For information about subscribing or purchasing offprints of the published version, with commentaries and author's response, write to: journals_subscriptions@cup.org (North America) or journals_marketing@cup.cam.ac.uk (All other countries).
Opioids; analgesia; nociception; spinal cord; peptides; excitatory amino acids; hypersensitivity; cholecystokinin; development of pain
The spinal mechanisms of action of opioids under normal conditions are reasonably well understood. Of importance to the control of different pain states is the realization that the spinal effects of opioids can be enhanced or reduced depending on pathology and activity in other segmental and non-segmental pathways. This account will consider this plasticity in relation to the control of different pain states using opioids as a theme.
The complex and contradictory findings on the supraspinal actions of opioids are explicable in terms of heterogeneous descending pathways to different spinal targets using multiple transmitters and receptors - opioids can therefore both increase and decrease activity in descending pathways. One could envisage that these pathways exhibit considerable plasticity.
There is increasing evidence that delta opioid receptor agonists have the potential to replace morphine as major analgesics with reduced side-effect profiles.
The concept of pre-emptive analgesia, based on preventing the induction of some of the negative plastic influences on opioid controls and the detrimental effects of pain, is correct but experimental verification of the hypothesis in the clincial setting is difficult. One of the reasons is a delayed compensatory upregulation of inhibitory systems, particularly in inflammation, which may counter persistent painful inputs. Combination therapy with opioids may be beneficial in many pain states where either negative influences are blocked or inhibitory controls enhanced.
Finally, the developmental aspects of these systems is discussed with regard to the treatment of pain in young children, where inhibitory systems in the spinal cord are immature..
PLASTICITY IN PAIN
The two major types of clinical pain arise from distinct events in the periphery. Inflammatory pain arises from tissue damage such as that produced by trauma, surgery, childbirth or invasion of tissue by a tumour. The second type of pain is termed neuropathic pain and results from damage to a nerve; trauma, surgrey and cancer can also cause this type of pain. In the case of inflammatory pain, the damage to tissue causes the local production of a number of chemical mediators which sensitize and/or activate the peripheral endings of nociceptive C-fibres (Dray, Urban & Dickenson, 1994). In the case of neuropathic pain, activity is set up in the nerve itself. Whatever the origin of the pain, impulses are conveyed to the first synapse in the dorsal horn of the spinal cord where the interplay between excitatory and inhibitory events determines the ascending messages which are transmitted to higher centres. However, descending controls from the brain stem can be triggered and further alter processing in the spinal cord (Besson & Chaouch 1987, Fields, Barbaro & Heinrecher 1987). Opioid analgesics can exert controls on these events by direct actions on the spinal cord but can also interact with systems at the origins of the descending controls (Duggan & North 1984, Dickenson 1994a). Thus at a number of levels there is the potential for alteration in the messages that ultimately give rise to the final sensation of pain. This article is an attempt to bring together some of the interactions between excitatory and inhibitory events to explain some of the different characteristics of inflammatory and neuropathic pain and to investiagte how alterations in these systems can give rise to difficulties in treating certain pains, especially neuropathic pains.
As a basis for this it is well established that the repertoire of the adult central nervous system is not fixed and immutable. Plasticity, the ability of central nervous function to change in response to internal and external events, can be due to alterations in connectivity ( Woolf & Doubell 1994); (Dray, Urban, & Dickenson 1994). However, of great relevance to the control of these different pain states (McQuay & Dickenson 1990), plasticity can also result from a relatively rapid induction and activation of different pharmacological systems under different circumstances (Dickenson 1994a,b); (McMahon, Lewin & Wall 1993); (Price, Mao & Mayer 1994); (Woolf 1994). This account will concentrate on interactions between spinal pharmacological systems and opioid analgesia. States of increased pain transmission, central hypersensitive states (Woolf 1983), can result from activation of spinal systems which do not participate in the responses to brief stimuli. In these cases where the level of excitatory transmission is augmented opioid inhibitions will need to be increased to compensate. Consequently, the actions of opioids are not fixed but highly dependent on activity in other transmitter systems which in turn appear to be influenced by different types of pain. This account will consider this plasticity in relation to the control of different pain states using opioids as a theme. The main emphasis will be on results from models of acute and more prolonged pains such as that arising from inflammation and neuropathy and will attempt to discuss why the actions of opioids may differ in different pain states. Thus, pain arising from tissue damage (inflammatory nociceptive pain) can respond well to opioids whereas neuropathic pain (arising from peripheral or central nerve damage) and allodynia (where touch is perceived as pain) can show poor opioid sensitivity.
OPIOID RECEPTORS
Opioids act by activating three opioid receptors; the mu, the delta and the kappa ( Kosterlitz 1985). The endogenous opioid peptides, the natural ligands for the receptors, namely the enkephalins, dynorphins and endorphin family are not entirely specific for any one of these receptors but a number of synthetic agents with high selectivity are available to study the individual receptors (Kosterlitz 1985); (Dickenson 1994a). Neurones producing the different opioids can be now unequivocally identified, using messenger RNA probes for the precursor propeptides. Opioid peptide synthesis can be altered in animal models of different pain states. For example, dynorphin levels in the spinal cord during inflammation are increased enormously due to the switching on of the gene for the synthesis of the parent propeptide (Dubner & Ruda 1992). The biological life-times of the endogenous opioids, particularily the enkephalins, are brief due to rapid peptidase degradation. It is now possible to protect the enkephalins from breakdown by the use of peptidase inhibitors, some of which are now active by systemic routes (Roques, Noble, Dauge, Fournie-Zaluski & Beaumont 1993). The use of these agents together with the synthesis of stable analogues of the endogenous opioids themselves with far greater selectivity than the endogenous opioids has provided the means for the study of the roles and function of opioids and their receptors.
MECHANISMS OF OPIOID ANALGESIA
There are three key mechanisms of action of opioids. The underlying events by which opioids interfere with the transmission of pain are the same as the mechanisms by which opioids cause their other actions including side-effects (Duggan & North 1984); (Dickenson 1994a). These are:
a. A pre-synaptic action on the terminals of neurones whereby transmitter release is reduced by activation of opioid receptors. In tissues where relative receptor location has been gauged the number of presynaptic opioid sites predominates over post-synaptic locations.
b. There are significant numbers of post-synaptic opioid receptors and after activation the resultant hyperpolarisation reduces evoked activity in the neuronal pathways. The post-synaptic effects can be on cell bodies of output neurones, interneurones or dendrites.
c. An alternative post-synaptic action involves disinhibition so that in a circuit of two neurones, where the second cell is held in check by an inhibitory neurone, by opioid inhibition of the first neurone allows the second cell to become active.
These actions occur at a number of sites in the nervous system to produce analgesia. The two key sites would be appear to be a spinal and a midbrain/brain stem action in normal circumstances but an additional peripheral site in inflammed tissue can also be induced.
SPINAL ANALGESIA
Opioid receptors in the spinal cord are a critical site in the production of analgesia. Spinal opioid analgesia demonstrates how basic research in animals can have a rapid and important application to the clinical relief of pain. Opioid inhibition of nociceptive neurones in spinal animals and then evidence for analgesia following epidural and intrathecal opioids in animals was soon followed by clinical usage (Yaksh & Nouiehed 1985); (Besson & Chaouch 1987).
PRE-SYNAPTIC ACTIONS
The highest levels of opioid receptors in the spinal cord are around the C-fibre terminal zones in lamina 1 and the substantia gelatinosa with lower levels found in deeper layers. The best current estimates suggest that the mu receptor forms 70%, the delta 24% and the kappa 6% of the total opioid sites in the rat spinal cord (Besse, Lombard, Zakac, Roques & Besson 1990). The idea that kappa levels are higher in the mouse and guinea pig spinal cord has been put forward. However, studies in species other than the rat have not been carried out with the most selective ligands for the receptors and so may not represent the true relative distribution of the receptors. We lack systematic quantitative studies in a number of species on the relative distribution of the receptors at a variety of CNS sites. Recent studies using probes for the selected sequences of the delta receptor have unequivocally shown that many of these receptors are located pre-synaptically on afferents and in close apposition to enkephalin containing cells (Dado, Law, Loh & Elde 1993). In addition, spinal application of antisense to the delta receptor has shown that this leads to a marked reduction in delta mediated analgesia without alteration of the effects of morphine (Uhl, Childers & Pasternak 1994).
The relative numbers of pre- and post-synaptic receptors can be calculated after nerve section and the former predominate. The proportions of pre-synaptic opioid receptors in the spinal cord varies from 70 to 50%, with over 70% of the total mu receptor sites being on the afferent terminals (Besse et al. 1990) along with large numbers of delta receptors (Dado et al.1993). Given the number of receptors it is not surprising that there is much evidence for a pre-synaptic action of opioids from studies of opioid inhibition of C-fibre evoked release of transmitters (substance P and glutamate) and from in vitro and in vivo electrophysiological studies (Yaksh & Nouiehed 1985); (Yaksh & Malmberg 1994); (Dickenson 1994a). However, studies with other approaches have failed to demonstrate this (Lang, Duggan & Hope 1991) .
Pre-synaptic actions on transmitter release result from an opening of potassium channels (mu and delta receptors) or a closing of calcium channels (kappa) both of which lead to a reduction in calcium influx into C-fibre terminals so diminishing transmitter release (North 1989). C-fibres are believed to release a number of coexisting transmitters including the tachykinins, excitatory amino-acids and a number of excitatory peptides which act on multiple receptors (Besson & Chaouch 1988); (Dickenson 1994a); (Dray, Urban & Dickenson 1994). Consequently, the pre-synaptic action opioid action of reducing the release of many transmitters will be a highly effective route to analgesia since it will be equivalent to the block of multiple post-synaptic receptors. It is therefore not that likely that any single antagonist of one of these post-synaptic receptors will have sufficient efficacy to compete with the opioids as powerful analgesics in acute and chronic pains. An exception to this is likely to be agents acting directly, or indirectly, to modulate the N-methyl-D-aspartate (NMDA) receptor in pains where central hypersensitivity, an augmented spinal response to a low or moderate afferent input is generated.
Section of a peripheral nerve will lead to degeneration of the nerve and the pre-synaptic receptors synthesized in the dorsal root ganglion will be lost (reducing mu opioid receptor sites in the dorsal horn by 70%) (Besse et al 1990). It would be interesting to know if less severe nerve damage impairs the production of functional opioid receptors. However, in animal models where the pre-synaptic opioid receptors have been removed by prior rhizotomy whether electrophysiological with systemic dosing in spinal preparations, (Lombard & Besson 1989) or behavioural with spinal application (Xu & Wiesenfeld-Hallin 1991) morphine is still effective, although higher doses than in normal animals are required in the former case. Thus when there is a loss of opioid receptors, there will be a reduction in opioid actions that should predictably be overcome by dose escalation (Lombard & Besson 1989) If nerve damage does lead to a loss of some of the spinal opioid receptors one could predict that opioid delivery to supraspinal sites would target the normal population of supraspinal receptors and in a animal study intraventricular morphine was effective in a neuropathic pain model where spinal was not. In clinical practice in neuropathic pains there is a reduced systemic opioid effectiveness (Arner & Meyerson 1988) which can be overcome by increasing the dose in some circumstances ( Portenoy, Foley, Inturrisi 1990 ); (Jadad., Carroll, Glynn, Moore & McQuay 1992). Dose increases may not always be possible since side effects may become intolerable. Where the side effects can be controlled, if this increase in dose does not overcome the pain we have to suspect that opioid receptor loss or dysfunction is not the only factor.
POST-SYNAPTIC ACTIONS
Evidence for functional opioid actions at post-synaptic receptors is based on electrophysiological and behavioural approaches (Duggan & North 1984); (Lombard & Besson 1989). Post-synaptic hyperpolarisations again result via the opening of K+ channels or the closing of calcium channels (North 1989). These receptors could hyperpolarise the dendrites of projection neurones, interneurones (both would be selective for noxious transmission) or the cell body of projection cells which may not be selective for nociceptive inputs since many but not all, neurones in the dorsal horn receive both nociceptive and tactile inputs.
An important indirect post-synaptic action is the opioid disinhibitory effect mediated via GABA and enkephalin neurones in the substantia gelatinosa which in turn leads to an inhibition of output neurones. Thus here, neurones can be recorded in the substantia gelatinosa that are facilitated by opioids, an action that requires GABA-A receptor function. There is both morphological and electrophysiological evidence to support this action (see Magnuson & Dickenson 1991).
These post synaptic actions of opioids present some problems of interpretation since any direct hyperpolarisation of a cell soma would inhibit all responses of the cell including the innocuous inputs onto convergent or multireceptive cells. However many of the opioid receptors in the substantia gelatinosa could be on the dendrites of the deep cells penetrating into the C-fibre terminal zone; inhibitory effects here would also be selective since they are likely to be spatially distinct from the large fibre inputs. Another possibility is that the post-synaptic disinhibitory effects of opioids selectively feed onto nociceptive circuitry. When allodynia and some of the hyperalgesias are transmitted through A-fibre afferents it would not be surprising to find a poor sensitivity to opioids (Yaksh 1989) since the only opioid control of A-fibre inputs is via the relatively small number of post-synaptic receptors on the output neurones (Duggan & North 1984);(Besson & Chaouch 1987). Doses of morphine that abolish C-fibre evoked responses in normals animals have only minor effects on A-fibre activity (Dickenson & Sullivan 1986). In these pain states, novel non-opioid therapy may have to be considered, directed at the spinal systems generating the tactile evoked pain (Yaksh 1989).
ALTERNATIVES TO MORPHINE?
It is surprisingly unclear whether different opioids may have slightly different ranges of pharmacological actions which could allow a choice of opioid for different pains. Clinical studies comparing different opioids in different pain states are needed. Morphine, at physiological doses, probably acts only at the mu receptor. This drug has high affinity for the mu receptor and has a relative affinity of 50 times less for the delta receptor with minimal affinity for the kappa receptor (Kosterlitz 1985). Thus relatively non-selective effects could occur with very high doses such as those achieved in neurochemical and in binding studies where non-mu effects of morphine have been reported. However, in vivo, doses of spinal morphine which are sufficient to abolish the C-fibre evoked responses of dorsal horn nociceptive neurones are probably entirely mu receptor mediated. In support of this, there is no evidence for mu-delta cross tolerance from physiological studies (Kalso, Sullivan, McQuay, Dickenson & Roques 1993). Since morphine is the standard opioid for clinical practice plasticity related to mu receptor mechanisms is of great importance (Dickenson 1994a).
Opioid drugs which act on receptors other than the mu receptor for morphine site could be analgesics with reduced morphine like side effect profiles. As is the case with any transmitter system, the greater the number of receptors the greater the chance that the desirable effects can be separated from the unwanted effects. In the case of the opioid receptors a further division of the receptors from the main three, the mu, delta and the kappa, has been proposed. The mu receptor has been suggested to consist of a mu 1 and a mu 2 subtype (Pasternak & Wood 1986), the delta has been also subdivided and the kappa receptor has been divided into three sub-types (Jiang, Takemori, Sultana, Portoghese, Bowen, Mosberg & Porreca 1991); (Traynor 1989). Whether these subtypes have functional consequences remains to be seen since as yet the physiological consequences are hazy except for the delta subtypes where there is evidence for differential effects of the two receptors (Jiang et al 1991). The recent cloning of the opioid receptors (Uhl et al 1994) will further facilitate this task since probes based on the receptor sequence will provide unequivocal proof of location of the particular receptor and important insights into the mechanisms of opioid actions and the existence of sub-types. At the present time there is no evidence from the cloning studies for receptor subtypes: the receptors isolated whether mu, delta or kappa have been single identical species. It may be that there is alternative splicing to produce the sub-types or that local neuronal tissue environments allow the subtypes to be expressed and that the cell lines used so far may underestimate the variability within the opioid receptor family (Uhl et al 1994). What is known, however, is that the rat and mouse opioid receptors are, as far as can be said, identical in structure and pharmacology to the human receptors further verifying the important links between animal studies and clinical practice (Uhl et al 1994).
DELTA OPIOIDS
Antagonists for the opioid receptors have demonstrated the independence of mu, delta and kappa receptors in terms of antinociception although there have been problems in some studies in demonstrating kappa receptor agonist effects (Millan 1990). The independent analgesic effects following activation of non-mu receptors indicates potential for opioid analgesics which are delta or kappa agonists. Kappa opioids are not always particularly effective analgesics in animals and this appears to be reflected in the initial early clinical studies with these drugs in humans. The delta receptor may well be an important target for novel opioid therapy. Animal studies have shown that opioids selective for the delta receptor can equal the analgesic effects of morphine by actions at both spinal and supraspinal sites in a number of nociceptive tests (Sullivan, Dickenson & Roques 1989); ( Jiang et al 1991). One would also predict reduced respiratory depression and gastrointestinal effects of delta as compared to mu opioids. There are now reports of potent and selective non-peptide delta opioids which have been tested in a number of paradigms in animals. A highly selective delta opioid, SNC 80 which was effective by central and systemic (including oral administration) routes has been produced. The analgesic effects were reversed by a number of delta but not mu opioid receptor antagonists. Importantly, in tests of respiratory function, SNC 80 stimulated rather than reduced respiratory rate in sheep (Porreca, personal communication). The potential of delta opioids is therefore high and there may eventually be delta opioids as clinical alternatives to morphine. As will be discussed later, there are suggestions that some of the reduced effects of morphine in neuropathy may be due to negative effects of cholecystokinin on mu receptors - these effects do not extend to delta mediated actions. Consequently, it is possible that delta opioids may be better analgesics than mu opioids in the treatment of neuropathic pains.
ENDOGENOUS OPIOIDS
What about endogenous opioids? The enkephalins are rapidly degraded by membrane bound peptidases. The synthesis of peptidase inhibitors has been a successful strategy so that kelatorphan, a mixed peptidase inhibitor, inhibiting at least two of the important breakdown enzymes, affords almost complete protection to the enkephalins (Roques et al 1993). The spinal application of the inhibitor produces a reduction of nociceptive responses of cells with the pool of enkephalins protected by the inhibitor likely to be derived from both a segmental release and from descending pathways activated by the stimulus. The inhibitions are reversed by a selective delta antagonist (Dickenson, Sullivan, Fournie-Zaluski & Roques 1987). The very recent reports of a systemically active mixed peptidase inhibitor, RB101, is the next stage towards the clinical application of this novel approach to pain relief. Interestingly, the side effect profile of RB 101 appears to be unlike that of morphine in terms of physical and psychological dependence (Roques et al 1993).
NOVEL PERIPHERAL ANALGESIA
Opioids lack peripheral actions in undamaged tissue but there is now good evidence that the consequences of inflammation can reveal a novel site of opioid action which appears rapidly (Stein 1994). The synthesis of opioid receptors occurs in dorsal root ganglion cells amongst other locations in the body. These receptors are transported in the fine afferent fibres in both directions; the centrally directed receptors become the pre-synaptic receptors and the peripherally transported receptors somehow become active only following inflammation. The relative effectiveness of mu, delta and kappa receptor activation to elicit peripheral analgesia varies between models but in arthritic states all three are active (Stein, Millan, Shippenberg, Peter & Herz 1989). Alongside the appearance of the functional opioid receptors on afferent nerves, the arrival of endogenous opioid peptides at the injury site seems to be related to immune cell proliferation. Thus opioids unable to penetrate the CNS and so being devoid of central side-effects may be good analgesics in inflammatory states via these peripheral sites. There have now been a number of clinical studies on this effect, the bulk of which have been been positive. Thus the local application of morphine into the knee joint in patients has been shown to produce a local analgesic effect. Interestingly, a recent study has shown that the degree of analgesia can be related to the amount of tissue damage and so presumably the degree of inflammation (Stein1994). Peripherally acting opioids may then have potential analgesic effects in inflammation (Stein, Millan, Shippenberg, Peter, Herz 1989); (Stein 1994).
SUPRASPINAL ANALGESIA
The first ever demonstration of opioid actions within the central nervous system was the analgesia seen following intraventricular morphine. Numerous supraspinal sites of opioid analgesia have been established (Besson & Chaouch 1987). These have now been localized to areas in the medial brain stem around the nucleus raphe magnus and extending rostrally to periaqueductal and periventricular grey and other areas with the monoamines appearing to be critical transmitters in these pathways (Yaksh, Al-Rodhan & Jenson 1988). The roles of these areas in morphine analgesia has been based on microinjection studies and the ability of naloxone locally applied into these areas to reduce the effects of systemic morphine.
The mechanisms of action of opioids at these supraspinal levels and in particular with respect to how they interact with descending inhibitory controls is still unclear. Opioid induced increases and decreases in descending inhibitory controls have been reported. The roles of these descending pathways in different models of various pain states are unknown. We need much information on the physiological and pharmacological bases for supraspinal analgesia in animal models of persistent pain to form a basis for the potential use of manipulation of the monoamines in difficult clinical pains.
Examination of the anatomy and the pharmacology of the descending systems may provide a basis for these disparate results with regard to opioid interactions with descending controls. Firstly, descending controls originate from many different areas of the brain stem and the midbrain and a complex pharmacology exists in these descending pathways ( Yaksh et al 1988). Noradrenaline, 5HT, enkephalin and substance P are involved with the latter three co-existing in some neurones projecting from the brain stem and midbrain to the spinal cord. In addition to interactions between these transmitters there are a number of local transmitter systems (cholinergic, GABAergic and opioid) in the nuclei where the descending controls originate. Direct opioid inhibitions or indirect disinhibitions could result from local opioid actions in these supraspinal areas. These opposite effects could themselves be on inhibitory and/or excitatory pathways. A further complexity is that the monoamines have a number of receptors at the spinal terminal sites which when activated could inhibit or excite depending on the receptor - an additional complication is the presence of auto- or heteroreceptor control of the release of transmitter at these terminal sites. Whether or not the post-synaptic receptors are on excitatory or inhibitory elements will also be important. Consequently the problem is not the direction of effect of opioids on these systems but to understand the physiological roles and consequences of the mixed opioid actions on these multiple pathways (Dickenson. 1994a).
The consideration of the direction of effect of the spinal monoamine receptors serves to illustrate these points. In the case of noradrenaline, there is a general consensus that notwithstanding a possible role of the alpha-1 receptors, the predominant spinal target for the transmitter are alpha-2 receptors, located in the spinal cord, post-synaptically to the noradrenergic terminals. In a similar manner to the opioid receptors and the afferent nociceptive fibres, these receptors are located both pre-synaptically and post-synaptically on spinal sensory circuits and there is ample evidence for alpha-2 agonists being effective analgesics in a number of animal models of acute and more persistent pains. In addition, there is little doubt that alpha-2 agonists synergize with morphine, probably as a result of dual activation of separate receptors with similar locations and effector mechanisms (Dickenson & Sullivan 1993); (Yaksh & Malmberg 1994). However, relatively little is known about the driving force behind pain related changes in noradrenergic activity in these models. An exception is a report of increased alpha-2 mediated activity in inflammation which does not however, contribute to enhanced spinal opioid effectiveness (Stanfa & Dickenson 1994). There may however be a supraspinal site of action of noradrenaline in enhancing opioid actions (Hylden, Thomas, Iadorola, Nahin & Dubner 1991).
Possibly the effectiveness of tricyclic antidepressants (TADS) for pain relief in humans relates to enhancement of the availability of noradrenaline and serotonin. This is where the problems arise. The number of receptors for serotonin or 5-hydroxytrptamine (5HT) increases on a regular basis. Presently, there are at least 7 major receptors with over 20 subtypes. The receptor that underlies descending antinociception at the spinal level is unknown. Thus increases in 5HT levels in the cord will activate all the receptors irrespective of whether they are excitatory and inhibitory. Potentially, knowledge of the particular roles of the monoamine receptors may lead to better therapeutic efficacy by agents (or combinations thereof) acting on particular receptors rather than the indirect indiscriminate activation of multiple receptors produced by the TADS (Max 1994).
Bearing in mind these complexities, how do opioids interact with descending pathways? Since many of the sites of opioid actions at supraspinal sites overlap with areas where descending inhibitory controls originate the simplest situation is that supraspinal opioids increase these descending monoamine inhibitions and these in turn block spinal pain transmission by actions at inhibitory spinal receptors. For opioids to increase descending inhibitions, the mechanism will have to be via disinhibitions (Fields, Barbaro & Heinricher M 1988) . The clearest demonstration of supraspinal descending inhibitory controls being increased by morphine is the finding that the spinal induction of c-fos, used as a marker of noxious evoked activity, is very clearly reduced by intraventricular morphine (Gogas, Presley, Levine & Basbaum 1991).
There are however, other findings that do not find this direction of effect. Diffuse Noxious Inhibitory Controls (DNIC), are descending controls induced by heterosegmental noxious stimulation and partly involve both opioid and serotoninergic mechanisms. Morphine, either given directly into supraspinal tissues or at low systemic doses without direct spinal actions, reduces these descending controls (Le Bars & Villaneuva 1988).
There are therefore considerable difficulties in arriving at a simple consensus as to the direction of effect of opioids on descending control systems. However, as discussed above, the multiplicity of these descending controls in terms of their anatomy, their pharmacology and their spinal projections forms a framework within which the various directions of effect of opioids can be incorporated. There is no doubt that whatever the mechanism, supraspinal opioids produce behavioural analgesia (Yaksh et al 1988); (Besson & Chaouch 1987).
It is highly likely that both the level of pain transmission and the effectiveness of opioids in different pain states is determined by alterations in descending control pathways. Other than studies on alpha-2 adrenoceptors we are ignorant of the extent of plasticity in these systems.
PLASTICITY IN OPIOID CONTROLS
At this stage, having provided a framework of opioid effects and mechanisms on which to consider plasticity in opioid systems it is pertinent to consider particular pain states where there is evidence for changed opioid effectiveness. Why are opioids sometimes poorly effective in neuropathic pain states in man and animals (Arner & Meyerson 1988); (Portenoy et al 1990); (Jadad et al 1992)? On the other hand, there is good evidence that in a number of inflammatory models opioids are more effective than in normal animals (Dickenson 1994a). Finally, on the basis of there being transmitter systems in the CNS that can reduce opioid effectiveness, pre-emptive analgesia should stop the induction of these systems and so provide better pain relief. Why has it been so difficult to provide clear and marked clinical benefits of this approach (McQuay 1994)?
The analgesic effects of morphine can vary in different pain states. The mechanisms behind these changes have been elusive, but their identification and eventual manipulation may be of considerable clinical benefit. Firstly, let us consider the pharmacological systems that can interfere with opioid effectiveness bearing in mind that pathology can also play a role as discussed earlier with regard to nerve section and the loss in number of pre-synaptically located opioid receptors. There appears to be four major pharmacological factors
a. interference with mu receptor function by the metabolite of morphine, morphine-3-glucuronide (M3G). This has been proposed on the basis of behavioural studies but is not supported by electrophysiological and clinical studies.
b. changes in the levels of the non opioid peptides, FLFQPQRFamide and/or cholecystokinin (CCK), either spinally where there is very strong evidence for CCK as a regulator of morphine analgesia, or supraspinally as a more global negative influence on opioid actions.
c. increased levels of the opioid peptide, dynorphin which has been shown to occur after persistent pain. In theory this peptide can reduce mu opioid analgesia but the physiological role of dynorphin as an opioid modulator is not good. Finally,
d. an excess of excitatory activity, so that a spinally generated hypersensitive state is induced, against which opioid controls are insufficiently efficacious. The N-methyl-D-aspartate (NMDA) receptor is a very strong candidate for the final common path for generation of this state and there is poor opioid sensitivity of a number of electrophysiological and behavioural measures of pain where NMDA receptor activation has been induced.
These four possibilities are not mutually exclusive (Dickenson. 1994a). Thus in a particular pain state where opioids are being used to treat the pain, NMDA mediation of spinal transmission may be occurring at the same time as elevated spinal CCK and dynorphin levels with high plasma M3G levels (Dickenson 1991b). The evidence for and against these systems altering opioid analgesia will be considered in turn.
MORPHINE METABOLITES
The actions of morphine do not end with metabolism. It is now well established that the glucuronidation of morphine produces two major metabolites, morphine-3-glucuronide and morphine-6-glucuronide with remarkably different actions. The 6-glucuronide is more potent than morphine itself, and although the degree of this enhanced action is variable from study to study, it is at least 10 to 30 fold more effective in tests of analgesia (Sullivan, McQuay, Bailey & Dickenson 1989) . The reasons for this are not obvious since the affinity of morphine-6-glucuronide for the mu receptor is not appreciably greater than morphine itself although it has more delta and less kappa affinity. However, as discussed earlier, as with morphine, it is likely that predominant mu activity underlies the analgesia with morphine like opioids at therapeutic doses. However, the other metabolite, M3G, morphine-3-glucuronide, has no affinity for the mu receptor and being unable to bind to the receptor has no opioid actions. Nevertheless, results from behaviour after administration of M3G have led to the suggestion that morphine-3-glucuronide is a factor that contributes to reduced opioid sensitivity (Smith, Watt & Cramond.1990); (Gong, Hedner, Bjorkman & Hedner 1992). The metabolite given by the intraventricular route caused marked behavioural agitation which interfered with the tests. By contrast to these studies there is electrophysiological (where non-specific effects are less likely to interfere with the results) and behavioural evidence that even with dose ratios of 100 of the metabolite to morphine, morphine-3-glucuronide has absolutely no effect on the spinal antinociceptive effects of morphine (Hewett, Dickenson & McQuay 1993). Since a) the spinal site of action of morphine is a major contributor to systemic analgesia, b) morphine-3-glucuronide does not bind to opiate receptors and c) in renal insufficiency where the metabolite will accumulate opiate effects tend to be enhanced and so it is therefore highly unlikely that morphine-3-glucuronide is an important factor in cases of opioid poorly responsive pain. M3G should not, at present, be used as an excuse not to persevere with or increase the dose of morphine in pain states where opioid responsivity is poor. Patient controlled analgesia has revealed that neuropathic pain patients can gain relief with morphine although within this patient group, pain control is not as good as in patients with nociceptive pains (Jadad et al 1992). Dose escalation can also be effective (Portenoy et al 1990).
ANTI-OPIOID PEPTIDES
Amongst the numerous factors influencing morphine analgesia, accumulating evidence points to the non-opioid peptide, cholecystokinin (CCK) as an important physiological modulator of analgesic mechanisms. The exogenous spinal application of CCK and another peptide, FLFQPQRFamide, both non-opioid peptides that are found within intrinsic neurones in the spinal cord will prevent mu but not delta mediated neuronal inhibitions (Baber, Dourish & Hill 1989); (Dickenson 1994a), and both reduce intrathecal morphine analgesia in behavioural studies (Zhou et al 1993). Thus in situations where there is a release of these peptides one would expect a reduction in morphine effects without requiring any change in opioid receptor number. In fact, the ability of these peptides to interfere with analgesia is not restricted to the effects of opioids but also includes alpha-adrenoceptor agonist actions.
CCK has been shown to reduce the analgesic effects of morphine at a number of CNS sites and has also been implicated in the development of opioid tolerance (Baber et al 1989). Interestingly spinal and supraspinal delta opioid mediated analgesias are not altered by CCK so that if there are physiological situations where CCK reduces mu opioid actions future clinical delta agonists could be effective.
One of the key sites for these interactions is the spinal cord. Negative results with cDNA probes within dorsal root ganglia in the normal rat make it very unlikely that genuine CCK is found in nociceptive C-fibres in normal animals. Intriguingly, induction of the peptide in afferents occurs after pathological damage to the afferents and the consequences of this are discussed later with regard to neuropathic pain(Xu, Puke, Verge, Wiesenfeld-Hallin, Hughes & Hokfelt 1993). Endogenous CCK in the dorsal horn under non-pathological conditions is thought to originate from both intrinsic neurones found in superficial laminae and descending fibres. The receptors are found both pre- (approximately 50-60%) and post-synaptic to the primary afferent fibres, mirroring the mu opiate receptor distribution in the rat spinal cord. The post-synaptic CCK receptors are mainly of the CCKB type in the rat spinal cord but of the A-type in the primate. Interestingly, the pre-synaptic receptors are of the CCKB type in all species. (Ghilardi, Allen, Vigna, McVey & Mantyh 1992). Thus CCKB receptor antagonists will be critical in testing whether CCK influences morphine analgesia in humans (Stanfa, Dickenson, Xu & Wiesenfeld-Hallin 1994).
The mechanism by which CCK attenuates the antinociceptive effect of morphine is not on opioid receptors but via activation of CCK receptors which may then interfere with opioid actions via post-receptor mechanisms. Key sites for these CCK-opioid interactions is likely to be the spinal terminals of C-fibres. Here, one possibility is that CCK mobilizes calcium from intracellular stores. This will counter the opioid suppression of the rise in internal calcium produced by depolarization, the basis for opioid reductions in transmitter release. Again, CCK only reverses the suppression of the induced rise in [Ca2+]i produced by mu but not delta opioid agonists (Wang, Ren & Han 1992).
At the same time there is good evidence for another mechanism for the CCK-opioid interaction which involves the endogenous enkephalins acting on the delta opioid receptor. Here, both CCK antagonists and the prescence of inflammation (see below) enhance morphine analgesia, an effect that is prevented by delta opioid antagonists. The theory is then that CCK inhibits the release of enkephalins-removal of this control allows the increased levels of the enkephalins to cause a delta receptor mediated synergy with the mu receptor (Vanderah et al, 1994, Ossipov et al 1995). Both theories are not mutually exclusive.
In keeping with CCK reducing opioid analgesia, the ability of morphine to inhibit spinal nociceptive processing is enhanced in the presence of selective CCKB antagonists demonstrating physiological antagonism of morphine antinociception by endogenous CCK under conditions of acute nociception. There is now evidence for this interaction in animal models more relevant to clinical situations such as inflammatory and neuropathic pain models (Stanfa et al 1994, Ossipov et al 1995).
As discussed, in neuropathies, morphine tends to have a reduced effectiveness whereas after inflammation, morphine has enhanced actions. In fact, a few hours after carrageenan inflammation there are mild increases in the potency of delta and kappa opioid effects but marked increases in the effects of morphine (Stanfa, Sullivan & Dickenson 1992, Ossipov et al 1995). One reason is the novel peripheral action of opioids described earlier and systemic dosing studies will be confounded by this additional site of action. However, spinal morphine is almost 20 fold more potent than in normal rats after carrageenan inflammation. The mechanisms for this latter effect must be central and rapidly induced since the increased opioid actions occur within one hour of the inflammation, ruling out receptor upregulation (Stanfa et al1990). In this model, exogenous CCK still attenuates the antinociceptive effects of morphine but CCK receptor antagonism no longer produces an enhancement of the antinociceptive effect of morphine. The most likely basis for these results is a decreased availability of CCK within the spinal cord following carrageenan inflammation, either due to a decreased release of CCK or reduced content within the dorsal horn. This reduced functional activity of CCK in inflammation is therefore a major factor in the enhanced potency of spinal morphine seen in these animals (Stanfa & Dickenson 1993). Interestingly, in exactly the same model there is an increased alpha-2 inhibitory tone in the spinal cord but in this case, antagonist studies have shown that noradrenergic activity is not a factor in the altered opioid sensitivity although it may well reduce inflammation induced nociception (Stanfa & Dickenson 1994 ).
Neuropathic models reveal that CCK plays an entirely opposite role. In nerve damage, increases in CCK systems have been shown to underlie observed reductions in spinal opioid sensitivity. It has been shown that an increase in spinal CCK (likely to be due to novel synthesis of the peptide in primary afferent fibres (Xu et al 1993) leads to a reduction in the potency of spinal morphine in a rat model of neuropathic pain following peripheral nerve injury. If the increased CCK is derived from induction of the peptide in the afferents, the interference would seem to be directed at the pre-synaptic mu receptor. The opioid responsiveness of this model was restored by CCKB antagonism. Different pain states may then lead to changes in the levels and synthesis of CCK that can shift opioid sensitivity in either direction (Stanfa et al 1994). An important point that arises from these studies is that the attenuation of opioid analgesia by CCK is not global but selective for mu but not delta opioid receptor events. The prediction would be that delta opioids may have efficacy in pain states where morphine is poorly effective due to enhanced CCK levels.
CCK may then be an endogenous "brake" applied to the antinociceptive actions of morphine. In addition to the alterations in CCK induced by different pain states, behavioural studies have suggested that the release of endogenous CCK is even governed by the environment to which an animal is exposed. It is suggested that CCK released in "safe" situations prevents the acute antinociceptive effects of mu agonists and so reduces the effects of morphine (Wiertelak, Maier & Watkins 1992). Findings such as these may provide a basis for events such as stress and anxiety altering opioid efficacy. Whatever the case, these studies serve to indicate that the release of CCK is not fixed but varies, in both directions, from its normal state according to both external and internal events.
It therefore appears that CCK can act to control spinal morphine analgesia. Attenuation of this negative influence leads to augmented spinal opioid controls. This finding, along with the novel peripheral actions of opioids in inflammation and enhanced descending controls could be an adaptation of intrinsic inhibitory systems to balance enhanced nociception during inflammatory state. By contrast to the natural physiological processes of inflammation, the pathological changes in neuropathic pain states results in counterproductive increases in CCK. CCK antagonists being developed by the pharmaceutical industry may well enhance morphine analgesia in non-pathological pain states and restore morphine analgesia in humans with neuropathic pains. The predicted anxiolytic effects of these antagonists would be a bonus when used as analgesic adjuncts especially in states of chronic pain where anxiety commonly accompanies pain (Stanfa et al 1994).
DYNORPHIN
In inflammatory states there is an increase in the mRNA in the spinal cord for dynorphin and to a lesser extent, for enkephalin with all the cells increasing dynorphin synthesis having a preceding rise in c-fos, a protooncogene (Dubner & Ruda 1992). Dynorphin can mimic some of the increases in excitability seen after inflammation, such as the increased nociceptive responses of neurones, whilst inhibiting others (Knox & Dickenson 1987). Kappa opioids can functionally antagonize the mu receptor in the spinal cord so potentially contributing to a decreased morphine effectiveness yet opioids are more potent in inflammatory models (Stanfa et al 1992). Furthermore, increases in spinal dynorphin levels also occur in neuropathic states where opioid actions tend to be reduced. These general increased dynorphin levels in different physiological pain models where opioid actions can be increased or decreased depending on the model makes kappa antagonism of morphine unlikely to be of physiological significance. Quite what the functional consequences of the increases in dynorphin mean to the spinal cord remain to be seen (Stanfa & Dickenson 1994b) and since dynorphin can elicit NMDA receptor mediated effects as well as opioid actions (see Dubner & Ruda 1992) the picture is complex.
CENTRAL HYPERSENSITIVITY
One of the most important new concepts related to pain is the idea that the ascending and propriospinal pain messages from the dorsal horn are not the same under all circumstances. We are nearing explanations for the extreme aberrations of pain transmission such as phantom limb pains and hyperalgesias and allodynias where the relations between the stimulus and the response are markedly perturbed. The basis for this lack of strict concordance between stimulus and response appears to be the generation of central hypersensitivity (Woolf 1994); (Woolf & Thompson 1991); (Dickenson 1994b); (McMahon et al 1993 ); (Price et al 1994).
There are two key observations on this subject. Firstly, high frequency C fibre stimuli results in an marked and prolonged increase in the flexion withdrawal reflex in rats recorded from motoneurones in spinal animals (Woolf 1983). Thus, noxious stimuli can enhance spinal excitatory events. Secondly, the repetition of a constant intensity C-fibre stimulus induces the phenomenon of wind-up whereby the responses of certain dorsal horn nociceptive neurones suddenly increase markedly (both in terms of magnitude and duration) despite the constant input into the spinal cord (Dickenson 1994b). Volatile general anaesthesia such as with halothane fails to prevent this type of activity indicating that the treatment of post-operative pain states needs to take into potential priming events occurring during the operations. The object of this account is to discuss the possible pharmacological substrates underlying these changes.
SUBSTRATES FOR CENTRAL HYPERSENSITIVITY
PEPTIDES
Historically Substance P (SP) was the first transmitter to be related to the transmission of pain. SP release can be detected in the spinal cord following high but not low intensity peripheral stimulation. The use of antibody microprobes to detect the spatial release of SP has shown that it is essentially restricted to the zones where the C-fibres terminate (Duggan, Hendrey, Morton, Hutchison & Zhao 1988). In addition to substance P, the release of neurokinin A and CGRP following C-fibre activation has been demonstrated. However when CGRP is present the subsequent release of SP is now extended to cover much of the dorsal horn. The interpretation of this finding is that the degradation of SP is reduced by CGRP binding to the peptidase that also cleaves SP and so SP can diffuse in the active form over considerable distances (Schaible, Hope, Lang & Duggan 1992). The concept of actions at a distance from the release site, so-called volume transmission, has attracted interest as a basis for non-synaptic transmitter actions. Events such as these may have relevance to pain in that the peptides may diffuse to distant receptors, avoiding both peptidases and spatially restricted inhibitory influences. The induction of inflammation is accompanied by enhanced release of these peptides centrally which may then contribute to the central hypersensitivity (Sluka, Dougherty, Sorkin, Willis &Westlund 1992); (Dray, Urban & Dickenson 1994); (Todd & Spike 1993).
The post-synaptic receptors for the neurokinin family of peptides, substance P and neurokinins A and B are now well characterized (Otsuka & Yoshioka 1993). Cloning and sequencing has been achieved. Currently there are held to be three subclasses of tachykinin receptors; the neurokinin 1,-2, and -3 receptors. Early studies on the role of SP in neuronal events in nociception were bedeviled by poorly selective antagonists and non-specific effects of the drugs. More recent studies have indicated a role of SP at the NK1 receptor in different types of more prolonged nociceptive transmission including slow excitatory post synaptic potentials induced by repetitive C-fibre stimulation and C-fibre induced reflex facilitation. Likewise, the NK1 receptor antagonist RP67580 has only weak actions on acute responses but produces marked inhibitions of the formalin response of the dorsal horn neurones with equivalent effects on the two phases of the response. These recent studies would indicate that the ability of NK1 receptor antagonists to reduce the activation of dorsal horn neurones depends on the type of stimulation used (Otsuka & Yoshioka 1993). A consensus is that the conditions for the release of substance P from the fine afferents include a sufficiently long stimulus at an intensity sufficient to activate C-fibres (Urban, Thompson & Dray 1994). The acute responses of the neurones must therefore include some other transmitter and the evidence implicates glutamate and aspartate. We lack antagonists for the other peptides but some such as galanin and neuropeptide Y are induced in afferents after nerve damage (Todd & Spike 1993); (Urban, Thompson & Dray 1994).
EXCITATORY AMINO-ACIDS
A large proportion of peripheral sensory fibres including both small and large fibres contain glutamate and aspartate (Battaglia & Rustioni 1988). In the case of the C-fibres, the coexistence of glutamate with peptides (Battaglia & Rustioni 1988) would make it highly likely that a noxious stimulus releases both peptides and excitatory amino-acids from the afferent nociceptive fibres. Thus in clinical pain states post-synaptic activation of both neurokinin and other peptide receptors together with the receptors for the excitatory amino-acids on nociceptive neurones will occur. The development of selective agents for the the receptors, the N-methyl-d-aspartate (NMDA), the metabotropic and the alpha-amino-3-hydroxy-5-methyl-isoxazole (AMPA) receptors has enabled their roles in spinal processing to be studied.
The metabotropic receptor has as yet an ill-defined role in pain states but may well contribute by acting to enhance NMDA and AMPA receptor function via intracellular actions. Use of AMPA receptor antagonists indicate that acute noxious but also innocuous stimuli seem to be transmitted via AMPA receptor activation (Dougherty et al 1992); (Neugebauer et al 1993). The widespread roles of AMPA receptors in CNS function and the lack of nociceptive selectivity mean that the receptor as a therapeutic target looks doubtful. By contrast, the NMDA receptor has become an increasingly important target site as evidence accumulates for a role of the receptor in the enhancement of spinal processing of painful messages (see recent reviews (Dickenson 1990); (Dickenson 1994b) (Price et al 1994)) and in many long term events in the brain (Collingridge & Singer 1990); (Daw, Stein & Fox 1993). In the spinal cord, the NMDA receptor may play a similar role especially in more prolonged pain states involving hypersensitivity where functional alterations in central transmission processes may occur.
The complexity of the NMDA receptor-channel is striking: to operate, certain particular conditions need to be met: the release and binding of the coagonists for the receptor, glycine and glutamate, are needed together with a non-NMDA induced depolarization to remove the resting magnesium block of the channel (Dickenson 1994b). C-fibre induced release of excitatory peptides either in a restricted spatial zone or via volume transmission may provide the required depolarization to remove the block since neurokinin receptor antagonists can reduce NMDA mediated responses in the spinal cord (Urban et al1994). For these reasons, the NMDA receptor-channel complex is not a participant in "normal" synaptic transmission, but when the correct conditions are achieved the complex will suddenly become activated and add a powerful depolarizing or excitatory drive to transmission of pain in the spinal cord which then appears to lead to enhanced synaptic transmission or hypersensitivity (Dickenson 1990); (Woolf &Thompson 1991); (Neugebauer et al 1993); (Dickenson 1994); (McMahon et al 1993); (Dubner & Ruda 1992); (Price et al 1994). Increased release of afferent peptides in, for example, inflammation, could facilitate NMDA transmission by more effective removal of the magnesium block of the receptor channel or by increasing the release of the excitatory amino-acids themselves (Kangra & Randic 1990). It is now well established that wind-up and the reflex hypersensitivity are NMDA receptor mediated. Further experiments with formalin indicate that "pathological pain" where inflammation is present, can be distinguished from the acute phase response where there is no damage, on the basis of the sensitivity of only the former to NMDA antagonism. Both the induction and the subsequent maintenance of these responses are dependent on NMDA processes (Haley, Sullivan & Dickenson 1990); (Neugebauer et al 1993); (Price et al 1994).
However, NMDA receptor activation can also influence inhibitory interneurones in the spinal cord. There is some evidence for this proposition from carrageenan inflammation where excessive NMDA receptor activation subsequently induces inhibitory influences (Stanfa et al 1992). Excessive NMDA activation in the CNS is one mechanism behind excitotoxicity so that elevated NMDA activation may trigger inhibitory systems as an auto-limiting device to prevent over excitation and maybe even cell death. Possibly, the loss of inhibitions to counter NMDA excitatory mechanisms (Woolf & Doubell 1994) leads to some of the problems of neuropathic pain. In this regard, NMDA mediated allodynia can be induced by blockade of spinal inhibitory tone in normal animals (Yaksh 1989). Furthermore, failure of inhibitions could underlie the transition from acute to chronic pain.
Other approaches have revealed roles of the NMDA receptor in spinal pain processes including ischaemia and neuropathic pain states where NMDA antagonists have beneficial effects weeks after induction of the injury against the hyperalgesia and spontaneous pain. Thus there is evidence for an involvement of the NMDA receptor in inflammatory pain, neuropathic pain, allodynia and ischaemic pain. Of critical importance, not only can wind-up be demonstrated in elegant psychophysical studies in humans (Price, Mao, Frenk & Mayer 1994) but recent evidence has shown an NMDA dependency of allodynias and wind-up pains in controlled clinical studies (Eide, Jorum, Stubhaug, Bremnes & Breivik 1994).
NITRIC OXIDE AND ARACHIDONIC ACID
Central plasticity can also involve a gas. Nitric oxide (NO), a diffusible gas, is produced in response to NMDA receptor activation and so may well mediate some or all of the consequences of NMDA receptor activation in nociception (Meller & Gebhart 1993). Blockers of nitric oxide synthase (NOS) are effective against inflammatory and neuropathic nociception in animals by spinal actions. There are hints that NO may feed back and enhance the release of the afferent transmitters and so set up a positive feedback loop (Sorkin 1993). In addition, an induction of NO in the afferents has been reported after nerve damage. Spinal production of arachidonic acid in response to C-fibre stimulation and NMDA receptor activation may achieve the same end and prevention of this could underlie some of the central analgesic effects of NSAIDS (Malmberg & Yaksh 1992).
The evidence for spinal actions of NSAIDS continues to grow based not only on the demonstrations of C-fibre and more importantly NMDA evoked release of prostanoids but that the spinal action of NSAIDS can be shown by intrathecal administration of these agents. In addition, it has been shown that the hyperalgesia produced by NMDA and substance P is reduced by spinal NSAIDS (Malmberg & Yaksh 1992). Since there is both evidence for prostanoid, NO and NMDA mediated release of glutamate and substance P it would appear that the production of novel mediators by NMDA receptor activation underlies retrograde messenger control of transmitter release (Sorkin 1993).
There will be problems with NOS blockers in therapy since NO is endothelium derived relaxing factor and systemic administration may induce analgesia but will be accompanied by severe hypertension. However, since it has been demonstrated that neuronal NOS differs from that in the endothelium it may be possible to separate these effects.
CONSEQUENCES OF CENTRAL HYPERSENSITIVITY
These are the presently characterized systems, operating in the spinal cord which appear to mediate central hypersensitivity, a state where amplification and prolongation of the afferent barrage occurs. The established roles of these different transmitter systems may offer novel targets for therapy. This may be important since wind-up, the hypersensitized reflex and several measures of NMDA dependent measures in neuropathy and allodynia models can be poorly sensitive to opioids (Dickenson 1994a). As discussed earlier, the reasons may be nociception through channels not controlled by opioids (large fibre induced allodynias), high levels of excitability (NMDA mediated amplification) or pathological loss of the opioid receptors.
TREATING OPIOID POORLY RESPONSIVE PAIN
The first approach is based on the animal studies showing that loss or dysfunction of pre-synaptic opioid receptors can be overcome by increasing the dose of opioid (Lombard & Besson 1989); (Xu & Wiesenfeld-Hallin 1991). In addition, in pains where the NMDA receptor is operating and there is reduced opioid sensitivity such as in some of the models for inflammatory and neuropathic pain, this too can be overcome by dose escalation (Dickenson & Sullivan 1989); (Chapman & Dickenson 1992). The simple augmentation of the dose of morphine should be first tried although side effects may confound this tactic. Another approach may be to use high efficacy opioids such as alfentanyl or sufentanil but data is lacking on this point. If opioids cannot produce the desired effects, different pharmacological approaches are possible and in the case of the NMDA receptor there are many experimental drugs which effectively block the receptor, the channel or associated sites. Some of these are in development as potential drugs but there clearly is a need for agents to be tested now. In fact, ketamine blocks the channel associated with the NMDA receptor and has current use in the relief of pain. Dextrophan and dextromethorphan are also antagonists at this site and are currently used in humans for their anti-tussive effects. Both have been shown to not only to reduce wind-up itself (Dickenson et al 1992) but to be effective in the Bennett model of neuropathic pain after spinal application (Mao et al 1993); (Tal & Bennett 1993) and in humans (Price, Mao, Frenk & Mayer 1994). Recently, the anti-Parkinson drug, memantine has been shown to be an effective NMDA antagonist. All could be used to test the clinical effectiveness of NMDA blockade in opioid poorly responsive pains. However, the NMDA antagonists would only be effective in reducing hyperalgesia not abolishing the pain (Dickenson 1994). These agents may turn out to be especially useful in the allodynias which sensitive to NMDA receptor antagonists but not to opioids. In fact not only is there psychophysical evidence for `wind-up' pain in humans being mediated by NMDA receptors, based on studies with dextromethorphan (Price et al1994), but clinical trials showing that ketamine can reduce allodynias, hyperalgesias and cause pain relief in circumstances where opioids had poor or restricted efficacy (Eide et al 1994).
One practical application of the poor opioid responsiveness of NMDA mediated pains is that the co-administration of morphine with low doses of an NMDA antagonist should be beneficial in these pain states. This is indeed the case and the combination has been shown to synergize in one study (Chapman & Dickenson 1992) and be additive in another model (Yamamoto & Yaksh 1992). However both studies have shown that the additional NMDA antagonism restores the opioid sensitivity of the responses. Furthermore, spinal local anaesthetics synergize with spinal morphine (Akerman, Arwestrom & Post 1988) partly due to the ability of the former to reduce NMDA mediated activity (Fraser, Chapman & Dickenson 1992). The spinal release of prostaglandins affords another target and centrally acting NSAIDS have been shown to reduce persistent inflammatory nociception and the behavioural hyperalgesia produced by spinal substance P and NMDA. Predictably, NSAIDS will synergize with opioids (Yaksh & Malmberg 1994).
It has been suggested that once these central hypersensitivity states have been induced, they remain active in the absence of peripheral inputs (Coderre et al 1993). There is counter-evidence from both animal studies and in human pain states where there is clear evidence for central changes that these are entirely dependent on peripheral inputs for maintenance (Dickenson and Sullivan 1987b, Gracely, Lynch & Bennett1993). Thus it would seem that the central pain hypersensitivity generators are continually triggered by afferent activity. Consequently there is a place for peripheral local anaesthetics although the symptoms may well reappear once the block wears off. In addition, there is evidence that systemic local anaesthetics may have selective effects on ectopic foci in a damaged peripheral nerve at doses which do not alter conduction in the nerve (Devor, Wall & Catalan 1992). Finally, as stated earlier, spinal sites of action of local anaesthetics include a reduction in wind-up (Fraser et al 1992).
In all of these studies on opioid poorly responsive pain, the emphasis has been on mu opioids, especially morphine, but mu and delta opioids and alpha-2 agonists all have similar effects on wind-up, reducing the initial responses of the cells but with wind-up breaking through the inhibitions as the stimulation continues and restoring the cell responses (Dickenson 1991); (Dickenson 1994) ). It is therefore unlikely that these three systems would have differential effects on NMDA receptor mediated events making drugs such as clonidine unlikely to be alternatives to morphine, at least with regard to spinal events where the NMDA receptor is implicated. In addition, the negative effects of CCK on analgesia are not only directed against mu opioids but also alpha-2 adrenoceptors. However, it is possible that in cases where pre-synaptic opioid receptors are reduced such as in cases of peripheral nerve pathology, alpha-2 receptors may persist at post -synaptic sites and so provide a therapeutic target. In addition, by systemic routes but also by the spinal route, there may well be sympathetic blocking effects of clonidine which will be of importance in sympathetically maintained pains. Yet again, alpha-2 adrenoceptor agonists produce powerful potentiations of opioid analgesia (Dickenson & Sullivan 1993); (Yaksh & Malmberg 1994).
PRE-EMPTIVE ANALGESIA
The idea of pre-emptive analgesia has arisen due to the potential for induction of hypersensitivity, genes and negative influences on opioid controls in addition to the well-established detrimental effects of the stress and hormonal responses to pain. Thus treating pain before it arises rather than waiting for it to develop appears to have a good rationale basis (Woolf 1994). Animal studies lend support to this in that several of the measures of central hypersensitivity are less sensitive to opioids given as a post-treatment as compared to pre-emptive administration (Chapman, Haley, & Dickenson 1994). Restricting comments to the use of opioids, there are several reasons why the clinical studies on pre-emptive analgesia have either been negative or have showed relatively weak benefits (McQuay 1994); (Dahl 1994). Most have used post-operative pain measures and invariably operative procedures will induce inflammation.
The key issues in comparing pain relief pre- and post-operatively are the exact timing of administration and that the baseline opioid sensitivity remains the same (McQuay 1994). With regard to the former, it may be that a pre-treatment is the same as an early post-treatment since they both will pre-empt late developing central hypersensitivity. Studies with varied timing of opioids on the formalin response support this idea (Chapman et al 1994). With regard to opioid sensitivity, it is well-established in animal models of carrageenan inflammation and arthritis that not only is novel peripheral opioid receptor mediated analgesia rapidly revealed but spinal opioid sensitivity is enhanced just as rapidly (Stanfa et al 1992); (Stein 1994). Thus post-treatment with opioids will impinge upon enhanced opioid systems so that comparisons with the effectiveness of pre-treatments will be biased. In neuropathic pains, where in general opioid sensitivity is reduced, inhibitions are lost and central and sympathetic processing is aberrant, the impact of pre-treatment is far more obvious.
THE ROLES OF INHIBITIONS
When NMDA mediated central events leading to hypersensitivity are active, in the shorter term models (formalin and the hypersensitive reflex) there is a reduced sensitivity to opioids whereas once several hours have elapsed (carrageenan and arthritic inflammation) opioid sensitivity is now found to be increased. In the case of the former models dose-escalation can overcome the reduced opioid sensitivity (Chapman et al 1994). The reasons for these differences could reside in the unchecked NMDA receptor mediation of activity in the shorter term models, eg. formalin, prior to the induction of the slower developing inhibitory changes overcomes opioid inhibitions. The compensatory increases in spinal opioid sensitivity (via altered CCK) and other inhibitions means that the longer term acute pains respond well to opioids, partly to the enhanced opioid effectiveness per se but also because the increased non-opioid (GABA and alpha-2 adrenoceptor mediated) inhibitory events will reduce the NMDA driven level of excitability (Stanfa et al 1994); (Castro-Lopes, Tavares, Tšlle, & Coimbra, 1994). A good example of the potential controlling influence of inhibitions is seen in the formalin response. NMDA receptor activation in the spinal cord amplifies a low level of C-fibre input (Heapy, Jamieson & Russell 1987) to generate the characteristic response. The resultant behavioural and neurophysiological responses to the peripheral injection of formalin last for one hour but the C-fibre inputs continue for over two hours and the peripheral inflammation for even longer (Porro & Cavazzuti 1993). The central responses must surely be curtailed by inhibitory controls. If the plastic changes are leading to compensatory increased central inhibitions such as these then pre-emptive approaches may prevent both this beneficial plasticity as well as the target of central hypersensitivity mechanisms. If this is the case, as the pre-emptive agent wears off the pain may return in the absence of compensatory inhibitions so that the pain intensity is heightened.
In general, the roles of inhibitions has received less attention than the excitatory systems but in an animal model of allodynia, NMDA antagonists are effective against the tactile evoked nociception whereas morphine is not (Yaksh 1989 ). In this model the NMDA mediated allodynia is induced by blockade of spinal GABA or glycine inhibitory tone in normal animals. This has bearing on the point made previously that inhibitions are important in controlling and limiting the extent of NMDA receptor participation in nociceptive processing in the spinal cord. Increasing GABA function by the administration of benzodiazepines reduces NMDA mediated hyperalgesia after ischaemia (Cartmell & Mitchell 1993) and another intriguing example of this is the enhanced NMDA mediated nociception seen when spinal glycine inhibitions are blocked in a neuropathic model in the rat (Seltzer et al 1991). By contrast, GABA upregulation may be an intrinsic compensatory mechanism in longer term inflammation (Castro-Lopes et al 1994); (Stanfa & Dickenson 1994).
Conceivably, failure of these and other inhibitory controls may lead to chronicity of pain. In fact there is strong evidence that the hyperalgesia seen in neuropathic models is as much a consequence of loss of inhibitions and reorganization as excess excitations (Woolf & Doubell 1994); (Dray et al 1994) which could be due to a destructive loss of inhibitory interneurones, itself exacerbated by pharmacological block of these inhibitory systems (Sugimoto et al 1990).
DEVELOPMENTAL ASPECTS
Examination of events in the neonatal spinal cord are of great relevance to the role of modulatory systems in controlling excitation since the maturation of the inhibitory systems is slow. Paediatric pain control has to take into account the findings that the development of the nervous system is accompanied by marked changes in many transmitters and receptors over time. A number of studies on the anatomical and functional development of the excitatory and inhibitory pharmacology of the rat and human spinal cord suggest that the rodent provides a good model for investigating clinical questions (Fitzgerald 1991)
Studies using the neonatal spinal cord have shown that all the excitatory and inhibitory receptors covered in this account are functional at day 1 in the rat and many of these neuropharmacological systems are present even before birth. Full maturation of the endogenous transmitter pathways and connections is much slower (Marti, Gibson, Polak, Facer, Springall, Van Aswegen,& Koltzenburg 1987), particularly the local inhibitory and long descending controls. The levels of excitatory transmitters such as substance P and excitatory amino-acids including glutamate tend to increase during development At the same time NK and NMDA receptor location in young animals is far more exuberant than in adults and their numbers decline with time as the receptors shrink back to assume the discrete adult form (Charlton & Helke 1986). Dendritic development, interneurones (Bicknell & Beal 1984) and descending inhibitory controls (Fitzgerald & Koltzenburg 1986) are slowly developing so that modulation of excitability is delayed. The conclusion from these studies is that the transmission of pain in the spinal cord of the young is likely to be exaggerated compared to the adult as a result of excess early development of excitation and delayed maturity of the intrinsic controlling inhibitory systems.
This can be seen, again using the formalin response but now in neonatal rats (Guy & Abbot 1992). In the first week of life, the response to the peripheral algogen is dramatic and disruptive to normal behaviour. As development proceeds the response declines until at about 3 weeks, when inhibitory systems have matured, it resembles the more discrete adult form of response.
Opioids and their three receptors also change with time and the relative arrival of the mu, delta and kappa receptors differs as does the development of the endogenous opioids. Interestingly, although the opioid receptor affinities resemble the adult at very early stages the numbers of receptors decline over time (Attali, Saya & Vogel 1990); (Sales, Charnay, Zajac, Dubois, & Roques 1989) . The early ontogeny of the opioid receptors means that there is a substrate present for the production of analgesia by exogenous opioids whereas the controlling influences of endogenous opioid neuronal systems may only appear later in development.
Thus the systems in the adult that generate spinal hypersensitivity may be more effective in the young nervous system as a result of less inhibitory influences as a result of immaturity of these systems. Exogenous activation of opioid receptors allows pain in the young to be controlled by adequate analgesia either with opioids alone or by some of the combinations discussed previously. Since activity dependent plasticity is important in determining maturation of connectivity in the developing nervous system it has to be considered that uncontrolled pain in the young may have long term consequences for the neurobiology of pain. In this respect, fears regarding the consequences of opioid use in young children, which are probably unfounded anyway based on a lack of any dramatic effect of early opioid exposure on later function (Bardo, Bhatnagar & Gebhart 1982), pale into insignificance compared to the possible permanent alterations in sensory processing that uncontrolled pain in the young could produce.
CONCLUSIONS
There still remain many questions to be answered with regard to the events in different pain states that can alter, in either direction, the analgesic effects of opioids. As has been discussed, this recent accumulating knowledge of plasticity provides a rational basis for combination therapy (Dickenson & Sullivan 1993); (Dickenson. 1994b ) (Yaksh & Malmberg 1994) and examples given here fall into two categories. Firstly, the combination of opioids with other inhibitory agents such as alpha-2 adrenoceptor agonists and secondly, dual therapy where the non-opioid acts to reduce excitability or to control interfering systems (CCK and NMDA receptor antagonists, NSAIDS, local anaesthetics). Dual therapy such as this has been shown to result in restoration, additivity or potentiation of opioid analgesia (Dickenson & Sullivan 1993); (Yaksh & Malmberg 1994). It is also becoming increasingly clear that particular pain states have different plasticities; despite some similarities (eg. NMDA mediation of hypersensitivity), inflammatory and neuropathic pains are not only very different from each other but cannot be viewed as uniform syndromes. For example, in neuropathic pains the human and animal studies on opioid sensitivity do not reach a consensus; animal studies suggest both good (Attal, Chen, Kayser & Guilbaud 1991); (Yamamoto & Yaksh 1992) and poor opioid sensitivity and the clinical studies are not totally in agreement on this point either (Arner & Meyerson 1988); (Portenoy et al 1990); (Jadad et al 1992). Particular monotherapies and combination therapies will likely be appropriate for different states within a pain syndrome (McQuay & Dickenson 1990). Other key points that need investigation are the following: Does a repeated noxious insult alter the plasticity? Knowledge is building up for the neonate but what happens in the ageing nervous system? The answers to these and other questions relating to plasticity have important consequences for the clinical treatment of pain.
ACKNOWLEDGEMENTS
The funding of research from the authors laboratory is from the Medical Research Council and The Wellcome Trust. Major contributions to this account, both experimental and conceptual, have come from Louise Stanfa, Vicky Chapman, Ann Sullivan, Alison Reeve, Henry McQuay, Eija Kalso, Maria Fitzgerald and Jane Haley. Their enthusiasm, dedication and assistance is gratefully acknowledged.
REFERENCES
Akerman, B., Arwestrom, E. & Post, C. (1988) Local anesthetics potentiate spinal morphine analgesia. Anesthesiology and Analgesia 67: 943-948.
Arner, S. & Meyerson, B.A. (1988) Lack of analgesic effect of opioids on neuropathic and idiopathic forms of pain. Pain 33: 11-23.
Attal, N., Chen, Y.I., Kayser, V. & Guilbaud, G.L. (1991) Behavioural evidence that systemic morphine may modulate a phasic pain-related behaviour in a rat model of peripheral mononeuropathy. Pain 47: 65-70.
Attali, B., Saya, D. & Vogel, Z. (1990) Pre- and postnatal development of opiate receptor subtypes in the rat spinal cord. Developmental Brain Research 54: 97-102.
Baber, N.S., Dourish, C.T. & Hill, D.R. (1989) The role of CCK, caerulein, and CCK antagonists in nociception. Pain 39: 307-328.
Bardo, M.T., Bhatnagar, R.K. & Gebhart, G.F. (1982) Differential effects of chronic morphine and naloxone on opiate receptors, monoamines, and morphine- induced behaviours in preweanling rats. Developmental Brain Research 4: 139-147.
Battaglia, G. & Rustioni, A. (1988) Coexistence of glutamate and substance P in dorsal root ganglion cells of the rat and monkey. J Comparative Neurology 277:302-312.
Besse, D., Lombard, M.C., Zakac, J.M., Roques, B.P. & Besson, J.M. (1990) Pre- and postsynaptic distribution of mu, delta and kappa opioid receptors in the superficial layers of the cervical dorsal horn of the rat spinal cord. Brain Research 521:15-22.
Besson, J.M. & Chaouch, A. (1987) Peripheral and Spinal Mechanisms of Nociception. Physiological Reviews 67: 67-186.
Bicknell, H.R. & Beal, J.A. (1984) Axonal and dendritic development of substantia gelatinosa neurons in the lumbosacral spinal cord of the rat. The Journal of Comparative Neurology 226: 508-522
Cartmell, S.M. & Mitchell, D. (1993) Attenuation of reperfusion hyperalgesia in the rat by systemic administration of benzodiazepines, British Journal of Pharmacology 110: 1067-1072.
Castro-Lopes, J.M., Tavares, I., Tšlle, T.R. & Coimbra, A. (1994) Carrageenan- induced inflammation of the hind foot provokes a rise of GABA- immunoreactive cells in the rat spinal cord that is prevented by peripheral neurectomy or neonatal capsaicin treatment. Pain 56:193-201.
Chapman, V & Dickenson, A.H. (1992) The combination of NMDA antagonism and morphine produces profound antinociception in the rat dorsal horn. Brain Research 573: 321-323.
Chapman, V & Dickenson, A.H. (1994) Enhanced responses of rat dorsal horn neurones after ultraviolet irradiation of the hindpaw; roles of the NMDA receptor, Neuroscience Letters 176, 41-44.
Chapman, V., Haley, J.E & Dickenson, A.H. (1994) Electrophysiological analysis of preemptive effects of spinal opioids on N-Methyl-D-Aspartate receptor mediated events. Anesthesiology 81, 1429-35
Charlton, C.G. & Helke, C.J. (1986) Ontogeny of substance P receptors in rat spinal cord: quantitative changes in receptor number and differential expression in specific loci. Developmental Brain Research 29: 81-91.
Collingridge, G. & Singer, W. (1990) Excitatory amino acid receptors and synaptic plasticity. Trends in Pharmacological Sciences 11: 290-296.
Dado, R.J., Law, P.Y, Loh, H.H. & Elde, R. (1993) Immunofluorescent identification of a delta (d)-opioid receptor on primary afferent nerve terminals. NeuroReport, 5:341-344.
Dahl, J. (1994) Neuronal plasticity and pre-emptive analgesia: implications for the management of postoperative pain. Danish Medical Bulletin 41: 434-441.
Daw, N.W., Stein, P.S.G. & Fox, K. (1993) The role of NMDA receptors in information processing. Annual Rev iew of Neuroscience 16: 207-222.
Devor, M., Wall, P.D & Catalan, N. (1992) Systemic lidocaine silences ectopic neuroma and DRG discharge without blocking nerve conduction. Pain 48: 261-268.
Dickenson, A.H. (1990) A cure for wind-up: NMDA receptor antagonists as potential analgesics. Trends in Pharmacological Sciences 11:307-309.
Dickenson, A.H. (1991) Mechanisms of the analgesic actions of opiates and opioids. British Medical Bulletin 47: 690-702.
Dickenson, AH (1994a) Where and How Opioids Act Proceedings of the 7th World Congress on Pain, Progress in Pain Research and Management Vol. 2. eds G.F.Gebhart, D.L. Hammond and T.S. Jensen. IASP Press, Seattle. pp. 525- 552.
Dickenson, A.H. (1994b) NMDA receptor antagonists as analgesics. In: Progress in Pain research and Management, ed. H.L. Fields & J.C. Liebeskind. Seattle: IASP Press, 173-187.
Dickenson, A,H, & Sullivan, A.F. (1986) Electrophysiological studies on the effects of the effects of intrathecal morphine on nociceptive neurones in the rat dorsal horn. Pain 24: 211-222.
Dickenson, A.H. & Sullivan, A.F. (1993) Combination therapy in analgesia; seeking synergy. Current Opinion in Anaesthesiology 6: 861-865.
Dickenson, A.H., Sullivan. A.F., Fournie-Zaluski, M.-C. & Roques, B. (1987) Prevention of degradation of endogenous enkephalins produces inhibition of nociceptive neurones in rat spinal cord. Brain Research 408: 185-191.
Dickenson, A.H., Sullivan, A.F., Knox, R.J., Zajac, Z. & Roques, B.P. (1987) Opioid receptor types in the rat spinal cord: electrophysiological studies with m and d opioid receptor agonists in the control of nociception. Brain Research 413: 649-652.
Dougherty, P.M., Palecek, J., Paleckova, V., Sorkin, L. & Willis, W.D. (1992) The role of NMDA and non-NMDA excitatory amino-acid receptors in the excitation of primate spinothalamic tract neurons by mechanical, chemical, thermal and electrical stimuli. J.ournal of Neurosci.ence 12: 3025-3041.
Dray, A., Urban, L. & Dickenson, A.H. (1994) Pharmacology of Chronic Pain. Trends in Pharmacological Sciences 15: 190-197.
Dubner, R. & Ruda, M.A. (1992) Activity dependent neuronal plasticity following tissue injury and inflammation. Trends in Neurosciences 15: 96- 103.
Duggan, A.W., Hendrey, I.A., Morton, C.R., Hutchison, W.D. & Zhao, Z.O. (1988) Cutaneous stimuli releasing immunoreactive substance P in the dorsal horn of the cat. Brain Research 451: 261-273.
Duggan, A.W. & North, R.A. (1984) Electrophysiology of opioids. Pharmacological Reviews35 : 219-281.
Eide, P.K., Jorum, E., Stubhaug, a., Bremnes, J. & Breivik, H. (1994) Relief of post- herpetic neuralgia with the N-methyl-D- aspartate receptor antagonist ketamine: a double blind, cross-over comparison with morphine and placebo. Pain 58: 347-354.
Fields, H.L., Barbaro, N.M. & Heinricher, M. M. (1988) Brain stem neuronal circuitry underlying the antinociceptive action of opiates. In: Progress in Brain Research, Pain modulation Vol. 77 eds. Fields H.L and Besson JM. Elsevier, Amsterdam pp 245-258
Fitzgerald, M.(1991) Development of pain mechanisms. British Medical Bulletin 47 667-675.
Fitzgerald, M. & Koltzenburg, M (1986) The functional development of descending inhibitory pathways in the dorsolateral funiculus of the newborn rat spinal cord. Developmental Brain Research 24: 261-270.
Fraser, H., Chapman, V. & Dickenson, A.H. (1992) Spinal local anaesthetic actions on afferent evoked responses and wind-up of nociceptive neurones in the rat spinal cord: combination with morphine produces marked potentiation of antinociception. Pain 49: 33-41.
Ghilardi, J.R., Allen, C.J., Vigna, S.R., McVey, D.C. & Mantyh, P.W. (1992) Trigeminal and dorsal root ganglion neurons express CCK receptor binding sites in the rat, rabbit, and monkey: possible site of opiate-CCK analgesic interactions. Journal of Neuroscience 12: 4854-4866.
Gogas, K.R., Presley, R.W., Levine, J.D. & Basbaum, A.I. (1991) The antinociceptive effects of supraspinal opioids results from an increase in descending inhibitory control: correlation of nociceptive behaviour and c-fos expression. Neuroscience 42: 617-628.
Gong, Q.L., Hedner, J., Bjorkman, R. & Hedner, T. (1992) Morphine-3-glucuronide may functionally antagonize morphine-6-glucuronide induced antinociception and ventilatory depression in the rat. Pain 48: 249-255.
Gracely, R., Lynch, S.A. & Bennett, G.J. (1993) Painful neuropathy: altered central processing maimtained dynamically by peripheral input. Pain 52: 251-253.
Guy, E.R. & Abbot, F.V. (1992) The behavioral response to formalin in preweanling rats. Pain, 51: 81-90.
Haley, J.E., Sullivan, A.F. & Dickenson, A.H. (1990) Evidence for spinal N- methyl-D-aspartate receptor involvement in prolonged chemical nociception in the rat. Brain Research 518: 218-222.
Heapy, C.G., Jamieson, A. & Russell, N.J.W. (1987) Afferent C-fibre and A-delta activity in models of inflammation. British. Journal of Pharmacology.90: 164P.
Hewett, K., Dickenson, A.H. & McQuay, H.J. (1993) Lack of effect of morphine-3- glucuronide on the spinal antinociceptive actions of morphine in the rat: an electrophysiological study. Pain 53: 59-63.
Hylden, J.L.K., Thomas, D.A., Iadorola, M.J., Nahin, R.L. & Dubner, R. (1991) Spinal opioid analgesic effects are enhanced in a model of unilateral inflammation/hyperalgesia: possible involvement of noradrenergic mechanisms. European Journal of Pharmacology 194: 135-143.
Jadad, A.R., Carroll, D., Glynn, C.J., Moore, R.A. & McQuay, H.J. (1992) Morphine responsiveness of chronic pain: double-blind randomized crossover study with patient-controlled analgesia. The Lancet 339: 1367-1371.
Jiang, Q.A.E., Takemori, A.E., Sultana, M., Portoghese, P.S., Bowen, W.D., Mosberg, H.I. & Porreca, F. (1991) Differential antagonism of opioid delta antinociception by [D-Ala2, Leu5, Cys6] enkephalin and naltrindole 5'- isothiocynanate: evidence for delta receptor subtypes. Journal of Pharmacology & Experimental Therapeutics 257: 1069-1077.
Kalso, E.A., Sullivan, A.F., McQuay, H.J., Dickenson, A.H. & Roques, B.P. (1993) Cross-tolerance between mu opioid and alpha-2 adrenergic receptors, but not between mu and delta opioid receptors in the spinal cord of the rat. Journal of Pharmacology & Experimental Therapeutics 265: 551-558.
Kangra, I. & Randic, M. (1990) Tachykinins and calcitonin gene-related peptide enhance the release of endogenous glutamate and aspartate from rat spinal dorsal horn slice. Journal of. Neuroscience. 10: 2026-2038.
Kosterlitz, H.W. (1985) Opioid peptides and their receptors. Proceedings of the Royal Society of London. B 225: 27-40.
Knox, R.J. & Dickenson, A.H. (1987) Effects of selective and non-selective k opioid agonists on cutaneous C-fibre evoked responses of rat dorsal horn neurones. Brain Research 415 : 21-29.
Lang, C.W., Duggan, A.W & Hope, P.J., (1991) Analgesic doses of morphine do not reduce the noxious evoked release of immunoreactive neurokinins in the dorsal horn of the cat spinal cord. British Journal of Pharmacology 103: 1871-1876.
Le Bars D, Villaneuva L. (1988) Electrophysiological evidence for the activation of descending inhibitory controls by nociceptive afferent pathways, In: Progress in Brain Research, Pain modulation. Vol. 77.eds Fields H.L and Besson JM eds.Elsevier, Amsterdam pp 275-300.
Levine, J.D., Fields, H.L. & Basbaum, A.I. (1993) Peptides and the primary afferent nociceptor. Journal of Neuroscience 13: 2273-86.
Lombard, M.C.& Besson, J.M. (1989) Attempts to gauge the relative importance of pre-and postsynaptic effects of morphine on the transmission of noxious messages in the dorsal horn of the rat spinal cord. Pain 37: 335-346
Magnuson, D.S.K. & Dickenson, A.H. (1991) Lamina-specific effects of morphine and naloxone in dorsal horn of rat spinal cord in vitro. J Neurophysiology 66: 1941-1950
Malmberg, A.B. & Yaksh, T.L. (1992)Antinociceptive effects of spinal nonsteroidal antiinflammatory on the formalin test in the rat. J. Pharmacology and Experimental Therapeutics. 263: 136-146.
Marti, E., Gibson, S.J., Polak, J.M., Facer, P., Springall, D.R., Van Aswegen, G. & Koltzenburg, M. (1987) Ontogeny of peptide- and amine-containing neurones in motor, sensory, and autonomic regions of rat and human spinal cord, dorsal root ganglia, and rat skin. The Journal of Comparative Neurology 266: 332-359.
Maves, T.J. & Gebhart, G.F. (1992) Antinociceptive synergy between intrathecal morphine and lidocaine during visceral and somatic nociception in the rat. Anesthesiology76: 91-99
Max, M.B. (1994) Antidepressants as analgesics In: Progress in Pain research and Management, ed. H.L. Fields & J.C. Liebeskind. Seattle, IASP Press pp229-246
McMahon, S.B., Lewin, G.R. & Wall PD. (1993) Central excitability triggered by noxious inputs. Current Opinion in Neurobiology 3: 602-610
McQuay, H.J. (1994) Do preemptive treatments provide better pain control? In: Proceedings of the 7th World Congress on Pain, Progress in Pain Research and Management Vol. 2. eds G.F.Gebhart, D.L. Hammond and T.S. Jensen. IASP Press, Seattle. pp. 709-723.
McQuay H, Dickenson AH. (1990) Implications of central nervous system plasticity for pain management. Anesthiesiology 45:101-2.
Meller S.T. & Gebhart G.F. (1993) Nitric oxide (NO) and nociceptive processing in the spinal cord. Pain 52: 127-136
Millan MJ. (1990) k opioid receptors and analgesia. Trends in Pharmacological Sciences 11: 70-76
Neugebauer, V., Lucke, T. & Schaible, H-G. (1993). N-methyl-D-aspartate(NMDA) and non-NMDA receptor antagonists block the hyperexcitability of dorsal horn neurons during development of acute arthtitis in rat's knee joint. Journal of Neurophysiology 70: 1365-1377.
North RA. (1989) Drug receptors and the inhibition of nerve cells. British Journal of Pharmacology 98: 13-28.
Ossipov, M.H., Kovelowski, C.J. & Porreca, F. (1995) The increase in potency produced by carrageenan-induced inflammation is blocked by naltrindole, a selective d opioid antagonist. Neuroscience Letters 184: 173-176.
Otsuka M. & Yoshioka K. (1993) Neurotransmitter functions of mammalian tachykinins. Physiological Reviews 73: 229-3037.
Pasternak G.W. & Wood PJ. (1986) Multiple opioid receptors. Life Sciences 28:1889-1893
Portenoy R.K, Foley, K.M & Inturrisi CE. (1990) The nature of opioid responsiveness and its implications for neuropathic pain: new hypotheses derived from studies of opioid infusions. Pain 43: 273-286
Porro, C.A & Cavazzuti, M. (1993) Spatial and temporal aspects of spinal cord and brainstem activation in the formalin pain model. Progress In Neurobiology 41, 565-607.
Price D.D, Mao, J & Mayer D.J. (1994) Central neural mechanisms of normal and abnormal pain states In: Progress in Pain research and Management, ed. H.L. Fields & J.C. Liebeskind. Seattle, IASP Press 61-84.
Price, DD., Mao, J., Frenk, H. & Mayer, D.J. (1994) The N-methyl-D-aspartate antagonist dextromethorphan selectively reduces temporal summation of second pain. Pain in press
Roques B.P, Noble, F, Dauge, V, Fournie-Zaluski, MC, Beaumont, A. (1993) Neutral endopeptidase 24.11: structure, inhibition and experimental and clinical pharmacology. Pharmacological Reviews 45: 88-146
Sales, N., Charnay, Y., Zajac, J.-M., Dubois, P.-M. & Roques, B.P. (1989) Ontogeny of m and d opiod receptors and of neutral endopeptidase in human spinal cord: an autoradiographic study. Journal of Chemical Neuroanatomy 2: 179-188.
Schaible, H.G., Hope, P.J., Lang, C.W. & Duggan, A.W (1992) Calcitonin gene related peptide causes intraspinal spreading of substance P released by peripheral stimulation. European Journal of Neuroscience, 4: 750-757.
Seltzer, Z., Cihin, S., Ginzburg, R. & Beilin, B., (1991) Modulation of neuropathic pain behavior in rats by spinal disinhibition and NMDA receptor blockade of injury discharge. Pain 45: 69-75.
Sluka, K.A., Dougherty, P.M., Sorkin, L.S., Willis, W.D. and Westlund, K.N. (1992) Neural changes in acute arthritis in monkeys. III. Changes in substance P, calcitonin gene-related peptide and glutamate in the dorsal horn of the spinal cord. Brain Research Reviews 17: 29-38.
Smith, M.T, Watt, J.A, Cramond, T. (1990) Morphine-3-glucuronide - a potent antagonist of morphine analgesia. Life Sciences 47: 579-585
Sorkin, L. (1993) NMDA evokes an L-NAME sensitive spinal release of glutamate and citrulline. NeuroReport 4: 479-482.
Stanfa, L.C & Dickenson A.H. (1993) Cholecystokinin as a factor in the enhanced potency of spinal morphine following carrageenan inflammation. British Journal of Pharmacology 108: 967-973
Stanfa L.C & Dickenson A.H. (1994a) Enhanced alpha-2 adrenergic controls and spinal morphine potency in inflammation. Neuroreport 5: 469-72.
Stanfa L.C & Dickenson A.H. (1994b) Electrophysiological studies on the spinal roles of endogenous opioids in carrageenan inflammation. Pain 56: 185-191
Stanfa, L.C, Dickenson A.H, Xu, X-J & Wiesenfeld-Hallin, Z (1994) Cholecystokinin and morphine analgesia: variations on a theme. Trends in Pharmacological Sciences 15: 65-6.
Stanfa LC, Sullivan AF, Dickenson AH. (1992) Alterations in neuronal excitability and the potency of spinal mu, delta and kappa opioids after carrageenan induced inflammation. Pain 50: 345-54.
Stein, C. (1994) Interaction of immune-competent cells and nociceptors Proceedings of the 7th World Congress on Pain, Progress in Pain Research and Management Vol. 2. eds G.F.Gebhart, D.L. Hammond and T.S. Jensen. IASP Press, Seattle.pp 285-297.
Stein C, Millan, M.J, Shippenberg, T.S, Peter, K & Herz, A. (1992) Peripheral opioid receptors mediating antinociception in inflammation. Evidence for involvement of mu. delta and kappa receptors.Journal of Pharmacology & Experimental Therapeutics248: 1269-1275
Sugimoto, T, Bennet, G. & Kajander, K.C. (1990) Transynaptic degeneration in the superficial dorsal horn after sciatic nerve injury, transection and strychnine. Pain 42: 385-393.
Sullivan, A.F, Dickenson, A.H & Roques, B.P. (1989) d- opioid mediated inhibitions of acute and prolonged noxious-evoked responses in rat dorsal horn neurones. British Journal of Pharmacology 98:1039-1049
Sullivan, A.F, McQuay, H.J, Bailey, D & Dickenson, A.H. (1989) The spinal antinociceptive actions of morphine metabolites morphine-6-glucuronide and normorphine in the rat. Brain Research 482: 219-224
Todd A.J & Spike, RC. (1993) The localization of claasical transmitters and neuropeptides within neurons in laminae I-III of the mammalian spinal dorsal horn. Progress in Neurobiology 41: 609-645.
Traynor, J. (1989) Subtypes of the k opioid receptor: fact or fiction? Trends in Pharmacological Sciences 10: 52-53
Uhl, G.R, Childers, S & Pasternak, G. (1994) An opiate receptor gene family reunion.Trends in Pharmacological Sciences 17: 89-93
Urban, L., Thompson, S.W.N. & Dray, A. (1994) Modulation of spinal excitability: co-operation between neurokinin and excitatory amino acid neurotransmitters. Trends in Neurosciences. 17: 432-438.
Vanderah, T.W., Lai, J., Yamamura, H.I. & Porreca, F. (1994) Antisense oligonucleotide to the CCKB receptor produces naltrindole- and [Leu5] enkephalin antiserum-sensitive enhancement of morphine antinociception. Neuroreport 5: 2601-2605.
Wang. J. Ren, M & Han, J. (1992) Mobilization of calcium from intracellular stores as one of the mechanisms underlying the antiopioid effects of cholecystokinin octapeptide. Peptides 13: 947-951
Wiertelak, E.P. Maier, S.F. & Watkins, L.R. (1992) Cholecystokinin antianalgesia: safety cues abolish morphine analgesia. Science 256: 830-833
Woolf, C.J. (1983) Evidence for a central component of post-injury pain hypersensitivity. Nature 306: 686-688.
Woolf, C.J. (1994) A new strategy for the treatment of inflammatory pain. Drugs Suppl. 5: 1-9.
Woolf, C.J. & Doubell, T.P. (1994) The pathophysiology of chronic pain - increased sensitivity to low threshold Ab-fibre inputs. Current Opinion in Neurobiology. 4: 525-534.
Woolf, C.J & Thompson, S.W.N. (1991) The induction and maintenance of central sensitization is dependent on N-methyl-D-aspartic acid receptor activation; implications for the treatment of post-injury hypersensitivity states. Pain 44: 293-299
Xu, X.J. Puke, M.J.C. Verge, V.M.K. Wiesenfeld-Hallin, Z. Hughes, J. & Hokfelt, T. (1993) Up-regulation of cholecystokinin in primary sensory neurons is associated with morphine insensitivity in experimental neuropathic pain in the rat. Neuroscience Letters 152:129-132
Yaksh, T.L. (1989) Behavioural and autonomic correlates of the tactile evoked allodynia produced by spinal glycine inhibition: effects of modulatory receptor systems and excitatory amino acid antagonists. Pain 37:111-23.
Yaksh, T.L. AL-Rodhan, N. & Jenson,T.S. (1988) Sites of action of opiates in production of analgesia. In: Progress in Brain Research eds HL Fields, JM Besson Amsterdam Elsevier pp371-394
Yaksh, T.L. & Nouiehed, R. (1985) Physiology and Pharmacology of spinal opiates. Annual Review of Pharmacology & Toxicology 25: 433-462
Yaksh, T.L. & Malmberg, A.B. (1994) Interaction of spinal modulatory systems.In: Progress in Pain research and Management, ed. H.L. Fields & J.C. Liebeskind. Seattle: IASP Press 151-171.
Yamamoto, T. & Yaksh, T.L. (1992) Studies on the spinal interaction of morphine and the NMDA antagonist MK-801 on the hyperesthesia observed in a rat model of sciatic mononeuropathy. Neuroscience Letters 135: 67-70
FIGURE LEGEND
The diagram depicts the interactions between the different pahrmacological systems descrribed in the chapter. Activity generated in peripheral sensory nerves releases a number of transmitters into the dorsal horn of the spinal cord. Their receptor actions and interactions, subject to control by local and supraspinal pathways, determines the output from dorsal horn projection neurones.
Many of these systems are subject to plasticity.