Therapeutic Potential for Metabotropic Glutamate Receptor 7 Modulators in Cognitive Disorders

Abstract

Metabotropic glutamate receptor 7 (mGlu₇) is the most highly conserved and abundantly expressed mGlu receptor in the human brain. The presynaptic localization of mGlu₇, coupled with its low affinity for its endogenous agonist, glutamate, are features that contribute to the receptor’s role in modulating neuronal excitation and inhibition patterns, including long-term potentiation, in various brain regions. These characteristics suggest that mGlu₇ modulation may serve as a novel therapeutic strategy in disorders of cognitive dysfunction, including neurodevelopmental disorders that cause impairments in learning, memory, and attention. Primary mutations in the GRM7 gene have recently been identified as novel causes of neurodevelopmental disorders, and these patients exhibit profound intellectual and cognitive disability. Pharmacological tools, such as agonists, antagonists, and allosteric modulators, have been the mainstay for targeting mGlu₇ in its endogenous homodimeric form to probe effects of its function and modulation in disease models. However, recent research has identified diversity in dimerization, as well as trans-synaptic interacting proteins, that also play a role in mGlu₇ signaling and pharmacological properties. These novel findings represent exciting opportunities in the field of mGlu receptor drug discovery and highlight the importance of further understanding the functions of mGlu₇ in complex neurologic conditions at both the molecular and physiologic levels.

SIGNIFICANCE STATEMENT

Proper expression and function of mGlu₇ is essential for learning, attention, and memory formation at the molecular level within neural circuits. The pharmacological targeting of mGlu₇ is undergoing a paradigm shift by incorporating an understanding of receptor interaction with other cis- and trans- acting synaptic proteins, as well as various intracellular signaling pathways. Based upon these new findings, mGlu₇’s potential as a drug target in the treatment of cognitive disorders and learning impairments is primed for exploration.

Introduction

Structure and Function of mGlu₇

Of the eight known metabotropic glutamate (mGlu) receptors, mGlu₇ is the most widely distributed in the brain, as well as the most evolutionarily conserved (Flor et al., 1997; Niswender and Conn, 2010). mGlu₇ is a class C G-protein coupled receptor and a member of the group III metabotropic glutamate receptors, along with mGlu₄, mGlu₆, and mGlu₈ (Niswender and Conn, 2010). This group of mGlu receptors canonically signals through the G_i/o G-protein pathway to attenuate adenylate cyclase activity and regulate ion channel activity (Fig. 1) (Houamed et al., 1991; Tanabe et al., 1992; Okamoto et al., 1994; Niswender and Conn, 2010; Iacovelli et al., 2014). mGlu₇ is localized predominantly presynaptically within the brain and is activated by high μM to mM concentrations of the primary excitatory neurotransmitter of the central nervous system (CNS), glutamate (Okamoto et al., 1994). This localization is further consolidated to the active zone of the presynaptic terminal (Somogyi et al., 2003), where the protein organizes in dense clusters partially dependent on interactions with the transsynaptic adhesion proteins Extracellular leucine-rich repeat and fibronectin type III domain containing 1 and 2 (ELFN1/2), as well as intracellular scaffolding protein interacting with C kinase 1 (PICK1) (El Far et al., 2000; El Far and Betz, 2002; Perroy et al., 2002; Bertaso et al., 2008; Niswender and Conn, 2010; Tomioka et al., 2014; Stachniak et al., 2019; Park et al., 2020). Additionally, the localization and downstream signal transduction of mGlu₇ is dependent on a number of other cytosolic proteins, such as calmodulin (Nakajima et al., 1999; O’Connor et al., 1999; El Far and Betz, 2002) and Munc 13-1 (Martín et al., 2010). mGlu₇ exhibits widespread expression in the CNS but exhibits particularly high expression within the amygdala, hippocampus, and hypothalamus (Kinoshita et al., 1998; Fisher et al., 2018). The receptor is also detected at low levels in peripheral tissues in the reproductive system, endocrine tissues, and smooth muscle (Uhlén et al., 2015; Alam et al., 2020; Sjöstedt et al., 2020).

Fig. 1.

Cryo-electron microscopy image of mGlu₇ homodimer with accompanying downstream molecular pathways associated with receptor activation in response to high synaptic levels of glutamate. Structure PDB ID: 7EPC from Du et al. (2021). Structures of human mGlu₂ and mGlu₇ homo- and heterodimers (Du et al., 2021). Figure created with BioRender.com. CaMKII, calmodulin-dependent kinase II; CREB, cAMP response element binding protein; ERK, extracellular-regulated kinase; GIRK, G-protein coupled inwardly rectifying potassium channel; JNK, c-Jun N-terminal kinase; MEK, mitogen-activated protein kinase kinase; MEKK, mitogen-activated protein kinase kinase kinase; mGlu_7, metabotropic glutamate receptor 7; MKK4, mitogen-activated protein kinase kinase 4; MKK7, mitogen-activated protein kinase kinase 7; PICK1, protein interacting with C kinase 1; PKCα, protein kinase C alpha.

At the level of protein structure, mGlu₇ is embedded within the cell membrane by a 7-transmembrane domain and flanked by an N-terminal glutamate-binding domain, also known as the Venus flytrap domain, a cysteine-rich domain, and a C-terminal cytoplasmic domain (Pin et al., 2003; Ferraguti and Shigemoto, 2006; Du et al., 2021). The transmembrane domain and C-terminal domain regions are responsible for associating with heterotrimeric G-proteins for signal transduction, while the N-terminal Venus flytrap domain and cysteine-rich domain are responsible for ligand binding and also play roles in receptor dimerization (Niswender and Conn, 2010; Kniazeff et al., 2011; Levitz et al., 2016; Koehl et al., 2019; Fisher et al., 2021; Habrian et al., 2023). This dimerization of mGlu₇, as well as all mGlu receptors, is essential for proper signaling and necessary for trafficking from the endoplasmic reticulum to the plasma membrane (Conn and Pin, 1997; Kunishima et al., 2000; Jingami et al., 2003; Pin et al., 2003; Niswender and Conn, 2010). The mGlu₇ receptor monomer is also capable of heterodimerization with members of the group II or group III mGlu receptor family (Fig. 2) (Doumazane et al., 2011; Levitz et al., 2016; Du et al., 2021); heterodimerization has distinct effects on receptor pharmacology and can further alter the characteristics of intracellular signaling downstream of mGlu activation (Doumazane et al., 2011; Kammermeier, 2012; Yin and Niswender, 2014; Liu et al., 2017; Moreno Delgado et al., 2017; Habrian et al., 2019; Lee et al., 2020, 2023; Thibado et al., 2021; Xiang et al., 2021; Lin et al., 2022; Wang et al., 2023).

Fig. 2.

Currently known mGlu heterodimers containing mGlu₇ and the functional/pharmacological effects of heterodimerization in comparison with mGlu_7/7 homodimer activity. mGlu_2/7 and mGlu_3/7 heterodimers have shown pharmacological changes in ligand binding, as well as functional changes in response to ligands compared with mGlu₇ homodimers (Du et al., 2021; McCullock and Kammermeier, 2021; Habrian et al., 2023). Changes in mGlu₇ allosteric modulator function have also been reported at the Schaffer collateral–Cornu Ammonis 1 synapse region of the hippocampus that are consistent with the expression of an mGlu_7/8 heterodimer (Lin et al., 2022). Figure created with BioRender.com

While downstream signaling increases in complexity in the context of mGlu₇ heterodimers, canonical mGlu₇ signaling is primarily mediated by activity of G_i/o G-proteins (Fig. 1) (Okamoto et al., 1994). Both G_α and βγ subunits play distinct roles in mGlu₇-mediated synaptic release control, which work together to modulate neurotransmitter release. For example, G_α activation leads to adenylate cyclase inhibition, reduction in cAMP levels, and reduced protein kinase A (PKA) activity (Millán et al., 2002; Pelkey et al., 2008). While it is known that cAMP levels and PKA activation are important for synaptic release control (Trudeau et al., 1996; Kaneko and Takahashi, 2004; Nagy et al., 2004; Cho et al., 2015), the function of these effectors in modulating neurotransmitter release via mGlu₇ activation is not fully understood. cAMP reduction contributes to decreases in presynaptic calcium ion (Ca²⁺) flux, along with essential functions of the βγ subunits in inhibiting presynaptic Ca²⁺ channels and activating protein kinase C (PKC) (Perroy et al., 2000; Martín et al., 2007, 2010, 2011). This inhibition of Ca²⁺ flux is also dependent on mGlu₇’s interaction with PICK1 (Perroy et al., 2002; Bertaso et al., 2008). While it has been shown that this reduction in Ca²⁺ flux is capable of reducing synapsin-1 phosphorylation, modulating synaptic release (Hay et al., 2000), it is also hypothesized that mGlu₇ activation reduces phosphorylation of synapsin-1 through a cAMP/PKA-dependent mechanism, which is also predicted to reduce neurotransmitter release (Fig. 1) (Hay et al., 2000; Patzke et al., 2019, 2021). Additionally, there is evidence to suggest that mGlu₇ can directly regulate synaptic release dynamics through translocation of Munc 13-1, an essential component of synaptic release machinery (Martín et al., 2010). This variety of direct and indirect mechanisms could give mGlu₇ considerable control over synaptic function depending upon its agonist-mediated or constitutive activation status.

βγ-mediated mGlu₇ signaling also activates potassium efflux through G protein-coupled inwardly rectifying potassium channels (Saugstad et al., 1996; Niswender et al., 2010), causing hyperpolarization of the presynaptic terminal to provide an additional mechanism of neurotransmitter release control (Fig. 1). Although activation of mGlu₇ is generally associated with depression of synaptic release (Gigg et al., 1994; Okamoto et al., 1994; Saugstad et al., 1996; Schoepp, 2001; Pelkey et al., 2005; Martín et al., 2007; Summa et al., 2013; Klar et al., 2015; Gogliotti et al., 2017; Girard et al., 2019), its localization and stimulation patterns can determine the type of plasticity induced after its activation. As an autoreceptor, short-term activation of mGlu₇ leads to synaptic depression, while repeated mGlu₇ stimulation at the presynapse can lead to synaptic potentiation (Pelkey et al., 2005, 2008). Additionally, mGlu₇ functions as a heteroreceptor on γ-aminobutyric acid (GABA)-releasing interneurons of the hippocampus and amygdala, and its activation at these locations is essential for neuronal plasticity (Gee et al., 2014; Klar et al., 2015). Activation of mGlu₇ at these heteroreceptor locations reduces GABA-mediated depression of excitatory neurons, allowing for enhanced plasticity and activity of local excitatory neurons and synapses (Gee et al., 2014; Klar et al., 2015). This function is directly contrary to mGlu₇’s canonical role when expressed at excitatory presynapses, which represses continued glutamate release.

In addition to these mechanisms, mGlu₇-expressing synapses can also be controlled via noncanonical signaling. For example, orthosteric activation of mGlu₇ can induce signaling through β-arrestins 1 and 2 following receptor phosphorylation by G-protein coupled receptor kinases (GRKs). These proteins activate (β-arrestin 1) or inhibit (β-arrestin 2) the mitogen-associated protein kinase cascade, respectively (Fig. 1) (Iacovelli et al., 2014). Evidence also suggests that mGlu₇ exhibits constitutive activity that contributes to the regulation of tonic neurotransmitter levels (Kammermeier, 2015), and mGlu₇ has been shown to potentiate glutamate release upon prolonged stimulation (Martín et al., 2010, 2011). These extensive intracellular signaling pathways highlight the variety and complexity of mechanisms by which mGlu₇ can modulate synaptic release and provide evidence that these mechanisms may play a role in vital cognitive processes.

mGlu₇ is encoded by the GRM7 gene, which is located at Chr3p26 and consists of 880,419 base pairs and 10 exons (O’Leary et al., 2016). GRM7 has many splice variants containing 3′ alterations to the mRNA (Schulz et al., 2002), but the main isoforms expressed in the human brain are splice variants 1 and 2, encoding mGlu_7a and mGlu_7b, respectively (O’Leary et al., 2016). mGlu_7b mRNA includes a 92-nucleotide insertion at the 3′ end of the gene, which translates to a 7 amino acid addition at the C-terminus of the protein (Flor et al., 1997). Although mGlu_7a and mGlu_7b have slightly different temporal and spatial expression patterns (Flor et al., 1997; Monn et al., 1997; Linden et al., 2004), to date, this insertion has not been observed to confer any functional changes to the mGlu₇ receptor. This extensive gene also contains numerous, clinically significant single nucleotide polymorphism (SNP) or mutation sites within both introns and exons (O’Leary et al., 2016). These nucleotide alterations are often correlated with neurodevelopmental disorders, psychiatric disorders, or learning disabilities, even when present in gene introns where it is not clear how/if they affect protein sequence or structure (Elia et al., 2011; Park et al., 2013; Tomioka et al., 2014; Liu et al., 2015; Bedogni et al., 2016; Charng et al., 2016; Reuter et al., 2017; Dark et al., 2018; Noroozi et al., 2019; Marafi et al., 2020; Fisher et al., 2021; Zhang et al., 2021; Song et al., 2021; Chaumette et al., 2022; Januel et al., 2024).

Cognitive Disorders and Dysfunction

Cognitive processes are applications of attention, memory, and learning and the ability to call upon formed memories in the future (Millan et al., 2012). Control of cognitive function arises from a variety of sources at the molecular, cellular, and histologic levels within the brain. This is driven by various mechanisms including phosphorylation imbalance, redox states, and neurotransmitter release/tone (Cui et al., 2008; Giralt et al., 2012; Yabuki et al., 2014; Kaur and Sharma, 2022). Cognitive dysfunction (used interchangeably with the terms cognitive deficit or cognitive impairment) encompasses impacts on cognitive processes associated with underlying conditions such as psychiatric disorders, neurodevelopmental disorders, neurodegenerative diseases, and aging. These conditions and their effects on cognition have been reviewed on numerous occasions (Levy, 1994; Gräff and Mansuy, 2009; Giralt et al., 2012; Millan et al., 2012; Yabuki et al., 2014; Kim et al., 2015; Kaur and Sharma, 2022; Dhakal and Bobrin, 2023; Hoglund et al., 2023). Cognitive disorders differ from cognitive dysfunction in that they are standalone diagnoses unassociated with an underlying condition (Dhakal and Bobrin, 2023). The current International Classification of Diseases (ICD) codes identify various levels of cognitive disorder severity, each categorized with its own respective code (https://www.cdc.gov/nchs/icd/icd10cm_browsertool.htm). However, there is often difficulty in translation of these codes internationally, as each nation adopts and revises them to fit their own medical system (Jetté et al., 2010). Despite these differences, patients diagnosed with cognitive disorders often have variable or unknown origins of the disorders, which complicates treatment strategies (Srivastava and Schwartz, 2014). mGlu₇ has been directly associated with effects on cognition in mouse models (Bushell et al., 2002; Hölscher et al., 2004; O’Connor et al., 2010; Fisher et al., 2020), and, as mentioned, SNPs or mutations in the GRM7 gene result in intellectual disability and developmental delay in humans. These observations suggest that the receptor is an underexploited target in the treatment of these disorders. Here we focus on how mGlu₇ specifically modulates cognition.

Current State of the Field

Models of mGlu₇-Related Cognitive Impairments

The consequences of knockout of the Grm7 locus have been studied in animal models for their effects on cognitive function both alone as well as in the context of other phenotypes related to various disease states. Grm7^−/− mice exhibit diminished anxiety behaviors, as well as deficits in associative fear learning (Masugi et al., 1999; Sansig et al., 2001; Bushell et al., 2002; Fisher et al., 2020), suggesting that abnormalities induced by Grm7 deletion stem from receptor activity in both the hippocampus and amygdala. Attenuation of mGlu₇ function or expression also alters behavioral and electrochemical function in the thalamocortical network (Tassin et al., 2016). Reduced mGlu₇ expression during mouse development increases phospho-cAMP response element binding protein (CREB) and Yes-associated protein (YAP) levels, leading to increased neural progenitor cell proliferation and reduced terminal differentiation (Xia et al., 2015). Mutations in the mouse Grm7 gene have also been reported to reduce the GABA-synthesizing enzymes Gad65 and Gad67 and increase reelin expression in the hippocampal formation (Wierońska et al., 2010a), which suggests dysfunctional GABA signaling and increased synaptic connectivity in these animals (Sansig et al., 2001; Bushell et al., 2002; Gogliotti et al., 2017; Fisher et al., 2020, 2021). Interestingly, proper mGlu₇ expression and intracellular protein interactions have been shown to be essential in preventing epileptic activity in the brain of mice (El Far et al., 2000; Perroy et al., 2002; Bertaso et al., 2008; Zhang et al., 2008; Girard et al., 2019), suggesting a mechanism for mGlu₇ dysregulation in other contexts that display severe abnormalities in neuronal excitability. Mouse models in which mGlu₇ function is altered by disrupting protein-protein interactions with Pick1 and Elfn1 also display deficits in cognitive domains (Bertaso et al., 2008; Zhang et al., 2008; Tomioka et al., 2014).

While the face validity of Grm7 knockout models is strong, these models lack construct validity regarding GRM7 mutations seen in patient populations. A clinically relevant Grm7 mutation induced in an exon sequence in rodents results in reduced surface expression of mGlu₇, as well as diminished axon outgrowth and presynaptic terminal development (Fisher et al., 2021; Song et al., 2021). This mutation (I154T) appears to cause severe developmental delay and other neurologic phenotypes via a reduction in receptor dimerization (Fisher et al., 2021). This decreased dimerization results in reduced surface expression of mGlu₇ receptor protein, although expression levels of Grm7 mRNA remain normal (Fisher et al., 2021). There have been other GRM7 mutations associated with neurodevelopmental disorders as well, such as R622Q, R658W, and T675K, which have been examined for neuronal morphology and function (Song et al., 2021). Of these observed mutations, a dual mutant R658W/T675K mGlu₇ receptor has been tested in vitro using isolated rat neurons, and these mutations cause degradation of the mGlu₇ protein as well as diminished axon outgrowth that is not able to be rescued by cell treatment with mGlu₇ agonists (Song et al., 2021).

Other translational models, such as mouse models of autism spectrum disorder (ASD; associated with numerous cognitive deficits in humans), exhibit reduced mGlu₇ expression in the prefrontal cortex, which can be ameliorated by virally increasing mGlu₇ expression in this area (Wang et al., 2021). Furthermore, patients with the neurodevelopmental disorder Rett syndrome (RTT) express reduced levels of mGlu₇ in the brain, and mice modeling RTT also express low levels of mGlu₇ protein (Bedogni et al., 2016; Gogliotti et al., 2017; Freitas and Niswender, 2023). An additional nongenetic model of pain-induced cognitive deficits has also been used to assess the effects of mGlu₇ on cognition, as well as algesia (Palazzo et al., 2015). Interestingly, this study identified dramatic changes in mGlu₇ expression patterns in response to chronic pain, as well as analgesic, anxiolytic, and procognitive effects of an mGlu₇ negative allosteric modulator. These mouse models are useful in probing the functional and behavioral effects of Grm7 coding mutations, neurodevelopmental-related dysfunctions, and changes in mGlu₇ expression or signaling in response to external effects. However, mice have limited ability to recapitulate models of intronic SNPs in the human GRM7 gene, which have been correlated with cognitive impairments (Park et al., 2013, 2014; Chaumette et al., 2022).

Clinical Examples of mGlu₇-Related Cognitive Impairments

GRM7 SNPs Associated with Cognitive Impairments

Various clinical studies have identified GRM7 SNPs and correlated their possible connection to cognitive deficits and disorders (Park et al., 2013; Yang and Pan, 2013; Liu et al., 2015; Charng et al., 2016; Reuter et al., 2017; Marafi et al., 2020; Zhang et al., 2021; Chaumette et al., 2022; Januel et al., 2024). SNPs rs9826579 and rs3792452 have been associated with attention deficit hyperactivity disorder (ADHD) in Chinese and Korean sample populations, respectively (Park et al., 2013; Zhang et al., 2021), and an association of rs3792452 with methylphenidate responses in ADHD patients has also been reported (Park et al., 2014). While the mechanisms of how these SNPs contribute to the development of ADHD have yet to be fully elucidated, it is hypothesized that they alter mGlu₇ expression, dysregulating the balance of GABA and glutamate within the brain (Zhang et al., 2021). Novel de novo GRM7 intronic SNPs, chromosomal deletions, and coding mutations have also been correlated with ASD and developmental delay in patient populations (Yang and Pan, 2013; Liu et al., 2015; Fisher et al., 2018; Marafi et al., 2020). The intronic SNP rs1396409 was also recently shown to be associated with cognitive deficits in schizophrenia patients early in their diagnosis (Chaumette et al., 2022). Marafi et al. conducted an extensive study of GRM7 variants associated with neurodevelopmental disorders, identifying three novel variations, as well as characterization of the clinical phenotypes associated with 11 affected individuals (Marafi et al., 2020). All affected patients suffered from developmental delay, early-onset epilepsy, and microcephaly, with a majority of cases also presenting with cerebral atrophy and hypomyelination as revealed by neuroimaging. Recently, five additional patients have been reported with similar phenotypes (Januel et al., 2024).

RTT

RTT is caused by pathogenic mutations in the methyl CpG-binding protein 2 (MECP2) gene and is characterized by a period of normal development for approximately 6 months after birth, followed by developmental regression accompanied by numerous cognitive, behavioral, and physical impairments (Amir et al., 1999; Neul et al., 2010). As a transcriptional modulator through binding of methylated DNA, MeCP2 regulates the expression of numerous genes (Chahrour et al., 2008). mGlu₇ receptor levels are reduced in RTT patient brain samples, and Grm7 transcript is reduced in Mecp2^+/− mouse models of RTT (Bedogni et al., 2016; Gogliotti et al., 2017; Freitas and Niswender, 2023). Our laboratory has shown that, in both male (Mecp2^−/y) and female (Mecp2^+/−) mice, hippocampal long-term potentiation (LTP) deficits, behavioral abnormalities, and cognitive deficits can be corrected with mGlu₇ positive allosteric modulators (Gogliotti et al., 2017). These findings directly implicate mGlu₇ dysregulation in the cognitive pathophysiology of RTT and demonstrate evidence that potentiation or activation of mGlu₇ could be used to rescue these deficits.

ADHD

ADHD is a common learning disability in children, with an incidence of approximately 5% (Polanczyk et al., 2007). This disorder results in reduced attention, as well as hyperlocomotion in some individuals (Dark et al., 2018); these phenotypes reduce children’s effectiveness in school and can greatly impact learning if left untreated during childhood (Pfiffner and Haack, 2014). Genome-wide association studies have established a connection between GRM7 expression and ADHD incidence in the human population (Elia et al., 2011), which is hypothesized to be connected to mGlu₇’s role in synaptic plasticity as an essential part of memory and learning (Hölscher et al., 2004; Klar et al., 2015; Bushell et al., 2002). There is also evidence to implicate the ELFN1 gene in ADHD development, as Elfn1 knockout model mice exhibit seizures and hyperactivity phenotypes and mutations in ELFN1 are correlated with ADHD in clinical populations (Tomioka et al., 2014). In mice, Elfn1 plays a major role in recruitment of mGlu₇ to the presynaptic active zone (Park et al., 2020; Matsunaga and Aruga, 2021). The similar phenotypes seen in mice and humans with either GRM7 or ELFN1 mutations suggest a crucial level of regulation imparted by the mGlu₇/Elfn1 transynaptic interaction. Elfn1 has also been shown to act as an allosteric regulator of group III mGlu receptor activity in vitro (Dunn et al., 2018; Stachniak et al., 2019); therefore, loss of Elfn1, dysregulation of mGlu₇ activity, or a combination of these factors could play an essential role in ADHD. Further, it will be important to determine if the interactions of Elfn1 with mGlu₇ in vivo complicate the pharmacology of mGlu₇ ligands when the two are coexpressed, not only in ADHD but in any scenario where mGlu₇ is a potential drug target.

Current Research Areas and Future Directions

Receptor Dimerization and Localization

As previously mentioned, receptor dimerization of mGlu₇ is essential for surface expression and function (Pin et al., 2003; Habrian et al., 2019; Lee et al., 2020, 2023; Du et al., 2021; Fisher et al., 2021; Thibado et al., 2021; Lin et al., 2022; Wang et al., 2023). The critical nature of dimerization for mGlu₇ is evident from our findings that the mGlu₇ I154T point mutation, one of the first mutations observed in patients (Charng et al., 2016; Marafi et al., 2020; Fisher et al., 2021), results in reduced surface expression in vitro due to impaired dimerization and almost no receptor expression in mice (Fisher et al., 2021). These effects were studied solely in the context of mGlu₇ receptor homodimerization. As recent research has begun to elucidate patterns of heterodimerization of mGlu₇ and its effects on receptor function and pharmacology (Yin and Niswender, 2014; Liu et al., 2017; Habrian et al., 2019; Du et al., 2021; McCullock and Kammermeier, 2021; Xiang et al., 2021; Lin et al., 2022), it will be interesting to revisit the effects of various clinical mGlu₇ mutations on heterodimerization of the receptor with other mGlus.

The ability of mGlu receptors to heterodimerize was first characterized by the Pin laboratory in 2011 (Doumazane et al., 2011), where it was determined that group I receptors could heterodimerize within their group (mGlu₁ and mGlu₅), while group II and group III mGlu receptors could heterodimerize between groups. These heterodimers have high biologic relevance, as they have been physically (Meng et al., 2022) and pharmacologically (Yin and Niswender, 2014; Moreno Delgado et al., 2017; Habrian et al., 2019, 2023; Xiang et al., 2021; Lin et al., 2022) observed throughout the brain. These patterns of heterodimerization also have implications for further therapeutic development, as they confer pharmacological changes to the receptors compared with their homodimeric counterparts (Fig. 2). For example, mGlu_2/7 heterodimers have been observed in the hippocampus (Habrian et al., 2019), and these receptors display cooperativity that is not present in homodimers. Additionally, L-2-amino-4-phosphonobutyric acid (L-AP4), a group III mGlu receptor agonist, exhibits threefold higher efficacy at an mGlu_2/7 heterodimer when compared with an mGlu_7/7 homodimer (Du et al., 2021). An mGlu₇ negative allosteric modulator (NAM) can even restrict downstream signaling of heterodimer activation in the presence of an mGlu₂ positive allosteric modulator (PAM) at mGlu_2/7 heterodimers when treated with group 2-specific agonist LY354740 (Du et al., 2021). These findings suggest there is biased association of G-proteins to heterodimer subunits. Separate studies have also concluded that mGlu_3/7 heterodimers have higher basal signaling and increased activation kinetics compared with either mGlu_3/3 or mGlu_7/7 homodimers, and this activity is sensitive to orthosteric antagonists (Habrian et al., 2019; Kukaj et al., 2023). Recently, we identified a signature consistent with mGlu_7/8 heterodimers at Schaffer collateral-CA1 (SC-CA1) synapses of the hippocampus that segregates compounds into those that block mGlu_7/8 heterodimers and those that do not (Lin et al., 2022), providing additional challenges in understanding mGlu₇ pharmacology and drug discovery efforts to target this receptor therapeutically.

There are also numerous post-translational modifications recently identified that contribute to mGlu₇ surface expression and synaptic localization. For example, N-linked glycosylation at 4 asparagine residues of mGlu₇ is essential for its release from the endoplasmic reticulum and further trafficking to the cell membrane (Park et al., 2020). Once expressed in the outer cell membrane, the activities of neuronal precursor cell-expressed developmentally downregulated 4 (Nedd4) E3 ubiquitin ligase, as well as β-arrestins, are essential for endocytosis and degradation of the receptor, both tonically and in response to agonist stimulation (Lee et al., 2019). Recent advances have been made in understanding active zone clustering of mGlu₇, which depends on both cis- and trans-acting mechanisms (Boudin et al., 2000; Suh et al., 2008; Tomioka et al., 2014; Lee et al., 2019; Park et al., 2020; Kang et al., 2021). As discussed earlier, Elfn1 functions trans-synaptically to facilitate mGlu₇ clustering, as well as modulate its pharmacology and inducing constitutive activity (Stachniak et al., 2019). This localization provided by Elfn1 is exceptionally important considering mGlu₇’s low affinity for glutamate (Okamoto et al., 1994; Niswender and Conn, 2010). Interestingly, the trans-synaptic interaction of Elfn1 with mGlu₇ enhances constitutive activity of mGlu₇, reducing synaptic glutamate release (Stachniak et al., 2019). This anchoring function of Elfn1 has been observed to occur with various group III mGlu receptors, reducing cAMP concentrations and modifying agonist pharmacology, as well as increasing constitutive activity of these receptors (Dunn et al., 2018). These dimerization and localization mechanisms are essential to mGlu₇ surface expression and function, implicating numerous proteins as alternative mechanisms potentially modulating mGlu₇ signaling in various disease states.

Intracellular Signaling Pathways

It has recently been elucidated that neuromodulatory and retrograde synaptic signaling in the CNS and peripheral nervous system is capable of increasing phosphorylation of synapsin-1 through cAMP-dependent pathways, directly modulating vesicular release potentials (Patzke et al., 2019; Polishchuk et al., 2023). These findings suggest that synapsin-1 phosphorylation could therefore be attenuated by reduction of cAMP via neuromodulatory signals (such as via mGlu₇ activation) as well. Coincidentally, mGlu₇ has been previously implicated in modulation of synapsin-1 phosphorylation; this study attributed this effect to reductions in Ca²⁺ channel activation, and likely reduction of CaMKII activity, although the specific effector was not identified (Hay et al., 2000). These findings suggest a potentially significant role of mGlu₇ downstream signaling in the direct modulation of synaptic release through synapsin-1 via two separate pathways. These mechanisms warrant further investigation and pharmacological exploitation, as synapsin-1 has been shown to play a role in learning, memory, and developmental deficits in animal models (Kao et al., 2002; Cui et al., 2008). Evidence suggests that mGlu₇ can also directly modulate synaptic release through interactions with Munc 13-1, a protein required for synaptic vesicle priming (Martín et al., 2010). These pathways, paired with mGlu₇’s known and extensive effects on various other cellular and synaptic regulatory mechanisms discussed previously, may help fully elucidate mGlu₇’s observed effects within the hippocampus in attenuating GABA release and facilitating LTP, which is essential at this location for memory formation (Klar et al., 2015). The impact of mGlu₇ activation on synaptic release machinery function is a novel area of research requiring further investigation. The use of models knocking out or manipulating various components of the synaptic release machinery, or studies employing mGlu₇ ligands to study specific aspects of synaptic release mechanisms, could provide new insight into how mGlu₇ activation has a direct effect on synaptic vesicle release.

Novel findings also suggest differential roles of β-arrestin-mediated signaling and internalization regarding the mGlu family (Lee et al., 2023), which have identified different mechanisms of internalization, recycling, degradation, and signaling based on C-terminal sequences of mGlu receptors. Findings also show that these effects are further modulated by the formation of mGlu heterodimers. While mGlu₇ was not extensively interrogated in this study, previous results indicate a major role of β-arrestins in mGlu₇ expression and function (Iacovelli et al., 2014; Lee et al., 2019). This suggests unknown β-arrestin-mediated modulation of mGlu₇ heterodimer function that could have implications on pharmacological activity at these receptors. Utilizing models to interrogate patterns of G protein-coupled receptor kinase phosphorylation and β-arrestin binding of mGlu₇ heterodimers would provide more insight into how stimulation patterns can affect surface presence and downstream signaling of these heterodimers and build upon current findings reported for other mGlu heterodimers (Lee et al., 2023), as well as mGlu₇ homodimers (Iacovelli et al., 2014). Elucidating these processes would increase understanding of how mGlu₇-targeting compounds exert their effects in cell-based assays as well as in in vivo models.

mGlu₇ Specific Ligands

The field of mGlu pharmacology relies heavily on the use of tool compounds to agonize, antagonize, and modulate the receptors to observe their physiologic function and validate their potential as drug targets (Fig. 3). Agonists have been used for decades to stimulate the mGlu receptors, with the prototypical group III mGlu receptor agonist being L-AP4 (Graham and Burgoyne, 1994; Schoepp et al., 1996; Frauli et al., 2007). A series of compounds was later developed to probe the effects of group III mGlus both in vitro and in vivo. LSP1-2111, and the related LSP1-3801, were the first of these compounds, displaying anxiolytic effects in mouse models (Acher et al., 2010, 2012; Wierońska et al., 2010b; Amalric et al., 2013; Cajina et al., 2013). These compounds were followed by LSP4-2022 and LSP3-2156 (Goudet et al., 2012; Podkowa et al., 2015) and, most recently, LSP2-9166 (Hajasova et al., 2018), which is selective for mGlu₄ and mGlu₇ compared with the broad group III activity of the previous iterations of LSP compounds as well as L-AP4. The previously mentioned agonists act orthosterically, which imparts difficulty in terms of group III mGlu targeting selectivity. Allosteric agonists have also been developed to increase this selectivity. An mGlu₇-specific allosteric agonist, AMN082, has been extensively investigated (Mitsukawa et al., 2005), although it undergoes rapid metabolism in vivo and exhibits significant off-target activity (Sukoff Rizzo et al., 2011). Two additional studies have recently reported allosteric agonists of the mGlu₇ receptor as well. Cid et al. developed a series of allosteric agonists utilizing computational modeling, which have comparable potencies to AMN082 (Cid et al., 2019). Dickson et al. have also reported an mGlu₇ allosteric agonist, CVN636, with high potency, selectivity, and favorable CNS penetrance (Dickson et al., 2023).

Fig. 3.

Metabotropic glutamate receptor group III and mGlu₇-selective tool compounds organized by their mode of pharmacology. All orthosteric agonists are active at all group III receptors, aside from LSP2-9166, which is active at mGlu₄ and mGlu₇ and only active at mGlu₈ at high concentrations (Acher et al., 2012; Amalric et al., 2013; Hajasova et al., 2018). AMN082 is a selective allosteric agonist (Mitsukawa et al., 2005), while CVN636 (Dickson et al., 2023) and the series of compounds published by Cid et al. (Cid et al., 2019) are also selective, with more favorable in vivo parameters. NAMs shown are active at mGlu₇ homodimers, with some displaying activity at mGlu_7/8 heterodimers as well (Reed et al., 2020; Lin et al., 2022). VU0155094 and VU0422288 are group III mGlu receptor-selective PAMs (Jalan-Sakrikar et al., 2014), VU6005649 is a dual mGlu₇/mGlu₈ PAM, and VU6027459 is an mGlu₇-selective PAM (Reed et al., 2020). VU6046980 is a highly mGlu₇-selective PAM derived from VU6027459 (Kalbfleisch et al., 2023). Figure created with Biorender.com.

The development of allosteric modulators has also provided new tools and therapeutic opportunities to target specific mGlu receptor subtypes. A breakthrough for the group III mGlu receptors was the report of the PAM, N-phenyl-7-(hydroxyamino) cyclopropa[b]chromen-1a-carboxamide, published in 2003 (Maj et al., 2003; Marino et al., 2003), which was the first reported modulator of mGlu₄. Two PAMs with activity at mGlu₇ were reported in 2014: VU0155094 and VU0422288 (Jalan-Sakrikar et al., 2014). However, these compounds are not selective for mGlu₇, with activity at mGlu₄ and mGlu₈ as well. Further compound screening and optimization has highlighted the difficulty of producing useful in vivo PAM tool compounds (Reed et al., 2020; Kalbfleisch et al., 2023). We recently reported several PAM ligands for mGlu₇: VU6027459 (Reed et al., 2020) and the optimized VU6046980 (Kalbfleisch et al., 2023). While these compounds are CNS penetrant and display the necessary pharmacokinetic properties to serve as tool compounds, the scaffold appears to exhibit an off-target sedation activity that is retained in Grm7^−/− mice (Kalbfleisch et al., 2023). These challenges in the development of PAMs suggest that further work is needed to identify selective mGlu₇ PAMs for preclinical use.

A NAM with specific mGlu₇ activity was reported in 2007 with the development of MMPIP (Suzuki et al., 2007). These early allosteric compounds laid the groundwork for the current field of subtype-specific mGlu allosteric modulators. ADX71743 was later reported as an mGlu₇-specific NAM, with characterization in vitro as well as in vivo (Kalinichev et al., 2013). We reported VU6010608, a structurally distinct NAM from either the MMPIP or ADX71743 scaffolds, which had potent mGlu₇ NAM activity but was metabolically labile (Reed et al., 2019a; Reed et al., 2019b). VU6012962, derived from VU6010608, displays favorable CNS penetrance and high potency (Reed et al., 2019b). VU6012962 was further optimized using iterative libraries of compounds replacing labile amide linkers with oxadiazole bioisosteres characterized through structure-activity relationships to identify VU6019278, a potent mGlu₇-specific NAM with improved pharmacokinetic properties for use in vivo (Reed et al., 2019a).

Context-Selective Ligands

Traditional investigation of small molecules acting on mGlu₇ has focused on the targeting of mGlu_7/7 homodimers. However, numerous mGlu heterodimers have been pharmacologically characterized, and these various combinations exhibit differential pharmacological responses when compared with homodimers (Kammermeier, 2012; Yin and Niswender, 2014; Wang et al., 2023). Considering the extensive expression of mGlu₇ across the brain, and the propensity of mGlu receptors to form heterodimers (Okamoto et al., 1994; Conn and Pin, 1997; Doumazane et al., 2011; Du et al., 2021; Lin et al., 2022; Meng et al., 2022; Habrian et al., 2023), tool compounds selective for specific dimer structures of mGlu₇ could be exceptionally helpful in elucidating the function of heterodimers versus homodimers, as well as their regional expression and specificity. Two identified mGlu₇ NAMs have been shown to attenuate receptor activation in vitro and block hippocampal LTP ex vivo regardless of dimerization with mGlu₈, while two similar compounds were shown to only act on mGlu_7/7 homodimers in vitro and have no effect on LTP measurements (Lin et al., 2022). These data suggest that allosteric modulators can be used to differentiate the function of mGlu₇ homodimers and heterodimers, allowing for a deeper understanding of complex heterodimer expression and function as well as the opportunity to specifically target therapeutic effects to unique receptor conformations with spatial selectivity.

Conclusions

mGlu receptors, including mGlu₇, have been well studied for decades as drug targets for numerous neurologic diseases. mGlu₇, as the most conserved and highly expressed of the family, poses unique opportunities (and challenges) as a drug target. Observed behavioral alterations in animal models, as well as in humans with knockout or loss-of-function mutations, have identified mGlu₇ function as essential in proper neural development and cognitive function, an area that is affected in the pathophysiology of a multitude of diseases yet has few efficacious treatments. The numerous tools for exploring mGlu₇ pharmacology have been useful in understanding its basic signaling characteristics and potential therapeutic effects, but there has not yet been an mGlu₇-targeting compound studied in the clinic. This is likely a result of our limited understanding of the wide expanse of complex interactions implicating mGlu₇ in vivo, an area we are just beginning to understand more in depth. Novel findings in these understudied areas, as well as an increased focus on translational research in the biologic sciences as a whole, create a unique opportunity for more selective, efficacious, and safe treatments to be developed at the bench and transitioned to the bedside where they can benefit patients suffering from debilitating cognitive conditions associated with mGlu₇ dysregulation.

Abbreviations

ADHD

attention deficit hyperactivity disorder

Ca²⁺

calcium ion

CNS

central nervous system

ELFN1/2

extracellular leucine-rich repeat and fibronectin type III domain-containing 1/2

GABA

γ-aminobutyric acid

L-AP4

L-2-amino-4-phosphonobutyric acid

LTP

long term potentiation

mGlu

metabotropic glutamate

NAM

negative allosteric modulator

PAM

positive allosteric modulator

PKA

protein kinase A

RTT

Rett syndrome

SNP

single-nucleotide polymorphism

Authorship Contributions

Wrote or contributed to the writing of the manuscript: Parent, Niswender.