The analysis of pain mechanisms in preclinical animal models of inﬂammatory and neuropathic pain has shown that iGluRs and mGluRs play important roles in peripheral sensitization, central sensitization, and descending pain mod- ulation (Fisher et al. 2000; Fundytus 2001; Carlton 2001; Willis 2001; Parsons
2001; Neugebauer and Carlton 2002; Neugebauer 2002; Varney and Gereau
2002; Ruscheweyh and Sandkuhler 2002; Lesage 2004; Woolf 2004).
Ionotropic Glutamate Receptors
Peripheral administration of NMDA or non-NMDA receptor antagonists in- hibits nociceptive behavior in the second phase of the formalin test and thermal and mechanical hyperalgesia associated with inﬂammation of the hindpaw (complete Freund’s adjuvant CFA- or carrageenan-induced) and knee joint (kaolin/carrageenan-induced) (Carlton 2001; Parsons 2001; Du et al. 2003). NMDA receptor antagonists tested in these studies include the clinically used compounds dextrorphan (a metabolite of dextromethorphan), ketamine, and memantine. Peripheral NMDA receptors might also be involved in visceral pain, as memantine was shown to inhibit the responses of single ﬁbers in the decentralized pelvic nerves to colorectal distention (McRoberts et al. 2001). In a model of neuropathic pain, block of peripheral NMDA receptors (by MK-
801), but not non-NMDA receptors (by NBQX), attenuated and delayed the onset of mechanical hyperalgesia (Jang et al. 2004). The potential therapeutic value of peripheral NMDA, AMPA, and kainate receptor antagonists still awaits a systematic analysis in models of inﬂammatory and, particularly, neuropathic pain.
Spinal administration of NMDA or non-NMDA antagonists is well known to inhibit central sensitization in various models of inﬂammatory and neuro- pathic pain (Fisher et al. 2000; Fundytus 2001; Willis 2001; Parsons 2001). NMDA or non-NMDA receptor antagonists in the spinal cord reversed the enhanced responses of dorsal horn neurons, including spinothalamic tract cells, in inﬂammatory pain states induced by intradermal capsaicin, intra- plantar carrageenan, or intraarticular kaolin and carrageenan (Stanfa and Dickenson 1999; Fisher et al. 2000; Fundytus 2001; Willis 2001; Parsons 2001; Schaible et al. 2002). Compounds tested in these studies include the clini- cally used NMDA receptor antagonists ketamine and memantine, a somewhat AMPA receptor-preferring antagonist (NBQX), and a GluR5 kainate receptor- selective antagonist (LY382884). Central sensitization of dorsal horn neurons, including spinothalamic tract cells, in models of neuropathic pain has also been shown to be inhibited by spinal administration of NMDA (memantine, dextrorphan, and MK-801) and non-NMDA (CNQX and LY382884) receptor antagonists (Carlton et al. 1997; Carlton et al. 1998; Parsons 2001; Palecek et al.
2004; Willis and Coggeshall 2004).
Behavioral data also show that NMDA and non-NMDA receptor antagonists are antinociceptive in inﬂammatory and neuropathic pain models (Fisher et al.
2000; Fundytus 2001; Willis 2001; Parsons 2001; Soliman et al. 2005). Spinally or systemically administered NMDA receptor antagonists, including memantine and dextromethorphan, inhibit nociceptive behavior, thermal hyperalgesia, and mechanical allodynia associated with peripheral inﬂammation as well as visceral hypersensitivity. NMDA receptor antagonists also reduce heat hyper- algesia and, less consistently, mechanical allodynia associated with peripheral nerve injury. Interestingly, intrathecal administration of an NR2B subunit- selective antagonist (CP-101606) inhibited mechanical allodynia in a model of postoperative pain (Nishimura et al. 2004) but had no effect on mechanical allodynia in a neuropathic pain model whereas memantine was antinocicep- tive in that model (Nakazato et al. 2005), suggesting differential roles of spinal NR2B in different forms of pain. Systemic administration of another agent with NR2B-selective antagonistic properties (ifenprodil) had antinociceptive effects in inﬂammatory and neuropathic pain models but, unlike MK-801 and memantine, did not inhibit NMDA-evoked responses of dorsal horn neurons (Nakazato et al. 2005). Thus the role of spinal NR2B subunit-containing NMDA receptors in nociceptive processing remains unclear.
Intrathecally or systemically administered non-NMDA receptor antagonists, including NBQX, have antinociceptive effects in animal models of inﬂamma- tory and neuropathic pain, but they also produce side effects such as ataxia.
A selective AMPA receptor antagonist (GYKI53655), however, did not produce antinociception in the formalin pain test at doses that did not cause ataxia (Sim- mons et al. 1998), suggesting that kainate rather than AMPA receptors may be
useful targets. Indeed, more recent studies focused on kainate receptors and showed that systemic administration of a GluR5/6-selective antagonist (SYM
2081; 2S,4R-4-methylglutamic acid; Sutton et al. 1999) and GluR5-selective
antagonists such as LY382884 (Simmons et al. 1998) and orally active com- pounds LY467711 and LY525327 (Dominguez et al. 2005) inhibited nociceptive behavior in the formalin test, thermal hyperalgesia, and mechanical allodynia induced by carrageenan or capsaicin, and thermal hyperalgesia and mechani- cal allodynia in a model of peripheral nerve injury. The site or sites of action of these systemically administered drugs remain to be determined.
In the brainstem (RVM), NMDA receptor activation is involved in descend- ing facilitation whereas inhibition involves activation of non-NMDA receptors (Ren and Dubner 2002; Gebhart 2004; Vanegas and Schaible 2004). NMDA- dependent descending facilitation appears to be particularly important for the development of secondary thermal hyperalgesia and mechanical allodynia in the early stages of inﬂammatory pain; it may be important for maintenance rather than initiation of neuropathic pain states as well (Ren and Dubner
2002; Porreca et al. 2002; Heinricher et al. 2003; Gebhart 2004; Vanegas and
Schaible 2004). Microinjection of NMDA receptor antagonists such as AP5 into the RVM inhibited hyperalgesic behavior produced by injection of formalin into the hindpaw, carrageenan-induced hindpaw inﬂammation, application of mustard oil to the hindlimb skin (topical), and colon inﬂammation. In contrast, block of non-NMDA receptors in the RVM further enhanced the hyperalgesic response, suggesting the presence of descending inhibition. Differential effects of NMDA and non-NMDA receptor antagonists on antinociceptive “off-cells” and pronociceptive “on-cells” in the RVM also suggest that distinct phar- macological proﬁles of inhibitory and facilitatory circuits may be a neural mechanism of bidirectional descending control (Heinricher et al. 2003). De- scending modulation may undergo time-dependent functional changes such that facilitation decreases or inhibition increases (or both) at the later stages of inﬂammatory pain. These changes and their role in primary versus secondary hyperalgesia await a detailed analysis.
The role of iGluRs in the PAG in prolonged or chronic pain states is less well understood. Block of NMDA receptors in the PAG inhibited pain in the formalin
test, but not the hotplate test, suggesting that NMDA receptors are involved in tonic, but not phasic, pain (Vaccarino et al. 1997). In another study, however, NMDA receptor antagonists in the PAG were ineffective in the formalin test but activation of NMDA receptors produced antinociception (Berrino et al. 2001).
The role of non-NMDA receptors in the PAG remains to be determined.
In higher brain areas, intrathalamic injection of an NMDA receptor antago- nist (AP5) reduced thermal and mechanical hyperalgesia in the acute and sub- acute phases of the carrageenan-induced hindpaw inﬂammation model. The effects were conﬁned to the withdrawal responses evoked from the injected paw. Pretreatment with intrathalamic injections of antisense oligodeoxynucleotides directed against the NR1 subunit prevented the development of thermal hyper- algesia and attenuated the development of mechanical hyperalgesia (Kolhekar et al. 1997). The pharmacology of thalamic pain mechanisms has yet to be analyzed at the single cell level in preclinical pain models.
In the amygdala both NMDA (AP5) and non-NMDA (NBQX) receptor antag- onists inhibited the nociceptive responses of CeLC neurons in the kaolin/carra- geenan-induced arthritis pain model (Li and Neugebauer 2004b). However, in- creased function of NMDA receptors rather than non-NMDA receptors appears to be a key mechanism of pain-related synaptic plasticity recorded in brain slices from arthritic animals compared to naïve controls (Bird et al. 2005).
In the anterior cingulate cortex (ACC), bilateral microinjections of an NMDA receptor antagonist (AP5) signiﬁcantly inhibited bilateral mechanical allody- nia in the CFA-induced inﬂammatory pain model (posttreatment on day 1; Wei et al. 2002). Conversely, overexpression of NR2B in the forebrain, including the ACC and insular cortex, increased formalin-induced pain behavior and me- chanical allodynia associated with CFA-induced hindpaw inﬂammation (Wei et al. 2001). However, pretreatment with bilateral injections of NMDA (AP5) or non-NMDA (DNQX) receptor antagonists into the ACC had no effect on
acute nociceptive behaviors in the formalin test whereas affect-related behav- ior (formalin-induced conditioned place avoidance) was effectively eliminated by intra-ACC microinjection of AP5 but not DNQX (Lei et al. 2004).
Metabotropic Glutamate Receptors
There is good evidence to suggest the involvement and endogenous activation of peripheral group I mGluRs in models of inﬂammatory and neuropathic pain but not in normal nociception (Carlton 2001; Neugebauer 2001, 2002; Neugebauer and Carlton 2002; Varney and Gereau 2002; Lesage 2004). Periph- eral injections of antagonists for mGluR1 (including CPCCOEt) or mGluR5 (MPEP) inhibited nociceptive behavior in the second, but not the ﬁrst, phase of the formalin test. Similarly, peripheral administration of MPEP reversed mechanical allodynia in carrageenan- and CFA-induced inﬂammatory pain models. The antagonists did not inhibit normal nociception and had no ef- fect when injected into the contralateral non-inﬂamed paw. Peripheral group I mGluRs may also play a role in neuropathic pain. In the spinal nerve ligation model, peripheral injections of an mGluR5 antagonist (SIB-1757, 6-methyl-2- (phenylazo)-3-pyridinol) into the injured hindlimb, but not the contralateral paw, reversed thermal hyperalgesia, but not mechanical allodynia, and had no effect in sham-operated animals.
Peripheral injection of a group II mGluR agonist (APDC) inhibited no- ciceptive behavior in the ﬁrst and the second phase of formalin-induced pain (Neugebauer and Carlton 2002). APDC also blocked prostaglandin E2 (PGE2)-induced thermal hyperalgesia and PGE2- and carrageenan-induced mechanical allodynia (Yang and Gereau 2002; Yang and Gereau 2003). How- ever, APDC had no effect on basal thermal and mechanical thresholds in naïve animals. The antinociceptive effects of APDC were antagonized by a selec- tive group II antagonist (LY341495) when tested. Importantly, blockade of peripheral group II mGluRs (LY341495) prolonged PGE2- and carrageenan- induced mechanical allodynia, suggesting that peripheral group II mGluRs mediate endogenous anti-allodynic effects (Yang and Gereau 2003). Simi-
larly, peripheral injection of a group II/III antagonist (MSOPPE, (RS)-α- methylserine-O-phosphate monophenyl ester) enhanced glutamate-induced
mechanical allodynia (Neugebauer and Carlton 2002). In the spinal nerve le- sion model of neuropathic pain, pre-treatment with APDC delayed the on- set of mechanical allodynia but post-treatment with APDC had no effect (Jang et al. 2004). The contribution of peripheral mGluRs to nociceptive
processes associated with inﬂammatory and neuropathic pain would sug- gest that peripherally acting group I antagonists and group II agonists may
have a therapeutic value in the treatment of pain states arising from pe- ripheral tissue injury. The role of group III mGluRs remains to be deter- mined.
The endogenous activation and involvement of spinal group I mGluRs, par- ticularly mGluR1, in prolonged nociception and persistent pain states is well documented in behavioral and electrophysiological studies using antagonists, antibodies, and antisense oligonucleotides in models of inﬂammatory pain (in- duced by formalin, intradermal capsaicin, intraplantar carrageenan or CFA, and intraarticular kaolin/carrageenan) and neuropathic pain (Neugebauer et al. 1999; Fundytus 2001; Karim et al. 2001; Neugebauer 2001; Neugebauer
2002; Varney and Gereau 2002; Lesage 2004; Soliman et al. 2005). Electro- physiological studies of spinal dorsal horn neurons, including spinothalamic
tract cells, showed antinociceptive effects of spinally administered group I
antagonists, including mGluR1-selective compounds (CPCCOEt), or antisense oligodeoxynucleotides directed against mGluR1, in pain-related central sen- sitization following intradermal capsaicin, topical application of mustard oil to the skin, or intraarticular kaolin/carrageenan injections (Neugebauer et al.
1999; Fundytus 2001; Neugebauer 2001; Varney and Gereau 2002). The involve- ment and intrinsic activation of mGluR5 in spinal neurons in pain models is
not entirely clear yet.
Intrathecal administration of group I antagonists, including CPCCOEt- (mGluR1) and MPEP (mGluR5), inhibited nociceptive behavior in the forma- lin test, thermal hyperalgesia and mechanical allodynia following intraplantar carrageenan, and capsaicin-induced mechanical, but not thermal, hypersensi- tivity (Fundytus 2001; Neugebauer 2001; Neugebauer 2002; Varney and Gereau
2002; Lesage 2004; Soliman et al. 2005). Group I antagonists had also antinoci- ceptive effects in models of neuropathic pain. Whereas mGluR1-selective an- tagonists inhibited thermal hyperalgesia and mechanical allodynia, the role
of mGluR5 in neuropathic pain is controversial, but a number of studies have shown antinociceptive effects of systemically or intrathecally applied blockers of mGluR5 (Varney and Gereau 2002; Lesage 2004). When directly compared,
systemic administration of an mGluR1 antagonist [LY456236, (4-methoxy- phenyl)-(6-methoxy-quinazolin-4-yl)-amine HCl] reversed mechanical allo- dynia in a neuropathic pain model whereas mGluR5 antagonists (including
MPEP) only attenuated it; however, both strongly inhibited formalin-induced nociceptive behavior (Varty et al. 2005). It is possible that mGluR5 receptors play a more prominent role in inﬂammatory than neuropathic pain states, but mGluR1 appears to be important for both.
The roles of group II and group III mGluRs in spinal pain mechanisms are less clear. Intrathecal administration of a group III agonist (LAP4) produced
antinociceptive effects in the second phase of the formalin test whereas in- trathecally administered group II agonists such as LY379268 were ineffective in this pain model (Neugebauer and Carlton 2002; Varney and Gereau 2002; Jones et al. 2005). Intrathecal administration of group II (APDC) and III (LAP4) ago- nists inhibited mechanical allodynia following intradermal capsaicin but had little effect on thermal hyperalgesia (Soliman et al. 2005). Electrophysiological studies showed that activation of spinal group II mGluRs inhibited electrically evoked nociceptive responses of dorsal horn neurons in carrageenan-induced hindpaw inﬂammation whereas mixed effects (inhibition and facilitation) were observed in control rats. A selective group II agonist (LY379268) inhibited capsaicin-induced central sensitization of spinothalamic tract cells but had no effect on the responses of non-sensitized neurons (Neugebauer et al. 2000). In contrast, a group III agonist (LAP4) inhibited the responses of spinothalamic tract cells to brief innocuous and noxious stimuli under normal condition as well as capsaicin-induced central sensitization (Neugebauer et al. 2000). These data suggest a dramatic change in the functional role of spinal group II, rather than group III, mGluRs in inﬂammatory pain.
In the brainstem, administration of group I (DHPG) and II (LCCG) agonists into the PAG decreased the nociceptive response in the second phase of the formalin test. These antinociceptive effects were antagonized by the respec- tive antagonists (CPCCOEt, group I mGluR1; EGLU, group II) (Maione et al.
2000). A group III agonist (LSOP) increased the formalin-evoked nociceptive response. This facilitatory effect was antagonized by a group III antagonist (MSOP). Interestingly, an antagonist for mGluR5 (MPEP) potentiated per se the early nociceptive phase of the formalin test, suggesting mGluR5 in the PAG is endogenously activated to produce descending inhibition (Berrino et al.
2001). EGLU and MSOP had no effects on their own. It has been suggested that group I and group II mGluRs in the PAG positively modulate descending pain inhibition whereas group III mGluRs inhibit this antinociceptive pathway (Maione et al. 2000; Berrino et al. 2001).
In higher brain areas, the function of mGluRs in prolonged or chronic pain states is still largely unknown. Recent electrophysiological and behavioral stud-
ies in the amygdala suggested an important role for group I mGluRs, partic- ularly mGluR1, in nociceptive plasticity and pain behavior (Neugebauer et al.
2004). In the kaolin/carrageenan arthritis pain model, administration of an- tagonists selective for mGluR1 (CPCCOEt) or mGluR5 (MPEP) into the amyg- dala inhibited the increased responses of sensitized CeLC neurons whereas the responses of these neurons under normal conditions before arthritis were
inhibited by MPEP but not by CPCCOEt (Li and Neugebauer 2004a). Accord- ingly, CPCCOEt had no effect on basal synaptic transmission in CeLC neurons
recorded in slices from normal animals but inhibited pain-related synaptic plasticity in slices from arthritic rats. MPEP inhibited basal synaptic transmis- sion as well as synaptic plasticity (Neugebauer et al. 2003). Thus, enhanced receptor activation of mGluR1 appears to be a key mechanism of pain-related synaptic plasticity in the CeLC. Importantly, behavioral studies that addressed the signiﬁcance of mGluR1 and mGluR5 function in the amygdala showed that CPCCOEt administration into the CeLC inhibited higher integrated behavior organized in the amygdala (vocalization afterdischarges) and in the brainstem (vocalizations during stimulation) as well as spinal nociceptive withdrawal reﬂexes. MPEP inhibited only vocalizations organized in the amygdala. It ap- pears that activation of mGluR1 in the amygdala contributes to pain produc- tion through descending facilitation whereas mGluR5 is involved in intrinsic amygdala processes (Han and Neugebauer 2005). The roles of group II and III mGluRs in nociception and plasticity in the amygdala remain to be determined.