Imaging of Metabotropic Glutamate Receptors (mGluRs)

The ubiquitous amino acid L-glutamate is thought to act as a neurotransmitter at the majority of synapses in the brain. It mediates the major excitatory pathways in the brain, and is referred to as an excitatory amino acid (EAA). The EAA plays a role in a variety of physiological processes, such as long-term potentiation (learning and memory), the development of synaptic plasticity, motor control, respiration, cardiovascular regulation, emotional states and sensory perception (Bliss & Collingridge, 1993). The excessive or inappropriate stimulation of EAA receptors leads to neural cell damage or loss by a mechanism known as excitotoxicity (Lucas & Newhouse, 1957; Oney, 1978). EAA receptors are classified in two general types (Kornhuber & Weller, 1997). Receptors that are directly coupled to the opening of cation channels in the cell membranes of the neuron are termed ‘ionotropic’, which include NMDA, AMPA, and kainate receptors. The second type of receptors are the G-protein or second messenger-linked ‘metabotropic’ EAA receptors. This second type is coupled to multiple second messenger systems that lead to enhanced phosphoinositide hydrolysis, activation of phospholipase D, increase or decrease in cAMP formation, and changes in ion channel function (Kozikowski et al., 1998). Metabotropic glutamate receptors belong to Class C of a superfamily of G-protein coupled receptors (GPCRs). Class C GPCRs possess a large extracellular domain that is responsible for endogenous ligand recognition (Pin et al., 2003), in addition to the seven strand transmembrane domain, which is characteristic of all GPCRs. The mGluRs possess a large bi-lobed extracellular N-terminus of ~560 amino acids which has been shown by mutagenesis studies to confer glutamate binding, agonist activation of the receptor, and subtype specificity for group selective agonists (Schoepp et al., 1999). Since mGluRs have neuromodulatory role in the control of both glutamatergic and GABAergic neurotransmission, there has been much interest to develop novel mGluR ligands for therapeutic purposes of a variety of neurological and psychiatric conditions. The mGluRs have been proposed to be involved in physiological and pathophysiological processes of a number of CNS disorders, including anxiety, pain, depression, neurodegenerative disorders, schizophrenia, epilepsy, and drug abuse. In order to


500
characterize the role of mGluRs in different physiological processes there is a need to identify novel compounds, which are highly potent and specific for an mGluR group or a subtype. Such compounds are needed to further investigate mGluR function, and as potential therapeutic agents for a variety of neurological diseases, which are associated with the abnormal activation of mGluRs. A large amount of pharmacological agents acting at metabotropic glutamate receptors have been described in the literature (Guitart & Khurdayan, 2005;Kew, 2004;Layton, 2005;Marino et al., 2005;Rudd & McCauley, 2005;Schoepp et al., 1999;Slassi et al., 2005;Williams & Lindsley, 2005;Yang, 2005). According to the mode of binding, these mGluR pharmacological agents can be classified into competitive and non-competitive agents. Based on the mode of action, they can be classified into agonists, antagonists, and positive/negative/neutral modulators (Layton, 2005). Competitive agonists and antagonists bind to the same orthosteric binding site as endogenous glutamate (Niswender et al., 2005;Ritzen et al., 2005;Rudd & McCauley, 2005), which is a cleft between the two lobes in the extracellular N-terminus. Their binding ability depends on how much they can stabilize the closed conformation (Kew, 2004). These ligands received earliest research interest and have been well developed (Schoepp et al., 1999). They are all glutamate analogs or substituted glycines, which imply that they have poor selectivity within their group. In addition, competitive agonists and antagonists have structural carboxyl and amino groups, which make them too polar to penetrate the blood brain barrier (BBB) (Kew, 2004). Starting from 1996(Annoura et al., 1996, a number of different types of non-competitive negative, positive and neutral allosteric modulators have been developed as mGluR ligands (Niswender et al., 2005;Ritzen et al., 2005). These ligands modulate mGlu receptor activity by binding to allosteric binding sites that are located in the seven strand transmembrane domain. The allosteric binding sites are structurally distinct from the classical agonist orthosteric binding site (Williams & Lindsley, 2005). Positive and negative modulators thus offer a potential for improved selectivity for individual mGluR family members compared to competitive agonists and antagonists at the glutamate site (Kew, 2004). Positive allosteric modulators (PAM)s have little or no effect on the receptor but can significantly enhance the effect of endogenous ligand. Correspondingly negative allosteric modulators inhibit the activity of orthosteric agonists in a noncompetitive manner. These ligands are structurally diverse and not amino acid derivatives. They are lipophilic and have much better CNS penetrating ability. Thus, positive and negative modulators with high subtype selectivity, and appropriate lipophilicity are good candidates for mGluR radiotracer development. There will be no competitive binding of this kind of tracers with endogenous glutamate, which might otherwise decrease the availability in vivo, and thus decrease the sensitivity of potential ligands. During the last fifteen years the subtype selective modulators have been identified for mGluR1, mGluR2, mGluR3, mGluR4, mGluR5, mGluR7 and mGluR8. Based on these modulators, several positron emission tomography radiotracers have been developed for in vivo imaging of specific mGluRs. Presently, three mGluR ligands have been used for human studies. They have been developed as negative allosteric modulators for mGluR5. In this review we intend to summarize the radiotracers which have characteristics to be developed as tracers for in vivo PET imaging to investigate modulation of mGluRs in normal and pathological conditions. Emphasis will also be given to the highly potent and subtype selective allosteric modulators which are candidates for radiolabeling with 18 F or 11 C.

Metabotropic glutamate receptors and their physiological function
Recent molecular cloning studies have revealed the existence of eight different subtypes of mGluRs. The mGluR subtypes can be divided into three different groups according to their sequence similarities, signal transduction mechanism, and pharmacological profiles to agonists (Pin & Duvoisin, 1995). The first group comprising mGluR1 and mGluR5 is coupled to stimulating of phosphoinositide hydrolysis/Ca2+ signal transduction (Schoepp et al., 1994). The second group, consisting of mGluR2 and mGluR3, is negatively coupled through adenylate cyclase to cAMP formation (Tanabe et al., 1997). The third group, containing mGluR4, mGluR6, mGluR7 and mGluR8, is also negatively linked to adenylate cyclase activity but shows a different agonist preference (Conn & Pin, 1997;Tanabe et al., 1997). The neuroanatomical localization of Group I and Group II mGluRs in the rodent brain, as assessed by immunohistochemical or in situ hybridization techniques, has revealed overlapping, yet distinct patterns of expression of these receptors. In order to better characterize the roles of mGluRs in physiological processes, there is a need to identify novel compounds that are highly potent and specific for an mGluR group or a subtype. Such compounds are needed as pharmacological tools for further investigation of mGluR function, and as potential therapeutic agents for the treatment of diseases or conditions including epilepsy, cerebral ischemia, pain, spinal cord injury, 'neurotoxicity' and chronic neurodegenerative diseases (e.g. Parkinson's and Huntington's disease), which are associated with abnormal activation of mGluRs (Aguirre et al., 2001;Blakely, 2001;Calabresi et al., 1999;Keyvani et al., 2001;Marino et al., 2001;O'Neill, 2001;Popoli et al., 2001;Rao et al., 2000;Rouse et al., 2000). It is known that glutamate can act as a neurotoxin when energy supplies are compromised. This has stimulated a hypothesis that injury to neurons in some neurological conditions may be caused, partly, by over stimulation of glutamate receptors and/or glutamate transporters. These neurological conditions may be acute insult like stroke or chronic neurodegenerative states like Parkinson's or Huntington's disease or dementia. To better explore the roles of mGluRs in physiological and pathological processes, there is a need to learn more about functional behavior of these receptors in vivo.

PET radiotracer development
Positron emission tomography (PET) has become an important clinical diagnostic and research modality, and also a valuable technology in drug discovery and development (Cai et al., 2008). PET tracers have been used for the imaging and quantification of biochemical processes. PET tracers play a critical role for assessing in vivo distribution of specific receptors in normal and disease conditions to understand underlying mechanisms of physiology and pathology. Moreover, PET tracers serve as invaluable biomarkers during the clinical development of potential therapeutic mGluR modulators, in which the receptor occupancy of potential drug candidates in the brain is measured (Passchier et al., 2002;Sharma & Lindsley, 2007). In vivo receptor occupancy can help to answer many vital questions in the drug discovery and development process such as whether potential drugs reach their molecular targets, the relationship between therapeutic dose and receptor occupancy, the correlation between receptor occupancy and plasma drug levels, and the duration of time a drug remains at its target (Passchier et al., 2002). In PET imaging a small amount of tracer is injected into a living object. The tracer is labeled with a short-lived radioisotope, which emits positrons as it decays. The positrons collide with electrons resulting in high-energy photons that escape from the object and are detected by the PET scanner. Carbon-11 (t 1/2 = 20.4 min) and fluorine-18 (t 1/2 = 109.7 min) are the most commonly used radionuclides in PET imaging (Miller et al., 2008). The characteristics of successful PET tracers include high affinity, high selectivity over other mGluR subtypes as well as other receptors, suitable pharmacological properties including lipophilicity, metabolic stability, no radiolabeled metabolites that can penetrate into the brain, and the chemical structure of the precursor to allow fast labeling.

Allosteric modulators and radiotracers for Group I mGluRs
The group I receptors mGluR1 and mGluR5 exhibit different patterns of expression in the CNS. The distribution of mGluR1 is found throughout the human brain with high levels in the olfactory bulb, thalamus, hippocampus, lateral septum, superior colliculus and cerebellum (Olive, 2009). Inhibition of mGluR1 has been suggested as potential treatment for various psychiatric disorders including schizophrenia, anxiety, and neuropathic pain. The mGluR5 is usually found in postsynaptic neurons with moderate to high density in the frontal cortex, caudate, putamen, nucleus accumbens, olfactory tubercle, and hippocampus, whereas in contrast to expression patterns of mGluR1, the density in the cerebellum is low (Olive, 2009). Dysfunction of mGluR5 is implicated in a variety of diseases in the CNS, including anxiety, depression, schizophrenia, Parkinson's disease, and drug addiction or withdrawal.

Allosteric modulators and radiotracers for mGluR1
A variety of mGluR1 modulators have been reported in the literature. Competitive mGluR1 agonists and antagonists historically have been amino acid derivatives, which display poor potency, lack of selectivity and unsatisfactory BBB penetration (Layton, 2005). Although a number of selective competitive mGluR1 ligands appear in literature, they are not good candidates for potential PET tracers. None of the existing orthosteric ligands has a binding affinity (or potency) of IC 50 /K i /K d less than 20 nM with an acceptable selectivity over other members in the same group. There is a consensus that identification of highly potent and subtype selective competitive mGluR ligands has been difficult due to a high degree of sequence similarity at the orthosteric binding site to which the endogenous agonist binds (Layton, 2005;Williams & Lindsley, 2005). Alternatively, several structural types of mGluR1 allosteric modulators have been reported in literature, including negative and positive allosteric modulators which show high binding affinity, high selectivity and good lipophilicity (Layton, 2005). CPCCOEt (1) was the first reported mGluR1 negative allosteric modulator (Fig.1). Before 2008, only compound 4 (3,5-dimethyl PPP) (Micheli et al., 2003b) and a quinoline derivative 5 (JNJ16259685) (Lavreysen et al., 2004b;Mabire et al., 2005) had reported binding affinity (or potency) less than 20 nM (Table 1). 2,4-Dicarboxy-pyrrole ester 4 (3,5-dimethyl PPP), as a racemic mixture, is a highly potent and subtype-selective noncompetitive antagonist of mGluR1, having IC 50 of 16 nM at rat mGluR1 and > 1000-fold selectivity over mGluR 2, 4, and 5 (Micheli et al., 2003b). Pharmacological studies of its two enantiomers showed that the S-enantiomer had the same activity as the racemic mixture, while the R-enantiomer was less potent (40 nM). Although compound 4 had a poor stability to rat plasma esterase (t 1/2 =12 min versus 2.8 h in mice), a good CNS accumulation was observed 5 min after intravenous administration with a brain/plasma ratio of 20 (Micheli et al., 2003b). Compound 5 (JNJwww.intechopen.com

Imaging of Metabotropic Glutamate Receptors (mGluRs)
503 16259685) demonstrated high specificity over other mGlu receptor subtypes and a fast brain penetration with high receptor occupancy after subcutaneous administration (Lavreysen et al., 2004b). In addition to 5, a series of quinoline derivatives have been synthesized. The in vitro pharmacological data showed that they are highly potent noncompetitive mGluR1 antagonists (Mabire et al., 2005) with high binding affinity. However, the quinoline derivatives have issues of poor aqueous solubility and poor stability to human liver microsomes (Layton, 2005;Mabire et al., 2005). Since 2008, many new compounds ( Fig. 1 and Table 1) have been reported having binding affinity (or potency) less than 20 nM and high selectivity over other mGluRs. These compounds are diverse heterocyclic compounds including mono-, di-and tri-cyclic structures. Some of these compounds or their derivatives are amenable to radiolabeling with fluorine-18 or carbon-11. For example, a series of potent 2-fluoro-3-pyridyl-triazol derivatives such as FTIDC (10) and FPTQ (11) have been developed.. These derivatives are relatively easy to label with fluorine-18 at 2-pyridine position. Other compounds such as 12 are amenable to radiolabeling with carbon-11. MGluR1 expression is localized throughout the nervous system (Layton, 2005;Spooren et al., 2003). The distribution of mGluR1 in the peripheral nervous system (Bhave et al., 2001;Lesage, 2004;Skerry & Genever, 2001) and in the CNS has been studied using various methods including radioligand autoradiography and immunohistochemical techniques (Lavreysen et al., 2003;Lavreysen et al., 2004a;Shigemoto & Mizuno., 2000;Simonyi et al., 2005). MGlu1 receptors have been observed in the cerebellum, thalamus, hippocampus and  (Karakossian & Otis, 2004;Lavreysen et al., 2003;Shigemoto & Mizuno, 2000;Spooren et al., 2003). Tritium-labeled highly potent and subtype-selective radioligands were used earlier in mapping mGluR1 ex vivo (Yang, 2005). Presently, demands on PET radioligands are increasing due to the advantage of in vivo noninvasive imaging techniques to investigate pathophysiological processes. In 2002, a carbon-11 labeled CPCCO-Me analog was described in the literature (Yu & Brownell, 2002), but no animal studies were conducted. In the series of quinoline derivatives (represented by 5), several compounds are amenable to radiolabeling with either fluorine-18 or carbon-11. Carbon-11 labeling would not be preferred in the methyl ether positions, in spite of methyl ether position is very popular in 11 C-methylation, since O-demethylation of the methoxy groups on the quinoline moiety and the cyclohexyl ring are the major metabolic pathways (Mabire et al., 2005). Therefore, practical methods must be developed to label the methyl groups elsewhere in the molecule. Accordingly, Huang et al. successfully labeled a quinoline derivative, providing the first PET tracer, [ 11 C]JNJ-16567083, suitable for in vivo imaging of mGluR1 (Huang et al., 2005). [ 11 C]JNJ-16567083 (17) is an analog of JNJ-16567083 (5). In vitro binding experiments showed that JNJ16567083 (cold compound) possesses high affinity for rat mGluR1 (K i = 0.87 nM) and low affinity for mGluR5 (K i = 2366 nM). Ex vivo biodistribution studies in rats showed that [ 11 C]JNJ-16567083 has high brain uptake and its binding in brain is specific to mGluR1. MicroPET imaging experiments in rats indicated that radioactivity entered the brain rapidly and was localized over time in brain regions with high densities of mGluR1, such as the cerebellum and striatum. Activity in cerebellum peaked at ~10 min after intravenous injection. Radioactivity uptake was highest in the cerebellum, followed by striatum and hippocampus. However, evaluation of this PET tracer in higher species has not been reported. Yanamoto et al. have labeled an mGluR1 antagonist YM-202074 (12, K i = 4.8 nM) with 11 C and evaluated its potential as a PET ligand for mGluR1 (Yanamoto et al., 2010). In vitro autoradiographic study demonstrated that [ 11 C]YM-202074 (21, Fig.2) had high specific binding with mGluR1 in the rat cerebellum and its regional distribution was consistent with the distribution pattern of mGluR1 in the brain. However, the total accumulation of [ 11 C]YM-202074 in the brain was very low including lipophilic radiometabolites hampering its usefulness for in vivo imaging. Prahakaran et al. have reported the synthesis for in vitro and in vivo evaluation of [ 11 C]MMTP (20) as a potential PET ligand for mGluR1 (Prabhakaran et al., 2010). Synthesis of the corresponding desmethyl precursor was achieved by demethylation of the methoxyphenyl compound MMTP in 90% yield. Methylation using [ 11 C]MeOTf in presence of NaOH afforded [ 11 C]MMTP in 30% yield (EOS) with >99% chemical and radiochemical purities and with a specific activity of 3-5 Ci/µmol (n = 6). The total synthesis time was 30 min from EOB. In vitro autoradiography using phosphor imaging demonstrated that the radiotracer bound selectively mGlu1 receptors in slide-mounted sections of postmortem human brain containing cerebellum, hippocampus, prefrontal cortex and striatum. PET studies in anesthetized baboon showed that [ 11 C]MMTP penetrates the BBB and accumulates in cerebellum, a region of high expression of mGluR1. Recently, a 18 F-labeled triazole analog [ 18 F]FTIDC (19, Ki = 3.9 nM) (Ohgami et al., 2009) was presented for imaging of mGluR1 showing high uptake in the rat brain. In addition, Fujinaga et al. have labeled a triazole analog, FPTQ (11, IC 50 = 3.6 nM and 1.4 nM for human and mouse mGluR1, respectively) (Fujinaga et al., 2011 In vitro autoradiography showed that [ 18 F]FPTQ had high specific binding with mGluR1 in the rat brain. Biodistribution study using a dissection method and small-animal PET showed that [ 18 F]FPTQ had high uptake in the rat brain. The uptake of radioactivity in the cerebellum was reduced by unlabeled FPTQ and mGluR1-selective ligand JNJ-16259685 (Fujinaga et al., 2011), indicating that [ 18 F]FPTQ had in vivo specific binding to mGluR1. However, because of a low amount of radiolabeled metabolite present in the brain, this compound may have limiting use for in vivo imaging of mGluR1 by PET. Hostetler et al. have reported a PET radioligand, [ 18 F]MK-1312 (18), which was radiolabeled with fluorine-18 via nucleophilic displacement of the corresponding 2-chloropyridine precursor with [ 18 F]potassium fluoride (Hostetler et al., 2011). [ 18 F]MK-1312 was synthesized (n = 25) in good yield (46 ± 15%) with >98% radiochemical purity and high specific activity (2.5 ± 1.4 Ci/µmol). In vitro autoradiographic studies with [ 18 F]MK-1312 in rhesus monkey and human brain tissue slices revealed an uptake distribution consistent with the known distribution of mGluR1, with the highest uptake in the cerebellum, moderate uptake in the hippocampus, thalamus, and cortical regions, and the lowest uptake in the caudate and putamen. In vitro saturation binding studies in rhesus monkey and human cerebellum homogenates confirmed that [ 18 F]MK-1312 binds to a single binding site with a Bmax/Kd ratio of 132 and 98, respectively. PET studies in rhesus monkey with [ 18 F]MK-1312 showed high brain uptake and a regional distribution consistent with in vitro autoradiography results. Blockade of [ 18 F]MK-1312 uptake with mGluR1 allosteric antagonist MK-5435 dosedependently reduced tracer uptake in all regions of gray matter. These results show that [ 18 F]MK-1312 is a promising PET tracer for clinical studies to determine mGluR1 occupancy of MK-5435. In summary, several PET radioligands have been developed using highly potent and subtype-selective mGluR1 negative allosteric modulators. Although they showed efficacy in studying the distribution of mGluR1, some compounds may have limited applications because of low brain uptake and/or brain penetrating radiometabolites. [ 18 F]MK-1312 is the most advanced mGluR1 PET tracer, which has demonstrated efficacy in rhesus monkey. Although all the published mGluR1 PET tracers are radiolabeled mGluR1 negative allosteric modulators, mGluR1 positive allosteric modulators can also be used for developing mGluR1 PET tracers. Several papers have been published about the functional differences between antagonist and agonist tracers in imaging G-protein coupled receptors, including dopamines D2 receptor (Hwang et al., 2004;Wilson et al., 2005), serotonin receptors (Kumar et al., 2006;Prabhakaran et al., 2006) and mGlu receptors. GPCRs have been postulated to exist in interconvertible high-affinity and low-affinity states. The high-affinity sites are G-protein coupled, whereas the low-affinity sites are those uncoupled with G-protein. Antagonist radiotracers bind with equal affinity to both the high-and low-affinity forms of the receptor, and they do not provide information about in vivo affinity of the receptor for antagonist. On the contrary, agonist radioligands bind only to high-affinity form of the receptor, thus giving valuable information about in vivo affinity of the receptor for agonists in normal and abnormal states. Concerning the binding sites of allosteric modulators in the seven strand transmembrane domain, there is no evidence for difference between negative and positive modulators in terms of their binding to high-affinity or low-affinity states of mGlu receptors (Kew & Kemp, 2005). Fig. 3 illustrates structures of some representative positive allosteric mGluR1 modulators reported (Knoflach et al., 2001;Layton, 2005;Wichmann et al., 2002).

Allosteric modulators and radiotracers for mGluR5
Since the first selective mGluR5 antagonist was identified in 1999 (Varney et al., 1999), a large number of potent, subtype selective and structurally diverse allosteric modulators have been described. SIB1757 (28) and SB1893 (29) were discovered through random screening. Subsequent optimization by replacement of the trans-olefinic tether in SIB1893 (29) with a C≡C triple bond led to MPEP (30), which demonstrated a dramatically improved mGluR5 antagonist activity . Various structure-activity relationship (SAR) studies have been done on MPEP, in which chemical modifications were done to each of the three regions of the lead molecule, identifying a series of highly potent and selective diaryl (heteroaryl) acetylenes as mGluR5 noncompetitive antagonists. By assumption that the (2-methyl-1,3-thiazo-4-yl)ethynyl group is one of the best structural parts to achieve mGluR5 antagonist activity further SAR studies on MTEP (31) identified more high-profile ligands containing thiazole moiety as mGluR5 noncompetitive antagonists such as (33) (Iso et al., 2006). Many PET tracers have been synthesized by radiolabeling on the derivatives of MPEP and MTEP. A major concern with acetylenes in potential drugs is the possibility of chemical or metabolic reactivity (Milbank et al., 2007). Terminal acetylenes are well known to be mechanism-based CYP-inactivators (Testa & Jenner, 1981) and there is an increasing body of information suggesting that internal acetylenes can be activated by CYPs (Fontana et al., 2005;Foroozesh et al., 1997;Shimada et al., 2007) or even undergo uncatalyzed addition of glutathione (Chen et al., 2002;Mutlib et al., 1999). Mutlib et al. reported that incubation of MPEP with triple-labeled glutathione gave compounds with molecular weights and fragmentations consistent with both activated and unactivated addition of GSH to the alkyne (Mutlib et al., 2005). These events are potential sources for hepatic or idiosyncratic toxicity. To avoid a potential metabolic liability, many research groups have designed and synthesized mGluR5 negative allosteric modulators without the acetylene structure. Some structures such as 37 to 48 are given in Fig. 4, which may be useful for development of a new PET tracer.  (Kinney et al., 2005;Lindsley et al., 2004), respectively. Subsequent structure-activity relationship study on CDPPB identified several nanomolar potent pyrazole ligands (De Paulis et al., 2006). Although these compounds are potent with an EC 50 value of less than 20 nM, their poor binding affinity (K i ) and high lipophilicity  The discoveries of noncompetitive allosteric modulators with high binding affinity and subtype-selectivity entitle the exploration of the physiological functions of mGluR5 in normal and pathological states. Although in vitro and ex vivo studies using selective mGluR5 allosteric antagonists labeled with tritium Gasparini et al., 2002) have played important roles in elucidating the distribution and functions of mGluR5, PET tracers are needed for the in vivo quantitative visualization of mGluR5 in a living body and to conduct longitudinal studies of modulation of mGluR5 expression.

PET imaging studies of mGluR5 function
MPEP and MTEP have provided leads to some radioligand candidates for imaging human mGlu5 receptors with PET in vivo. Great effort was done to identify suitable positronemitting radiotracers for noninvasive imaging of mGluRs. To date, more than 15 mGluR5selective PET ligands labeled with 18 F or 11 C have been reported (Fig. 6) (65) into the brain was fast and the highest accumulation was reached in 1-5 min followed by fast washout, suggesting little retention by high affinity receptor binding. This creates limitation to obtain statistically meaningful imaging data without overdosing the object with radiation or saturating the receptor binding sites with accompanying cold compound. These ligands have limitation, due to high lipophilicity, unfavorable brain uptake kinetics, or a high rate of metabolism, though they possess favorable in vitro pharmacological profiles. For PET ligands to be used in the central nervous system, a postulated lipophilicity coefficient (logD or logP) value should be between 2 and 3 for good brain accumulation.  (Sime´on et al., 2007). [ 18 F]-SP203 has high affinity (IC 50 = 36 pM) and potency in a phosphoinositol hydrolysis assay (IC 50 = 0.71 pM) for mGluR5. It demonstrates a high uptake in mGlu5 receptor rich regions of the rat and rhesus brain. The major advantage of this tracer over [ 18 F]F-MTEB is its high radiochemical yield (87%) and easy radiosynthesis. This ligand is presently in NIH administrated clinical trial.
[ 11 C]M-FPEP (69, K D 1.2 nM and B max 84.5 f mol /mg) has an even biodistribution in all brain regions demonstrating that this tracer lacks specific binding (Ametamey et al., 2003).
Compound 70 showed little retention by the receptor (Krause et al., 2003). Compound 71 (rat K i 0.23 nM) had a good brain uptake and slow washout, with high concentration in striatum, frontal cortex and cerebellum of monkey    In PD monkey accumulation of [ 18 F]FPEB was enhanced compared to naïve monkey in several brain areas including caudate, putamen, accumbens and SN/VTA. Distribution of [ 18 F]FPEB accumulation is illustrated at 60-70 min after administration of radioligand (1.2-1.5 mCi iv., specific activity 1.9 Ci/mol).
cerebellum is an area with fairly low mGluR5 expression indicating that 71 may have nonspecific binding. Four derivatives, 72-75, were developed of ABP688. PET imaging with 72 (Lucatelli et al., 2009) did not allow visualization of mGluR5-rich brain regions in the rat brain due to fast washout and rapid defluorination. Compound 73 (Baumann et al., 2010a) was reported to have the high binding affinity to mGluR5. Further in vitro evaluation and in vivo imaging are needed for characterization of this ligand. Baumann et al. (Baumann et al., 2010b) reported that although [ 18 F]-FTECMO (74) displayed optimal lipophilicity (log D pH7.4 = 1.6 ± 0.2) and high stability in rat and human plasma as well as sufficient stability in rat liver microsomes, PET imaging with [ 18 F]-FTECMO in Wistar rats showed low brain uptake. Uptake of radioactivity into the skull was observed suggesting in vivo defluorination. Honer et al. reported that [ 18 F]-FE-DABP688 (75) have optimal lipophilicity (logD 2.1±0.1) and high plasma stability (Honer et al., 2007). Saturation assays of [ 18 F]-FE-DABP688 revealed a single high affinity binding site with a dissociation constant (K d ) of 1.6±0.4 nM and a B max value of 119±24 fmol/mg protein. PET scanning indicated radioactivity uptake in mGluR5-rich regions such as the hippocampus, striatum and cortex, and radioactivity accumulation in the cerebellum, a region with negligible mGluR5 density, was significantly lower. Biodistribution studies showed a similar distribution pattern of [ 18 F]-FE-DABP688 binding in the brain. The hippocampus-to-cerebellum and striatum-to-cerebellum ratios were 1.81±0.16 and 1.93±0.36, respectively. Blocking studies using coinjection of [ 18 F]-FE-. DABP688 and unlabeled M-MPEP (1 mg/kg) revealed more than 45% replacement in the hippocampus and striatum, thus demonstrating the in vivo specificity of tracer binding. This result shows that [ 18 F]-FE-DABP688 may be a useful PET tracer for imaging mGluR5

Allosteric modulators and radiotracers for group II mGluRs
Group II mGluRs have been shown to be expressed in several brain areas. The expression patterns of Group II receptors in the rodent brain parallel those of mGluR5, although the overall abundance of mGluR2/3 receptors appears slightly reduced as compared with that of mGluR5 (Olive, 2009). Expression levels of mGluR2/3 receptors are high in the olfactory bulb and hippocampus, and moderate in the dorsal striatum, nucleus accumbens, amygdala, anterior thalamic nuclei, cerebral cortex and cerebellum. Low levels of mGluR2/3 are found in the pallidum, colliculi, ventral midbrain and hypothalamus. Group II mGluRs act in the hippocampus to decrease synaptic transmission and glutamate release when activated. These receptors have been targeted extensively by potential neuroprotective agents to develop treatments for anxiety, schizophrenia, Alzheimer's disease, Parkinson's disease, pain, drug withdrawal, and epilepsy (Rudd & McCauley, 2005).

Allosteric modulators for mGluR3
Eli Lilly and Company reported the first series of compounds, 1-(heteroaryl)-3-(2,4dichlorobenzyl)amino-pyrolidine, acting as mGluR3 negative allosteric modulators (Britton et al., 2006). Figure 11 shows the chemical structures of two most potent ligands reported in the patent. Compounds, 94 and 95, have an IC 50 value of 77 nM, which is insufficient for in vivo detection of the receptors. Further SAR studies are needed to find more potent ligands. No PET radioligands have been identified for mGluR3 so far.

Allosteric modulators and radiotracers for Group III mGluRs
Group III metabotropic glutamate receptors are mGluR4, mGluR6, mGluR7 and mGluR8. There is no publication reporting mGluR6 allosteric ligands. MGluR4 has received much attention lately due to its implication in several diseases, such as PD, epilepsy, and anxiety. There has been substantial progress in identifying positive allosteric modulators for mGluR4. The compound PHCCC (96, Fig. 12), a partial selective mGluR4 potentiator, has been studied for many years. Unfortunately PHCCC and other early disclosed mGluR4 PAMs such as 98-101 (Fig.12) are deficient in their BBB penetration (Engers et al., 2009). The potencies of these compounds are also relatively low (EC 50 : 0.65 -5.0 M) and SAR studies around these structures have given 'flat' results. Addex Pharma disclosed a series of heteroaromatic compounds (102 in Fig. 12) as positive allosteric modulators for mGluR4, with many compounds having EC 50 < 0.5 µM (Bolea & Celanire, 2009). However, no other information was reported about these compounds. Two research groups; Addex Pharma (Bolea, 2009) and Vanderbilt University (Engers et al., 2009), have independently disclosed a series of small arylamide compounds as a new class of mGluR4 PAMs. Engers et al (Vanderbilt University) found from a high-throughput screening that there were a number of small arylamide compounds having mGluR4 PAM activity (Engers et al., 2009). They reported studies on SAR and in vitro and in vivo pharmacokinetic parameters in rat. The most potent compound in this series was 103 shown in Fig. 12. Researchers at Merck presented two new compounds, 104A and 105, with improved activity (Reynolds, 2008). phenylacetamide derivatives and found that 104B was the most potent (19.8 nM) mGluR4 positive allosteric modulator reported to date (Engers et al., 2010). Doller and co-workers (Lundbeck Research USA) have recently reported on a series of tricyclic thiazolopyrazole derivatives including compound 106, which was identified as a very potent and orally available compound with excellent brain penetration and good physicochemical properties (Hong et al., 2011). Fig. 13. Distribution of [ 11 C]methyl-PHCCC between 10-20 min after administration of radioligand in a control (1.2 mCi iv.) and PD (1.1 mCi iv.) rat brain. Coronal and axial views localize cortex at the level of S1 and S2 areas. It is noticeable that the accumulation of [ 11 C]methyl-PHCCC is enhanced in PD rat in the areas of subthalamic nucleus and spinal cord. The motor neurons in the ventral horn in the spinal cord express mGluR4 and the observed enhanced accumulation of mGluR4 ligand, [ 11 C]methyl-PHCCC is an indication of excess glutamate. This is the first time, when this aspect has been demonstrated in vivo in a PD model.

Allosteric modulators for mGluR7
It is reported that mGluR7 is widely expressed in the central nervous system and is primarily located on presynaptic terminals in brain regions such as the hippocampus, amygdala, and locus coeruleus. Mitsukawa et al. developed the first selective allosteric agonist of mGluR7, AMN082 (107), which has an EC 50 value of 64-290 nM and it is brain penetrating (Mitsukawa, 2005). However, converting it to a PET tracer is not straightforward. Researchers of Banyu Pharmaceutical Co reported a series of isoxazolopyridone derivatives as allosteric mGluR7 antagonists (Suzuki et al., 2007b). Compound MDIP (108) that was identified by random screening displayed mGluR7 antagonistic activity (IC 50 = 20 nM) and had no detectable activity on other mGluRs at 1000 nM. However, MDIP showed poor metabolic stability (predicted F H : 34%) on rat hepatocyte assay and low aqueous solubility (0.17 µg/mL, pH 7.4). It is assumed that poor metabolic stability and low aqueous solubility may be due to its high lipophilicity (clogD 7.4 : 3.5). Recently, Nakamura et al. have identified some isoxazolopyridone derivatives with potent mGluR7 antagonistic activity and metabolic stability, in which MMPIP (109) with improved physicochemical properties and metabolic stability showed good oral bioavailability and brain penetrability in rats (Nakamura et al., 2010).

Allosteric modulators for mGluR8
Recently, AstraZeneca developed a positive allosteric modulator for mGluR8 (Duvoisin et al., 2010;Duvoisin et al., 2011). The compound AZ12216052 as injected into the amygdale, reduced measures of anxiety. There is no PET ligand available and AZ12216052 does not cross blood brain barrier.

Conclusion
Glutamate is an interesting transmitter since it can participate also on glutamate metabolism to be converted to glutamine and its function as a neurotransmitter can be investigated based on its receptor functions. To understand the diverse physiological effects of glutamate it is important to know molecular identity of mGluRs expressed in distinct subpopulations of neurons. For instance, group I mGluRs are coupled to phospholipase C and subsequent production of inositol triphosphates and induces intracellular calcium release in Purkinje cells and hippocampal CA1 neurons, but the same receptor types are also coupled to inhibition of voltage-dependent calcium channel in hippocampal neurons without intracellular diffusible messengers (Choi & Lovinger, 1996). Group II mGluRs can be coupled to inhibition of cyclic AMP cascade in neural and glial cells while they are also linked to rapid-onset regulation of various channels including calcium channels and G-protein. The group III mGluRs-mediated effect is inhibition of neurotransmission through suppression of presynaptic voltage-dependent calcium channels (Pekhletski et al., 1996). This basic functional information of mGluRs has been obtained with in situ hybridization, immunohistochemistry and ex vivo studies with tritium labeled antibodies. While ex vivo studies can provide accurate endpoint information in steady state, they cannot provide information of the active inhibitory or stimulating effects in the system or interplay with other systems. To obtain functional information in real time, the investigation has to be done by using in vivo imaging methods. However, a lack of specific agonists and antagonists has limited the precise characterization of the role of individual metabotropic glutamate receptors in glutamatergic neurotransmission and hampered progress in identifying the physiological and pathological roles of mGluRs in vivo.
Recently, the modern computational chemistry has opened a wide range of technical approaches to design and construct molecules for imaging and to simulate their molecular targets. This technology has been used to design molecules for tracking different mGluRs. Especially, approach of allosteric compounds relies on sophisticated design of threedimensional arrangement of the tracer molecules responsible for the biological activity. Pharmacophore models can be constructed based on known biological activity. Design of novel allosteric modulators is an iterative process where structure-activity relationship information generated in the biological assays guides how to make structural alternations towards the optimal compound. Recently several non-competitive structurally diverse mGluR ligands have been published. These ligands, positive, negative and neutral modulators, bind to the allosteric binding sites located in the seven strand transmembrane domain. Based on these modulators, a number of radiotracers useful for imaging specific metabotropic glutamate receptors have been developed and their in vivo biological properties have been characterized. Development of metabotropic glutamate receptor ligands will open a new perspective for molecular imaging. Modulation of receptor functions might be used as diagnostic tools as well as to follow progression/regression of neural diseases. Presently, three mGluR ligands have been used in human studies. They are developed as negative allosteric modulators for mGluR5. For example, concerning PD, the death of dopamine neurons in the substantia nigra pars compacta causes a loss of dopamine in the basal ganglia. Dopamine modulation of neurotransmission in the striatum and other basal ganglia structures is crucial to gate cortical and thalamic excitatory input through the direct and indirect pathways. By using in vivo PET imaging studies and [ 18 F]FPEB we have found an upregulation of mGluR5 expression following dopamine denervation in animal models of PD (Figures 7 & 13), which probably represents a local compensatory mechanism, directed to dampen an excessive excitability of striatopallidal neurons. Drugs targeting the mGluR5 might provide new approaches by selectively reducing glutamate transmission in the areas where it is abnormally enhanced. In addition, we and others have found enhanced mGluR5 expression in several brain areas related to the indirect pathway in models of L-DOPA induced dyskinesias and some studies have shown promising therapeutic results after using mGluR5 antagonists. In gut glutamate is the main energy source and its neurotransmission is conducted by vagal afferents. The gut expresses also mGlu5 receptors and we have localized them with [ 18 F]FPEB. This phenomenon has raised a hypothesis that gut-brain axis as well as interplay with dopamine transmission might contribute to obesity. Even mGlu2 receptors had the earliest interest as targets for drug development and Eli Lilly developed several potent ligands targeted to mGluR2 there is not yet any specific allosteric www.intechopen.com Imaging of Metabotropic Glutamate Receptors (mGluRs) 521 modulators available for imaging purposes of mGluR2 function. The earlier compounds were missing receptor selectivity and sensitivity for imaging purposes since sequence similarity at the orthosteric binding site to which endogenous agonists bind. Present application of glutamate transmission has evoked an active drug development especially to develop allosteric modulators for neurodegenerative disorders, pain and schizophrenia. It should be noted that these disorders are affected also by modulation of dopaminergic system supporting hypothesis of interplay of these powerful transmitter systems. Future pharmacological and imaging studies will show which specific ligands acting at individual receptor subtypes could be used as sensitive indicators for diagnostic imaging. Therefore, there is an urgent need for development of allosteric modulators as imaging ligands for different of mGluRs for human use.

Acknowledgement
This work was supported by the NIH grant NIBIB-EB12864 to A-LB.