Novel Pharmacological Targets for GABAergic Dysfunction in ADHD

1. Introduction

1.1. Attention-deficit/hyperactivity disorder (ADHD) prevalence, symptoms, and emerging treatment avenues

Attention-deficit/hyperactivity disorder (ADHD) is a heterogenous neurodevelopmental disorder that affects approximately 5% of the population worldwide (Faraone et al., 2015, Faraone et al., 2021). The disorder more commonly affects males than females and persists into adulthood in about half of all individuals diagnosed (Faraone et al., 2015). There are several genetic and environmental risk factors associated with ADHD, and the polygenic causation of the disorder is reflected by its high rates of comorbidity with other neurodevelopmental and psychiatric disorders (Hegvik et al., 2021). Hallmark symptoms include attentional deficits, impulsivity, and hyperactivity, which are commonly associated with neurophysiological abnormalities in the prefrontal cortex (PFC) and frontostriatal brain circuits (Arnsten, 2009; Bush et al., 2005; Castellanos & Proal, 2012; Hart et al., 2013). Stimulant medications are the most common pharmacological treatments prescribed for ADHD, and more recently developed non-stimulant medications targeting α_2A-adrenoceptors and norepinephrine transporters (NETs) most likely exert efficacy through enhancing PFC function via increased catecholamine levels (reviewed in Arnsten & Pliszka, 2011).

Stimulants remain the most effective front-line treatment for ADHD, with mixed evidence suggesting that adjunctive treatment with non-pharmacological interventions such as cognitive training and behavioral therapy are more effective compared to stimulant treatment alone (Cai et al., 2023; Cortese et al., 2015; Posner et al., 2020). However, stimulant treatments are not recommended for children aged 5 years and younger and approximately 30% of ADHD patients do not respond to stimulant treatments; progress towards developing pharmacological treatments beyond targeting dopaminergic and noradrenergic systems has also been stagnant (Faraone et al., 2015; Franke et al., 2018). Non-stimulant treatments, including the α_2A adrenoceptor receptor agonists guanfacine and clonidine and the selective norepinephrine reuptake inhibitor atomoxetine, are typically reserved for patients who are poor responders to stimulants or experience intolerable side effects (Posner et al., 2020). Relative to stimulants, however, non-stimulant medications have lower response rates (Schwartz & Correll, 2014; Posner et al., 2020). Therefore, exploring novel pharmacological targets for ADHD based on recent findings underpinning the biological mechanisms of ADHD disease etiology is essential for progressing and diversifying treatment options (Ugarte et al., 2023). More specifically, these treatments may cater to subpopulations of ADHD patients with specific comorbidities associated with overlapping biomarkers and diagnostic criteria, as well as provide alternative pharmacotherapies for patients whom stimulants and catecholamine-targeting treatments are ineffective for and/or confer intolerable side effects (Hegvik et al., 2021).

Genetic evidence from patients reveals that ADHD is associated with mutations in genes encoding for various neurotransmitter systems, transcription factors, synaptic adhesion molecules (SAMs) and microRNAs (Elia et al., 2023; Cabana-Domínguez et al., 2022; Demontis et al., 2023; Demontis et al., 2019; Elia et al., 2010; Glessner et al., 2023; Kessi et al., 2022; Sharp et al., 2009), and there is a plethora of pharmacological targets beyond neurotransmitter systems that may exert therapeutic efficacy for treating ADHD based on human genetic evidence as well as biological evidence from ADHD preclinical models. Due to this broad heterogeneity and polygenicity, designing novel pharmacological treatments for ADHD may not be a “one-size fits all” approach. Rather, new medicines may be needed to treat subpopulations of ADHD patients with specific symptom clusters and/or comorbidities reflected by biomarkers associated with individual neurotransmitter systems, proteins, genes, or brain regions (Michelini et al., 2022). Considering the prominent role of excitatory/inhibitory (E/I) balance in maintaining healthy brain function during neurodevelopment, pharmacologically targeting receptors associated with GABAergic and glutamatergic pathways underlying E/I imbalance in ADHD may provide patients with precision medicine care and correct detrimental developmental trajectories in ADHD (Mamiya et al., 2021). Specifically, highly druggable G protein-coupled receptors (GPCRs) involved in regulating E/I balance and synaptic plasticity during development have emerged as promising therapeutic targets for a wide variety of central nervous system (CNS) disorders (Conn et al., 2009; Selten et al., 2018). By selectively modulating receptor subtypes associated with specific ADHD symptoms, these compounds would modify synaptic activity in a subtle fashion to reduce negative side effects compared to compounds that fully inhibit or potentiate pathways that alter normal brain function.

Emerging evidence has highlighted the neurotransmitter Gamma-Aminobutyric Acid (GABA) as a key neurotransmitter implicated in ADHD pathophysiology (Figure 1) (Edden et al., 2012; Puts et al., 2020). While preclinical investigations into how disruptions in GABAergic signaling underly abnormalities in genetic and behavioral models of ADHD have been limited, examining GABA levels in ADHD patient populations reveals that this neurotransmitter plays an underappreciated role in ADHD pathophysiology (discussed in section 2). Currently, it is unclear whether ADHD pathophysiology related to abnormalities in the dopaminergic system is preceded by or contribute to disruptions in GABAergic signaling in early development related to E/I deficits in synaptic plasticity (Castellanos & Proal, 2012). Therefore, exploiting pharmacological targets underlying disruptions in GABAergic signaling pathways related to cortical dysfunction in ADHD represents a critical and currently overlooked strategy for developing novel therapies for subpopulations of ADHD patients.

Figure 1:

Overview of ADHD pathophysiology related to GABAergic and cortical dysfunction. Reduced GABA concentrations in the cortex and striatum of ADHD have been reported using MRS (top box). Several genes identified in genome-wide association studies (GWAS) linked to GABAergic function are associated with ADHD (middle box). Altered GABA concentrations have been reported in the anterior cingulate cortex (ACC) and prefrontal cortex (PFC) during behavioral tasks related to ADHD symptoms of impulsivity and inattention (bottom box).

2. Reduced GABA signaling in PFC associated with ADHD in clinical populations

2.1.1. Genetic evidence of GABA dysfunction in ADHD

Genetic evidence linking the GABAergic system to ADHD is revealed by single nucleotide polymorphisms (SNPs) in the glutamic acid decarboxylase (GAD1) gene, which encodes a cortical enzyme involved in GABA biosynthesis. Specifically, GAD1 SNPs rs3749034 and rs11542313 are associated with increased susceptibility to ADHD and likely contribute within the hyperactive/impulsive domain (Bruxel et al., 2016). Additional studies identifying SNPs in genes associated with ADHD have been indirectly linked to GABA function. For example, SNPs in the COMT gene, encoding for catechol-o-methyltransferase, an enzyme that plays a key role in regulating synaptic dopamine in the cortex, is associated with ADHD in humans (Lasky-Su et al., 2008), and SNPs both in GAD1 and COMT are associated with modulation of GABA levels in the anterior cingulate cortex (ACC) of healthy volunteers (Marenco et al., 2010). Furthermore, mutations in the CDH13 gene, encoding for cadherin-13, are a risk factor for ADHD (Lesch et al., 2008; Neale et al., 2008, 2010; K. Zhou et al., 2008). Cadherin-13, a member of the cadherin family of cell adhesion molecules, disrupts GABAergic interneuron function in the mouse hippocampus, leading to cognitive deficits and altered E/I balance (Rivero et al., 2015); in humans, CDH13 mutations are associated with cognitive deficits in working memory in adults with ADHD (Arias-Vásquez et al., 2011; Ziegler et al., 2021). While several studies reveal that genetic variants in ADHD patients are directly or indirectly linked to GABAergic function, more studies are required to strengthen the association between GABAergic gene sets and ADHD symptom severity (Naaijen et al., 2017).

2.1.2. Altered basal GABA levels in select brain regions of ADHD patients

Neuroimaging studies reveal that basal GABA levels are reduced in specific brain regions of ADHD patients (Edden et al., 2012; Puts et al., 2020). Using magnetic resonance spectroscopy (MRS) at 3 Tesla, lower GABA concentrations have been reported in the primary somatosensory and motor cortices of ADHD children relative to age-matched (8–12 years) neurotypical controls (Edden et al., 2012). Additionally, GABA levels are reduced in the striatum of ADHD children aged 5–9 years relative to control subjects using MRS at 7 Tesla and reduced striatal GABA levels correlate with poor inhibitory control in behavioral testing (Puts et al., 2020). However, this study did not report basal differences in GABA levels in the ACC, dorsolateral PFC (dlPFC), or premotor cortex. In another study, GABA levels were found to be significantly lower in the ACC of adult female patients with ADHD, but not patients with bipolar disorder (BPD), and low ACC GABA levels were associated with increased inattention (Ende et al., 2016). Additionally, self-reported measures of impulsivity and aggression in this study negatively correlated with ACC GABA levels. Interestingly, one study found that higher GABA concentrations correlated with ADHD in adults; however, no differences in GABA concentrations were reported between ADHD children relative to control subjects (Bollmann et al., 2015). Importantly, low GABA concentrations in ADHD patients may be specific to subregions of the cortex (Sumner et al., 2010), which may relate to how expression patterns of proteins associated with ADHD and GABA function differ among brain regions.

Increased glutamate/glutamine (Glx) concentrations have also been reported in the striatum and cortical brain regions of ADHD patients (Carrey et al., 2007; Courvoisie et al., 2004; MacMaster et al., 2003; Moore et al., 2006), suggesting that deficits in inhibitory control in ADHD are mediated through disrupted E/I balance resulting from a lack of GABAergic and/or increased glutamatergic transmission. While glutamate concentrations in the brain cannot be distinguished from glutamine using lower resolution MRS scanners (<3 Tesla), abnormalities in brain Glx concentrations may provide insight into how glutamate metabolism is altered in ADHD patients (Maltezos et al., 2014). Future studies using MRS with higher field strength should be employed to understand how the glutamate:glutamine ratios in specific brain regions reflect alterations in the rate of glutamate synthesis and turnover in ADHD patients, as well as the potential contributing role of GABA-glutamine cycling (Bak et al., 2006; Albrecht et al., 2010). Although reports using MRS to measure brain GABA levels in ADHD are limited, MRS is a powerful technique that may be used in future studies to resolve discrepancies and better understand the role basal GABA concentrations, as well as the glutamate/GABA–glutamine cycle, in select brain regions are associated with ADHD diagnosis in both children and adults (Schür et al., 2016).

2.1.3. Abnormal brain GABA concentrations associated with ADHD behavioral phenotypes

Despite limited clinical literature investigating altered brain GABA levels in ADHD patients, it has been proposed that GABA plays a significant role in impulsivity and cognitive symptoms of ADHD (see Hayes et al., 2014). While some studies report no significant association between basal brain GABA levels in ADHD patients (Mamiya et al., 2022; Schür et al., 2016), smaller increases in GABA concentrations in the ACC correlate with impaired attentional control in ADHD patients measured using the Stroop task relative to control subjects (Mamiya et al., 2022). These correlations are not confounded by differences in GABA concentrations associated with percentage of gray matter in the frontal cortex, and this effect is specific to functional magnetic resonance imaging (fMRI) voxels in the frontal lobe, as no differences in GABA concentrations were detected in the visual cortex during the task. Impaired attentional alternating and executive function in ADHD children (8–12 years), measured using the Attention Network Test (see Fan et al., 2002), were found to be consistent with lower levels of ACC activation (Konrad et al., 2006). Importantly, lower GABA levels in the ACC correlated with impairments in response inhibition and greater impulsivity in a Go/No-Go behavioral task in adolescent male subjects (Silveri et al., 2013), providing direct evidence for lower GABA levels in the ACC associated with ADHD-related behaviors (Bezdjian et al., 2009). Additional evidence linking increases in cortical GABA concentrations to enhanced cognitive performance has been demonstrated in healthy adults and increases in GABA levels in the dlPFC and ACC are associated with improved performance in working memory and attentional tasks in healthy individuals (Bezalel et al., 2019; Yoon et al., 2016). Furthermore, reduced GABA levels are consistent with impaired inhibitory control in adults (Hermans et al., 2018), highlighting the importance of identifying pharmacological targets that rescue deficits in cortical GABAergic signaling associated with impulsive behavior (Ajram et al., 2017). Although one study found that GABA levels in the dlPFC positively correlate with rash impulsivity in healthy men (Boy et al., 2011), GABA levels in the ACC did not significantly predict behavioral traits of impulsivity in this study. Nonetheless, differences among studies showing that either high or low brain GABA levels are correlated with ADHD behavioral phenotypes highlight the importance of examining individual brain regions at baseline as well as during behavioral tasks to elucidate the potential bidirectional influence of GABA on ADHD diagnosis and behavioral phenotypes (Mamiya et al., 2021).

Treatment with methylphenidate, a commonly prescribed stimulant treatment for ADHD, beginning in childhood is associated with long-lasting decreases in GABA levels in the medial PFC (mPFC) persisting into adulthood, compared with methylphenidate treatment starting in adulthood which produces no differences in GABA levels (Solleveld et al., 2017). While the age-related decline in cortical GABA levels has been established in healthy adults (Ferguson & Gao, 2018), the precise implications of these long-lasting adaptations in GABA levels on behavioral phenotypes in ADHD patients as they age remain unclear.

2.2. Abnormal brain volume and functional biomarkers in the ADHD cortex

With recent clinical advancements in brain imaging techniques, researchers have correlated both functional and anatomical abnormalities in cortical regions with ADHD diagnosis and behavioral phenotypes. Clinical readouts of alterations in inhibitory signaling biomarkers in ADHD can be reflected by altered power of gamma oscillations measured using electroencephalography (EEG) or magnetoencephalography (MEG). Specifically, ADHD is associated with increased power, and synchrony of EEG and MEG gamma oscillations correlates with enhanced excitatory transmission in the cortex (Kamida et al., 2016; Karch et al., 2012; Lenz et al., 2008; Dor-Ziderman et al., 2021), pointing to a prominent role of reduced GABAergic and/or increased glutamatergic signaling associated with ADHD pathophysiology in clinical populations (Edden et al., 2012; Ende et al., 2016; Mortimer et al., 2019; Puts et al., 2020; Solleveld et al., 2017). Importantly, pathophysiological alterations in EEG gamma oscillations are driven by disruptions in GABAergic interneuron signaling (reviewed in Gonzalez-Burgos et al., 2010), suggesting a prominent role of GABA in mediating EEG abnormalities in ADHD patients.

Interestingly, adults with ADHD have significantly smaller volumes of gray matter in the PFC and ACC, which are essential brain regions required for attentional and decision-making processes (Seidman et al., 2006). Significantly smaller gray matter volumes are also reported in the ACC of children with ADHD, and this finding correlates with scores of selective inattention (Bonath et al., 2018). Therefore, understanding the role of GABAergic mechanisms underlying smaller gray matter volumes in cortical subregions of ADHD patients warrants further investigation.

2.2.2. Techniques for ADHD brain imaging studies

In addition to fMRI, near-infrared spectroscopy (NIRS) has been used as a novel biomarker to predict patient responsiveness to acute and chronic methylphenidate treatment (Ishii-Takahashi et al., 2015). While NIRS has the limitation of only being applicable to cortical brain tissue, a significant advantage of this technique over fMRI is that it can be used wirelessly in free-moving subjects. RIFC may be particularly attractive for clinical researchers studying patterns of cortical activation related to behaviors in children with ADHD, especially considering the comfort, safety, and portability of this method. Furthermore, EEG measures represent an additional non-invasive, inexpensive method to readily measure brain activity in infants that may be susceptible to an ADHD diagnosis during neurodevelopment (Karalunas et al., 2022). While gamma-band oscillations are a commonly reported EEG metric that have been shown to be elevated in ADHD patients (see section 2.2), the aperiodic EEG power spectrum has recently been emphasized as a unique biomarker for ADHD (Karalunas et al., 2022; Robertson et al., 2019). Specifically, the aperiodic exponent, which is characterized by an exponential decrease in power across increasing frequencies (i.e. power is highest in low frequencies and gradually decreases in higher frequencies), may be used to predict ADHD susceptibility across development (Karalunas et al., 2022). In ADHD patients, the aperiodic power spectral slope across development is characterized as flatter when compared to neurotypical controls, which show a steeper spectral slope (Karalunas et al., 2022). However, other studies reporting spectral slope as an EEG metric in ADHD patients have been mixed: while some report a steeper slope in ADHD children at rest (Robertson et al., 2019), others have found that alterations in the spectral slope occur during specific aspects of a cognitive task (Pertermann et al., 2019). In these studies, a flatter slope is reported for ADHD children relative to controls, and the initiation of methylphenidate treatment improves the steepness of this slope (Pertermann et al., 2019). Although a steeper spectral slope seen in ADHD patients would be consistent with increased theta/beta ratios that have been reported in ADHD patients (Markovska-Simoska et al., 2017), more studies examining the spectral slope and theta/beta ratios in ADHD patients are required to resolve discrepancies. Additionally, it is unclear whether this aperiodic power spectral slope metric is consistent with increased gamma oscillations found in ADHD, as the spectral slope analysis does not always encompass high frequency gamma oscillations. It is also unclear whether this broader spectrum analysis of EEG power changes would be preferred over conventional EEG metrics used for clinical assessments, which may be sensitive to disruptions arising from patient-to-patient variability in EEG frequencies (Gibson et al., 2022). Nonetheless, changes in the aperiodic exponent of the EEG power spectrum are reflective of alterations in cortical E/I balance (Gao et al., 2017), which allows for this technical analysis to be used to measure responses to novel treatments targeting E/I balance in ADHD. Although there have been few studies correlating EEG readouts with brain GABA concentrations using MRS (Wyss et al., 2017), future studies combining EEG and MRS methods in ADHD patient populations may be an effective strategy to correlate brain GABA levels with EEG deficits in specific brain regions with functional deficits in ADHD patients.

2.2.3. Hemispheric differences in ADHD brains

Interestingly, multiple clinical studies identifying disrupted cortical activity patterns in patients with ADHD only reveal detectable differences in the right hemisphere (Almeida et al., 2010; Almeida Montes et al., 2013; Hart et al., 2013; Rubia et al., 1999; Shaw et al., 2009). For example, transcranial magnetic stimulation (TMS)-evoked potentials and event-related potentials measured in adults with ADHD reveal that right PFC excitability is associated with ADHD severity and behavioral impulsivity (Hadas et al., 2021). ADHD children without a history of prior methylphenidate treatment showed lower levels of right inferior frontal cortex (RIFC) activation relative to healthy control subjects, and acute methylphenidate treatment enhanced RIFC activation in these untreated ADHD patients compared to untreated patients given an acute placebo dose (Ishii-Takahashi et al., 2015). Interestingly, while chronic methylphenidate treatment had no effect on RIFC activation in untreated ADHD patients, left inferior frontal cortex activation correlated with improvement of ADHD symptoms in untreated patients as measured by Clinical Global Impressions-Severity scores. Furthermore, targeting the right PFC with repeated TMS ameliorated ADHD symptoms, and this was associated with improvement of EEG biomarkers in the right PFC (Alyagon et al., 2020). Therefore, investigating biological mechanisms driving these hemispheric differences in rodent models using EEG is important for understanding how novel pharmacological targets could potentially correct atypical hemispheric asymmetries measured with EEG in ADHD patients (Mundorf & Ocklenburg, 2023). Additionally, techniques using MRS to identify how hemispheric differences in brain GABA levels correlate with alterations in EEG patterns in ADHD patients may help shed light on how biological mechanisms driving hemispheric differences in ADHD are potentially modulated via GABAergic activity.

2.2.4. Postmortem analysis of ADHD brains and neurodevelopmental disorders

In addition to functional evidence, postmortem analysis of brains of individuals with ADHD reveal that numerous differentially expressed genes in the ACC are associated with GABA activity, as well as glutamate, dopamine, and serotonin activity (Sudre et al., 2023). Overall, transcriptomic analysis performed in this study found cortico-striatal neurotransmitter abnormalities consistent with current developmental and neurotransmitter-based models of ADHD. Additionally, postmortem brain tissues of patients with neurodevelopmental and psychiatric disorders reveal abnormalities in proteins associated with GABAergic signaling pathways within the cortex (Fatemi et al., 2009, 2011; Guidotti et al., 2000). For example, in autism patients, GABA_A receptors are downregulated in the superior frontal cortex and the ACC (Fatemi et al., 2009; Oblak et al., 2009).

2.3. GABA dysfunction may underlie specific symptoms and comorbidities in ADHD subpopulations

Overall, converging evidence suggests ADHD is a disorder characterized by increased excitatory signaling and decreased inhibitory signaling in specific subregions of the frontal cortex associated with increased impulsivity and poor inhibitory control. Inconsistencies among studies relating to reduced GABA levels in ADHD patients warrants further investigation into how altered brain GABA levels are associated with ADHD in specific subpopulations. These subpopulations include childhood ADHD versus adult ADHD, as well as ADHD patients with a comorbid autism spectrum disorder (ASD), Tourette’s syndrome, or epilepsy diagnosis, as each of these disorders are also hallmarked by reduced brain GABA levels and disruptions in E/I balance (Fan et al., 2023; Fritschy, 2008; Lo-Castro & Curatolo, 2014; Openneer et al., 2020; Puts et al., 2015; Selten et al., 2018). Additionally, measuring GABA levels at rest and acutely during behavioral tasks in select cortical regions may serve as a critical biomarker to help show how cortical inhibition is impaired in ADHD at specific timepoints when symptoms emerge (Harris et al., 2021). Thus, investigating how alterations GABAergic transmission dictate performance on discrete behavioral tasks related to various ADHD symptoms will also help inform preclinical studies examining how pharmacological manipulations to the GABAergic system alter performance in genetic and behavioral models of ADHD. In addition to impulsive behavior, elucidating how GABA signaling in the PFC regulates behaviors related to working memory, attention, and cognitive flexibility will be critical for understanding how pharmacologically enhancing GABAergic transmission in the PFC may alleviate multiple cognitive symptoms of ADHD.

3. GABA-mediated signaling regulates important processes underlying ADHD phenotypes in preclinical models

3.1. Translational gaps in defining subregions of the cortex across rodent and primate species

Despite the link between reduced GABA signaling in ADHD patients in the clinic, investigation into how disrupted GABAergic transmission promotes behavioral phenotypes related to ADHD in preclinical models is limited. However, preclinical models of neurodevelopmental and neuropsychiatric disorders often encompass disease phenotypes across multiple disorders and can highlight precise biological phenomena underlying specific symptoms and comorbidities. When examining the role of the PFC in preclinical research models, there are often gaps in translating findings related to specific subregions of the PFC in rodent models to their corresponding subregions in primates and humans. Subregions of the frontal cortex in humans, such as the dlPFC and ACC, play discrete roles in ADHD pathophysiology (section 2.1). In rodents, the ACC is often defined as a subregion of the mPFC along with the infralimbic (IL) and prelimbic (PL) cortices (Fincham & Anderson, 2006; Heidbreder & Groenewegen, 2003; Hui & Beier, 2022; Laubach et al., 2018; Marusak et al., 2016; Seidman et al., 2006). In humans, the ACC and PFC are anatomically and functionally distinct. Additionally, there are discrepancies in anatomical definitions of the ACC in rodents, which highlights a need to adopt consistent nomenclature of the ACC to improve the translational relevancy of preclinical findings related to ACC function in ADHD (Heukelum et al., 2020).

3.2. GABAergic interneurons are critical regulators of E/I balance

Abnormalities in GABAergic interneurons have been thoroughly examined as a key component driving E/I balance, with brain network excitation and inhibition being primarily controlled by glutamate and GABA, respectively (Ferguson & Gao, 2018; Fritschy, 2008; Maksymetz et al., 2021; Scheyltjens & Arckens, 2016; Selten et al., 2018). Importantly, both glutamatergic and GABAergic metabotropic class C GPCRs receptors play a critical role in early CNS development and neurotransmitter specification (Root et al., 2008). One principal role of GABAergic interneurons is to inhibit pyramidal neurons to fine-tune network excitability in various brain regions. These interneurons can be subdivided into three non-overlapping categories based on distinct expression markers characterizing their function and morphologies: parvalbumin (PV), somatostatin (SST), and serotonin receptor 3A (5-HT_3A) subtypes (Kepecs & Fishell, 2014; Kupferschmidt et al., 2022; Selten et al., 2018). Additionally, interneurons may be classified as calretinin (CR)-expressing interneurons, which partially overlap with SST- and 5-HT_3A-expressing interneurons in the mouse (Lee et al., 2010). Notably, while CR interneurons comprise a significantly larger proportion of cortical interneurons in primates relative to other mammals (Hladnik et al., 2014), they remain largely unaffected in neurodevelopmental disorders such as schizophrenia relative to PV and SST interneurons (Prkačin et al., 2023). PV interneurons comprise about 40% of all GABAergic interneurons, with 30% of SST and 30% of 5-HT_3A accounting for majority of the rest of the population (Rudy et al., 2011). PV interneurons are hallmarked by their fast-spiking firing patterns and classified into two subtypes: basket cells and chandelier cells. While PV basket cells typically form multipolar synapses onto the soma and proximal dendrites of pyramidal neurons, PV chandelier cells synapse onto the axon initial segment of pyramidal neurons (see Nahar et al., 2021). Therefore, PV interneurons rapidly exert robust and precise inhibition of pyramidal neurons. SST interneurons have long been associated with Martinotti cells and display a regularly adapting firing pattern in which they initially fire bursts of two to three spikes (Kepecs & Fishell, 2014). Importantly, they target the distal dendrites of pyramidal neurons and act by gating signals from incoming excitatory projections. PV and SST GABAergic interneurons differ in their firing properties that result from differences in short term dynamics when excitatory postsynaptic currents (EPSCs) are evoked from presynaptic excitatory pyramidal neurons. SST interneurons exhibit strongly facilitating currents with low initial release probability followed by high release probability with increasing trains of stimulation. In contrast, PV interneurons exhibit depressing currents in which initially high release probability is followed by weaker EPSCs with subsequent stimulation (Ali et al., 1998; Ali & Thomson, 1998; Pouille & Scanziani, 2004). The 5-HT_3A subtype can further be subdivided into vasoactive intestinal peptide (VIP)-positive interneurons and VIP-negative interneurons. VIP interneurons act to inhibit other interneurons, primarily SST interneurons, to provide local disinhibition of cortical brain regions. However, the role of the VIP subtype in neurodevelopmental and psychiatric disorders has been understudied relative to SST and PV subtypes. While GABAergic interneurons are expressed widely throughout the brain and send either local or long-range projections to other neurons, their role in regulating network activity has been most thoroughly studied in the cortex and hippocampus.

3.3. Preclinical evidence of GABAergic interneuron dysfunction in manifestation of ADHD cognitive deficits

Neurodevelopmental disorders may have unique presentations yet share similar underlying genetic mechanisms. For example, ADHD is typically diagnosed in childhood, and schizophrenia is most often diagnosed in late adolescence and early adulthood (Faraone et al., 2015; Lawrence & Bernstein, 2024), yet these two disorders share similar underlying genetic abnormalities, deficits in E/I balance, and cognitive impairments (Bonvicini et al., 2016; Meredith, 2015; Schür et al., 2016; Slaby et al., 2022). Therefore, when investigating neurobiological mechanisms of GABAergic dysfunction underlying schizophrenia-like cognitive deficits in rodents, results from these experiments may also be translationally relevant to ADHD depending on the genetic, behavioral, or pharmacological models employed (Ugarte et al., 2023). For example, pharmacological blockade of GABA_A receptors in the PFC via direct infusions of bicuculline impairs cognition and increases impulsivity in a 5-choice serial reaction time (5CSRT) task, and this is not due to altered locomotor activity and/or motivation caused by bicuculline (Paine et al., 2011). The 5CSRT task employed in this study has high pharmacological predictive validity for ADHD when assessing the translational relevance of amphetamine treatment on improving cognition in both mice and humans (MacQueen et al., 2018). The effect of GABA on attentional deficits in ADHD preclinical models appears to be bidirectional, as both PFC hypoactivation induced by muscimol (GABA_A agonist) and PFC disinhibition induced by picrotoxin (GABA_A antagonist) impaired attention through reduced accuracy and increased omissions in the 5CSRT task (Pezze et al., 2014). Additionally, infusions of the GABA_A antagonist bicuculline directly into the dlPFC of monkeys has been shown to be associated with cognitive impairments (Pouget et al., 2009). Overall, behavioral tasks that model multiple cognitive symptoms of ADHD, such as the 5CSRT, reveal that disrupted GABA transmission in the PFC leads to impaired attention and increased impulsivity underlying ADHD cognitive deficits in preclinical species.

3.3.2. GABAergic interneurons mediate deficits in cognitive flexibility

Behavioral readouts of cognitive flexibility (e.g. set shifting and reversal learning) have been associated with the PFC and executive function in both ADHD patients (Halleland et al., 2012) and preclinical models (Chen et al., 2023). GABAergic blockade in the mPFC of rats via acute infusions of bicuculline was associated with impaired working memory in the delayed non-match to sample (DNMS) task as well as sociability deficits in the social preference test (Hernan et al., 2014). The GABAergic dysfunction model used in this study to induce focal neocortical interictal spikes measured by EEG revealed that GABA-mediated alterations in E/I balance in the PFC may underlie cognitive and social deficits in ADHD associated with childhood epilepsy. In another study using a Dlx5/6^+/− mouse model to examine the role of PV-interneurons during neurodevelopment, mice exhibited impairments in cognitive flexibility associated with reduced task-evoked gamma oscillations relative to wild-type (WT) control mice (Cho et al., 2015). Importantly, silencing PV interneurons in WT mice mimicked impaired cognitive flexibility and EEG deficits in Dlx5/6^+/− mice, and optogenetic stimulation of PV interneurons rescued these abnormalities in Dlx5/6^+/− animals. Interestingly, the impaired cognitive and EEG readouts in Dlx5/6^+/− mice were age-dependent, as 7-week-old mice did not exhibit deficits compared to 10-week-old mice. This is especially interesting considering that children with ADHD have higher gamma oscillations (Dor-Ziderman et al., 2021) and impaired executive function related to cognitive flexibility (Elosúa et al., 2017; Roberts et al., 2017). Therefore, examining how PV interneurons mediate GABAergic dysfunction in rodent models of ADHD may help elucidate how drug targets expressed in this subtype of GABAergic interneuron modulate EEG biomarkers relevant to ADHD.

Recent evidence has elucidated a more precise role of GABAergic interneurons in the ACC during the 5CSRT task using a chemogenetic approach: activation of PV-interneurons in the ACC, and to a lesser extent SST-interneurons, has been shown to improve sustained attention and impulsivity in the 5CSRT task (Jendryka et al., 2023). Additionally, chemogenetic activation of both PV and SST interneurons improved hyperactivity phenotypes in mice (Jendryka et al., 2023). Importantly, this study highlighted unique roles of GABAergic interneurons in fine-tuning symptoms of impulsivity, attention, and hyperactivity in ADHD preclinical models.

3.3.4. Deficits in working memory mediated through GABAergic interneuron dysfunction

Cognitive deficits related to working memory have also been implicated in ADHD, and strengthening working memory may be an effective therapeutic strategy to ameliorate cognitive symptoms in children with ADHD (Klingberg et al., 2002, Klingberg et al., 2005). Using the N-methyl-D-aspartate (NMDA) receptor antagonist MK-801 to create an NMDA receptor hypofunction model of cognitive impairments, pharmacological potentiation of the mGlu₁ subtype of metabotropic glutamate (mGlu) receptors reversed cognitive deficits in working memory during a y-maze behavioral task to examine spatial working memory performance in WT mice (Maksymetz et al., 2021). Importantly, pharmacological rescue of working memory deficits in this model was mediated through mGlu₁ expression on SST interneurons in the PFC, as reversal of MK-801-induced working memory deficits through systemic administration of the mGlu₁ positive allosteric modulator (PAM) VU6004909 was blocked via selective optogenetic inhibition of SST interneurons in the PFC (Maksymetz et al., 2021). Additionally, pharmacological activation of mGlu₁ improved cognitive deficits in short-term memory caused by impaired inhibitory signaling in the PFC, which was also induced through this NMDA receptor hypofunction model in mice (Luessen et al., 2022). Functionally, mGlu₁ activation enhances inhibitory transmission onto pyramidal neurons to ameliorate MK-801-induced PFC hyperactivation and cognitive impairments by correcting deficits in E/I balance and promoting cognition through inhibitory feedback (Maksymetz et al., 2021; Luessen et al., 2022).

3.3.5. Cognitive and affective deficits associated with mGlu ₅

Clinical studies have shown that SNPs in the GRM5 gene encoding for mGlu₅ are associated with ADHD (Elia et al., 2010; Glessner et al., 2023) and mGlu₅ has been demonstrated to play a key role in pathophysiological adaptations underlying plasticity of GABAergic interneuron networks (Joffe et al., 2022). While acute restraint stress impairs working memory during the y-maze task in WT mice, mice with selective deletion of mGlu₅ from SST interneurons did not show working memory impairments (Joffe et al., 2022). Moreover, evidence that mGlu₅ promotes cognitive flexibility was shown by pharmacological potentiation of mGlu₅ using the mGlu₅ PAM, VU0409551, which corrected stress-induced deficits in reversal learning (Joffe et al., 2019). VU04095551 has also been shown to reverse cognitive deficit-related novel object recognition performance in two separate studies using an NMDA receptor hypofunction model in either mice or rats, and these pro-cognitive effects of the mGlu₅ PAM are likely dependent on mGlu₅ signaling in the PFC (Brown et al., 2022; Ghoshal et al., 2017). Interestingly, VU0409551 is a biased mGlu₅ PAM, with preferential activation of mGlu₅ G_q-mediated pathways over mGlu₅-mediated modulation of NMDA receptor currents, and VU0409551 does not induce NMDA receptor-dependent long-term potentiation (LTP) in the hippocampus (Rook et al., 2015).

Furthermore, evidence linking emotional dysregulation in ADHD to mGlu₅ is shown by a study describing genetic mutations in the LRRC7 gene, encoding for the leucine-rich repeat containing protein 7 (LRRC7), a synaptic scaffolding protein located in the postsynaptic density (PSD) (Chong et al., 2019). Mutations in scaffolding proteins are often associated with ADHD (see 4.1), and LRRC7 SNPs are linked to high scores on the Child Behavior Checklist–Dysregulation Profile (CBCL–DP) in ADHD patients, conferring increased risk for these ADHD patients to develop affective disorders (Mick et al., 2011). Potentiation of mGlu₅ ameliorated deficits in affective behaviors seen in in Lrrc7 mutant mice and also correct anatomical abnormalities related to spine density and neurite outgrowth (Chong et al., 2019). Therefore, converging evidence suggests that potentiating mGlu₅ may serve as a novel pharmacological approach to restore working memory deficits as well as emotional dysregulation in ADHD by enhancing GABAergic transmission in the PFC through select interneuron subtypes.

3.3.6. Additional evidence of GABAergic dysfunction in ADHD cognitive deficits

Preclinical evidence from ADHD models highlights an important role of GABA synthesizing and metabolizing enzymes in mediating ADHD phenotypes. For example, both acute and chronic methylphenidate treatment in female rats increases glutamic acid decarboxylase (GAD) enzyme mRNA expression in the PFC, specifically GAD65 and GAD67, which are enzymes that convert glutamate to GABA (Freese et al., 2012). This increase in GAD enzyme mRNA expression would be anticipated to lead to increased GABA PFC levels. While GABA and dopamine can function as co-transmitters in monoaminergic neurons (Tritsch et al., 2012), the exact mechanisms of methylphenidate leading to increased GAD mRNA in the PFC remain unclear. In addition to the PFC, this disinhibitory phenomenon of attentional deficits also occurs in the hippocampus (McGarrity et al., 2017).

G protein-coupled receptor kinase-interacting protein-1 (GIT1) functions as a GTPase-activating protein domain for ADP ribosylation factor (ARF) small GTPases (Premont et al., 1998). Deleterious SNPs in GIT1 in humans are associated with ADHD, and Git homozygous KO mice show several ADHD-like impairments (Won et al., 2011). Behavioral impairments in Git KO mice include hyperactivity, impaired cognition in a novel object recognition task and reduced spatial learning and memory in the Morris water maze task; all behavioral impairments have been shown to be reversed by amphetamine treatment (Won et al., 2011). Additionally, Git KO mice show enhanced EEG theta rhythms, which are reduced by amphetamine treatment. Importantly, Git KO mice also exhibited reduced inhibitory transmission in the hippocampus which was associated with decreased GAD67 vesicular GABA transporter (vGAT) protein expression. Anatomical and behavioral deficits in Git KO mice are likely related to alterations in PV interneurons, as staining for PV interneuron markers was reduced in the hippocampus of these animals (Won et al., 2011). Thus, human mutations in GIT1 associated with ADHD are proposed to result in impaired E/I balance through GABAergic interneuron dysfunction.

Due to differences in the incidence of ADHD among males and females in human populations (Faraone et al., 2015), it is important to investigate sex differences related to ADHD in preclinical research. Sex differences in GABAergic systems in the mouse mPFC have been reported in which GABA_B receptor-dependent activation of G protein-gated inwardly-rectifying K+ (GIRK) channels in the mPFC is stronger in adolescent male mice relative to female mice (Marron Fernandez de Velasco et al., 2015). This may suggest that higher incidence of ADHD in males may be mediated through exacerbated GABAergic dysfunction in the PFC, although more research is needed to confirm this hypothesis.

4. ADHD mouse models from human mutations to study mechanisms of synaptic impairments and identify drug targets

4.1. Genetic models derived from patient SNPs used to study synaptic impairments in ADHD

In neurodevelopmental disorders, synapse formation, maintenance, and plasticity are regulated by an extensive variety of pre- and post-synaptic proteins that carefully maintain a delicate balance of glutamate and GABA signaling (Jang et al., 2017). Several mutations in genes encoding for SAMs and PSD proteins are associated with ADHD, including CDH13, CNTN4, ELFN1, LPHN3, LRRC7, NRNX1, PCDH7, SNAP25 among others (Dark et al., 2018; Kessi et al., 2022). The discovery of SNPs in SAMs and PSD proteins in clinical populations with ADHD has led to the development of many genetic mouse models of ADHD that have greatly improved our understanding of the pathophysiology underlying the disorder, especially in relation to ADHD comorbidities. For example, mutations in the LPHN3 (ADGRL3) gene encoding for the adhesion GPCR latrophilin 3 (Lphn3) has resulted in the study of a Lphn3 KO mouse model in the context of ADHD phenotypes (Moreno-Alcázar et al., 2021; Mortimer et al., 2019; Orsini et al., 2016). Not only do mutations in LPHN3 predict ADHD severity and responsivity to stimulant treatment in humans (Acosta et al., 2016; Arcos-Burgos et al., 2010; Arcos-Burgos et al., 2019; Moreno-Alcázar et al., 2021; Ribasés et al., 2011), but identification of specific LPHN3 mutations has also helped improve our understanding of molecular characteristics of Lphn3 receptor pharmacology (Moreno-Salinas et al., 2022). Furthermore, Lphn3 KO mouse models have assisted in the elucidation of the precise neurobiology underlying ADHD pathophysiology related to deficits in neuroplasticity and cortical maturation, mechanisms of dysregulated striatal dopamine, and differentially expressed genes encoding for SAMs (Mortimer et al., 2019; Orsini et al., 2016; Regan et al., 2021). Therefore, the high construct validity of these genetic ADHD rodent models may be used to determine face validity related to specific biomarkers in ADHD patients with these mutations, as well as assess predictive validity related to drug development efforts targeting proteins coupled to SAMs that show SNPs in ADHD patients. Furthermore, translational models can target precise symptoms that overlap in various neurodevelopmental disorders, effectively treating symptoms at discrete developmental timepoints when they emerge and informing researchers on how specific targets/proteins drive symptoms related to comorbidity and pathophysiology underlying ADHD and comorbid neurodevelopment disorders with common underlying genetic mechanisms. In turn, this can inform clinicians as to appropriate treatment regiments and improve the likelihood of responsivity to treatments (Arcos-Burgos et al., 2010; Labbe et al., 2012).

4.2. Selectively targeting mGlu receptors to restore E/I balance in CNS disorders

As previously mentioned, mutations in several GPCRs, SAMs, and other proteins result in GABAergic dysfunction leading to E/I imbalance in ADHD. Although the gene encoding for the GABABR1 metabotropic GPCR was not found to be significantly mutated in ADHD populations (Barr et al., 2000), mutations in several other mGlu genes, including GRM7, are significantly associated with ADHD (Glessner et al., 2023). The mGlu receptors are expressed in both excitatory pyramidal neurons as well as inhibitory interneurons and play key roles in regulating brain network activity to maintain a healthy E/I balance. Therefore, pharmacologically targeting select subtypes of these receptors with selective ligands such as allosteric modulators may represent novel treatment strategies with high potential to fine-tune network activity and alleviate multiple symptoms associated with GABAergic dysfunction in several CNS disorders (Connolly et al., 2015; Foster & Conn, 2017; Luessen & Conn, 2022). For ADHD patients, allosteric modulators of mGlu₇ are promising candidates as discussed below.

4.2.2. SNPs in GRM7 persistently emerge in ADHD populations

The GRM7 gene has repeatedly been shown to be mutated in clinical populations with ADHD (Elia et al., 2012; Glessner et al., 2023; Mick et al., 2008; S. Park et al., 2013; Q. Zhang et al., 2021). mGlu₇ is a presynaptic G_i/o coupled dimeric GPCR (Niswender & Conn, 2010) and its activation inhibits glutamate and GABA release (Freitas & Niswender, 2023). Previous work from our lab has demonstrated that activation of mGlu₇ that is expressed presynaptically on GABAergic interneuron terminals at hippocampal Schaffer Collateral-CA1 (SC-CA1) synapses inhibits GABA release to induce LTP through disinhibition of pyramidal neurons (Klar et al., 2015). Due to its low affinity for glutamate, mGlu₇ activation is predicted to occur only constitutively or under situations of intense stimulation, which would be anticipated to increase glutamate levels to the levels needed for agonist-mediated receptor activation (high μM to mM range). mGlu₇ also functions as an autoreceptor on excitatory pyramidal neuron terminals, where potentiation of the receptor leads to decreases in glutamate release probability (Baskys and Mealenka, 1991; Ayala et al., 2008). Therefore, mGlu₇ function is anticipated to serve as a critical regulator of GABAergic and glutamatergic mechanisms underlying E/I balance in ADHD brain circuitry. In one study identifying GRM7 SNPs in a Korean population of ADHD patients, the authors found that GRM7 mutations are associated with poor performance during a measure of sustained and selective attention (Park et al., 2013), suggesting that mGlu₇ may play a more prominent role in phenotypes of cognitive control in ADHD. Additionally, this study reported that GRM7 SNPs are associated with higher anxiety scores in ADHD patients, even after excluding subjects with an anxiety disorder diagnosis during analysis. It is also plausible that mGlu₇ could play a direct role in regulating dopamine release in the striatum, as group III metabotropic glutamate receptors have been shown to tonically inhibit dopamine release in rodents (Hu et al., 1999; Mao et al., 2000) and are expressed in rat VTA dopamine neurons (Phillips et al., 2022). However, whether this effect is mediated by mGlu₇ or other group III mGlu receptors has yet to be elucidated.

4.2.3. ELFN1 mutations in ADHD are associated with mGlu₇ function

Intriguingly, extracellular leucine rich repeat and fibronectin type III domain containing 1 (Elfn1), a postsynaptic scaffolding protein preferentially expressed in SST interneurons (Dolan & Mitchell, 2013; Keijser & Sprekeler, 2023; Sylwestrak & Ghosh, 2012), has recently been shown to physically anchor and recruit mGlu₇ and other group III mGlu receptors into place at presynaptic terminals onto pyramidal neurons and modulate their activity (Dunn et al., 2018; D. Park et al., 2020; Stachniak et al., 2019; Sylwestrak & Ghosh, 2012; Tomioka et al., 2014). Notably, human mutations in the ELFN1 gene are also associated with ADHD, as well as epilepsy and Tourette’s syndrome (Tomioka et al., 2014; Vaccarino, n.d. (online data)). It is hypothesized that mutations in ELFN1 contribute to SST interneuron dysfunction associated with ADHD and neurodevelopmental disorders (Matsunaga & Aruga, 2021). Taken together, genetic evidence from ADHD subpopulations with specific comorbidities related to PFC GABAergic interneuron dysfunction, such as epilepsy and Tourette’s syndrome, highlight the trans-synaptic role of mGlu₇/Elfn1 as an important pharmacological target to explore for potential novel ADHD treatments in specific subpopulations.

4.2.4. Grm7 and Elfn1 KO mouse models reveal ADHD-like behavioral phenotypes

Preclinical evidence reveals that Grm7 KO mice exhibit blunted responses to amphetamine-induced hyperlocomotion, as well as a blunted effect of amphetamine treatment on increased gamma power and lowered delta EEG power spectra patterns seen in littermate WT controls (Fisher et al., 2020). However, amphetamine treatment in Grm7 KO mice had no significant effect on theta, beta, or alpha EEG power spectra in this study, which are more commonly used to predict stimulant treatment responsivity in ADHD patients (Clarke et al., 2003; Loo et al., 1999; Loo & Makeig, 2012).

Further evidence of dysfunctional mGlu₇ signaling in ADHD-like behavioral phenotypes is indicated by the role of Elfn1 in physically recruiting and promoting constitutive activity of mGlu₇. Importantly, Elfn1 KO mice exhibit baseline hyperlocomotive behavior relative to wild-type controls, and they also show a blunted locomotor response to amphetamine (Dolan & Mitchell, 2013; Tomioka et al., 2014). These studies warrant further investigation into how Elfn1 regulates locomotor behaviors through modulation of group III metabotropic glutamate receptors, especially considering that Elfn1 is expressed in the globus pallidus of the basal ganglia, and that group III mGlu receptors play an intriguing role modulating GABA release in this brain region (MacInnes & Duty, 2008). Moreover, the parallels between Grm7 and Elfn1 KO mice related to the paradoxical “calming” effect of low doses of amphetamine may have important implications for the GABAergic origins of ADHD, as both mGlu₇ and Elfn1 are associated with modulating GABAergic brain circuits (Klar et al., 2015; Tomioka et al., 2014).

It is also worth noting that Grm7 and Elfn1 KO mice both exhibit an onset of seizures at almost the same developmental time point (3–4 months) (Sansig et al., 2001; Fisher et al., 2020; Dolan & Mitchell, 2013; Tomioka et al., 2014) further highlighting the role of this synaptic GPCR complex as a potential underlying contributor to the high rate of comorbid ADHD and epilepsy. One hypothesis is that the emergence of seizures at this specific timepoint is due to excitotoxic glutamate release onto SST interneurons, which eventually leads to excessive pyramidal neuron activity that may result in seizure activity. Although Tomioka et al. report that the number of SST interneurons were reduced at 4 months of age, this effect was not significant (p=0.08). Also, despite reduced mGlu₁ staining (a surrogate for SST cell density) in dendritic spines, there was no reduction in total mGlu₁ staining at 4 months. Moreover, these alterations in SST neurons that occur in Elfn1 KO mice at 4 months may be a consequence of the onset of seizures, rather than the cause. An alternative explanation to the parallel emergence of seizures in Grm7 and Elfn1 KO mice is that SST neurons lose their ability to provide sustained feedback inhibition, thus causing pyramidal neurons to become hyperexcitable over time leading to runaway activity. Considering that overactivation of cortical pyramidal neurons due to loss feedback inhibition leads to epileptic states (Silberberg and Markram, 2007), it is likely that this mechanism of GABAergic dysregulation is the underlying cause of seizures, rather than SST excitotoxicity that may occur only after the onset of seizures.

4.2.5. mGlu₇ receptor and Elfn1 scaffolding proteins selectively regulate GABAergic interneuron activity

Functionally, Elfn1 acts an endogenous allosteric modulator of mGlu₇, and in vitro Elfn1 acts as an allosteric inhibitor of group III mGlu receptor activity (Dunn et al., 2018). In native tissue, however, Elfn1 has been shown to enhance constitutive activity of mGlu₇ selectively at pyramidal cell to SST interneuron synapses (Stachniak et al., 2019), and this effect has been described in both the cortex and hippocampus (Stachniak et al., 2019; Tomioka et al., 2014). More specifically, Elfn1 is thought to play a critical role in suppressing initial release of glutamate from presynaptic pyramidal neurons onto postsynaptic SST interneurons, which are characterized by facilitating currents (Sylwestrak & Ghosh, 2012; Tomioka et al., 2014). Elfn1 KO mice exhibit strong initial EPSCs in SST interneurons relative to WT mice, which show weak initial EPSC amplitudes in these neurons (Stachniak et al., 2019; Sylwestrak & Ghosh, 2012; Tomioka et al., 2014). Additionally, pharmacological activation of mGlu₇ reduces spike probability and increases the number of stimuli needed to evoke half-maximal spiking for synaptically-evoked spikes in cortical SST interneurons (Stachniak et al., 2019). Therefore, it has been proposed that constitutive activity of mGlu₇, modulated by Elfn1, decreases initial release probability of glutamate onto SST interneurons. Remarkably, PV interneurons, which do not express high levels of Elfn1, switch from exhibiting depressing to facilitating currents when Elfn1 is artificially overexpressed in hippocampal PV interneurons using lentiviral-mediated transfection (Sylwestrak & Ghosh, 2012). The effect of mGlu₇/Elfn1 in suppressing initial glutamate release is present in both layer 2/3 and layer 5 of the cortex. Specifically in layer 2/3, strong facilitating currents onto SST interneurons are enhanced by recruitment of kainate receptors containing glutamate receptor, ionotropic, kainate 2 (GluK2-KARs), which play a critical role in facilitating late currents (Stachniak et al., 2019). Recent evidence reveals that GluK2-KAR activity, mediated through mGlu₇ and Elfn1 on pyramidal cell to SST interneuron projections, plays a critical role in late-phase sensory adaptation in neurodevelopment (Stachniak et al., 2023). Considering the role of SST interneurons in gating excitatory inputs onto pyramidal neurons, it is plausible that loss of either mGlu₇ or Elfn1 leads to functional abnormalities in SST interneurons that drive pathological hyperexcitability of pyramidal neuron networks underlying ADHD symptoms and comorbidities with epilepsy and Tourette’s syndrome (Figure 2).

Figure 2:

GABAergic SST interneurons (blue) express Elfn1 (green) which physically recruits and anchors mGlu₇ (purple) into place at presynaptic terminals of excitatory pyramidal neurons (yellow). Elfn1 promotes strongly facilitating currents in postsynaptic SST interneurons and Elfn1/mGlu₇ interactions create a delayed feedback loop that prevents excessive excitation of pyramidal neurons. Loss of Elfn1 diminishes late strongly facilitating currents in SST interneurons (inset), leading to impaired GABAergic signaling and hyperexcitability of pyramidal neurons. It is speculated that the additional loss of presynaptic ionotropic kainate receptors (orange) may contribute to loss of late strongly facilitating currents, leading to overall net loss of excitation of SST interneurons. It is hypothesized that ADHD patients with Single Nucleotide Polymorphisms (SNPs) in ELFN1 and GRM7 have impaired GABAergic signaling in the cortex that are mediated through GABAergic interneuron dysfunction, leading to pyramidal neuron hyperexcitability and disruptions in excitatory/inhibitory (E/I) balance. Representative traces of excitatory postsynaptic currents (EPSCs) following 5 trains of stimulation in wild-type (black) and Elfn1 knockout mice (red, adapted from Stachniak et al., 2019 and Stachniak et al., 2023).

While physical trans-synaptic interaction between mGlu₇ and Elfn1 has been well characterized in SST interneurons, additional evidence suggests that Elfn1 also regulates VIP interneuron activity (Stachniak et al., 2021). VIP interneurons primarily innervate neighboring SST interneurons and can be categorized into bipolar or multipolar subtypes based on morphology. Bipolar and multipolar VIP subtypes also show unique functional properties, and the transcription factor Prospero-related homeobox1 (Prox1) plays a role in determining physiological differences between these subtypes. Specifically, deletion of Prox1 from VIP interneurons selectively modulates short-term synaptic facilitation dynamics of incoming excitatory inputs onto the multipolar, but not bipolar, subtype. Similarly, loss of Elfn1 modulates synaptic facilitation selectively in multipolar and not bipolar VIP interneurons (Stachniak et al., 2021), supporting previous evidence that Elfn1 drives this physiological synaptic facilitation phenotype seen in select subtypes of GABAergic interneurons (Sylwestrak & Ghosh, 2012). Interestingly, both bipolar and multipolar VIP interneuron subtypes express high levels of Elfn1 mRNA, and Prox1 deletion reduces Elfn1 mRNA expression in both subtypes. Because deletion of Prox1 from VIP interneurons only modulates short-term synaptic facilitation in the multipolar and not bipolar VIP interneurons, and multipolar short-term synaptic facilitation is associated with Elfn1 protein expression, this suggests that there are post-translational modifications preventing Elfn1 protein expression in bipolar VIP interneurons, which drive physiological differences in synaptic facilitation between the bipolar and multipolar subtypes. Intriguingly, post-translation differences in Elfn1 expression, as well as mGlu₇, may also be modulated by microRNAs (Whipple et al., 2020). Overall, reports indicate that Elfn1 plays an important functional role in modulating group III mGlu receptor constitutive activity to inhibit presynaptic neurotransmitter release and promote strongly facilitating currents in select subtypes of GABAergic interneurons.

4.3. Additional avenues for targeting dysregulated group III mGlu receptors in ADHD: microRNAs, Elfn2 and mGlu heterodimers

It has recently been found that certain microRNAs (miR-652–3p, miR-148b-3p, and miR-942–5p) are differentially expressed in ADHD patients (Nuzziello et al., 2019); deciphering which proteins these microRNAs target may help establish or further strengthen findings related to GABAergic dysfunction and E/I imbalance in ADHD pathophysiology. For example, mir-34 regulates expression of mGlu₇, and chronic treatment with mood stabilizers lithium and valproate in rats decreases mir-34 levels, leading to upregulated Grm7 expression (R. Zhou et al., 2009). Therefore, differentially expressed microRNAs may regulate expression of specific proteins associated with ADHD; in turn, pharmacological treatments for ADHD may be designed to modulate microRNA expression to alleviate ADHD symptoms by regulating expression of proteins associated with ADHD pathophysiology.

In addition to Elfn1, the related scaffolding protein Elfn2 also binds and recruits group III mGlu receptors in place at presynaptic terminals in trans (Dunn et al., 2019). Although no studies to date have identified SNPs in ELFN2 associated with ADHD or other CNS disorders, Elfn2 KO mice display several phenotypes associated with neuropsychiatric disorders and exhibit unique expression patterns relative to Elfn1 (Dunn et al., 2019). Therefore, extrapolating on the role of Elfn2 relative to Elfn1 in relation to group III mGlu receptors may help elucidate how these two related scaffolding proteins play unique roles in synaptic architecture, mGlu receptor pharmacology, and potentially ADHD pathophysiology.

It has recently been shown that the mGlu receptors are capable of heterodimerizing (Du et al., 2021; J. Lee et al., 2020; Meng et al., 2022; Xiang et al., 2021; Yin et al., 2014; Doumazane et al., 2011). This finding suggests that investigating the role of mGlu heterodimers in ADHD pathophysiology may generate novel drug targets that selectively target heterodimers in specific brain circuits in which they are expressed (Lin et al., 2022; Xiang et al., 2021). For example, SNPs in GRM4 are associated with ADHD (Zhang et al., 2021), and mGlu₄ is associated with impulsive behaviors in mice (Piszczek et al., 2022). Furthermore, SNPs in GRM8, encoding for the mGlu₈ receptor, are implicated in ADHD and other neurodevelopmental disorders (Glessner et al., 2023), and a recent report suggests that pharmacological inhibition of mGlu_7/8 heterodimers mediates physiological deficits in LTP at SC-CA1 synapses that may be related to seizure activity in vivo (Lin et al., 2022). Therefore, understanding the role of mGlu receptor heterodimer formation may help elucidate how discrete brain circuits underlie specific behaviors related to ADHD and comorbid disorders by pharmacologically and specifically probing these heterodimers.

5. Conclusion

Although treatment options for ADHD are limited and progress towards novel therapeutics for the disorder have been stagnant, emerging evidence from GWAS studies identifying genes associated with ADHD has created a new frontier for developing personalized ADHD medicines that target specific symptoms and cases of ADHD with comorbidities. While ADHD is typically characterized as a disorder of dysregulated dopaminergic signaling, recent studies highlighting the role of E/I balance in ADHD reveal that this neurodevelopmental disorder is hallmarked by disruptions in GABAergic and glutamatergic signaling, with the PFC and ACC mediating cognitive deficits in ADHD among other symptoms. GABAergic interneurons likely play a critical role in ADHD pathophysiology related to E/I balance, specifically in relation to synaptic connections between subtypes of inhibitory neurons and pyramidal neurons. While the relationship between excitatory pyramidal neurons and inhibitory interneurons is best understood in cortical and hippocampal brain regions, investigating the role of the GABAergic interneurons in mesolimbic brain circuits may bridge the gap between interneuron dysfunction and abnormalities in striatal brain circuits. Nonetheless, genetic evidence suggests that symptoms of ADHD are mediated by deficits in synaptic plasticity and abnormal synaptic architecture. Although it is commonly reported that 50% of ADHD patients outgrow their symptoms by adulthood, a more careful longitudinal analysis reveals that only about 10% of patients become completely symptom free, while most others experience fluctuating symptoms that continue into adulthood (Sibley et al., 2022). In this small subset of patients that experience full remission, there are likely compensatory mechanisms related to synaptic plasticity and architecture that allow them to completely outgrow their ADHD symptoms. As these synaptic connections undergo changes into adulthood, ADHD symptoms may begin to decline for certain individuals, while other patients may experience fluctuating symptoms related to individual differences in compensatory mechanisms of synaptic reorganization and function. Additionally, ADHD symptoms persisting into adulthood may present in a manner that manifest into diagnosis of other neuropsychiatric disorders, such as substance use disorder and BPD. Therefore, treating ADHD during this critical window in neurodevelopment when symptoms emerge may help treat patients who are susceptible to developing these other neuropsychiatric disorders commonly associated with ADHD. It is also important to note that alterations in E/I balance may shift over the course of development, as well as day-to-day, depending on the prevalence of certain symptoms as well as specific tasks being performed. Overall, investigating neurobiological mechanisms that disrupt E/I balance in ADHD at a specific timepoint during neurodevelopment is anticipated to assist in generating precision medicines that effectively treat the underlying cause of the disorder and place patients on a healthy developmental trajectory.

Highlights:

• New precision medicine treatments are needed for ADHD

• Brain GABA levels are altered in ADHD patients and behaviors

• ADHD pathophysiology is characterized by disrupted E/I balance

• mGlu₇ and Elfn1 represent novel pharmacological targets for ADHD

Abbreviations:

ARF

ADP ribosylation factor

ACC

anterior cingulate cortex

ADHD

attention-deficit/hyperactivity disorder

ASD

autism spectrum disorder

BPD

bipolar disorder

CDH13

cadherin-13

COMT

catechol-o-methyltransferase

CNS

central nervous system

DNMS

delayed non-match to sample

dlPFC

dorsolateral PFC

EEG

electroencephalography

E/I

excitatory/inhibitory

EPSC

excitatory postsynaptic current

Elfn1

extracellular leucine rich repeat and fibronectin type III domain containing 1

5CSRT

five choice serial reaction time task

fMRI

functional magnetic resonance imaging

GABA

gamma-aminobutyric acid

GWAS

genome-wide association studies

GAD

glutamic acid decarboxylase

GPCR

G protein-coupled receptor

GIT1

G protein-coupled receptor kinase-interacting protein-1

GIRK

G protein-gated inwardly-rectifying K+

IL

infralimbic

Lphn3

latrophilin 3

GluK2-KARs

kainate receptors containing glutamate receptor, ionotropic, kainate 2

LRRC7

leucine-rich repeat containing protein 7

LTP

long-term potentiation

MRS

magnetic resonance spectroscopy

mPFC

medial PFC

mGlu receptor

metabotropic glutamate receptor

MPH

methylphenidate

NIRS

near-infrared spectroscopy

NET

norepinephrine transporter

NMDA

N-methyl-D-aspartate

PV

parvalbumin

PFC

prefrontal cortex

PSD

postsynaptic density

PL

prelimbic

Prox1

Prospero-related homeobox1

RIFC

right inferior frontal cortex

5-HT_3A

serotonin receptor 3A

SNP

Single Nucleotide Polymorphism

SST

somatostatin

SAM

synaptic adhesion molecule

TMS

transcranial magnetic stimulation

VIP

vasoactive intestinal peptide

vGAT

vesicular GABA transporter

WT

wild-type