Time course and effective sites for inhibition from midbrain periaqueductal gray of spinal dorsal horn neuronal responses to cutaneous stimuli in the cat

Neuroanatomy of the pain system and of the pathways that modulate pain

We review many of the recent findings concerning mechanisms and pathways for pain and its modulation, emphasizing sensitization and the modulation of nociceptors and of dorsal horn nociceptive neurons. We describe the organization of several ascending nociceptive pathways, including the spinothalamic, spinomesencephalic, spinoreticular, spinolimbic, spinocervical, and postsynaptic dorsal column pathways in some detail and discuss nociceptive processing in the thalamus and cerebral cortex. Structures involved in the descending analgesia systems, including the periaqueductal gray, locus ceruleus, and parabrachial area, nucleus raphe magnus, reticular formation, anterior pretectal nucleus, thalamus and cerebral cortex, and several components of the limbic system are described and the pathways and neurotransmitters utilized are mentioned. Finally, we speculate on possible fruitful lines of research that might lead to improvements in therapy for pain.

Functional characteristics of the midbrain periaqueductal gray

The major functions of the midbrain periaqueductal gray (PAG), including pain and analgesia, fear and anxiety, vocalization, lordosis and cardiovascular control are considered in this review article. The PAG is an important site in ascending pain transmission. It receives afferents from nociceptive neurons in the spinal cord and sends projections to thalamic nuclei that process nociception. The PAG is also a major component of a descending pain inhibitory system. Activation of this system inhibits nociceptive neurons in the dorsal horn of the sinal cord. The dorsal PAG is a major site for processing of fear and anxiety. It interacts with the amygdala and its lesion alters fear and anxiety produced by stimulation of amygdala. Stimulation of PAG produces vocalization and its lesion produces mutism. The firing of many cells within the PAG correlates with vocalization. The PAG is a major site for lordosis and this role of PAG is mediated by a pathway connecting the medial preoptic with the PAG. The cardiovascular controlling network within the PAG are organized in columns. The dorsal column is involved in pressor and the ventrolateral column mediates depressor responses. The major intrinsic circuit within the PAG is a tonically-active GABAergic network and inhibition of this network is an important mechanism for activation of outputs of the PAG. The various functions of the PAG are interrelated and there is a significant interaction between different functional components of the PAG. Using the current information about the anatomy, physiology, and pharmacology of the PAG, a model is proposed to account for the interactions between these different functional components.

Projections from the periaqueductal gray to the rostromedial pericoerulear region and nucleus locus coeruleus: anatomic and physiologic studies

Previous studies showed that the nucleus locus coeruleus (LC) receives two major afferent inputs from 1) nucleus paragigantocellularis and 2) nucleus prepositus hypoglossi, both in the rostral medulla. Recent reports suggested that the midbrain periaqueductal gray (PAG) projects to the rostromedial pericoerulear area and LC. Since the PAG is a major site for control of central antinociception, and since descending noradrenergic fibers have been implicated in pain modulation, we have investigated in detail the functional anatomy of projections from PAG to the dorsolateral pontine tegmentum. A combined anatomical and electrophysiological approach was used to assess the organization and synaptic influence of PAG on neurons in the rostromedial pericoerulear region and in LC proper. Injections of the tracer wheatgerm agglutinin conjugated to horseradish peroxidase encompassing LC proper and the rostromedial pericoerulear area retrogradely labeled neurons in PAG located lateral and ventrolateral to the cerebral aqueduct; injections restricted to LC proper did not consistently label PAG neurons. Deposits of the anterograde axonal tracer Phaseolus vulgaris leucoagglutinin into this same region of PAG labeled axons that robustly innervated the zone rostral and medial to LC. Only sparse fibers were observed in LC proper. Consistent with these results, focal electrical stimulation of LC antidromically activated only a few PAG neurons (6 of 100); all of these driven cells were located lateral and ventrolateral to the cerebral aqueduct. The majority of neurons in the rostromedial pericoerulear area were robustly activated by single pulse stimulation of PAG. In contrast, single pulse electrical stimulation of lateral PAG produced weak to moderate synaptic activation of some LC neurons; stimulation of ventrolateral PAG produced predominant inhibition of LC discharge, perhaps through recurrent collaterals subsequent to antidromic activation of neighboring LC cells. Taken together, these results indicate that PAG strongly innervates the region rostral and medial to LC, including Barrington's nucleus, but only weakly innervates LC proper. Although recent studies indicate that the dendrites of LC neurons ramify heavily and selectively in the rostromedial pericoerulear region, the results of the present physiological studies suggest that PAG preferentially targets rostromedial pericoerulear neurons rather than LC dendrites.

Viscerotopic organization of neurons subserving hypotensive reactions within the midbrain periaqueductal grey: a correlative functional and anatomical study

Microinjection of the excitatory amino acid D,L-homocysteic acid (40 nmol, in 200 nl) made into the ventrolateral part of the caudal half (A2.5-P1.5) of the midbrain periaqueductal gray (PAG) of the decerebrate cat evoked a hypotensive reaction associated with a slowing of the heart and a decrease in either external iliac or renal vascular resistance. The decrease in iliac vascular resistance was elicited from the pretentorial portion (A2.5-A0.6) of the PAG hypotensive area, whereas the decrease in renal vascular resistance was elicited from the subtentorial portion (A0.6-P1.5). Anatomical experiments using the method of retrograde transport of rhodamine-labelled microspheres or wheat germ agglutinin-horseradish peroxidase demonstrated topographically organized projections from the ventrolateral PAG to the subretrofacial (SRF) pressor nucleus in the rostral ventrolateral medulla. The pretentorial part of the ventrolateral PAG projected mainly to the caudal part of the SRF nucleus, which preferentially controls iliac vascular resistance. The subtentorial part of the ventrolateral PAG projected mainly to the rostral part of the SRF nucleus, which preferentially controls renal vascular resistance. Taken together, these findings suggest: (i) that neurons within the ventrolateral PAG are viscerotopically organized; and (ii) that their hypotensive function may be mediated by an inhibition of SRF pressor neurons. The results are discussed in relation to the recently described PAG hypertensive area which also is viscerotopically organized and projects to the SRF nucleus.

Flight and immobility evoked by excitatory amino acid microinjection within distinct parts of the subtentorial midbrain periaqueductal gray of the cat

Unilateral microinjections of the excitatory amino acid, D,L-homocysteic acid (DLH) made in the lateral and ventrolateral parts of the subtentorial (A 1.0-P 1.5) midbrain periaqueductal gray (PAG) of the freely moving cat evoked two distinct patterns of coordinated somatic changes. When DLH injection (80 nmol) was made within the lateral part of the subtentorial PAG it evoked a flight reaction, characterized by strong locomotion (running) and multiple jumps. This flight reaction was quite distinct from the defensive threat display previously described following DLH microinjection in the lateral part of the pretentorial PAG. When DLH injection (80 nmol) was made in the subtentorial PAG region, ventrolateral to the aqueduct, it elicited a cessation of both spontaneous locomotion and general movements (e.g. licking, scratching, grooming, head and limb movements), a reaction termed immobility. The subtentorial PAG regions from which flight and immobility were evoked are seemingly identical to the lateral and ventrolateral subtentorial PAG regions in which hypertensive and hypotensive reactions have been evoked previously by DLH microinjection. The present results together with our previous studies suggest that: (1) the lateral PAG of the cat contains at least two, topographically separable neuronal pools, which mediate different types of defense reactions (i.e. threat display--lateral part of the pretentorial PAG; flight reaction--lateral part of the subtentorial PAG); and (2) excitation of neurons in the ventrolateral PAG alters autonomic and somatic functions, but in a direction opposite to that of lateral PAG neurons, namely decreased somatomotor activity and hypotension.

Electrical stimulation of the rat ventral midbrain elicits antinociception via the dorsolateral funiculus

The pain-suppressive effects of focal electrical stimulation of sites throughout the ventral midbrain were examined in awake rats. Chronic bipolar electrodes were implanted in medial and lateral regions of the midbrain. Current thresholds for suppression of the tail-flick reflex in response to noxious heat were determined for both a biphasic and a monophasic stimulation parameter at each site. Stimulation of areas throughout the ventral midbrain produced tail-flick suppression (TFS), but no one area was consistently effective in all animals. Monophasic and biphasic stimulation were qualitatively equal in the duration of TFS and the distribution of effective sites. The production of TFS was not correlated with other behavioral reactions to brain stimulation. TFS appeared to be mediated by non-opiate pathways since naloxone administration (10 mg/kg) had no discernible effect on the production of TFS. The current threshold for producing TFS was extremely variable over both short (one half hour) and long (one week) intervals. The incidence of TFS from previously effective sites was significantly less following bilateral dorsolateral funiculus (DLF) lesions, indicating that the antinociceptive effects of ventral midbrain stimulation are mediated by this spinal pathway.

Brain stem terminations of the trigeminal and upper spinal ganglia innervation of the cerebrovascular system: WGA-HRP transganglionic study

The central projections of the nerve fibers innervating the middle cerebral and basilar arteries were investigated by transganglionic tracing of wheat germ agglutinin conjugated with horseradish peroxidase (WGA-HRP) in the rat. WGA-HRP was applied to the exposed basilar and/or middle cerebral arteries. Sections of the brain, trigeminal and upper spinal ganglia were reacted with tetramethylbenzidine for detection of the tracer. The results demonstrate that trigeminal neurons that innervate the middle cerebral artery project to the trigeminal main sensory nucleus, pars oralis, and the dorsocaudal two-fifths of pars interpolaris of the trigeminal brain stem nuclear complex. Terminals were also visible in the ipsilateral nucleus motorius dorsalis nervi vagi (dmnX) and in the lateral nucleus tractus solitarius (nTs) bilaterally at the level of the obex. The ventral periaqueductal gray, including the dorsal raphe and C2 dorsal horn, were also innervated by nerve fibers from the middle cerebral artery. Ipsilateral trigeminal rhizotomy prior to WGA-HRP application over the middle cerebral artery impeded the visualization of nerve terminations throughout the brain stem. Pretreatment with capsaicin reduced the density of labeled neurons and terminals within the trigeminal ganglion and the brain stem, respectively, following WGA-HRP application over the middle cerebral artery. Basilar artery fibers terminate in the C2 dorsal horn, the cuneate nuclei, dmnX, and nTs bilaterally. A few projections were also labeled in the ventral periaqueductal gray. Unilateral upper two spinal dorsal rhizotomy prior to WGA-HRP application over the exposed basilar artery resulted in terminal labeling within the C2 dorsal horn, the cuneate nucleus, dmnX, and nTs contralateral to the rhizotomy, whereas the ipsilateral side was devoid of any labeling. Bilateral superior cervical ganglionectomy prior to WGA-HRP administration to the middle cerebral and basilar arteries did not alter the visualization of nerve terminations throughout the brain stem.

The effects of periaqueductal gray and nucleus raphe magnus stimulation on the spontaneous and noxious-evoked activity of lateral reticular nucleus neurons in rabbits

In urethane-anesthetized rabbits the effects of periaqueductal gray (PAG) and nucleus raphe magnus (NRM) stimulation on the spontaneous and noxious-evoked activity of the lateral reticular nucleus (LRN) neurons were studied. The PAG and the NRM stimulating electrodes were located in the optimal sites for suppressing the jaw-opening reflex (JOR) evoked by the tooth pulp stimulation. It was found that the 12% of neurons tested were affected by one or both stimuli. A total of 80 responsive neurons (52% antidromically activated by the cerebellum) were analyzed. Out of these neurons, 31 showed a convergence to both stimuli, 43 responded only to PAG and 6 only to NRM. Noxious heat stimulation of the contralateral foot was effective in altering the activity of 60% of these neurons. The PAG and NRM stimuli modified the noxious-evoked responses in most of these units. While the excitation was the predominant effect on the spontaneous activity (52 cells), the inhibition was predominant on the noxious-evoked activity (29 cells). These results indicate the presence of connections from PAG and NRM to LRN, probably devoted to the processing of the nociceptive information.

Inhibition of the responses of neurons in the rat spinal cord to noxious skin heating by stimulation in midbrain periaqueductal gray or lateral reticular formation

Single units responsive to noxious heating of glabrous hindfoot skin were recorded in the lumbar spinal cord of rats anesthetized with sodium pentobarbital. Unit responses to heat stimuli (e.g. 50 degrees C, 10 s) delivered at 2-min intervals were stable and were markedly suppressed during stimulation (100-ms pulse trains at 100 Hz, 3/s, 25-400 microA) in the midbrain periaqueductal gray (PAG) or lateral reticular formation (LRF). Inhibition did not appear to outlast the midbrain stimulation period. By systematically varying the position of an array of 3 or 5 stimulating electrodes, we observed that stimulation in PAG and subjacent tegmental areas, as well as in widespread regions of the LRF bilaterally, suppressed unit responses to noxious skin heating. The degree of suppression of unit responses increased with graded increases in PAG or LRF stimulation intensity. LRF appeared to be more efficacious than PAG stimulation, based on lower currents at threshold, as well as on significantly greater slopes in current-inhibition plots for LRF compared to PAG stimulation. Unit responses increased linearly with graded increases in stimulus temperature from 42 to 54 degrees C. Slopes of temperature-response lines were reduced during PAG stimulation with no change in response threshold, while temperature-response lines were shifted in a parallel manner toward higher temperatures during LRF stimulation with an increase in response threshold. The results suggest that differential inhibitory systems are activated by PAG or LRF stimulation and are discussed in relation to previous findings in the cat and as a possible mechanism of stimulation-produced analgesia.

Response of brainstem trigeminal neurons to electrical stimulation of the dura

The extracellular response of medullary trigeminal neurons to electrical stimulation of the dura was studied in anesthetized cats. Fifty-six medullary trigeminal units were excited by stimulation sites near major dural vessels with an average latency of 11.0 ms. Many units also responded to infraorbital nerve shock and had cutaneous receptive fields that included the ipsilateral periorbital region. These cutaneous responses were either wide dynamic range or nociceptive specific in type. Electrical stimulation of the midbrain periaqueductal gray region suppressed the response of medullary trigeminal units to either dural stimulation or infraorbital nerve shock. Medullary trigeminal neurons that receive convergent inputs from dura and facial skin may provide a physiological substrate for the cutaneous referral of dural sensation.

Hypothalamic inhibition of rat dorsal horn neuronal responses to noxious skin heating

The responses of single lumbar dorsal horn neurons to noxious radiant heat stimuli (42-54 degrees C, 10 sec, 1/2 min) applied to glabrous hind paw skin were recorded in rats anesthetized with sodium pentobarbital. Unit responses to 50 or 52 degrees C stimuli were constant over time and were consistently and powerfully inhibited during bipolar stimulation (three 100 msec trains/sec at 100 Hz, 200 microA) in the medial hypothalamus. Inhibition was also evoked by stimulation in medial and ventrobasal thalamic nuclei, lateral hypothalamus and adjacent cerebral peduncle, and amygdala. Inhibition increased with graded increases in intensity of hypothalamic stimulation, with a mean inhibitory threshold of 71 +/- 43 (S.D.) microA for 13 units. The responses of dorsal horn units to graded increases in the temperature of noxious heat stimuli were inhibited during hypothalamic stimulation, such that slopes of the linear temperature-response functions were reduced with no change in response threshold (mean: approximately 44 degrees C). Inhibition was blocked or reduced in 4/7 units following systemic administration of the 5-hydroxytryptamine (5-HT) antagonist methysergide. The results confirm and extend previous work in the cat and are discussed in relation to analgesic mechanisms.

Periaqueductal gray inhibition of trigeminal subnucleus caudalis unitary responses evoked by dentine and nonnoxious facial stimulation

The possible pain inhibitory effects of periaqueductal gray (PAG) stimulation were investigated in cats anesthetized with Nembutal and immobilized with Flaxedil. Unitary responses evoked by electrical stimulation of the upper canine dentine and by cutaneous facial noxious and nonnoxious stimuli were recorded extracellularly from the trigeminal subnucleus caudalis. A bipolar electrode was introduced into the PAG to test the effects of PAG excitation on the trigeminal response to dentine (TRED) and cutaneous nonnoxious stimulation. In some experiments, a similar electrode was lowered into the contralateral posterior thalamus to study the antidromic activation of subnucleus caudalis cells and the effects of thalamic stimulation on the TRED. Dentine stimulation evoked brief (6- to 15-ms) bursts of 1 to 10 spikes with 3- to 25-ms latencies. Most units (88%) were also activated by cutaneous facial stimulation. Stimulation of the posterior thalamus had no effect on the TRED or on responses to cutaneous stimulation, but activated antidromically 10% of the units. In 71% of the units PAG stimulation inhibited the TRED. In some of those cases (12%), the inhibitory effect persisted 30- to 60 s. The PAG stimulation could produce paradoxical effects, potentiating the TRED evoked by threshold intensity and inhibiting the TRED elicited by suprathreshold stimulation. About one-half the PAG points evoked detectable effects. Their location had no clear topographical distribution, although ventral sites were more potent than dorsal sites. Responses evoked by nonnoxious facial stimulation were also inhibited by the PAG.

The location of brainstem neurones tonically inhibiting dorsal horn neurones of the cat

In an investigation of the origin of tonic descending inhibition of dorsal horn neurones by impulses in unmyelinated primary afferents, brainstem regions were electrolytically lesioned. With each neurone studied, tonic descending inhibition was measured before and after brainstem lesions by cooling a segment of spinal cord cephalic to the recording site. Such inhibition was not reduced by lesions of areas which, when stimulated, produce analgesia. These included the periaqueductal grey and the raphé areas of the midbrain and pons-medulla. Tonic descending inhibition was reduced by bilateral lesions of the ventrolateral caudal medulla in the region of the lateral reticular nuclei. Lateral reticular areas may have a functional role in the control of pain.

Serotonergic mediation of descending inhibition from midbrain periaqueductal gray, but not reticular formation, or spinal nociceptive transmission in the cat

Electrical stimulation in the midbrain periaqueductal gray (PAG) and lateral midbrain reticular formation (LRF) strongly suppresses the responses of spinal dorsal horn neurons to noxious heating of the skin. The possible role of serotonin (5-hydroxytryptamine, 5-HT) was investigated by quantitatively comparing certain parameters of descending inhibition from PAG and LRF in normal cats [14,15] and cats whose central 5-HT levels had been reduced by pretreatment with p-chlorophenylalanine (PCPA, 300 or 500 mg/kg i.p., 72 h prior to acute experiment). Single lumbar dorsal horn neuronal responses to noxious radiant heating of glabrous footpad skin(50 degrees C, 10 sec, 1/3 min) were recorded in normal and PCPA-pretreated cats anesthetized with sodium pentobarbital and N2O. Inhibition of neuronal heat-evoked responses during midbrain stimulation (mean frequency 30 Hz, up to 800 microA current intensity) was expressed as percent of the unit's control response in the absence of midbrain stimulation. Inhibition by PAG stimulation of units from cats pretreated with 300 mg/kg PCPA (mean inhibition at 450 microA to 60% of control in 12 units) was not detectably different from that in control (non-pretreated) cats. However, inhibition by PAG stimulation was significantly weaker in units from cats pretreated with 500 mg/kg PCPA (mean to 83.4% of control in 9 units). In the latter group, mean current threshold for inhibition was higher, and slope of current-intensity plots lower, than in the control and 300 mg/kg PCPA pretreatment groups. In contrast, mean inhibition by LRF stimulation was enhanced in the 300 and 500 mg/kg PCPA treatment groups in a dose-related manner. In normal (non-pretreated) cats, systemic administration of the putative 5-HT antagonist methysergide (0.07--1 mg/kg) reduced or abolished inhibition by PAG stimulation in each of 8 units. Low doses of methysergide had little or no effect on inhibition produced by LRF stimulation in 6 units. The results suggest pharmacologically distinct mechanisms of inhibition produced by stimulation in PAG and LRF.