Endocannabinoids Can Open the Pain Gate

+ See all authors and affiliations

Sci. Signal.  15 Sep 2009:
Vol. 2, Issue 88, pp. pe57
DOI: 10.1126/scisignal.288pe57


Endocannabinoids produced in the spinal cord can enhance pain by dampening the synapses of inhibitory interneurons that usually prevent the perception of innocuous stimuli as painful. This mechanism promotes pain responsiveness to normally innocuous mechanical stimuli in the skin surrounding a site of injury in both animals and humans subjected to sustained stimulation of pain-sensing nerves. The pain-promoting action of endocannabinoids wanes during the development of chronic pain that is induced by inflammation or nerve injury. This finding may partially explain why, in human trials, cannabinoid drugs have been negative for treatment of most types of acute and postsurgical pain but are effective for some chronic pain states.

Cannabis and cannabinoid drugs have an uncertain place in pain management. Animal studies consistently predict that cannabis and cannabinoids should relieve both acute and chronic pain states with an efficacy comparable to that of the gold standard, morphine (1, 2). Although cannabis has historically been recommended for pain (3), human trials have shown cannabinoids to be largely ineffective in relieving acute pain, and they can even enhance some pain responses (1, 410). Results are more promising regarding chronic pain, particularly neuropathic pain from nerve damage, although mixed outcomes have also been reported (1, 10, 11). In contrast to animal studies, the psychotropic side effects of cannabinoids in humans limit the doses that can be used to treat pain (1, 10, 11), and thus other explanations for poor analgesic efficacy may have been overlooked. Pernia-Andrade et al. (12) present compelling evidence for a pain-promoting mechanism of cannabinoid signaling in animals subjected to intense noxious stimuli. They report that cannabinoid drugs and endocannabinoids generated in the spinal cord can impair inhibitory control of pain-sensing neurons, thereby opening a “pain gate” that enhances neurotransmission of both painful and innocuous mechanical stimuli through “pain pathways” to higher centers in the brain.

These findings are provocative because most research over the past decade has focused on understanding the mechanisms of the pain-relieving properties of cannabis. In terms of what is already known about cannabinoid actions in the spinal cord, these findings imply differential modulation of distinct pain-sensing modalities by endocannabinoids: Primary pain pathways that transduce thermal pain sensations are inhibited, whereas pain sensitivity to both noxious (hyperalgesia) and innocuous (allodynia) mechanical stimuli in the area surrounding a primary injury site are simultaneously facilitated.

The findings of Pernia-Andrade et al. (12) are consistent with the known organization of cannabinoid signaling mechanisms in the dorsal horn of the spinal cord (Fig. 1). Primary afferent, pain-sensing neurons include small-diameter unmyelinated C fibers that largely transduce thermal pain and myelinated Aδ fibers that transduce mechanical pain sensation (13). Electrophysiological studies have established that activation of the type 1 cannabinoid (CB1) receptor, which is the major cannabinoid receptor found on nerves, selectively inhibits neurotransmission from C fibers onto pain transmission neurons in the superficial dorsal horn, whereas transmission by Aδ or Aβ fibers (Aβ fibers transduce innocuous mechanical sensation) is largely unaffected (12, 1416), even though CB1 receptors are found on a range of sensory neurons (1719).

Fig. 1

(A) Pain transmission neurons (green) in the spinal cord receive direct glutamatergic synaptic input (pink circles) from pain-sensing nerves (yellow and purple). Nerves that transmit innocuous mechanical stimuli (blue) synapse with inhibitory (gray) neurons that use glycine and GABA as neurotransmitters and with excitatory (green) interneurons, both of which innervate pain transmission neurons. Because innocuous stimuli simultaneously activate connected inhibitory and excitatory interneurons, the transmission of innocuous information to pain pathways is blocked. (B) When pain nerves receive intense, prolonged stimulation, excessive release of glutamate activates mGluRs to promote the formation of endocannabinoids that reduce inhibitory synaptic transmission onto all excitatory neurons. This opens the gate for innocuous mechanical stimuli to excite pain transmission neurons and produce pain.

Although often overlooked, CB1 receptors are also found on inhibitory interneurons in the dorsal horn (12), and their activation inhibits release of the neurotransmitters γ-aminobutyric acid (GABA) and glycine onto pain transmission neurons and excitatory interneurons in the spinal cord (12, 14). Dorsal horn neurons that use both glycine and GABA as neurotransmitters (glycinergic/GABAergic neurons) are activated by innocuous mechanical stimuli, such as light touch, and normally prevent the activation of pain transmission neurons by non-noxious information (20, 21). Moreover, the loss of glycinergic/GABAergic inhibition produces hyperalgesia and allodynia in pain states (2024).

Pernia-Andrade et al. (12) have directly shown that endogenously generated cannabinoids (endocannabinoids) impair neurotransmission from glycinergic/GABAergic synapses onto excitatory neurons in the dorsal horn to enhance pain. Endocannanabinoids, including 2-arachidonyl glycerol (2-AG) and anandamide, are lipid mediators synthesized “on demand” by strongly depolarizing neurons or the activation of group I metabotropic glutamate receptors (mGluR1 or mGluR5) (25, 26). Endocannabinoids are short-range retrograde messengers that depress neurotransmission in many regions of the central nervous system, particularly at inhibitory synapses (25, 26). Using mice that express enhanced green fluorescent protein (EGFP) exclusively in glycinergic neurons in the spinal cord (27), Pernia-Andrade et al. (12) showed that strong depolarization of excitatory, but not inhibitory, neurons, as well as activation of mGluR1 and mGluR5, can induce CB1 receptor–mediated inhibition of glycinergic/GABAergic neurotransmission. Thus, endocannabinoids are produced on demand by strong stimuli in excitatory neurons and signal in a retrograde fashion to inhibitory nerve terminals in the superficial dorsal horn.

The authors next injected mice intradermally with capsaicin to induce intense, persistent activation of pain-sensing neurons. Capsaicin is the active ingredient in hot chili peppers that selectively activates excitatory (TRPV1) ion channels, which are predominantly found on C-fiber, nociceptive neurons (28), to produce prolonged mechanical hyperalgesia after intradermal injection. Supporting their in vitro findings, both the electrophysiological signature of capsaicin-induced mechanical hyperresponsiveness in neurons deep in the dorsal horn (Fig. 1) and the sensitization to normally innocuous mechanical stimuli in the skin adjacent to the capsaicin injection site were blocked by the direct application of antagonists against mGluR1 and mGluR5 or CB1 receptors to the dorsal spinal cord. These actions of endocannabinoids were due solely to the activation of CB1 receptors on glycinergic/GABAergic interneurons because they were abolished in mice lacking CB1 receptors specifically in inhibitory (ptf1a-CB1−/− mice) spinal interneurons but not in nociceptive sensory neurons (sns-CB1−/− mice). In contrast, direct application of an agonist with affinity for both CB1 and CB2 receptors to the spinal cord in vivo produced a modest mechanical sensitization. Thus, CB1 receptors found on inhibitory glycinergic/GABAergic neurons are both necessary and sufficient for the pain-promoting, endocannabinoid-mediated actions of capsaicin.

The authors have also provided evidence that endocannabinoid signaling in humans mediates secondary hyperalgesia during prolonged electrical stimulation of C fibers. As would be expected from the animal studies, administration of the CB1 receptor antagonist rimonabant in humans inhibited secondary hyperalgesia and allodynia but had no effect on acute pain ratings. Conversely, similar studies in humans have reported that cannabinoid agonists have no effect on acute pain or are mildly hyperalgesic after painful electrical stimulation (5), capsaicin injection, or mild burns (4, 6).

It is difficult to reconcile the in vivo electrophysiological and behavioral findings of Pernia-Andrade et al. (12) with the extensive literature that has almost universally reported pain-relieving actions of cannabinoids (1, 2). One explanation may lie in the very specific behavioral endpoint, secondary hyperalgesia and allodynia, which is likely to be a hallmark of inhibitory interneuron dysfunction. If other sensory modalities such as thermal pain or spontaneous pain (29) had been examined, then different results might have been observed. The inhibition of acute thermal pain by CB1 receptor activation has been widely reported (1, 2), and CB1 receptor activation inhibits the release of glutamate from excitatory C fibers (15) that encode thermal pain (13). Other major signaling elements of the endocannabinoid system have also been identified in primary afferent synapses in the spinal cord, including mGluR5 receptors that are positioned in close proximity to the 2-AG synthesizing enzyme diacylglycerol lipase (30). Indeed, Nyilas et al. (30) reported endocannabinoid-mediated thermal analgesia after either spinal activation of mGluR5 receptors or footshock stress; however, they did not examine mechanical stimuli. In addition, CB1 receptors on pain-sensing nerves outside the spinal cord appear to be at least partly responsible for cannabinoid-mediated pain relief in animal models (17). The most parsimonious explanation of these divergent findings is that endocannabinoid release in the spinal cord can directly inhibit primary afferent C-fiber synapses to suppress thermal pain while simultaneously enhancing secondary mechanical hyperalgesia through inhibition of glycinergic and GABAergic synapses. Clarification of this discrepancy will require simultaneous investigation of both primary (thermal) and secondary mechanical responses in pain models that intensely stimulate C fibers, such as surgical incision models. It will also be important to examine the clinical efficacy of cannabinoids more carefully in pain conditions with ongoing or sporadic nociceptor activity associated with secondary hyperalgesia, as these may be unaffected or exacerbated by cannabinoid agonists.

Secondary hyperalgesia and allodynia facilitates protection of the site of acute injury by sensitizing the surrounding area to pain. In many chronic pain states, however, secondary hyperalgesia and allodynia become pathological. Pernia-Andrade et al. (12) found that in chronic pain models, an agonist of CB1 and CB2 receptors, but not CB1 antagonists or the deletion of CB1 receptors, reversed mechanical pain, a finding that is consistent with other studies (1, 2). It is not clear why the pro-nociceptive effects of cannabinoids are lost and cannabinoid analgesia retained or even enhanced in the persistent pain models. Adaptations to the properties of inhibitory synapses in the spinal cord could be pivotal and explain why cannabinoid drugs are effective in some chronic pain states but not others. We do know that inhibitory neuron activity is depressed in models of chronic pain, both in vitro (2224) and in vivo (31, 32). The adaptations that impair inhibitory synapses in chronic pain may render cannabinoid inhibition of glycinergic/GABAergic synapses ineffective or, alternatively, analgesic rather than pro-nociceptive. The latter possibility is consistent with the finding that glycinergic/GABAergic synapses switch from an inhibitory to an excitatory nature in chronic pain models because of disruption of the transmembrane chloride gradient (22, 23). In pain conditions where glycinergic/GABAergic synapses become excitatory, cannabinoids would be expected to dampen the aberrant excitatory activity generated by mechanical stimuli to produce analgesia. Although the contributions of endocannabinoids to chronic pain mechanisms in peripheral (17, 29) and central (33) nerves are more complex than discussed here, the study of Pernia-Andrade et al. (12) provides a more rational framework to examine both anti- and pro-nociceptive mechanisms of cannabinoids in pain states.


M.J.C. and C.M. are supported by National Health and Medical Research Council of Australia (NHMRC) program grant 351446. M.J.C. is also supported by NHMRC fellowship 511914.

References and Notes

View Abstract

Related Content

Navigate This Article