Group-I Preferring Antagonists

The first generation of selective antagonists described for metabotropic glutamate receptors were phenyl glycine derivatives, e.g., 4CPG, MCPG, and 4C3HPG (Fig. 2). Those compounds were widely examined in model systems of neurodegeneration/neuroprotection (see below). A very frequently used compound is MCPG which is an equally potent antagonist of mGluR1 and mGluR2 (IC₅₀ = 20–500 μM).^9,30 MCPG is less potent at mGluR5 (IC₅₀ = 200–1000mM), and showed little antagonist activity at mGluR3, −4, −6, −7, and −8 (IC₅₀ > 1000mM).⁹ 4CPG is a more potent antagonist for mGluR1 (IC₅₀ = 20–80mM) with little effects on mGluR2, −4, −5 (IC₅₀ >500mM), and no test results available for mGluR3, −6, −7, and −8;^9,30 4C3HPG shows a very similar profile but in addition to the effects of 4CPG it is an agonist at mGluR2 (EC₅₀ = 20mM).³⁰

Figure 2

Chemical structures and abbreviated names of most commonly used group-I mGluR antagonists. For all compounds shown, neuroprotective effects were reported.

The second generation of group-I mGluR antagonists showed higher potency and/or selectivity. LY367366 antagonizes mGluR1 and −5 (IC₅₀ = 3–6 μM), interacts with group-II and -III to a lower extent (IC₅₀ > 10mM) and shows no activity at iGluRs; the compound LY367385 shows a very similar profile except, it doesn't antagonize mGluR5 (Fig. 2).^9,31 AIDA is a mGluR1-selective antagonist (IC₅₀ = 4–200mM) with little activity at mGluR2, −4, and −5 (IC₅₀ > 1000mM); tests at mGluR3, −6, −7 and −8 were not performed.^9,30

More recently, amino acid-unrelated compounds with a high degree of selectivity for mGluR1 or mGluR5 were described. MPEP is a very potent and selective antagonist at mGluR5 with an IC₅₀ of 32 nM, but it has no activity at all other mGluRs up to 10mM; at this concentration it's also inactive at a broad selection of iGluRs (Fig. 2).³² CPCCOEt is a very selective antagonist for mGluR1 (IC₅₀ = 6.5mM) which shows up to 100 mM no activity at mGluR2, −4, −5, −7, and −8 (mGluR3 and −6 were not tested).^9,33 Similarly, BAY36–7620 was very recently described as a highly selective mGluR1 antagonist showing higher potency than CPCCOEt (IC₅₀ = 0.16mM; Fig. 2).³⁴

Agonists for Metabotropic Glutamate Receptors

The first mGluR agonist used on neuronal tissue were quisqualate, ibotenate, and trans-ACPD or its active isomer 1S,3R-ACPD.⁹ Although important insights into neuroprotective/neurodegenerative activity of mGluR agonists were gained (e.g., activation of group-I amplifies neuronal death),³⁵ more direct evidence for an involvement of specific groups/subtypes of mGluRs arose via the use of more selective agonists. The glutamate analogues L-CCG-I and DCG-IV (Fig. 3) were used to demonstrate group-II mGluRs as potential targets for neuroprotective drugs.³⁶ Both compounds show very potent agonist activity at mGluR2 and mGluR3 (EC₅₀-values = 0.1–0.9 mM), however at higher concentrations L-CCG-I is also an agonist at mGluR1, −4, −5, −6, −7, and −8 (EC₅₀-values = 2–9 μM), while DCG-IV antagonizes all group-III mGluRs (IC₅₀ = 22–40 mM); DCG-IV is also a potent agonist of NMDA receptors.^9,37 Considerably improved compounds were published more recently. 2R,4R-APDC shows agonist EC₅₀-values at mGluR2 and mGluR3 of 0.3–0.4 μM. The compound has no appreciable iGluR activity and doesn't activate or antagonize any group-I or -III mGluR up to 100 μM. Thus, 2R,4R-APDC is extremely specific for group-II receptors.⁹ The compounds LY354740 and LY379268 show much higher potency and equally good specificity for mGluR2 and mGluR3 (EC₅₀-values = 3–20 nM), when compared to 2R,4R-APDC (Fig. 3).^9,20 NAAG is an interesting molecule, because it activates only mGluR3 (EC₅₀ = 65 μM) but not mGluR2 (EC₅₀ > 300 mM), and it doesn't show interaction with group-I and group-III receptors.³⁸

Figure 3

Chemical structures and abbreviated names of most commonly used group-II and -III mGluR agonists. For all compounds shown, neuroprotective effects were reported.

An involvement of group-III mGluRs in neuroprotection has primarily been demonstrated by the use of the phosphono amino acids L-AP4 and (R,S)-PPG (Fig. 3). Both compounds are selective activators of group-III mGluRs with no appreciable activity at other mGluRs or iGluRs. The EC₅₀-values of L-AP4 at mGluR4 and −6 were reported to be 0.2–1 mM, 150–500 μM at mGluR7 and 0.06–0.9 μM at mGluR8. (R,S)-PPG is slightly less potent with EC₅₀-values at mGluR4, −6, −7, and −8 receptors of 5.2 mM, 4.7 mM, 185 mM and 0.2 μM, respectively.^5,9,18,21 The pharmacology of L-SOP at group-III mGluRs is very similar to (R,S)-PPG, but characterization of L-SOP at other receptors is not available to date.⁹

General Considerations

Glutamate and aspartate are the two acidic representatives among the natural amino acids. The presence of a distal carboxylic acid enhances their polar properties in comparison to the other amino acids. These polar properties limit considerably their capacity to cross membranes by passive diffusion. Therefore, glutamate is actively transported through biological membranes by specific transporter proteins. Up to date, almost all mGluR-selective ligands are amino acid derivatives. The only non-amino acid compounds discussed in this report (see Figs. 2 and 3) are MPEP, CPCCOEt and BAY36–7620 which are all group-I antagonists.^32,34,39,40 All other structures shown in Figure 2 and 3 share the same functional groups as glutamate, namely the amino acid functionality and the distal carboxylic acid (or phosphate moiety in the case of group-III ligands). Despite the structural analogy to glutamic acid the majority of these molecules fails to be actively transported through the cell membranes by the specific glutamate transporter systems and as a consequence have a poor oral bioavailability and do not cross the blood-brain-barrier (BBB). The two reported exceptions, LY354740 and its derivative LY379268, are orally bioavailable, enter the brain and display potent agonist activity at mGluR2, and −3.^20,41 In addition, these two ligands have a good water solubility, which is generally the case for glutamate analogs and of advantage for the formulation and the use in in vitro neuroprotection assays. The recently published non-amino acid compounds MPEP and BAY 36–7620 also demonstrate oral bioavailability and good BBB penetration^32,39,42 in contrast to the earlier published mGluR1 antagonist CPCCOEt, which neither shows bioavailability nor BBB penetration.

In Vivo Dosage and Side Effects

Most glutamate analogs such as MCPG, 4C3HPG and AIDA can only by studied in animal disease models after central administration (i.c.v.). The relatively low potency and lack of selectivity of such compounds usually doesn't allow conclusions on the side effect profile of antagonizing specific receptors. In contrast, MPEP, the mGluR5-selective antagonist at in vitro and in vivo conditions can easily be studied in animal models because it reaches the brain via oral and intravenous application (p.o. and i.v.).^32,42 Furthermore, MPEP shows a large window between the doses achieving the beneficial effects in animal models for inflammatory pain (10–100 mg/kg, p.o.) or anxiety (1–10 mg/kg, p.o.) and the dose inducing the first side effects (> 100 mg/kg, p.o.) such as inhibition of spontaneous locomotor activity.^32,42,43

In the case of group-II mGluRs, two orally active compounds have been investigated: LY354740 and LY379268. Both compounds are nanomolar potent agonists with similar potency at mGluR2 and mGluR3 and the investigations in vivo show that both compounds penetrate the BBB. In animal models, LY354740 showed anxiolytic like effects after oral administration (0.5–10 mg/kg, p.o.) without the sedative or motor impairment side effects of the classical benzodiazepine anxiolytics. Other group-II agonists, discussed here, are not bioavailable and do not cross the BBB, therefore have to be centrally administered.

None of the currently known group-III-selective agonists, e.g., L-AP4, L-SOP or (R,S)-PPG, show oral bioavailability or BBB penetration. Therefore, in vivo experiments necessitate a central administration of the drugs. The presence of a very polar acidic moiety such as the phosphate/phosphonate group is certainly a key element for the poor bioavailability of these molecules. In vivo (i.c.v. administration) and in vitro investigations using (R,S)-PPG in mGluR4- and mGluR7-deficient mice showed that mobilization of mGluR4 occurred to achieve neuroprotective effects (10 nmol, i.c.v.)^21,44 whereas activation of mGluR7 produces anticonvulsive effects (Herman van der Putten et al, personal communication). Interestingly, the sedative and respiratory side-effects induced by central administration of higher doses of PPG (200–2000 nmol, i.c.v. in mice)^21,45 are also seen in mGluR4- or mGluR7-deficient animals, clearly indicating that these side-effects are mediated by an interaction of (R,S)-PPG with target(s) distinct from mGluR4 and mGluR7 (Herman van der Putten, personal communication).

General and Introductory Considerations on Group-I Receptors

The role of mGluR1 and mGluR5 in neurodegeneration/neuroprotection has been debated for several years because of the contrasting effects of agonists (such as DHPG) in different in vitro models of neuronal toxicity.⁴⁶ However, the evidence that subtype-selective mGluR1 or −5 antagonists are consistently neuroprotective, suggests that endogenous activation of group-I receptors is permissive to cell death. mGluR1 and −5 are both coupled to polyphosphoinositide hydrolysis, but they differ for the kinetics of IP3-induced Ca²+ mobilization. In transfected cells, activation of mGluR1 induces a single-peaked Ca²⁺ response, whereas activation of mGluR5 promotes oscillatory increases in cytosolic free Ca²⁺ due to the PKC-dependent phosphorylation of a specific threonine residue that is just distal to the 7th trans-membrane domain of mGluR5.⁴⁷ Whether or not a similar difference exists under native conditions is unclear at present. Activation of both receptors regulates the activity of a variety of Ca²⁺ and K channels and turns on a not-yet-defined voltage-gated inward current,^48,49 with the net effect of enhancing neuronal excitability. Particular attention deserves the description of a positive modulation of NMDA-gated ion currents by mGluR5, which provides the most likely substrate for the permissive action of mGluR5 activators on neuronal toxicity.^50–55 Interestingly, NMDA receptor activation can also potentiate mGluR5 responses by inducing a calcineurin-dependent dephosphorylation of a PKC phosphorylation site that participates in mGlu5 desensitization.^56,57 The Shank family of postsynaptic density proteins may cross-link Homer and PSD-95, thus allowing the functional coupling between mGluR5 and NMDA receptors in the same region of the dendritic spine.⁵⁸ Studies performed in transfected oocytes suggest that only NMDA receptors containing NR2A or −2B subunits are positively modulated by mGluR5, whereas NMDA receptors containing the NR2C (or −2D) subunit are not. This may help explain why in mature cerebellar granule cells (that predominantly express the NR2C subunit) activation of group-I mGluRs enhances NMDA responses only when the NR2C subunit is knocked down by antisense oligonucleotides.⁵⁹ A question that awaits a final answer is whether or not group-I mGluR subtypes are presynaptically localized and modulate neurotransmitter release. This issue is fundamental because the extent of excitotoxic degeneration is strictly related to glutamate and GABA release in a number of disorders, such as ischemia-induced neurodegeneration and temporal lobe epilepsy. A series of studies carried out in the laboratory of J. Sanchez-Prieto indicate that a first application of group-I mGluR agonists facilitates glutamate release in cortical or hippocampal synaptosomes,^60–65 whereas a second drug application inhibits release. This functional change in the modulation of glutamate release depends on the PKC-dependent phosphorylation of the receptor (presumably mGluR5), leading to a switch in the G-protein coupling. A similar scenario is observed in cultured cortical cells, where a first exposure to DHPG amplifies NMDA toxicity, whereas a second exposure is neuroprotective.⁶⁶ As a secondary release in endogenous glutamate contributes to the progression of NMDA toxicity, the functional state of ‘presynaptic’ mGluR5 might be determinant for the final effect of group-I mGluR agonists on excitotoxic neuronal death. Electron microscopy studies have shown that group-I mGluRs are preferentially localized in the peripheral portion of postsynaptic densities rather than in presynaptic terminals.^22,23 However, the evidence that mGluR5 is found in the axon when co-transfected with Homer-1a raises the intriguing possibility that homer proteins regulate the dendritic or axon targeting of mGluR5.⁶⁷ Homer-1a is encoded by an early gene, which is switched on by neuronal hyperactivity. It will be interesting to examine whether changes in the subcellular distribution of mGluR5 are associated with synaptic hyperactivity that occurs during epilepsy or brain ischemia.

Effect of Group-I mGluR Agonists on Excitotoxic Neuronal Death

mGluR1/5 agonists may either amplify excitotoxic neuronal death or produce neuroprotection, depending on the neuronal type and the paradigm of toxicity. Possible explanations for contrasting data include (i) the absence or presence of the NR2C subunit in NMDA receptors; (ii) the existence of an activity-dependent “switch” between facilitatory and inhibitory mGluR subtypes (see above); and (iii) a role for glial cells expressing mGluR5. These aspects have been discussed in detail in a recent review.⁴⁶ It is noteworthy that excitotoxic neuronal death incorporates features of necrosis and apoptosis, and that apoptosis by trophic deprivation contributes to the overall neuronal death in brain ischemia.⁶⁸ One expects that, in experimental model of brain ischemia, pharmacological activation of group-I mGluRs amplifies necrotic death by further increasing cytosolic free Ca²⁺, but supports the survival of neurons undergoing apoptosis by trophic deprivation through the same mechanism.⁶⁹ Therefore one can easily conclude that pharmacological activation of group-I mGluR subtypes is not valuable in the experimental treatment of acute or chronic neurodegenerative disorders.

Neuroprotective Activity of mGluR1 Antagonists

A battery of competitive or non-competitive mGluR1 antagonists has been tested in in vitro and in vivo models of excitotoxic neuronal death. AIDA, LY367385, 4CPG, 4C3HPG, and CPCCOEt protect mixed murine cortical cultures against NMDA toxicity,^31,70,71 and are also effective in in vivo models of excitotoxic death.^31,71,72 LY367385, 4C3HPG and AIDA are also neuroprotective in the gerbil model of global ischemia and in murine cortical cultures and rat organotypic hippocampal cultures subjected to oxygen-glucose deprivation.^31,73,74 mGluR1 is preferentially localized in GABAergic neurons and its activation depresses inhibitory synaptic transmission.^22,75–78 Hence, it has been hypothesized that mGluR1 antagonists are neuroprotective by enhancing GABAergic transmission. Initial evidence in this line has been provided by the observation that local infusion of AIDA increases the basal output of GABA in the gerbil hippocampus.⁷⁴ In murine cortical cultures, neuroprotection by LY367385 or CPCCOEt (but not by MPEP) is occluded by GABA or by SKF89976A (an inhibitor of GABA transporter) and is abolished by a cocktail of GABA-A and GABA-B receptor antagonists.⁷¹ The same cocktail of antagonists prevents the protective activity of CPCCOEt against neurodegeneration induced by NMDA in in vivo studies.⁷¹ A role for GABAergic transmission is strengthened by the evidence that LY367385 and CPCCOEt enable NMDA to enhance GABA release in the striatum, and that activation of mGluR1 inhibits GABAergic IPSCs in cortico-striatal slices.⁷¹ Taken together, these data suggest that mGluR1 antagonists are neuroprotective by removing a tonic inhibitory control exerted by mGluR1 on GABA release. The finding that mGluR1 is localized also on GABAergic nerve terminals in the striatum is consistent with this hypothesis.⁷⁸ mGluR1 antagonists might be beneficial in CNS disorders characterized by an impairment of inhibitory synaptic transmission, such as Ammon's horn sclerosis.⁷⁹ Interestingly, the very recently introduced mGluR1 antagonist BAY36–7620 was shown to be protective in epilepsy, stroke and trauma models upon systemic administration.^34,39

Neuroprotective Effects of mGluR5 Antagonists

The development of subtype-selective mGluR5-specific antagonists has allowed to establish that endogenous activation of mGluR5 facilitates cell death in a variety of models of neurodegeneration. MPEP and its ancestors SIB-1757 and SIB-1893 (which all behave as non-competitive mGluR5 antagonists) are potent and effective in attenuating NMDA toxicity in culture, and at least MPEP can also reduce excitotoxic neuronal death in the rat striatum.⁸⁰ As opposed to mGluR1 antagonists, these effects do no involve changes in GABAergic transmission⁷¹ and might be explained with the ability of mGluR5 to positively modulate NMDA-gated ion currents (see above). In cortical cultures, neuroprotection by MPEP, SIB-1757 or SIB-1893 is observed at concentrations <10 μM, and, therefore, cannot be ascribed to any non-specific interaction with NMDA receptors.^32,42,80,81 In addition, the neuroprotective action of MPEP is not mimicked by its isomer iso-MPEP, which fails to antagonize mGluR5 although it shares most of the structural and physico-chemical features of MPEP.⁸⁰ No data are yet published on the effect of MPEP in experimental models of hypoxic/ischemic neuronal death. The expectation is unclear, because activation of mGluR5 is known to support the survival of developing neurons,⁸⁰ and apoptosis by trophic deprivation contributes to the overall neuronal death in brain ischemia (see above).

A series of studies outlines a potential use of mGluR5 antagonists in the experimental therapy of chronic neurodegenerative disorders. In cultured cortical cells, MPEP and its analogs are neuroprotective against β-amyloid toxicity at concentrations lower than those required for neuroprotection against excitotoxic death. In these studies, β-amyloid peptide is applied to the cultures in the presence of MK-801 and DNQX, thus excluding any involvement of ionotropic glutamate receptors in the neuroprotective activity of mGlu5 receptor antagonists.⁸⁰ Interestingly, inhibition of mGluR1 by AIDA does not reduce but rather exacerbates ß-amyloid toxicity in cortical cells.⁸³ Thus, at least in cultured neurons, endogenous activation of mGluR5 is required for the engagement of a death pathway in neurons exposed to ß-amyloid peptide. This encourages the study of mGluR5 antagonists in animal models of Alzheimer's disease. Finally, mGluR5 antagonists may be promising in the treatment of Parkinson's disease (PD). An optimal anti-Parkinsonian drug should combine neuroprotective properties with the ability to relief the cardinal symptoms of the disease (i.e., bradykinesia, rigidity and tremor). Recent unpublished data show that systemically administered MPEP or SIB-1893 (both at 10 mg/kg, i.p.) protect nigro-striatal dopaminergic terminals against metamphetamine toxicity and reduce the amount of reactive oxygen species produced by metamphetamine in the striatum of freely moving animals (G. Battaglia et al, manuscript in preparation). On the other hand, activation of mGluR5 induces direct excitation, and selectively potentiates NMDA currents in neurons of the subthalamic nucleus.⁵³ As an increased activity of the subthalamic nucleus is implicated in the pathophysiology of bradykinesia, one can predict that mGluR5 antagonists can improve Parkinsonian symptoms in experimental animals and humans. It will be interesting to test mGluR5 antagonists in classical experimental models of Parkinsonism, such as the MPTP model in mice or monkeys.

mGluR2/3 Agonists Are Protective in Several Experimental Paradigms of Neurodegeneration

In the last decade several reports have shown that pharmacological activation of group-II mGluRs is neuroprotective in a variety of models of neuronal degeneration, including neuronal cultures, brain slices and in vivo models of excitotoxicity. The non-selective group-II mGluRs agonists, DCG-IV, 4C3HPG, L-CCG-I and 1S,3R-ACPD, protect neurons against NMDA toxicity in mixed cultures of mouse cerebral cortex.^{36,66,84–89} This effect is substantially reduced by the group-II mGluR antagonists, PCCG-IV or MCCG-I, suggesting that activation of mGluR2/3 is responsible for neuroprotection and is supported by the finding that the highly selective mGluR2/3 agonist, 2R,4R-APDC, is neuroprotective in the same model.^90,91 DCG-IV or L-CCG-I are also neuroprotective against kainate toxicity in cultured cortical cells,³⁶ although Gottron et al⁹² found that, in the same cultures, protection by DCG-IV is restricted to the small percentage of neurons that respond to kainate with an enhanced influx of Co² . Neuroprotection by group-II mGluR agonists has also been shown in primary cultures of cerebellar granule cells,⁹³ as well as in primary cultures of mesencephalic neurons,⁹⁴ challenged with excitotoxins. Recently, the potent, highly selective and systemically active group-II mGluR agonists, LY354740, LY379268 and LY389795, have been shown to prevent excitotoxicity in both rat and mouse cortical neuronal cultures (see Kingston et al and D'Onofrio et al for support; but see also Behrens et al).^89,95–97 Interestingly, the effect of these compounds is enhanced by the presence of glial cells, suggesting an involvement of glial mGluR3 in the mechanism of neuroprotection.⁹⁶ Accordingly, NAAG, an endogenous selective mGluR3 agonist, protects cultured cortical neurons against NMDA toxicity.^38,86

Selective and non-selective group-II mGluR agonists exert neuroprotective activity also in further in vitro models of neuronal injury, i.e., oxygen-glucose deprivation- and staurosporine-induced neuronal death, in which neurons follow the apoptotic pathway of degeneration.^87,95,98,99

Group-II mGluR agonists have also been shown to be neuroprotective in in vivo models of neurodegeneration. DCG-IV, infused i.c.v., protects vulnerable neurons against local or systemic injection of kainate.¹⁰⁰ In addition, LY379268, locally or systemically injected, protects against NMDA-induced striatal GABAergic neuronal loss in rats.⁸⁹ Intracerebroventricular injection of 4C3HPG protects hippocampal CA1 neurons in a gerbil model of global ischemia (BCAO), both when administered prior to and after the induction of ischemia.¹⁰¹ The same degree of neuroprotection in hippocampal CA1 neurons, in the BCAO model, has been observed after systemic injection of LY354740 or LY379268.^102,103 In contrast, the infarct size following middle cerebral artery occlusion (MCAO), induced by endothelin-1 in rat, is unaffected by the drugs, suggesting that group-II mGluR agonists are likely to have more utility in global than in focal cerebral ischemia.^103,104 LY379268 exerts also neuroprotective activity in a neonatal rat model of hypoxia-ischemia.¹⁰⁵

Recently, group-II mGluR agonists have been shown to possess anti-seizure activity. LY354740 is anticonvulsant in mice when systemically administered in both the ACPD-induced limbic seizure model and the pentetrazole- and picrotoxin-induced seizures.^20,106,107 Moreover, group-II mGluR agonists attenuate both in vitro and in vivo models of traumatic neuronal injury, showing an additive effect with NMDA- and group-I mGluR antagonists.¹⁰⁸ In addition, L-CCG-I and 4C3HPG inhibit excitotoxic phenomena mediated by kainate on spinal cord motor neurons, a model resembling amyotropic lateral sclerosis.¹⁰⁹

Neuroprotection via mGluR2/3 Involves Glial-Neuronal Interactions

The mechanisms underlying group-II mGluR-mediated neuroprotection has been examined in cultured cortical cells exposed to a brief pulse with NMDA, a model of excitotoxicity in which the damage is induced by the influx of extracellular Ca²⁺ through the NMDA receptor channel during the pulse, and is amplified by the endogenous glutamate secondarily released after the pulse.¹¹⁰ Group-II mGluRs are preferentially localized in the preterminal region of the axon, far from the active zone of neurotransmitter release, and their activation inhibits the release of glutamate, but only in response to high concentrations of glutamate that spread back to the most remote region of the axon. Moreover, mGluR3 is also expressed by astrocytes throughout the brain.

It has been initially suggested that activation of mGluR2/3 is neuroprotective by reducing the release of endogenous glutamate. However, an additional mechanism might be responsible for neuroprotection since the protein synthesis inhibitor, cycloheximide, prevents the neuroprotective activity of group-II mGluR agonists in cortical cell cultures.^85,95,96 Thus, it has been proposed that group-II mGluR activation triggers a specific program that requires new protein synthesis and that is likely to occur in glial cells, as the conditioned medium collected from pure cultures of astrocytes, transiently exposed to DCG-IV, L-CCG-I, 4C3HPG, LY379268 (which activate both mGluR2 and mGluR3) or to NAAG (a selective mGluR3 agonist), is highly neuroprotective when transferred to mixed cortical cultures already challenged with NMDA.⁸⁵ Moreover, the neuroprotective activity of group-II mGluR agonists is almost completely lost in the absence of glial cells.^95,96 This effect is likely to be mediated by the production and release from astrocytes of transforming growth factor-β (TGF-β), induced by the activation of glial mGluR3.⁸⁶

In recombinant cells, mGluR2/3 are negatively coupled to adenylate cyclase through a Gi protein. However, it has been shown that in mixed cortical cultures exposed to NMDA, neither forskolin nor dibutyryl-cAMP counteract the neuroprotective activity of group-II mGluR agonists, excluding an involvement of the a subunits of the Gi protein coupled to group-II mGluRs. Therefore, to explain how activation of Gi-linked receptors leads to the production of neurotrophic factors, attention has been focused on the intracellular pathways activated by the βγ subunits released from the Gi protein (Fig. 4). A possible link is represented by the activation of mitogen-activated protein (MAP) kinase and phosphatidylinositol (PI)-3-kinase which occurs both in primary cultures and in recombinant cells (Fig. 4).⁸⁹ Accordingly, blockade of these pathways leads not only to a reduction of the TGF-βproduction induced by group-II mGluR activation, but also reverses the neuroprotective activity of mGluR2/3 agonists, an effect which also occurs in in vivo models of excitotoxicity.⁸⁹ MAP kinases, P90^rsk and P70^S6k, which regulate gene expression and protein synthesis, could represent the downstream effectors of the de novo synthesis of TGF-β.

Figure 4

Neuroprotection via activation of mGluR3. Schematic representation of the cascade of events leading to TGF-ß secretion by astrocytes. Protein kinases (e.g., TAK1) and SMAD proteins are likely to transduce TGF-ß signals from cell surface (more...)

However, the proposed mechanism of glial-neuronal interaction mediated by TGF-_ production in response to mGluR2/3 activation, observed both in cortical cultures challenged with NMDA and in rat striatum infused with NMDA, is not found in hippocampus after BCAO.¹¹¹ As Gi-linked receptors, such as group-II mGluRs and A1 adenosine receptors, have been shown to induce in cultured astrocytes the release of other factors, such as nerve growth factor and S-100b protein, other than TGF-β,¹¹² it cannot be excluded that in different brain regions other factors modulate the neuroprotective activity of group-II mGluRs.

mGluR2/3 Agonists and Potential Applicability in Acute and Chronic Neurodegeneration

Group-II mGluRs may be considered as a potential target for drugs aimed at reducing the progression of neuronal degeneration (Fig. 4). A safe neuroprotective drug with a favorable therapeutic window is particularly needed in the experimental therapy of ischemic brain damage. Group-II mGluR agonists might meet these criteria because: (i) they are neuroprotective against in vitro degeneration induced by oxygen-glucose deprivation and in in vivo models of global ischemia; (ii) they do not interfere with the normal excitatory synaptic transmission, in contrast to NMDA or AMPA receptor antagonists; (iii) they protect cultured neurons when applied after a toxic pulse with NMDA, at times at which NMDA antagonists are no longer protective, and they are even active in vivo when applied after the induction of an ischemic damage; (iiii) by locally increasing the production of neurotrophic factors such as TGF-β, they may provide a broad spectrum mechanism of protection, as TGF-β is known to exert neuroprotective activities against neuronal degeneration induced by excitotoxins, oxygen-glucose deprivation, β-amyloid peptide and the HIV capside protein gp120,^{86,98,99,113–124} and (iiiii) mGluR3 is localized on the vascular side of astrocytes, in proximity of endothelial cells, so they can be easily reached by drugs present in the blood stream and able to cross the blood-brain-barrier.

Recently, group-II mGluRs have been proposed as targets for the therapy of PD. Overactive glutamatergic afferents from the subthalamic nucleus could cause both an excitotoxic loss of dopaminergic neurons in the substantia nigra, and hyperactivation of GABAergic neurons in the globus pallidus, which leads to a reduction of motor activity.¹²⁵ Drugs, which can reduce glutamate release by acting at presynaptic group-II mGluRs might at one time delay the degeneration of substantia nigra neurons and improve motor activity.^126,127

Furthermore, it's interesting to add that group-II mGluR agonists could interfere with the pathophysiology of neuropsychiatric disorders such as anxiety and schizophrenia, and they may also exert anti-addictive effects.^{20,106,128–134}

Effects of Group-III mGluR Ligands in Experimental Paradigms of Neurodegeneration

There are numerous reports that describe activation of group-III mGluRs to be neuroprotective in vitro, and more recently, several in vivo observations support these findings. The agonists L-AP4, L-SOP and (RS)-PPG promote survival of rat cerebellar granule cells and protect cultured cortical and cerebellar neurons against toxic insults, such as prolonged β-amyloid peptide exposure, transient iGluR activation or mechanical damage.^{21,44,98,99,136–138} In rat hippocampal slices exposed to a severe hypoxic/hypoglycaemic insult, (RS)-PPG improves the recovery of population spike amplitude in CA1, a parameter for functional synaptic transmission and neuronal viability. Such acutely isolated hippocampal slices are only viable for a few hours, and electrophysiogical recordings can only be performed in a limited time window after the damaging insult when most probably excitotoxicity is still the predominant component in the pathophysiology of hypoxic/hypoglycaemic neuronal damage.^45,139

In a recent study we found (RS)-PPG to be neuroprotective against striatal lesions induced by local infusion of NMDA or quinolinic acid into the rat caudate nucleus and, to our knowledge, this provided the first in vivo evidence that activation of group-III mGluRs is neuroprotective in animal models.²¹ The use of such in vivo excitotoxic injury models, to produce neuronal depletion, reactive gliosis and alterations of neurotransmitter levels, has been highly valuable for examining pathological patterns reminiscent of Huntington's disease (HD). Even if the primary cause of HD is unrelated, excitotoxic injury mediated by iGluR activation may play a role in progressive neuronal depletion.¹⁴⁰

Extending the studies to further in vivo models of neurodegeneration, we and others confirmed neuroprotective effects of (RS)-PPG in rat and also mouse models of excitotoxicity (in vivo and in vitro), but, neither in focal cerebral ischaemia in mice, nor in global cerebral ischaemia in gerbils, nor in global cerebral ischaemia in rats (RS)-PPG had any significant influence on the extent of neuronal damage.^44,45

Neuroprotection via Group-III mGluRs Is Mediated by mGluR4 and/or mGluR7 and Involves a Presynaptic Regulation of Transmitter Release

In contrast to group-II mGluRs that require glial components for neuroprotection, Group-III agonists are equally protective in mixed cortical cultures (containing astrocytes and neurons, see Fig. 5) and in pure cultures of neurons. Therefore, this neuroprotection is not expected to be mediated by glial factors, but rather involves one or more receptor subtype(s) expressed on neuronal structures. Recently, we have searched for the identity of the neuroprotective group-III mGluR subtype using mixed cortical cultures. This model is particularly suitable for the study because cortical neurons of mixed cultures express all currently known group-III mGluR subtypes.¹³⁷ A possible role in neuroprotection for one or more group-III mGluR subtype(s) was initially suggested by the use of racemic (RS)-PPG in in vitro neuroprotection paradigms.²¹ More recently, the stereoisomers of (RS)-PPG were separated and it was found that all protective activity is harbored in the (+)-isomer.^44,141 (+)-PPG is neuroprotective against NMDA toxicity with an EC₅₀ value of about 5 μM. This value coincides with that found for the activation of recombinant mGluR4, but differs by at least 25-fold from the potency of PPG at mGluR7 and −8.²¹ Furthermore, there is strong evidence for a critical role of mGluR4 in mediating neuroprotection via the use of cortical cultures prepared from mGluR4 subtype-deficient mice (−/−, Fig. 5B), where all group-III agonists [i.e., L-AP4, (RS)-PPG and L-SOP] failed to protect against NMDA toxicity. This was in contrast to the protective effects of group-III agonists in wild-type (+/+) and heterozygous neurons and did not reflect a general refractoriness of −/− knockout neurons to mechanisms of protection, because the group-I/II mGluR ligands 4C3HPG, CPCCOEt and MPEP, retained their protective activity in those mGluR4 −/− cultures.^32,33,44,142

Figure 5

Phase-contrast photomicrographs of cultured cortical cells from CD1 (wild-type, wt, A) and mGluR4-deficient (knock-out, ko, B) mice. Dissociated cortical neurons prepared from fetal mice at 14–16 days of gestation after 13 days of in vitro culture (more...)

We extended the study to the in vivo model of excitotoxic degeneration by unilaterally injecting NMDA ± (RS)-PPG into the caudate nucleus of wild-type or mGluR4 −/− mice. This brain region has been selected because it receives an extensive glutamatergic innervation from the cerebral cortex. Low doses of (RS)-PPG (10 nmol), which should preferentially activate mGluR4 over mGluR7 receptors, were neuroprotective in +/+ mice, but were totally inactive in mGluR4 −/− mice. In contrast, the −/− mice were partially protected with higher doses of (RS)-PPG (100 nmol),⁴⁴ which are expected to recruit mGluR7 (unpublished observations by Herman van der Putten, based on experiments with mGluR7 −/− mice). The prominent role of mGluR4 and possibly mGluR7 in mediating group-III agonist-induced neuroprotection is also consistent with the high expression levels and broad distribution of mGluR4 and mGluR7 in many regions of mammalian brain, including basal ganglia, cortical areas and hippocampus.^23,143–145

Inhibition of glutamate release by presynaptic mGluR4 and mGluR7 may represent a common mechanism of neuroprotection in vitro and in vivo. Accordingly, an enhanced release of endogenous glutamate has been shown to facilitate the progression of NMDA toxicity in cortical cultures,¹¹⁰ and we have found that cortical cultures from mGluR4 −/− mice were more vulnerable to low concentrations of NMDA, and showed higher extracellular glutamate levels, as compared to wild-type +/+ cultures. Moreover, in vivo microdialysis studies showed that intrastriatal infusion of NMDA increased extracellular glutamate levels to a greater extent in mGluR4 −/− than in +/+ mice, supporting the hypothesis that the mGluR4 subtype is necessary for the maintenance of the homeostasis of extracellular glutamate levels.⁴⁴ In addition to the regulation of presynaptic glutamate release, an inhibition of NMDA receptors by postsynaptic group-III mGluRs (hypothetical) via a protein phosphorylation cascade may also be involved in their neuroprotective effects.¹⁴⁶ Furthermore, group-III mGluRs also exert protection in primary hippocampal neurons by modulation of the free radical nitric oxide and the cascade of programmed cell death.¹⁴⁷ Thus, activation of group-III mGluRs could open several novel strategies to interfere with the progressive course of neurodegenerative disorders.

mGluR4/7 Activators May Become Applicable in Basal Ganglia Disorders and Other Chronic Degenerative Diseases

The broadest analysis of group-III mGluR activation in animal models of neuronal damage was recently preformed by Henrich-Noack et al.⁴⁵ The results, as summarized above, support the notion that group-III mGluR agonists are a quite valuable class of drugs against pathologies closely related to excitotoxic cell damage, but are most probably not effective enough when damaged brain tissue has progressed into a multifactorial pathology as it appears after an ischaemic challenge, e.g., in human stroke. This view is supported by several clinical stroke trials that have evaluated the potential of glutamate transmission-modifying drugs, and to date, the results of these attempts have been disappointing.^148,149

In amyotropic lateral sclerosis (ALS) and Huntington's disease (HD), circumstantial evidence suggests that excitotoxicity may contribute to the pathogenic process. In fact, it was recently shown that an anti-glutamate drug, riluzole, provides some therapeutic benefit in the treatment of ALS and clinical trials aiming at neuroprotection in HD are in progress.¹⁴⁸ HD and Parkinson's disease (PD) are examples of neurodegenerative disorders where mitochondrial dysfunction may sensitize populations of basal ganglia neurons to excitotoxicity from synaptic glutamate, and recent evidence strongly suggests that the group-III receptors mGluR4 and mGluR7 modulate basal ganglia function at multiple sites.^143,144 Activation of mGluR4 and/or mGluR7 may increase viability of selective cell populations, e.g., GABAergic-projecting neurons in the striatum which are selectively lost in HD. Alternatively, modulation of group-III mGluR activity may balance basal ganglia circuits under pathological conditions, for instance in PD where subthalamic nuclei are overactive.¹⁴⁴ Thus, reduction of glutamatergic transmission by group-III mGluR activation may well be a novel and viable mechanism for experimental therapy of ALS, HD, and PD.

Evidence for an involvement of group-III mGluRs, or glutamatergic transmission in general, in Alzheimer's disease (AD) is more indirect and speculative than for ALS, HD, and PD. Experimental evidence, however, shows that various group-III agonists are highly protective against programmed cell death in cultured cortical cells induced by prolonged β-amyloid peptide exposure.^80,98 On the other hand, no in vivo data from transgenic models of AD have been published to date that would support a role of group-III receptors in AD.

Finally, it's very interesting to mention the effectiveness of group-III agonists in various experimental animal models of epilepsy.^21,150–152