A role for presynaptic NMDA receptors in central sensitization in the spinal cord dorsal horn?

Editor,—Buerkle and colleagues report that intrathecal administration of remifentanil in rats that had received intraplantar formalin could wholly abolish nociceptive behaviour during phase 1 of the formalin test but was associated with only partial inhibition of glutamate release as measured by microdialysis.1 Intraplantar injection of formalin results in a biphasic behavioural response, the second phase of which is considered to represent NMDA receptor-mediated central sensitization.2 3 Phase 1 is characterized by paw flinching and licking corresponding to an acute barrage of small primary afferent fibre discharge and is followed, after a brief quiescent period, by a more protracted phase 2 period of similar behaviour associated with a lower level of tonic afferent activity.4 Buerkle and colleagues contend that “supramaximal doses of intrathecal remifentanil sufficient to inhibit phase 1 behaviour still permit sufficient glutamate release to permit spinal facilitation” via presumed activation of a post-synaptic NMDA receptor during phase 1. Their data merit further discussion in the light of recent anatomical and functional demonstrations of presynaptic NMDA receptors in the spinal cord dorsal horn.5 6 The spinal action of remifentanil is mediated by -opioid receptors at both pre and postsynaptic loci within the dorsal horn, and as shown in this and previous studies,3 opioid administration can abolish both phase 1 and phase 2 of the formalin test. However, where opioid effects are limited purely to phase 1 by intrathecal naloxone7 8 or by the use of remifentanil,1 9 only a partial suppression or delay of phase 2 behaviour has been reported despite abolition of the phase 1 behavioural response. Therefore, sensitization still occurs during phase 1 of the formalin test, and crucially, such sensitization may occur despite electrophysiological evidence indicating that opioid suppression of dorsal horn neurone firing during phase 1 may be virtually complete.10 This “remifentanilresistant” component of central sensitization of the current study was considered by Buerkle and colleagues to be a consequence of incomplete opioid-mediated inhibition of glutamate release from primary afferent or interneurone nerve terminals during phase 1, similar to that reported previously with systemic morphine.11 If the initiation of central sensitization underlying phase 2 of the formalin test involves postsynaptic NMDA receptor activation, it would be expected to require sustained depolarization of nociceptive neurones in the dorsal horn.12 How can this be reconciled with the behavioural evidence of suppression of postsynaptic activity in this and other studies, and electrophysiological evidence of suppression of dorsal horn neuronal activation during phase 1 of the formalin test?10 13 Functional presynaptic NMDA receptors which potentiate the release of C-fibre primary afferent transmitters have been recently identified within the dorsal horn.5 6 It may be hypothesized that the enhanced release of glutamate during phase 1 demonstrated by Buerkle and colleagues may activate such presynaptic receptors which are depolarized by virtue of phase 1 afferent activity. This activation of presynaptic NMDA receptors would initiate a longlasting Ca2 -mediated enhancement of primary afferent transmitter release and increase the gain of C-fibre synaptic transmission in the dorsal horn during phase 2. Fundamentally, presynaptic sensitization could occur despite opioid suppression of postsynaptic neuronal activity and would only become manifest behaviourally after opioid suppression of postsynaptic neuronal activity within the dorsal horn was removed. Support for this hypothesis remains circumstantial in the absence of a selective antagonist of presynaptic NMDA receptors. However: