIO (500 μM) almost completely abolished cGMP immunoreactivity (not shown), indicating that most basal cGMP levels within the PVN are NO driven.
Heterogeneous distribution of cGMP expression within neurosecretory and preautonomic PVN neuronal populations.
Double-labeling fluorescent immunohistochemistry with retrograde tracing was performed to determine colocalization of cGMP within identified magnocellular neurosecretory or preautonomic PVN neurons. Representative examples are shown in Fig. 2. A total of 173 magnocellular and 860 preautonomic neurons were sampled for this study. Of the magnocellular neurons, 119 were shown to contain cGMP (68.8%). In contrast, only 80 preautonomic neurons expressed cGMP (9.3%), a proportion significantly less than that observed in magnocellular cells (P < 0.01, Fig. 3A).
Oxytocin and vasopressin magnocellular neurons express similar basal cGMP levels.
Analysis was undertaken to determine whether there was any difference in cGMP immunoreactivity between magnocellular neurons that expressed OT and those that expressed VP. Of the 173 cells counted, 90 and 83 neurons were OT and VP immunoreactive, respectively. Both populations showed similar levels of colocalization, as each subgroup showed 69% of magnocellular neurons also immunoreactive for cGMP (62/90 OT and 57/83 VP, P > 0.5, Fig. 3B), indicating that basal cGMP levels in magnocellular neurosecretory neurons are not dependent on their neurochemical identity.
Comparison of cGMP immunoreactivity among preautonomic neurons, according to their subnuclei distribution.
Previous data from our laboratory support the presence of subnuclei-dependent differences in functional and structural properties of preautonomic PVN neurons (61). Thus, we aimed to determine whether basal cGMP immunoreactivity in preautonomic PVN neurons was dependent on their subnuclei distribution, regardless initially of their projection target (see below). Thus, we compared the expression of cGMP among retrogradely labeled neurons located in the VM, DC, and PaPo subnuclei (Fig. 4A). Within the VM, only 6.3% (21/332) of preautonomic neurons expressed cGMP, while in the DC and PaPo subnuclei, 10% (24/240) and 12.15% (35/288) neurons, respectively, were found to express cGMP. Despite a tendency for a lower degree of colocalization in the VM subnucleus, our results indicate no significant difference in the distribution of cGMP among these topographically distinct preautonomic neuronal populations (P > 0.1).
Comparison of cGMP immunoreactivity among preautonomic neurons, according to their innervated target.
We next analyzed whether cGMP levels in preautonomic PVN neurons varied according to their innervated targets (i.e., NTS, RVLM, or IML). Within the NTS-, RVLM-, and IML-projecting PVN neurons, 7.9% (24/301), 2.4% (3/121), and 12.1% (53/438), respectively, were found to be cGMP immunoreactive. Results from a statistical analysis indicated that basal cGMP immunoreactivity within these target-specific PVN populations was not significantly different from one another (Fig. 4B).
Comparison of cGMP immunoreactivity among preautonomic neurons, according to their neurochemical phenotype.
Preautonomic PVN neurons are known to be neurochemically heterogeneous (19). Similar to magnocellular neurosecretory neurons, a proportion of preautonomic neurons express OT and VP peptides. Thus, we performed triple fluorescence labeling (i.e., OT/VP, retrograde labeling, and cGMP) to determine whether cGMP levels in preautonomic PVN neurons may be dependent, at least in part, on their neurochemical identity. Of the total number of preautonomic cells sampled, we were able to identify 253 and 185 as OT and VP immunoreactive, respectively. Within the OT population, 15.4% (39/253) were found to contain cGMP. Within the VP population, 7.5% (14/185) contained cGMP. Although there appeared to be twice as many OT preautonomic neurons expressing cGMP, these differences were not statistically significant (Fig. 4C). Importantly, when the incidence of cGMP immunoreactivity in neurochemically identified OT and VP neurons was compared between magnocellular and preautonomic PVN neurons, significantly higher levels were observed in both cases in magnocellular neurons (OT: 68.8% vs. 15.4% in magnocellular and preautonomic, respectively, P < 0.01; VP: 68.6% vs. 7.5% in magnocellular and preautonomic, respectively, P < 0.01). These results indicate that basal cGMP levels in major projecting PVN neurons are largely dependent on their overall function (i.e., magnocellular neurosecretory vs. preautonomic), rather than on their neurochemical phenotype, topographical distribution, and/or autonomic-related innervated target.
Cyclic GMP expression in PVN nNOS positive neurons.
In several CNS brain regions, cGMP has been shown not to be present within NO-producing (nNOS expressing) neurons (14, 49). However, recent studies demonstrated high nNOS expression in magnocellular neurosecretory but not in preautonomic PVN neurons (35, 62). Thus, on the basis of our present results showing high and low cGMP immunoreactivity within magnocellular neurosecretory and preautonomic populations, respectively, we would predict a high degree of colocalization of cGMP with nNOS in PVN neurons. To confirm this, we performed double-immunohistochemical studies to localize cGMP in nNOS-positive neurons. As previously reported (23), we found nNOS-immunoreaactive neurons in the PVN to be concentrated in magnocellular neurosecretory nuclei, including the LM and MM. As shown in the representative examples in Fig. 5, the great majority (>90%, 92/102) of nNOS-immunoreactive cells also displayed robust immunoreactivity for cGMP.
Cyclic GMP expression in PVN GABAergic neurons and terminals.
Numerous studies indicate that nitric oxide inhibitory actions in the PVN are mediated through activation of local GABAergic interneurons (11, 24, 72, 74) and that these actions are mediated through production of cGMP (33). Thus, we aimed to determine whether basal cGMP levels were also found within PVN GABAergic interneurons. To identify this scarcely distributed interneuronal population, we used a transgenic mouse that expresses eGFP driven by the glutamic decarboxylase (GAD67) promoter, rendering GABAergic somata readily fluorescent (42). As shown in the representative examples of Fig. 6, we found basal cGMP immunoreactivity in most GABAergic somata sampled (n = 28/35). Moreover, cGMP immunoreactivity was also readily identified within GFP-labeled PVN axonal terminals (Fig. 6).
The most common signaling cascade by which NO mediates its functional effects involves activation of soluble guanylate cyclase, which subsequently elevates intracellular cGMP (57). Indeed, a cGMP-dependent cascade was recently demonstrated to mediate inhibition of presympathetic PVN neuronal activity in response to SNAP, an NO donor (33). Moreover, we have shown that NO is constitutively produced within the PVN, where it restrains ongoing presympathetic neuronal firing activity (35, 62). In addition, constitutively produced NO was shown to inhibit PVN-driven renal sympathetic nerve discharge (76). Still, whether cGMP also mediates constitutively produced NO actions within the PVN is unknown, and a detailed characterization of NO-receptive, cGMP-producing PVN neuronal populations, in particular, preautonomic ones, is lacking. In this study, we evaluated cGMP expression within various PVN cell populations, including preautonomic IML-, RVLM-, and NTS-projecting neurons, magnocellular neurosecretory OT and VP neurons, nNOS-expressing neurons, and GABAergic inhibitory interneurons. Our results support a heterogeneous distribution of basal cGMP levels within PVN neuronal populations, with relatively high levels found in magnocellular neurosecretory, nNOS-expressing, and GABAergic inhibitory neurons. Conversely, a relatively small proportion of preautonomic PVN neurons were found to be cGMP immunoreactive. As discussed below, these studies provide important functional insights into the cellular sources and targets of NO within the PVN, as well as the cell-cell communication modalities by which NO could modulate neuroendocrine and autonomic outflow from this key homeostatic hypothalamic center.
Cyclic GMP is highly expressed in PVN magnocellular neurosecretory neurons.
Consistent with our recent studies in SON neurons (62) but different from those previously reported by Vacher et al. (68), we found the majority of PVN magnocellular neurosecretory neurons to contain high somatic basal levels of cGMP. These results suggest that magnocellular neurosecretory neurons are important functional targets of NO and that in this neuronal population, NO can act in a direct, postsynaptic manner. Direct effects of NO in magnocellular neurons could involve both rapid, short-term, as well as slower, longer-lasting actions. For example, a direct short-term inhibitory effect of NO on magnocellular membrane excitability has been demonstrated (3, 12), an effect that likely involves modulation of voltage-gated ion channel activity, as demonstrated in other neuronal types (17, 20, 40). In addition, the NO-cGMP pathway is also capable of inducing long-lasting actions, such as changes in gene expression, by regulating specific transcription factors, most typically CREB (9, 47). Importantly, CREB was recently reported to highly colocalize with nNOS-positive cells in the PVN (53), and its phosphorylated state (pCREB) to be regulated by hyperosmolarity (39, 53, 56), a condition known to upregulate NO production and nNOS activity (26, 58, 67). Thus, in the magnocellular system, pCREB could play an important role in hyperosmotic-mediated increase in nNOS expression (53), supporting a functional link between NO-cGMP-CREB-nNOS in this system. Future studies are warranted to further explore the functional significance of long-term actions of NO in magnocellular neurons. It is worth noting, however, that the NO-cGMP pathway has also been shown to mediate indirect, synaptically mediated inhibitory actions in magnocellular neurons (3, 61).
In addition to being major cellular targets of NO, magnocellular neurons constitute major sources of this molecule within the SON and PVN. This is supported by several lines of evidence, including the expression of high levels of neuronal NOS (1, 23, 62), and of NO bioavailability markers (62). Moreover, this is further supported by results from the present study showing a high degree of colocalization of nNOS and cGMP within this neuronal population. Thus, altogether, these data support the NO-cGMP cascade as an important autocrine pathway by which magnocellular neurosecretory neurons could autoregulate their own function.
Preautonomic PVN neurons express relatively low levels of basal cGMP.
In contrast to magnocellular neurosecretory neurons, our results indicate a very low incidence (less than 10%) of basal cGMP in preautonomic PVN neurons. This was the case regardless of their innervated target, their topographical distribution within the PVN, and their neurochemical identity. Thus, it is reasonable to conclude that heterogeneity in basal cGMP immunoreactivity in the PVN is largely dependent on neuronal function (i.e., neurosecretory vs. preautonomic), but not neurochemical phenotype. The low incidence of cGMP in preautonomic PVN neuronal somata is also in line with our previous finding, indicating a low expression of nNOS in these neuronal populations (35).
The low levels of endogenous cGMP reported in this study, along with previous reports indicating also low levels of nNOS expression in this neuronal population (29, 35, 70) indicate that preautonomic PVN neurons are neither direct targets nor major cellular sources of NO within the PVN. Nonetheless, it is well established that constitutive NO efficiently restrains their firing activity, and consequently autonomic outflow from the PVN (34–36, 76). Thus, it is likely that the NO modulating preautonomic PVN neuronal function originates from alternative cellular source/s. Moreover, these data suggest that NO actions on this subpopulation of PVN neurons are mediated either through a direct, cGMP-independent mechanism, and/or alternatively, indirectly mediated. While cGMP-independent mechanisms cannot be ruled out, several lines of evidence support that NO effects on preautonomic PVN neurons are indirectly mediated through presynaptic modulation of GABA release, the dominant inhibitory neurotransmitter in the PVN (15). First, results from our laboratory showed that constitutive NO inhibitory actions on PVN-RVLM and PVN-NTS projecting neurons require intact and functional local GABAergic terminals and that enhanced NO availability increased the frequency of GABAergic inhibitory currents (35) (see also Refs. 33, 71). Secondly, work from the Patel laboratory showed that when acting in the PVN, NO inhibition of renal sympathetic nerve activity also required the presence of functional GABAergic inputs (76). Finally, recent work indicates that NO-mediated increased GABAergic activity in the PVN is mediated through a cGMP signaling cascade (33, 71). Altogether, these data suggest, though not conclusively demonstrate, that NO inhibitory actions on preautonomic PVN neurons and autonomic outflow from the PVN involve a cGMP-dependent increase in local GABAergic inhibitory activity.
This is further supported by results from the present study showing high levels of basal cGMP both in GABAergic neuronal somata and axonal processes within the PVN. GABAergic interneurons are scarcely distributed within the PVN and in peri-PVN areas (66) and express relatively low somatic levels of glutamic decarboxylase (GAD) (a commonly used marker for the identification of this neuronal population). Thus, identifying this neuronal population is not efficiently obtained using conventional immunohistochemical approaches. To circumscribe this problem, we used a transgenic mouse that expressed eGFP driven by the GAD67 promoter (4, 42). A limitation of these studies, however, is that NO-cGMP signaling mechanisms could differ between rat and mice. However, our results showing high levels of cGMP in magnocellular neurons both in mice and rats, as well as cGMP localization in mouse eGFP-GAD67 somata and terminals (this study), as well as in GABAergic immunoreactive terminals in the rat SON (62), would indicate similarities in NO-receptive neuronal populations in hypothalamic centers in these two species. Despite these caveats, our results are generally consistent with the notion that the PVN GABAergic interneuronal population is a key NO-receptive, cGMP-producing neuronal group within the PVN.
Perspectives and Signficance
Despite the well-characterized role of NO in the regulation of neuroendocrine and autonomic function within the SON and PVN, a comprehensive understanding of the precise underlying cellular mechanisms is still missing. Unique factors influencing the actions of this unconventional gaseous neurotransmitter include the specificity of the cellular sources and their efficacy in producing NO, the distribution of the cellular targets and their sensitivity to NO, the spatial interrelationship among sources and targets, as well as the signaling mechanisms underlying NO-mediated effects (for a review, see Ref. 46). All of these factors are particularly important in the PVN, a region characterized by a complex neuroanatomical, neurochemical, and functional organization (2, 63). Therefore, to better understand the functional role of NO in the regulation of autonomic and neuroendocrine homeostasis, a thorough characterization of the cellular sources and targets, as well as the precise mechanism of action of NO within discrete cell populations in the PVN is needed. On the basis of the present studies characterizing NO-receptive neuronal populations in the PVN, along with our previous studies assessing cellular sources and actions of NO (35, 61, 62), it is reasonable to speculate that magnocellular neurosecretory neurons constitute important sources and functional targets of NO. In these neurons, NO could influence neurosecretory function either in a direct or indirect manner. Direct actions could mediate long-lasting effects associated with changes in gene expression, while indirect actions likely mediate short-term, GABA-mediated inhibition of neuronal activity. Conversely, while preautonomic neurons do not appear to constitute critical sources of NO within the PVN, their ongoing activity is still efficiently modulated by NO. In these neurons, NO likely arises from a different cellular source and acts in an indirect, synaptically mediated manner. Thus, NO modulatory actions on autonomic and neuroendocrine control within the PVN involve intricate cell-cell interactions and diverse communication modalities. Given NO's unique ability to freely diffuse in brain tissue and to affect targets at relatively long distance from its sources, it will be important in the future to determine whether and how these signaling modalities also mediate cross talk of information between these two information processing systems (60), thus playing a key role in the generation of complex patterns of homeostatic responses by the PVN.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2008 the American Physiological Society