mGluRs belong to family 3 of G protein-coupled receptors, which can trigger long-lasting intracellular processes and plastic changes. They are characterized by a seven-transmembrane domain topology and a large N-terminal extracel- lular domain, which contains important residues for ligand binding and forms two lobes that close like a Venus’ ﬂytrap upon ligand binding (Bockaert and Pin 1999). The second intracellular loop determines G protein speciﬁcity and the intracellular C-terminal interacts directly with intracellular proteins such as Homer, which are involved in receptor trafﬁcking, synaptic anchoring, cell signaling, and constitutive (basal) receptor activity (Bockaert and Pin 1999; De Blasi et al. 2001; Gasparini et al. 2002; Bhave et al. 2003). Eight mGluR subtypes have been cloned and are classiﬁed into groups I (mGluRs 1 and 5), II (mGluRs
2 and 3), and III (mGluRs 4, 6, 7, and 8) based on their sequence homology, signal transduction mechanisms, and pharmacological proﬁle (Schoepp et al.
1999; De Blasi et al. 2001; Neugebauer 2001; Gasparini et al. 2002; Varney and Gereau 2002; Lesage 2004; Swanson et al. 2005; see Table 1). Several splice vari- ants have been identiﬁed that may differ with regard to their pharmacology and G protein coupling.
Group I mGluRs couple through Gq proteins to the activation of phospholi- pase C (PLC), which leads to the formation of inositol-1,4,5-trisphosphate (IP3)
and diacylglycerol (DAG), resulting in calcium release from intracellular stores and activation of protein kinase C (PKC), respectively (Anwyl 1999; Schoepp et al. 1999; Neugebauer 2001; Gasparini et al. 2002). Tyrosine kinase activa- tion is another signaling pathway of group I mGluRs. Both PKC- and tyrosine kinase-dependent pathways can involve mitogen-activated protein (MAP) ki- nases such as the extracellular signal-regulated kinases 1/2 (ERK1/2) (Varney and Gereau 2002). Stimulation of adenylyl cyclase (AC) by group I mGluRs has been reported but the underlying mechanism is not yet known. Group II and group III mGluRs are negatively coupled to AC through Gi/Go proteins, thereby inhibiting cyclic AMP (cAMP) formation and cAMP-dependent pro- tein kinase (PKA) activation (Anwyl 1999; Schoepp et al. 1999; Neugebauer
2001; Gasparini et al. 2002).
Individual mGluR subtypes show distinct synaptic localization and function
(Schoepp et al. 1999; Neugebauer 2001; Gasparini et al. 2002; Varney and Gereau
2002; Lesage 2004; Swanson et al. 2005). Group I mGluR1 and 5 are most often localized postsynaptically whereas group II mGluR2 and group III mGluR8 are largely presynaptic. Group II mGluR3 and group III mGluR4 and 7 can be found both pre- and postsynaptically. In general, the predominant effect of group I mGluR activation is enhanced neuronal excitability and synaptic transmission whereas groups II and III typically mediate inhibitory effects, although excep- tions exist. The mGluRs can regulate neuronal excitability through direct or indirect effects on a variety of voltage-sensitive ion channels, including high
voltage-activated Ca2+ channels, K+ channels, and nonselective cationic chan- nels (Anwyl 1999; Schoepp et al. 1999; Neugebauer 2001). The modulation of
ligand-gated ion channels by mGluRs includes the group I mGluR-mediated enhancement of ionotropic glutamate receptor function, which likely involves
receptor phosphorylation (Anwyl 1999; Fundytus 2001; Neugebauer 2001). Group I mGluRs also potentiate the function of the capsaicin/vanilloid re- ceptor, TRPV1 (Hu et al. 2002). Convincing evidence suggests that mGluRs interact with the opioid system and play a role in the development of opioid
tolerance and dependence (Fundytus 2001). mGluRs can also modulate the re- lease of transmitters by acting as autoreceptors (glutamate) or heteroreceptors (GABA, substance P, serotonin, dopamine, and acetylcholine) (Cartmell and
Several potent and mGluR subgroup/subtype-selective compounds have been developed (Schoepp et al. 1999; Neugebauer 2001; Gasparini et al. 2002; Varney and Gereau 2002; Lesage 2004; Swanson et al. 2005; see Table 1). Currently available agonists are selective for subgroups I (DHPG, (S)-3,5- dihydroxyphenylglycine), II (LY354740, [1S, 2S,5R,6S]-2-aminobicyclo[3.1.0] hexane-2,6-dicarboxylic acid; LY379268, (−)-2-oxa-4-aminobicylco hexane-
4,6-dicarboxylic acid), and III (LAP4, l-(+)-2-amino-4-phosphonobutyric- acid). LY354740 has been tested in phase II clinical trials. Several subtype- selective agonists have become available. (RS)-2-chloro-5-hydroxyphenylgly- ine (CHPG) is selective for mGluR5. (S)-3,4-dicarboxyphenylglycine [(S)-3,4-
DCPG] activates mGluR8 with nanomolar potency and has greater than 280- fold selectivity over other mGluR subtypes (Thomas et al. 2001). Positive al- losteric activators have been identiﬁed for mGluR4 (N -phenyl-7-(hydoxylimi- no)cyclopropa[b]chromen-1a-carboxamide [PHCCC]; Maj et al. 2003; Shipe et al. 2005) and mGluR7 (N,N -dibenzhydrylethane-1,2-diamine dihydrochlo- ride [AMN082]; Conn and Niswender 2006; Mitsukawa et al. 2005). Antagonists for group I mGluRs include the competitive mGluR1 subtype-selective antag-
onist LY367385 [(S)-(+)-α-amino-4-carboxy-2-methylbenzeneacetic acid] and noncompetitive (allosteric) antagonists for mGluR1 [CPCCOEt, 7-(hydroxyimi- breakno) cyclopropa[b]chromen-1a-carboxylate ethyl ester] and mGluR5-
(MPEP, 2-methyl-6-(phenylethynyl)pyridine). Competitive group II and- group III mGluR antagonists are available; they are subgroup- but not subtype- selective (Table 1). The activity of mGluRs can be regulated not only by receptor
agonists and antagonists but also through receptor phosphorylation, including PKC-mediated desensitization of group I mGluRs and uncoupling of groups II and III mGluRs from G proteins by PKC and PKA (Karim et al. 2001; Neuge-
bauer 2001; Varney and Gereau 2002).
Glutamatergic Nociceptive Transmission in the Peripheral and Central Nervous System
The role of iGluRs and mGluRs in nociceptive transmission has been reviewed in detail in a number of recent articles (Fisher et al. 2000; Carlton 2001; Willis
2001; Neugebauer and Carlton 2002; Neugebauer 2002; Varney and Gereau 2002; Ruscheweyh and Sandkuhler 2002; Hewitt 2003; Garry et al. 2004; Lesage 2004). The following is a summary of glutamate receptor function in nociceptive processing in naïve animals; the involvement of iGluRs and mGluRs in pain- related plasticity will be discussed in the context of preclinical pain models (Sect. 3).
Ionotropic Glutamate Receptors
Functional NMDA, AMPA, and kainate receptors are present along the pain neuraxis from the peripheral nervous system to the brain.
It is ﬁrmly established now that iGluRs (particularly NR2B, GluR1, and GluR5) are localized in the periphery on nociceptive primary afferent terminals and on their cell bodies (Carlton 2001). Injections of NMDA, AMPA, or kainate into peripheral tissues excite and sensitize nociceptors in electrophysiologi-
cal studies and produce pronociceptive behavioral effects including thermal and mechanical hyperalgesia as well as spontaneous lifting and licking behav- iors (Lawand et al. 1997; Carlton 2001). These effects can be blocked by the appropriate antagonists. Although the details of iGluR-mediated signal trans- duction at the peripheral nerve terminals need to be determined, it is clear that neuronal as well as nonneuronal elements serve as the sources for glutamate (Carlton 2001). Importantly, glutamate levels are increased in the joints of pa- tients with arthritis, which may suggest a role of peripheral iGluRs in clinical pain associated with tissue damage (McNearney et al. 2000).
In the spinal dorsal horn, NMDA, AMPA, and kainate receptors have now been shown to be localized pre- and postsynaptically, particularly in the superﬁcial laminae I and II (Tolle et al. 1993; Willis and Coggeshall 2004; Lu et al. 2005). NMDA receptors (NR1) are present on the terminals of primary afferents and of GABAergic interneurons as well as on nearly all dorsal horn neurons. The distribution of the different NR2 subunits appears to be more complex, but NR2B is found predominantly in small-diameter primary afferents and in the superﬁcial dorsal horn whereas NR2A is highly concentrated in the deep dor- sal horn (Nagy et al. 2004). AMPA receptors are expressed on the terminals of unmyelinated (GluR1) and myelinated (GluR2) afferents and of GABAergic interneurons (GluR4) as well as on neurons in the superﬁcial (GluR1 and 2) and deep (GluR3 and 4) dorsal horn. Although the AMPA receptor subunit distribution is controversial, it has been suggested that unmyelinated C-ﬁbers are positioned to activate calcium-permeable AMPA receptors (lacking GluR2) on inhibitory (and some excitatory) interneurons in the superﬁcial dorsal horn whereas (small) myelinated ﬁbers make contact with GluR2-containing AMPA receptors on excitatory interneurons (Willis and Coggeshall 2004). Such an arrangement would limit the nociceptive transmission from C-ﬁbers under normal conditions. Kainate receptor subunits KA1 and KA2 are found presy- naptically on primary afferents and GABAergic interneurons as well as postsy- naptically on dorsal horn neurons (Willis and Coggeshall 2004; Lu et al. 2005). Importantly, essentially all of the kainate receptor-expressing primary affer- ents appear to be nociceptors, and GluR5 is the predominant kainate receptor subunit expressed in primary afferents (Carlton 2001).
Presynaptic NMDA, AMPA, and kainate receptors have been shown to act as autoreceptors on primary afferents to decrease the release of glutamate through
a mechanism that involves primary afferent depolarization, thus preventing the propagation of action potentials into the central terminals or, more likely, reducing their size (Lee et al. 2002; Bardoni et al. 2004). The consequence of
such inhibitory action for spinal nociceptive processing remains to be deter- mined, but inhibitory effects of kainate receptor activation on spinal excitatory
neurotransmission have been described (Ruscheweyh and Sandkuhler 2002; Youn and Randic 2004). The predominant effect, however, of spinal NMDA, AMPA, or kainate receptor activation is the excitation of spinal dorsal horn neurons and the increases of their responses to innocuous and noxious stimuli (Fundytus 2001; Ruscheweyh and Sandkuhler 2002; Youn and Randic 2004). Accordingly, intrathecal administration of NMDA, AMPA, and kainate recep- tor agonists produces spontaneous nociceptive behavior and thermal and me- chanical hyperalgesia (Fundytus 2001). Conversely, there is good evidence that spinal administration of NMDA, AMPA, and kainate receptor antagonists can inhibit nociceptive transmission in dorsal horn neurons and pain behavior (Stanfa and Dickenson 1999; Moore et al. 2000; Fundytus 2001; Willis 2001; Schaible et al. 2002; Ruscheweyh and Sandkuhler 2002; Garry et al. 2004). It should be noted, however, that AMPA receptor activation is also involved in the transmission of non-nociceptive information whereas blockade of NMDA and kainate receptors does not affect all forms of nociceptive transmission. NMDA receptors may be involved predominantly in tonic pain responses and persis- tent pain states (see Sect. 3). The role of kainate receptors is only beginning to emerge.
Activation of NMDA and AMPA receptors in the brainstem, including peri- aqueductal gray (PAG) and rostral ventromedial medulla (RVM), produces antinociceptive effects, which is consistent with the key role of these brain- stem areas in the descending inhibition of nociception and pain (Berrino et al. 2001; Ren and Dubner 2002; Gebhart 2004; Vanegas and Schaible 2004). More recently, however, it has been shown that glutamate or NMDA, but not AMPA, micro-injected into the RVM facilitates spinal nociceptive processing (Ren and Dubner 2002; Gebhart 2004). The dual facilitatory and inhibitory role of iGluRs in the brainstem reﬂects the bidirectional modulation of pain by the descending control system. These differential effects appear to be activity and time-dependent as is evident in pain models (see Sect. 3). Antagonists for non-NMDA, but not NMDA, receptors in the RVM produce facilitation of nociceptive responses in naïve animals, suggesting that a glutamatergic de- scending inhibitory system is tonically active (Urban et al. 1999). The role of kainate receptors in descending pain modulation is not yet known.
The contribution of iGluRs to nociceptive processing in higher brain areas has been determined in the thalamus (Salt 2002) and amygdala (Neugebauer
et al. 2004). Sensory thalamic relay nuclei such as the ventrobasal thalamus (VB) play an important role in gating and processing sensory information transmitted to the cerebral cortex. Sensory transmission to the VB complex
involves both NMDA and non-NMDA receptors. NMDA receptors contribute strongly to nociceptive responses of VB neurons and prolonged or repetitive
synaptic input. Non-NMDA receptors mediate non-nociceptive responses and the fast component of sensory-evoked synaptic responses but contribute little to nociceptive transmission. Kainate receptors do not appear to contribute to the responses of thalamic relay cells to sensory inputs, but studies using LY382884 (kainate GluR5 antagonist) suggest there is activation of kainate receptors on GABAergic axons from the thalamic reticular nucleus, which results in disinhibition through decreased recurrent inhibition. This has been proposed to play an important novel role in extracting sensory information from background noise (Salt 2002).
The latero-capsular part of the central nucleus of the amygdala (CeLC) represents the nociceptive amygdala, which is believed to be concerned with the emotional-affective component and modulation of pain (Neugebauer et al.
2004). Nociceptive responses of CeLC neurons in naïve animals are mediated by a combination of NMDA and non-NMDA receptors whereas non-nociceptive
inputs and basal synaptic transmission activate only non-NMDA receptors (Li and Neugebauer 2004b; Bird et al. 2005). The relative contribution of AMPA
and kainate receptors is not yet known. Enhanced NMDA receptor function appears to be an important mechanism in pain-related plasticity in the CeLC (see Sect. 3).
Metabotropic Glutamate Receptors
The important role of group I mGluRs in peripheral and central nociceptive processing and pain behavior is now well established whereas the functions of groups II and III mGluRs are less well known (Fundytus 2001; Neugebauer
2001; Neugebauer 2002; Varney and Gereau 2002; Lesage 2004).
Group I mGluR1 and mGluR5 have been demonstrated on a subset (∼20%–
30%) of unmyelinated and myelinated peripheral axons (Neugebauer and Carl- ton 2002; Lesage 2004). There is evidence for a colocalization of mGluR5 with the capsaicin/vanilloid receptor (TRPV1), while mGluR1 may also be present on sympathetic efferent ﬁbers. Group II mGluR2 and mGluR3 and group III mGluR7 and mGluR8 are also present in a subpopulation of small-diameter primary afferent neurons, and there is some colocalization of group I and group II mGluRs (Neugebauer and Carlton 2002; Varney and Gereau 2002). Exogenous activation of peripheral group I mGluRs (Neugebauer and Carlton
2002; Varney and Gereau 2002; Lesage 2004) produced long-lasting thermal hy- peralgesia in mice (but not rats), which was signiﬁcantly reduced by peripheral
injections of mGluR1 antagonists (CPCCOEt and LY367385) and completely blocked by an mGluR5 antagonist (MPEP). Peripheral injections of a selec-
tive mGluR5 agonist (CHPG) and the mGluR1/5 agonist (DHPG) were equally potent in producing mechanical allodynia in rats. The mechanical allodynia was blocked by a selective mGluR5 antagonist (MPEP) but not by a selec- tive mGluR1 antagonist (4CPG, (S)-4-carboxy-phenylglycine). These data may suggest a role of peripheral mGluR5 in mechanical allodynia whereas both mGluR1 and mGluR5 are involved in thermal hyperalgesia.
There is good evidence to suggest the involvement and endogenous acti- vation of peripheral group I mGluRs in persistent pain states (see Sect. 3) but not in normal nociception. Subcutaneous injection of a group II mGluR agonist (APDC; 2R,4R-4-aminopyrrolidine-2,4-dicarboxylate) into the mouse hindpaw did not alter mechanical and thermal sensitivity in behavioral studies (Yang and Gereau 2002; Yang and Gereau 2003), but very low concentrations of APDC inhibited the extracellularly recorded responses of nociceptive ﬁbers to heat stimuli in an in vitro skin-nerve preparation (Neugebauer and Carlton
2002). Peripheral group II mGluRs may be particularly useful targets in pain states (see Sect. 3). The role of peripheral group III mGluRs remains to be
Group I mGluR1 and mGluR5 are functionally expressed in the spinal dor- sal horn: mGluR5 expression is particularly strong in the superﬁcial laminae whereas high levels of mGluR1 are found in the deep dorsal horn. Anatom- ical and electrophysiological data further suggest that mGluR1 and mGluR5 are mainly postsynaptic but can also act presynaptically to facilitate spinal neurotransmission (Neugebauer 2001, 2002; Varney and Gereau 2002; Lesage
2004). Importantly, mGluR5 was identiﬁed on GABA and non-GABAergic dor- sal horn neurons. Activation of spinal group I mGluRs generally produces
pronociceptive effects in behavioral and electrophysiological assays, although mixed excitatory and inhibitory effects have been reported (Fundytus 2001; Neugebauer 2001; Varney and Gereau 2002; Lesage 2004). Intrathecal admin-
istration of group I agonists such as DHPG evokes spontaneous nociceptive behavior, thermal and mechanical hyperalgesia, and mechanical allodynia, which can be blocked with antagonists or antibodies for mGluR1 or mGluR5,
where tested (Fundytus 2001; Neugebauer 2001; Varney and Gereau 2002; Lesage 2004). Activation of spinal group I mGluRs can produce facilitation or inhibition of synaptic transmission and of dorsal horn neuron responses to
innocuous and noxious stimuli (Gerber et al. 2000a; Neugebauer 2001; Lesage
2004). Antagonists selective for mGluR1 (CPCCOEt) or mGluR5 (MPEP) can block these facilitatory effects but do not appear to affect nociceptive behav- ior under normal conditions (Fundytus 2001; Neugebauer 2001; Varney and Gereau 2002; Lesage 2004; Soliman et al. 2005). Spinal mGluR5 may mediate the inhibitory effects of groupI mGluR activation.
The roles of group II and group III mGluRs in spinal nociceptive processing are less clear (Neugebauer 2001; Neugebauer 2002; Varney and Gereau 2002). Both group II mGluR2/3 and group III mGluR4 and 7, but not mGluR6 and 8, are present in the spinal cord. Group III mGluRs are localized predominantly on presynaptic terminals in the dorsal and ventral horns whereas group II mGluRs have been detected on presynaptic terminals in the superﬁcial dor- sal horn as well as on postsynaptic elements in deeper laminae. Agonists of group II (LCCG, (2S,1 S,2 S)-2-(carboxycyclopropyl)glycine) and III (LAP4) mGluRs can inhibit synaptic transmission in the dorsal horn in spinal cord slices in vitro (Gerber et al. 2000b), but behavioral studies found little evi- dence for antinociceptive effects of group II or group III mGluR activation (Neugebauer 2001; Neugebauer 2002; Varney and Gereau 2002). Intrathecal administration of selective group II (APDC) and group III (LAP4) agonists had no effect on mechanical withdrawal response thresholds in the absence of tissue damage (Soliman et al. 2005). Similarly, systemic administration of a se- lective group II agonist (LY379268) had no signiﬁcant effects on the tail ﬂick or paw withdrawal tests of acute thermal nociceptive function (Simmons et al.
2002). Intraspinal administration of LY379268 had no effect on the responses of nociceptive dorsal horn neurons (spinothalamic tract cells) recorded under normal conditions whereas LAP4 inhibited the responses to brief noxious and
innocuous mechanical cutaneous stimuli (Neugebauer et al. 2000).
In the brainstem, activation of group I mGluR1/5 (DHPG) or mGluR5 (CHPG) in the PAG had antinociceptive effects in the hotplate test in naïve animals (Maione et al. 1998; Lesage 2004). Intra-PAG administration of an mGluR1 antagonist (CPCCOEt) attenuated the effect of DHPG whereas an mGluR5 antagonist (MPEP) blocked the effect of CHPG, indicating the involvement of both mGluR1 and mGluR5 subtypes. When administered without exoge- nous agonists, MPEP, but not CPCCOEt, had pronociceptive effects, suggesting the presence of tonic descending inhibition mediated by mGluR5 in the PAG. Activation of mGluR5 (CHPG) in the RVM, however, produced cold hypersen- sitivity that was prevented by MPEP (Lesage 2004). The possibly differential roles of mGluR1 and mGluR5 in descending inhibitory and facilitatory sys- tems await a detailed analysis. Different from group I mGluRs, activation of groups II and III mGluRs in the PAG appears to decrease the descending in- hibition of pain behavior (Maione et al. 1998). Agonists of group II (LCCG) and III (LSOP, l-serine-O-phosphate) receptors administered into the PAG decreased the latency of the nociceptive reaction in the hot plate test. Antag-
onists of group II [EGLU, (2S)-α-ethylglutamic acid] and III [MSOP, (RS)-α- methylserine-O-phosphate] receptors antagonized these pronociceptive effects
of LCCG and LSOP. Interestingly, MSOP, but not EGLU, alone increased the la-
tency in the hot plate test, perhaps suggesting the tonic activation of group III
mGluRs in the PAG.
In the VB, mGluR1 and mGluR5 are involved in the processing of nocicep- tive, but not non-nociceptive, information (Salt and Binns 2000; Lesage 2004). Intrathalamic administration of antagonists selective for mGluR1 (LY367385) and mGluR5 (MPEP) inhibited the responses of VB neurons to brief noxious heat but not innocuous stimuli. Intrathalamic MPEP had no effect on noxious mechanical stimuli; the role of thalamic mGluR1 in mechanonociception re- mains to be determined. Group II and III mGluRs mediate the presynaptic reduction of GABAergic inhibition in the VB, producing disinhibition of sen- sory (including nociceptive) processing (Salt et al. 1996; Salt 2002). Agonists of group II (LY354740 and APDC) and III (LAP4) decreased the inhibitory trans- mission from the thalamic reticular nucleus and the respective antagonists
[MCCG (α-methyl-CCG-I) and MAP4 (α-methyl-l-AP4)] blocked these disin- hibitory effects. Accordingly, administration of a group II antagonist (EGLU) in the thalamic reticular nucleus produced antinociceptive effects presumably
by blocking the activation of group II mGluRs by the noxious input, thus dis- inhibiting the GABAergic neurons that inhibit thalamic relay neurons (Neto and Castro-Lopes 2000). When evaluating the therapeutic potential of these drug targets, it is important to consider possible facilitatory effects of other-
wise inhibitory groups II and III mGluRs in brain areas such as the PAG and thalamus.
In the amygdala (Neugebauer et al. 2004), activation of group I mGluR1/5
(DHPG) or mGluR5 (CHPG) enhanced the responses of CeLC neurons to in- nocuous and noxious stimuli in naïve animals (Li and Neugebauer 2004a). Likewise, DHPG and CHPG potentiated normal synaptic transmission of pre- sumed nociceptive input from the pontine parabrachial area to the CeLC in brain slices (Neugebauer et al. 2003). An mGluR5 antagonist (MPEP) inhibited brief nociceptive responses of CeLC neurons under normal conditions whereas an mGluR1 antagonist (CPCCOEt) had no effect. Similarly, MPEP, but not CPC- COEt, inhibited basal synaptic transmission in CeLC neurons in slices from naïve animals. The role of group II and III mGluRs in nociceptive processing in the amygdala is not yet known.