Columns Of PAG Research Articles

Unravelling the dorsal periaqueductal grey matter NMDA receptors relevance in the nitric oxide-mediated panic‑like behaviour and defensive antinociception organised by the anterior hypothalamus of male mice.

Previous studies suggested that the dorsal column of the periaqueductal grey matter (dPAG) can be a target of neural pathways from hypothalamic nuclei involved in triggering fear-related defensive responses. In turn, evidence is provided suggesting that microinjection of the nitric oxide (NO) donor SIN-1 into the anterior hypothalamus (AH) of mice evokes panic-like behaviours and fear-induced antinociception. However, it is unknown whether the dPAG of mice mediates these latter defensive responses organised by AH neurons. This study was designed to examine the role of dPAG in mediating SIN-1-evoked fear-induced defensive behavioural and antinociceptive responses organised in the AH of mice. First, neural tract tracing was performed to characterise the AH-dPAG pathways. Then, using neuropharmacological approaches, we evaluated the effects of dPAG pretreatment with either the non-selective synaptic blocker cobalt chloride (CoCl2; 1mM/0.1 μL) or the competitive N-methyl-D-aspartate (NMDA) receptor antagonist LY235959 (0.1nmol/0.1 μL) on defensive behaviours and antinociception induced by microinjections of SIN-1 in the AH of male C57BL/6 mice. AlexaFluor488-conjugated dextran-labelled axonal fibres from AH neurons were identified in both dorsomedial and dorsolateral PAG columns. Furthermore, we showed that pre-treatment of the dPAG with either CoCl2 or LY235959 inhibited freezing and impaired oriented escape and antinociception induced by infusions of SIN-1 into the AH. These findings suggest that the panic-like freezing and oriented escape defensive behaviours, and fear-induced antinociception elicited by intra-AH microinjections of SIN-1 depend on the activation of dPAG NMDA receptors.

Functional activation of the periaqueductal gray matter during conditioned and unconditioned fear in guinea pigs confronted with the Boa constrictor constrictor snake.

The periaqueductal gray matter (PAG) is an essential structure involved in the elaboration of defensive responses, such as when facing predators and conspecific aggressors. Using a prey vs predator paradigm, we aimed to evaluate the PAG activation pattern evoked by unconditioned and conditioned fear situations. Adult male guinea pigs were confronted either by a Boa constrictor constrictor wild snake or by the aversive experimental context. After the behavioral test, the rodents were euthanized and the brain prepared for immunohistochemistry for Fos protein identification in different PAG columns. Although Fos-protein-labeled neurons were found in different PAG columns after both unconditioned and conditioned fear situations at the caudal level of the PAG, we found greater activation of the lateral column compared to the ventrolateral and dorsomedial columns after predator exposure. Moreover, the lateral column of the PAG showed higher Fos-labeled cells at the caudal level compared to the same area at the rostral level. The present results suggested that there are different activation patterns of PAG columns during unconditioned and conditioned fear in guinea pigs. It is possible to hypothesize that the recruitment of specific PAG columns depended on the nature of the threatening stimulus.

Neurons expressing estrogen receptor α differentially innervate the periaqueductal gray matter of female rats

The periaqueductal gray matter (PAG) is a brainstem site involved in distinct autonomic and behavioral responses. Among them, the motor control of female sexual behavior, including lordosis, is well described. Lordosis reflex is highly dependent on increasing levels of estradiol that occur in the afternoon of the proestrus day in normally cycling females. This effect is thought to be mediated primarily via actions in the ventromedial nucleus of the hypothalamus (VMH). By binding to estrogen receptor α (ERα), estradiol changes the activity of VMH neurons that project to the PAG. Evidence also exists for the coordination of PAG outputs by estradiol-responsive neurons outside the VMH. However, a comprehensive analysis of these circuitries is not available. Using stereotaxic injection of the retrograde tracer Fluorogold in distinct columns of the PAG we performed a systematic mapping of neurons innervating the PAG and those coexpressing ERα immunoreactivity. We found that the forebrain projections to PAG columns are largely segregated and that most of the ERα expressing neurons preferentially target the lateral and the ventrolateral columns. Dual labeled neurons were mostly found in the intermediate subdivision of the lateral septal nucleus, the posterior aspect of the medial bed nucleus of the stria terminalis, the medial preoptic nucleus, the striohypothalamic nucleus and the ventrolateral VMH. Few dual labeled neurons were also observed in the arcuate nucleus, in the posterodorsal subdivision of the medial nucleus of the amygdala and in the ventral premammillary nucleus. Our findings indicate that ERα modulates sexual behavior in female rats via an integrated neural network that differentially innervate the columns of the PAG.

µ1 -Opioid receptors in the dorsomedial and ventrolateral columns of the periaqueductal grey matter are critical for the enhancement of post-ictal antinociception.

Generalised tonic and tonic-clonic seizures are followed by significant increase in nociceptive thresholds in both laboratory animals and humans. The endogenous opioid peptides play a role in antinociceptive signalling, and the periaqueductal grey matter (PAG) is recruited to induce analgesia. Thus, the aim of this investigation was to evaluate the role of µ1 -opioid receptors in the dorsomedial (dm) and ventrolateral (vl) columns of PAG in post-ictal antinociception. Pentylenetetrazole (PTZ; 64 mg/kg), which is an ionotropic GABA-mediated Cl- influx antagonist, was intraperitoneally (IP) administered to induce tonic-clonic seizures in Wistar rats. The tail-flick test was used to measure the nociceptive threshold. Microinjections of naltrexone (5.0 µg/0.2 µL), which is a non-selective opioid receptor antagonist, in both dmPAG and vlPAG decreased the tonic-clonic seizure-induced antinociception in seizing animals from 10 to 120 min after seizures. Furthermore, microinjections of the µ1 -opioid receptor-selective antagonist naloxonazine (5.0 µg/0.2 µL) into the dmPAG decreased post-ictal antinociception immediately after convulsive reactions and from 10 to 90 min after seizures. However, vlPAG-pretreatment with naloxonazine at the same concentration decreased the post-ictal antinociception 30 min after the onset of tonic-clonic seizures and the nociceptive threshold returned to basal values 120 min after seizures. These findings indicate that µ1 -opioid receptor-signalling mechanisms in both dmPAG and vlPAG play a relevant role in the organisation of post-ictal antinociception. In addition, µ1 -opioid receptors in the dmPAG rather than in vlPAG seem to be more critically recruited during the antinociception induced by generalised tonic-clonic seizures.

The role of dorsomedial and ventrolateral columns of the periaqueductal gray matter and in situ 5-HT2Aand 5-HT2Cserotonergic receptors in post-ictal antinociception

The periaqueductal gray matter (PAG) consists in a brainstem structure rich in 5-hydroxytryptamine (5-HT) inputs related to the modulation of pain. The involvement of each of the serotonergic receptor subtypes found in PAG columns, such as the dorsomedial (dmPAG) and the ventrolateral (vlPAG) columns, regarding post-ictal antinociception have not been elucidated. The present work investigated the participation of the dmPAG and vlPAG columns in seizure-induced antinociception. Specifically, we studied the involvement of serotonergic neurotransmission in these columns on antinociceptive responses that follow tonic-clonic epileptic reactions induced by pentylenetetrazole (PTZ), an ionophore GABA-mediated Cl(-) influx antagonist. Microinjections of cobalt chloride (1.0 mM CoCl2 /0.2 µL) into the dmPAG and vlPAG caused an intermittent local synaptic inhibition and decreased post-ictal antinociception that had been recorded at various time points after seizures. Pretreatments of the dmPAG or the vlPAG columns with the nonselective serotonergic receptors antagonist methysergide (5.0 µg/0.2 µL) or intramesencephalic microinjections of ketanserin (5.0 µg/0.2 µL), a serotonergic antagonist with more affinity to 5-HT2A/2C receptors, decreased tonic-clonic seizure-induced antinociception. Both dmPAG and vlPAG treatment with either the 5-HT2A receptor selective antagonist R-96544 (10 nM/0.2 µL), or the 5-HT2C receptors selective antagonist RS-102221 (0.15 µg/0.2 µL) also decrease post-ictal antinociception. These findings suggest that serotonergic neurotransmission, which recruits both 5-HT2A and 5-HT2C serotonergic receptors in dmPAG and vlPAG columns, plays a critical role in the elaboration of post-ictal antinociception.

Efferent connections of the parvalbumin-positive (PV1) nucleus in the lateral hypothalamus of rodents.

A solitary cluster of parvalbumin-positive neurons--the PV1 nucleus--has been observed in the lateral hypothalamus of rodents. In the present study, we mapped the efferent connections of the PV1 nucleus using nonspecific antero- and retrograde tracers in rats, and chemoselective, Cre-dependent viral constructs in parvalbumin-Cre mice. In both species, the PV1 nucleus was found to project mainly to the periaqueductal grey matter (PAG), predominantly ipsilaterally. Indirectly in rats and directly in mice, a discrete, longitudinally oriented cylindrical column of terminal fields (PV1-CTF) was identified ventrolateral to the aqueduct on the edge of the PAG. The PV1-CTF is particularly dense in the rostral portion, which is located in the supraoculomotor nucleus (Su3). It is spatially interrupted over a short stretch at the level of the trochlear nucleus and abuts caudally on a second parvalbumin-positive (PV2) nucleus. The rostral and the caudal portions of the PV1-CTF consist of axonal endings, which stem from neurons scattered throughout the PV1 nucleus. Topographically, the longitudinal orientation of the PV1-CTF accords with that of the likewise longitudinally oriented functional modules of the PAG, but overlaps none of them. Minor terminal fields were identified in a crescentic column of the lateral PAG, as well as in the Edinger-Westphal, the lateral habenular, and the laterodorsal tegmental nuclei. So far, no obvious functions have been attributed to this small, circumscribed column ventrolateral to the aqueduct, the prime target of the PV1 nucleus.

Efferent connections of the parvalbumin‐positive (PV1) nucleus in the lateral hypothalamus of rodents

ABSTRACTA solitary cluster of parvalbumin‐positive neurons – the PV1 nucleus – has been observed in the lateral hypothalamus of rodents. In the present study, we mapped the efferent connections of the PV1 nucleus using nonspecific antero‐ and retrograde tracers in rats, and chemoselective, Cre‐dependent viral constructs in parvalbumin‐Cre mice. In both species, the PV1 nucleus was found to project mainly to the periaqueductal grey matter (PAG), predominantly ipsilaterally. Indirectly in rats and directly in mice, a discrete, longitudinally oriented cylindrical column of terminal fields (PV1‐CTF) was identified ventrolateral to the aqueduct on the edge of the PAG. The PV1‐CTF is particularly dense in the rostral portion, which is located in the supraoculomotor nucleus (Su3). It is spatially interrupted over a short stretch at the level of the trochlear nucleus and abuts caudally on a second parvalbumin‐positive (PV2) nucleus. The rostral and the caudal portions of the PV1‐CTF consist of axonal endings, which stem from neurons scattered throughout the PV1 nucleus. Topographically, the longitudinal orientation of the PV1‐CTF accords with that of the likewise longitudinally oriented functional modules of the PAG, but overlaps none of them. Minor terminal fields were identified in a crescentic column of the lateral PAG, as well as in the Edinger–Westphal, the lateral habenular, and the laterodorsal tegmental nuclei. So far, no obvious functions have been attributed to this small, circumscribed column ventrolateral to the aqueduct, the prime target of the PV1 nucleus.J. Comp. Neurol. 521:3133–3153, 2013. © 2013 Wiley Periodicals, Inc.

Chemical neuroanatomical and psychopharmacological evidence that κ receptor-mediated endogenous opioid peptide neurotransmission in the dorsal and ventral mesencephalon modulates panic-like behaviour

The chemical neuroanatomy and the effects of central administration of opioid antagonists on the innate fear-induced responses elicited by electrical (at escape behaviour threshold) stimulation of the midbrain tectum were determined. The aim of the present work was to investigate the interaction between the tecto-nigral endogenous opioid peptide-mediated disinhibitory pathways and nigro-tectal inhibitory links in the control of panic-like behaviour and their organisation in the continuum comprised by the deep layers of the superior colliculus (dlSC) and the dorsolateral columns of the periaqueductal grey matter (dlPAG). Beta-endorphin-labelled neurons and fibres were found in the dorsal midbrain and also in the substantia nigra. Opioid varicose fibres and terminal buttons were widely distributed in PAG columns and in all substantia nigra subdivisions. Microinjections of naltrexone (a non-selective opioid receptor antagonist; 5.0μg/0.2μl) or nor-binaltorphimine (a selective κ-opioid receptor antagonist; 5.0μg/0.2μl) in the dlSC/dlPAG continuum, in independent groups of animals, induced significant increases in the escape thresholds for midbrain tectum electrical stimulation. The microinjection of naltrexone or nor-binaltorphimine into the SNpr also increased the escape behaviour threshold for electrical stimulation of dlSC/dlPAG. These morphological and neuropharmacological findings support previous evidence from our team for the role played by the interaction between opioidergic and GABAergic mechanisms in the modulation of innate fear-induced responses. The present data offer a neuroanatomical basis for both intratectal axo-axonic/pre-synaptic and tecto-nigral axo-somatic opioid inhibition of GABAergic nigro-tectal neurons that modulate the dorsal midbrain neurons related to the organisation of fear-related emotional responses.

Ventrolateral periaqueductal gray lesion attenuates nociception but does not change anxiety-like indices or fear-induced antinociception in mice

The exposure of rodents to an open elevated plus-maze (oEPM: four open arms raised from the floor) elicits naloxone-insensitive antinociception. Midazolam infusion into the dorsal portion of the periaqueductal gray (dPAG), a structure of the descending inhibitory system of pain, failed to alter oEPM-induced antinociception. Chemical lesion of dorsomedial and dorsolateral PAG attenuated defensive behavior in the standard EPM (sEPM), an animal model of anxiety, but failed to change oEPM-induced antinociception. The present study investigated the effects of bilateral lesion, with the injection of NMDA ( N-methyl- d-aspartic acid), of the ventrolateral column of PAG (vlPAG) (i) on nociceptive response induced by 2.5% formalin injected into the right hind paw (nociception test) in mice exposed to the enclosed EPM (eEPM: four enclosed arms – a non-aversive situation) or to the oEPM and (ii) on anxiety indices in mice exposed to the sEPM without prior formalin injection. Results showed that oEPM-induced antinociception was not altered by lesion of vlPAG. Nevertheless, the lesion reduced the nociceptive response in mice exposed to the eEPM and increased general locomotor activity during the eEPM and oEPM exposure. Furthermore, vlPAG lesion did not alter anxiety-like indices in mice exposed to the sEPM. The results suggest that vlPAG does not play a role in oEPM-induced antinociception or in defensive reactions assessed in the sEPM. Moreover, vlPAG inactivation induces pain inhibition in mice not exposed to an aversive situation and seems to increase general activity.

GABA-immunoreactive neurons and terminals in the cat periaqueductal gray matter: A light and electron microscopic study

Immunocytochemical and electron microscopic methods were used to study the GABAergic innervation in adult cat periaqueductal gray matter (PAG). A mouse monoclonal antibody against gamma -aminobutyric acid (GABA) was used to visualize the inhibitory neuronal system of PAG. At light microscopy, GABA-immunopositive (GABA(IP)) neurons formed two longitudinally oriented columns in the dorsolateral and ventrolateral PAG that accounted for 36% of the neuronal population of both PAG columns; their perikaryal cross-sectional area was smaller than that of unlabeled (UNL) neurons found in the same PAG subdivisions. At electron microscopic level, patches of GABA immunoreactivity were readily detected in neuronal cell bodies, proximal and distal dendrites, axons and axon terminals. Approximately 35-36% of all terminals were GABA(IP); they established symmetric synapses with dendrites (84.72% of the sample in the dorsolateral PAG and 86.09% of the sample in the ventrolateral PAG) or with cell bodies (7-10% of the sample). Moreover, 49.15% of GABA(IP) axon terminals in the dorsolateral and 52.16% in the ventrolateral PAG established symmetric synapses with GABA(IP) dendrites. Immunopositive axon terminals and unlabeled terminals were also involved in the formation of a complex synaptic arrangment, i.e. clusters of synaptic terminals in close contact between them that were often observed in the PAG neuropil. Moreover, a fair number of axo-axonic synapses between GABA(IP) and/or UNL axon terminals were present in both PAG subdivisions. Several dendro-dendritic synapses between labeled and unlabeled dendrites were also observed in both PAG subdivisions. These results suggest that in the cat PAG there exist at least two classes of GABArgic neurons. The first class could exert a tonic control on PAG projecting neurons, the second could act on those GABAergic neurons that in turn keep PAG projecting neurons under tonic inhibition. The functional implications of this type of GABAergic synapse organization are discussed in relation to the dishinibitory processes that take place in the PAG.

Distribution of extracellular signal‐regulated kinase 1‐ and 2‐activated neurons in the rat periaqueductal gray matter after noxious stimulation

The periaqueductal gray matter (PAG), the midbrain region made up of neuronal columns encircling the cerebral aqueduct, plays a key role in nociception. As the extracellular signal-regulated kinases (ERKs) 1 and 2 are activated after noxious stimulation, we analyzed the distribution of ERK-activated neurons in the PAG after visceral noxious stimulation. Ether- and urethane-anesthetized rats received an intraperitoneal injection of acetic acid or were left untreated and were perfused after 2 hr. Serial sections immunoreacted with an antibody selective for the activated ERKs. Significant ERK activation occurred only in the ether-anesthetized noxious stimulated rats. In these rats, we evaluated the number of ERK-activated neurons and their density as the ratio of the number of immunolabeled neurons to the extension of the region where they were located. ERK-activated neurons were more numerous in the lateral (LPAG) and ventrolateral (VLPAG) columns, but without significant differences. No ERK activation was seen in neurons of the most rostral PAG. The ERK-activated neurons were significantly denser at the intermediate level of the PAG. At the caudal level, they were denser in the LPAG and VLPAG columns, and in the DPAG column at the intermediate and rostral level. These findings suggest that noxious stimulation activates ERKs in neurons involved in the different functional activities related to nociception, overlapping in the PAG columns, and strengthens the role of PAG in integration.

Periaqueductal gray matter projection to the parabrachial nucleus in rat

The efferent projections from the periaqueductal gray matter (PAG) to the parabrachial nucleus (PB) were studied in the rat following microinjections of the anterograde axonal tracer Phaseolus vulgaris-leucoagglutinin (PHA-L) into restricted regions of the PAG. The dorsomedial and dorsolateral PAG columns project almost exclusively to the superior lateral PB subnucleus, whereas the lateral and ventrolateral PAG columns project to five lateral PB sites: dorsal lateral subnucleus, medial and lateral crescent areas (which flank the dorsal lateral PB subnucleus), central lateral subnucleus (rostral portion), and superior lateral subnucleus. The PAG region lying near the cerebral aqueduct projects to five lateral PB sites: external lateral subnucleus (inner subdivision), medial and lateral crescent areas, central lateral subnucleus (rostral portion), and dorsal lateral subnucleus. The internal lateral PB subnucleus, which projects exclusively to the intralaminar thalamic nuclei, and the Kölliker-Fuse nucleus were not innervated by the PAG. The PAG selectively innervates individual PB subnuclei that may be part of the spino-parachio-forebrain pathway. All PAG columns, including the aqueductal region, project to the superior lateral PB subnucleus, a presumed nociceptive relay site that receives inputs from multiple spinal cord regions (laminae I, V, and VIII) and projects to the ventromedial and retrochiasmatic hypothalamic areas—two regions that have been implicated in complex goal-directed behavior (e.g., food intake and reproductive function). Earlier studies demonstrated that the dorsal lateral and external lateral PB subnuclei (inner division) receive overlapping inputs from the superficial dorsal horn (laminae I and II) and the nucleus tractus solitarius, and both PB subnuclei send projections to limbic forebrain areas (e.g., hypothalamus, preoptic region, amygdala). Because the PAG projects to both of these PB subnuclei, this projection system possibly functions as a behavioral state-dependent filter system that modulates ascending nociceptive and/ or visceral information as it is relayed through the PB to forebrain sites. J. Comp. Neurol. 401:437–454, 1998. © 1998 Wiley-Liss, Inc.

Periaqueductal gray matter projection to the parabrachial nucleus in rat

The efferent projections from the periaqueductal gray matter (PAG) to the parabrachial nucleus (PB) were studied in the rat following microinjections of the anterograde axonal tracer Phaseolus vulgaris-leucoagglutinin (PHA-L) into restricted regions of the PAG. The dorsomedial and dorsolateral PAG columns project almost exclusively to the superior lateral PB subnucleus, whereas the lateral and ventrolateral PAG columns project to five lateral PB sites: dorsal lateral subnucleus, medial and lateral crescent areas (which flank the dorsal lateral PB subnucleus), central lateral subnucleus (rostral portion), and superior lateral subnucleus. The PAG region lying near the cerebral aqueduct projects to five lateral PB sites: external lateral subnucleus (inner subdivision), medial and lateral crescent areas, central lateral subnucleus (rostral portion), and dorsal lateral subnucleus. The internal lateral PB subnucleus, which projects exclusively to the intralaminar thalamic nuclei, and the Kölliker-Fuse nucleus were not innervated by the PAG. The PAG selectively innervates individual PB subnuclei that may be part of the spino-parachio-forebrain pathway. All PAG columns, including the aqueductal region, project to the superior lateral PB subnucleus, a presumed nociceptive relay site that receives inputs from multiple spinal cord regions (laminae I, V, and VIII) and projects to the ventromedial and retrochiasmatic hypothalamic areas-two regions that have been implicated in complex goal-directed behavior (e.g., food intake and reproductive function). Earlier studies demonstrated that the dorsal lateral and external lateral PB subnuclei (inner division) receive overlapping inputs from the superficial dorsal horn (laminae I and II) and the nucleus tractus solitarius, and both PB subnuclei send projections to limbic forebrain areas (e.g., hypothalamus, preoptic region, amygdala). Because the PAG projects to both of these PB subnuclei, this projection system possibly functions as a behavioral state-dependent filter system that modulates ascending nociceptive and/or visceral information as it is relayed through the PB to forebrain sites.

Periaqueductal gray matter input to cardiac-related sympathetic premotor neurons

The periaqueductal gray matter (PAG) serves as the midbrain link between forebrain emotional processing systems and motor pathways used in the defense reaction. Part of this response depends upon PAG efferent pathways that modulate cardiovascular-related sympathetic outflow systems, including those that regulate the heart. While it is known that the PAG projects to vagal preganglionic neurons, including possibly cardiovagal motoneurons, no information exists on the PAG circuits that may affect sympathetically mediated cardiac functions and, thus, the purpose of this study was to use neuroanatomical methods to identify these pathways. First, viral transneuronal retrograde tracing experiments were performed in which pseudorabies virus (PRV) was injected into the stellate ganglion of rats. After 4 days survival, five PAG regions contained transynaptically infected neurons; these included the dorsomedial, lateral and ventrolateral PAG columns as well as the Edinger–Westphal and precommissural nuclei. Second, the descending efferent PAG projections were studied with the anterograde axonal marker Phaseolus vulgaris leuco-agglutinin (PHA-L) with a particular focus on determining whether the PAG projects to the intermediolateral cell column (IML). Almost no axonal labeling was found throughout the thoracic IML suggesting that the PAG modulates sympathetic functions by indirect pathways involving synaptic relays through sympathetic premotor cell groups, especially those found in the medulla oblongata. This possibility was examined by a double tracing study. PHA-L was first injected into either the lateral or ventrolateral PAG and after 6 days, PRV was injected into the ipsilateral stellate ganglion. After an additional 4 days survival, a double immunohistochemical procedure for co-visualization of PRV and PHA-L was used to identify the sympathetic premotor regions that receive an input from the PAG. The PAG innervated specific groups of sympathetic premotor neurons in the hypothalamus, pons, and medulla as well as providing reciprocal intercolumnar connections within the PAG itself (Jansen et al., Brain Res. 784 (1998) 329-336). The major route terminates in the ventral medulla, especially within the medial region which contains sympathetic premotor neurons lying within the raphe magnus and gigantocellular reticular nucleus, pars alpha. Both serotonergic and non-serotonergic sympathetic premotor neurons in these two regions receive inputs from the PAG. Weak PAG projections to sympathetic premotor neurons were found in the rostral ventrolateral medulla (including to C1 adrenergic neurons), locus coeruleus, A5 cell group, paraventricular and lateral hypothalamic nuclei. In summary, both the lateral and ventrolateral PAG columns appear to be capable of modulating cardiac sympathetic functions via a series of indirect pathways involving sympathetic premotor neurons found in selected sites in the hypothalamus, midbrain, pons, and medulla oblongata, with the major outflow terminating in bulbospinal regions of the rostral ventromedial medulla.

Local connections between the columns of the periaqueductal gray matter: a case for intrinsic neuromodulation

Chemical stimulation of the midbrain periaqueductal gray matter (PAG) in conscious animals produces opposite responses (viz., defensive behavior and pressor responses from the lateral PAG column vs. quiescence and depressor responses from the ventrolateral PAG column), raising the possibility that PAG columns are interconnected. To test this hypothesis, two types of anatomical experiments were performed in rats. First, the anterograde axonal marker Phaseolus vulgaris leuco-agglutinin (PHA-L) was injected into individual PAG columns or adjoining regions which included the Edinger–Westphal, dorsal raphe, and precommissural nuclei. Ten days later, the intrinsic PAG projections were studied. Our results show that each PAG column projects bilaterally to all of the other PAG columns, and also provides local connections within its own column. Furthermore, the Edinger–Westphal and precommissural nuclei project to all four PAG columns, while the dorsal raphe nucleus projects only to the ventrolateral and lateral columns. In a second experiment, we found that cardiovascular-related PAG projection neurons of both the lateral and ventrolateral columns receive inputs from the reciprocal PAG column. This was demonstrated by a double tracer neuroanatomical study in which PHA-L was first iontophoretically ejected into either the lateral or ventrolateral PAG columns and then, several days later the retrograde transneuronal viral tracer, pseudorabies virus, was injected into the stellate sympathetic ganglion. Intra-PAG circuits were visualized by a dual immunohistochemical procedure. These results suggest that during the fight-or-flight response when the `fight' program is activated, the `flight' PAG network may be inhibited and the converse situation may occur during the flight response.

Periaqueductal gray matter projection to vagal preganglionic neurons and the nucleus tractus solitarius

The periaqueductal gray matter (PAG) has been implicated in a variety of different functions, including autonomic regulation. Chemical stimulation of the lateral PAG produces hypertension and tachycardia while activation of the ventrolateral PAG produces the opposite effect. While these effects are the result of alterations in sympathetic activity, little is known about whether the PAG can modulate vagal functions as well. The anterograde axonal tracing method using the plant lectin Phaseolus vulgaris leucoagglutinin (PHA-L) was used to determine whether both of the lateral and ventrolateral PAG columns project to vagal preganglionic neurons and/or to the nucleus tractus solitarius (NTS). Highly restricted PHA-L injections were made in all four PAG columns throughout their rostrocaudal extent in rats. Labeled fibers were visualized by immunohistochemistry and studied in relationship with choline acetyltransferase (ChAT) immunostained parasympathetic preganglionic neurons of the dorsal motor vagal nucleus (DMV) and nucleus ambiguus (NA). The lateral PAG projects to the lateral DMV and to the caudal part of the external NA. The ventrolateral PAG innervates the same regions and also projects to the rostral part of the external NA – a site that contains cardiac parasympathetic preganglionic neurons. Both the lateral and ventrolateral PAG project to the NTS in a similar fashion innervating the medial, ventrolateral and commissural subnuclei. In summary, the lateral and ventrolateral PAG have similar patterns of innervation of the NTS and DMV, but their projection to the NA is different: the rostral external NA receives innervation only from the ventrolateral PAG and the lateral PAG innervates the caudal part.