Central Nervous System Mechanisms of Pain Modulation

SUMMARY

In this chapter we present an overview of pain-modulating systems with a focus on the properties of a network with major relays in the midbrain periaqueductal gray (PAG) and rostral ventromedial medulla (RVM). This network exerts bidirectional control over dorsal horn nociceptive transmission by means of separate anti- and pro-nociceptive output from the RVM. The PAG–RVM system is the central substrate for the actions of opioid and non-opioid analgesic drugs, but it also contributes to enhanced nociception and pain in a variety of conditions, including inflammation, nerve injury, stress, and sickness.

The PAG–RVM system integrates bottom-up and top-down influences to modulate nociception. It receives direct and indirect input from nociceptive transmission pathways. It has strong links with the hypothalamus and limbic forebrain, including the amygdala, anterior cingulate cortex, and anterior insula, which provide a mechanism through which cognitive and emotional factors can regulate pain processing. The discovery of brain stem pain-modulating systems—as well as our increasing understanding of the behavioral and sensory triggers that recruit them and the neurotransmitters that regulate and mediate their actions—has provided significant insight into the variability of pain that is seen in different behavioral or environmental contexts. The recognition that a dynamic balance between inhibition and facilitation defines the relationship between afferent input and sensory experience offers the promise of rationally developed treatments based on manipulation of psychological variables and on new, more selective drugs or drug combinations.

Introduction

The relationship between a peripheral stimulus and the resulting pain experience depends on a host of variables, including behavioral context and competing sensory input. As an extreme example, traumatic injuries sustained during athletic competition or combat are often initially reported as being relatively painless, although in other circumstances similar injuries are typically extremely painful ( , ). The modification of neural, behavioral, and subjective pain responses by behavioral context and cognitive and emotional factors results from the action of central nervous system (CNS) networks that dynamically modulate the transmission of nociceptive messages. Interplay among various networks and the net balance between descending facilitatory and inhibitory influences allow a range of responses to a given triggering stimulus. Presumably, these pain-modulating circuits exist because the ability to suppress or enhance nocifensive reflexes and other responses normally elicited by noxious stimuli enhances survival of the individual. For example, suppression of nocifensive reflexes and pain sensation might facilitate escape and reduce distraction in the face of a threat such as a predator or aggressive conspecific. Conversely, enhanced pain in the presence of tissue injury and inflammation could promote recuperative behavior and healing. However, abnormal or sustained engagement of descending facilitation is now recognized to be a factor contributing to chronic and abnormal pain states.

Descending Modulatory Control

As early as explicitly postulated modulatory influences on pain. They proposed that the thalamus is the center for the perception of pain and that the neocortex, the discriminative perception center, continuously modulates the responses of the thalamus to noxious stimuli. In their view, modulation of pain was a necessary part of the ongoing process of discriminative sensation. The first direct evidence that supraspinal sites control ascending (presumably sensory) pathways was provided by . subsequently demonstrated descending control of sensory input to the ascending pathways. However, the concept of a specific pain modulatory system was first clearly articulated by in their Gate Control Theory of Pain. Although there was limited evidence at that time for descending control of nociception, the next 10 years saw clear evidence of tonic descending inhibition of nociresponsive neurons in the dorsal horn ( ).

Stimulation-Produced Analgesia

The seminal observation pointing to selective pain-modulating systems was the discovery of “stimulation-produced analgesia,” which is a highly specific suppression of behavioral responses to noxious stimuli produced by electrical stimulation of the midbrain periaqueductal gray (PAG). During PAG stimulation, animals remained alert and active, and their responses to most environmental stimuli were unchanged. However, the expected responses to noxious stimuli, including orientation, vocalization, and escape, were absent. Stimulation-produced analgesia was also elicited from the rostral ventromedial medulla (RVM, ).

A critical step was the demonstration that stimulation-produced analgesia can be induced in humans. As in animals, stimulation-produced analgesia in humans is elicited by electrical stimulation of the PAG and more rostral periventricular structures. Importantly, motor function is not affected, and patients report subjective analgesia (see reviews by , ). Although this procedure is rarely performed at present, the specificity of the analgesic effect and the fact that it is consistently elicited from discrete brain sites that are homologous in a variety of species, including humans, speak to the biological importance of pain modulation.

Discovery of the pain-modulating role of the PAG was a decisive advance in understanding brain mechanisms of pain processing. Subsequent research demonstrated that the PAG is part of a central circuit that controls nociceptive transmission at the level of the spinal cord dorsal horn via a relay in the RVM ( Fig. 8-1 , ). In animals, simple noxious stimulus–evoked reflexes, considered to be organized primarily at the spinal level, are inhibited by stimulation of either the PAG or RVM. Furthermore, this inhibition is selective for nociceptive neurons in the spinal cord dorsal horn ( ). In addition, lesions of the descending projections of the RVM through the dorsolateral funiculus of the spinal cord block inhibition of both nociceptive dorsal horn neurons and nocifensive reflex responses. Importantly, suppression of noxious evoked behavior is not motor inhibition since affective responses are also modulated by PAG and RVM manipulation ( , , , ). Interest in this system was further heightened when it became clear that it is recruited by opioid analgesic drugs ( ).

Figure 8-1, Brain stem pain-modulating pathways.

Although other brain stem pain-modulating circuits will be mentioned (pontomedullary noradrenergic pathways, feedback via the nucleus reticularis dorsalis), this chapter emphasizes the PAG–RVM system. This focus is motivated by the demonstrated significance of this system in the actions of centrally acting analgesic drugs, its role in environmental analgesia, and evidence that it contributes to hyperalgesic states associated with inflammation, nerve injury, immune system activation, and stress.

PAG–RVM System and Facilitation of Pain Transmission

The seminal observation of stimulation-produced analgesia from the PAG and RVM and the recognition shortly thereafter that this network is recruited as part of the analgesic actions of opioids led to a widely held view of this circuit as an “analgesia system.” However, functional studies have clearly shown that brain stem modulatory systems exert bidirectional control and that pain facilitation is a major part of their function (for reviews see , , ).

Functional studies have implicated the RVM in hyperalgesia and chronic pain triggered through both “bottom-up” and “top-down” mechanisms. Thus the RVM contributes to hyperalgesia and allodynia in inflammatory and neuropathic pain models ( , ). Furthermore, facilitation of output from the RVM generates a tonic aversive state in rodents that can be demonstrated with place conditioning ( ). The RVM has also been implicated in deep muscle and visceral hypersensitivity and in headache-related pain ( , , , ). It may be a factor in post-surgical hypersensitivity, although this is apparently relatively minor ( , ). The examples just listed illustrate that descending facilitation is engaged as part of a positive feedback loop stimulated by noxious input. Top-down facilitating modulation is also ubiquitous. The RVM is required for hyperalgesia associated with naloxone-precipitated withdrawal or prolonged opioid administration ( , , ). The RVM is also engaged to produce hyperalgesia in models of mild or chronic stress ( , , ) and as part of the sickness response triggered by systemic immune activation ( , ). In these top-down examples, input from higher centers such as the amygdala and hypothalamus brings the PAG–RVM system into play to produce hyperalgesia.

Although relatively few studies have indicated a direct facilitating influence arising from the PAG, the observations outlined above indicate that the PAG–RVM system as a whole exerts true bidirectional control of nociceptive processing. Facilitation operates in parallel with descending inhibition, with differential control of input from various body regions or different modalities ( , , ). Discovering how this system is recruited to either inhibit or facilitate nociception under different physiological conditions and in different behavioral contexts is an important challenge for future understanding of descending modulation.

Organization of the PAG–RVM System

The influence of the PAG on the dorsal horn is relayed through the RVM (see Fig. 8-1 ). However, both the PAG and RVM receive significant input from higher structures, thereby providing a pathway through which cognitive and emotional factors can control pain processing. The PAG–RVM system also receives somatosensory, including nociceptive, information via the spinomesencephalic and spinoreticular tracts. This system is thus positioned to integrate top-down and bottom-up influences on pain transmission.

Ascending projections from both the PAG and RVM, though not as well studied as the descending pathway, have the potential to influence cortical processing of nociceptive input ( , ).

Connectivity of the PAG

The PAG receives direct projections from a number of medial prefrontal cortical areas, including the anterior cingulate, as well as from the agranular insular cortex ( , , ). The amygdala, a forebrain structure critically involved in processing emotionally significant input, is another major source of afferents to the PAG ( ). Cortical afferents to the amygdala largely target its basolateral component. The basolateral amygdala then projects to the central nucleus, which in turn projects densely to the PAG ( ). The central nucleus of the amygdala also receives nociceptive input, both directly from the spinal cord ( , ) and indirectly via a large projection from dorsal horn lamina I to the parabrachial nuclei ( ). The influence of the amygdala on pain is mediated through its connections with the PAG–RVM system ( , , , , , , , , ).

The hypothalamus and preoptic area also constitute a major source of afferents to the PAG ( , ). Manipulation of various hypothalamic regions can produce analgesia ( , , ) or hyperalgesia ( , , ).

Brain stem input to the PAG arises from the adjacent nucleus cuneiformis, the pontomedullary reticular formation, the locus coeruleus and other catecholaminergic cell groups, and the RVM ( , ). The PAG and adjacent nucleus cuneiformis also receive a significant projection from the dorsal horn, including spinal lamina I nociceptive neurons ( , ). The PAG can thus integrate input from the limbic forebrain and diencephalon with ascending nociceptive input from the dorsal horn ( ).

A major output from the PAG is to the RVM. Since the PAG itself projects only minimally to the spinal cord, this connection is critical for descending pain modulation. Thus, anatomical lesions, reversible inactivation with lidocaine (lignocaine), or microinjection of excitatory amino acid receptor antagonists into the RVM abolishes the analgesia produced by stimulation of the PAG ( ).

The PAG also projects rostrally to the medial thalamus and orbital frontal cortex and provides a possible substrate for ascending control of nociception ( , ). PAG projections to the dorsolateral and ventrolateral pontine tegmentum and to the ventrolateral medulla are likely to be involved in other aspects of PAG function, such as autonomic regulation ( ).

Connectivity of the RVM

Output from the RVM

The RVM includes the midline nucleus raphe magnus and the adjacent reticular formation ventral to the nucleus reticularis gigantocellularis. The major output of the RVM is to the spinal cord and includes serotonin-containing neurons of the B3 cell group, γ-aminobutyric acid (GABA), and several neuropeptides (see for review). The spinal terminals of RVM descending axons are most dense in dorsal horn laminae I, II (the substantia gelatinosa), and V. These laminae are targets of nociceptive primary afferents and include many nociresponsive neurons ( and see Chapter 5 ). Furthermore, activation of the RVM by electrical stimulation can exert both inhibitory and excitatory effects on nociceptive neurons in these laminae, including primate spinothalamic tract neurons ( ).

The details of synaptic connections of RVM terminals in the dorsal horn remain to be worked out, but there is morphological and electrophysiological evidence of a host of actions, including direct post-synaptic inhibition of projection neurons, inhibition of release of transmitters from primary afferents, excitation of inhibitory interneurons, and inhibition of excitatory interneurons ( ). Lumb and colleagues (see review by ) demonstrated differential regulation from the brain stem of dorsal horn wide–dynamic range neurons with and without significant C-fiber input. Dorsal horn neurons without significant C-fiber input were facilitated by brain stem stimulation. Lumb and colleagues argue that facilitation of these neurons results from descending inhibition of the neurons with C-fiber evoked responses (i.e., which function as inhibitory interneurons). However, other groups have found that descending inhibition and facilitation can have distinct time courses, with facilitation preceding inhibition ( ), and as outlined below, descending facilitation and inhibition from the RVM are exerted by different RVM cell populations. These latter findings imply that descending facilitation and inhibition are parallel and at least partially independent.

Like the PAG, the RVM has ascending projections, primarily non-serotonergic ( ). Targets include the hypothalamus, preoptic area, and central nucleus of the amygdala ( , ). The function or functions of these connections are unknown.

Input to the RVM

The PAG and the adjacent nucleus cuneiformis are a major source of input to the RVM. A substantial proportion of PAG afferents to spinally projecting RVM neurons are GABAergic, although non-GABAergic connections also exist ( ), including serotonin and neurotensin ( ). A large number of enkephalin and substance P neurons are found in the PAG, but few of these neurons project directly to the RVM ( ). Other brain stem afferents to the RVM include the parabrachial nuclei, Kölliker–Füse and intertrigeminal region, and rostral ventrolateral medulla. The RVM also receives substantial direct projections from higher centers implicated in emotion and autonomic regulation, including the hypothalamus, preoptic area, bed nucleus of the stria terminalis, central nucleus of the amygdala, and infralimbic and prelimbic cortex ( , , ).

Direct projections from the dorsal horn to the RVM are believed to be relatively sparse but have been demonstrated ( ). Additional possible sources of spinal information include the PAG and nucleus cuneiformis, as well as the medullary nucleus reticularis gigantocellularis, which receive direct spinomesencephalic and spinomedullary connections and project to the RVM.

Pharmacology of the PAG–RVM Pain-Modulating Network

Opioids (see also Chapter 30 )

Both the PAG and the RVM support the analgesic effect of μ-opioids ( ). That is, direct focal application of morphine or selective μ-receptor agonists in either region produces antinociception comparable to that seen following systemic administration. Conversely, inactivation of or microinjection of opioid antagonists into these sites reduces the analgesic effect of systemically administered opioids. μ-Opioid receptor agonists also act directly at the output target of the PAG–RVM system, in the spinal cord dorsal horn. With cloning of the three classic opioid receptors (μ, δ, and κ), discovery of a fourth related receptor (ORL1), and generation of selective antibodies for each, it became possible to map their distributions in the CNS. Each receptor is present in both the PAG and RVM, as well as in the spinal cord dorsal horn ( , ).

The effects of ligands selective for δ, κ, and ORL1 receptors are state dependent. For example, κ-opioid receptor agonists can have either an analgesic or anti-analgesic effect when microinjected into the RVM, depending on the test and the sex of the animal ( , ). δ-Opioid receptor agonists have a modest analgesic effect when microinjected into the PAG and RVM under basal conditions (see for review) but show significant antinociceptive potency during chronic inflammation ( ). Injection of the ORL1 agonist nociceptin (orphanin FQ) into the RVM interferes with the antinociceptive effect of μ-opioid agonists administered systemically or microinjected into the PAG, but it also attenuates the hyperalgesia associated with acute opioid withdrawal ( , ). This apparent paradox can be understood in the context of the RVM circuitry and is explained by the fact that nociceptin inhibits the firing of both pain-inhibiting and pain-facilitating neurons in the RVM ( ).

Endogenous opioid peptides also play an important role in the PAG–RVM system. The contribution of endogenous opioid peptides to pain modulation was first suggested by reports that stimulation-produced analgesia in animals and humans is reduced by the narcotic antagonist naloxone. Naloxone also worsens postoperative pain in patients who have not received exogenous opioid therapy, thus establishing the relevance of endogenous opioids to common clinical situations (see for review).

The different levels of the PAG–RVM circuit are linked in part through the release of endogenous opioids. Thus the antinociceptive actions of microinjected opioids (or electrical or chemical stimulation) at one site can be blocked by microinjection of opioid antagonists at a downstream site in the pathway. For example, the antinociceptive effect of a μ-opioid receptor agonist microinjected into the basolateral amygdala is reversed by injection of a μ-opioid receptor antagonist into the PAG ( ). Similarly, antinociceptive effects elicited from the PAG can be blocked by naloxone or selective μ-opioid receptor antagonists microinjected into the RVM ( , ). Moreover, endogenous opioid links are required for the antinociception triggered during conditioned fear, which is mediated by the PAG–RVM system ( , , ).

These data demonstrate that the PAG–RVM system is linked through the release of endogenous opioids and indicate that exogenous opioids produce their effects not only by direct binding to opioid receptors in the PAG–RVM circuit but also indirectly by triggering release of endogenous opioids acting at the same receptors. This may be a critical factor in the potentiated action of exogenous opioids during inflammation ( , ).

You're Reading a Preview

Become a Clinical Tree membership for Full access and enjoy Unlimited articles

Become membership

If you are a member. Log in here