GABA receptor subunit distribution and FMRP-mGluR5 signaling abnormalities in the cerebellum of subjects with schizophrenia, mood disorders, and autism

Abstract

Gamma-aminobutyric acid (GABA) is the main inhibitory neurotransmitter in the brain. GABAergic receptor abnormalities have been documented in several major psychiatric disorders including schizophrenia, mood disorders, and autism. Abnormal expression of mRNA and protein for multiple GABA receptors has also been observed in multiple brain regions leading to alterations in the balance between excitatory/inhibitory signaling in the brain with potential profound consequences for normal cognition and maintenance of mood and perception. Altered expression of GABA_A receptor subunits has been documented in Fragile X mental retardation 1 (FMR1) knockout mice, suggesting that loss of its protein product, fragile X mental retardation protein (FMRP), impacts GABA_A subunit expression. Recent postmortem studies from our laboratory have shown reduced expression of FMRP in brains of subjects with schizophrenia, bipolar disorder, major depression, and autism. FMRP acts as a translational repressor and, under normal conditions, inhibits metabotropic glutamate receptor 5 (mGluR5)-mediated signaling. In fragile X syndrome (FXS), absence of FMRP is hypothesized to lead to unregulated mGluR5 signaling, ultimately resulting in the behavioral and intellectual impairments associated with this disorder. Our laboratory has identified changes in mGluR5 expression in autism, schizophrenia, and mood disorders. In the current review article, we discuss our postmortem data on GABA receptors, FMRP, and mGluR5 levels and compare our results with other laboratories. Finally, we discuss the interactions between these molecules and the potential for new therapeutic interventions that target these interconnected signaling systems.

1. Introduction

This article reviews the current postmortem evidence for dysfunction of the gamma-aminobutyric acid (GABA) and the fragile X mental retardation protein (FMRP)-metabotropic glutamate receptor 5 (mGluR5) signaling pathways in four major psychiatric disorders: schizophrenia, bipolar disorder, major depression, and autism. GABAergic dysfunction has long been implicated in deficits associated with schizophrenia, mood disorders, and autism (Gonzales-Burgos et al., 2011; Luscher et al., 2011; Coghlan et al., 2012). Disruption of inhibitory GABAergic signaling and the resulting imbalance between excitatory and inhibitory circuitry in the brain may be associated with behavioral and cognitive deficits in each disorder. Postmortem findings from our laboratory (Fatemi et al., 2009a,b, 2010a, 2014) and others (Akbarian et al., 1995; Beneyto et al., 2011; Duncan et al., 2010; Glausier and Lewis, 2011; Hashimoto et al., 2008; Maldonado-Aviles et al., 2009) have identified changes in mRNA and protein for a number of individual GABA receptor subunits in brains of subjects with schizophrenia, mood disorders, and autism.

Data from animal models of fragile X syndrome (FXS), which lack FMRP (an RNA binding master protein that targets 5% of all brain genes including some of the GABA receptors) show altered expression of GABA receptor subunits in brain (Adusei et al., 2010; El Idrissi et al., 2005; D’Hulst et al., 2006; Gantois et al., 2006; Hong et al., 2012). Novel data from our laboratory has recently shown reduced expression of FMRP in schizophrenia, bipolar disorder, major depression, and autism (Fatemi and Folsom, 2011; Fatemi et al., 2010b, 2011a, 2013a), and represents the first data to show altered expression of FMRP in subjects who do not have a diagnosis of fragile X syndrome (FXS). These results have now been replicated in schizophrenia by other laboratories (Kelemen et al., 2013; Kóvacs et al., 2013).

FMRP negatively regulates mGluR5 activity. The mGluR5 theory of FXS proposes that in the absence of FMRP, mGluR5 activity is greatly increased ultimately resulting in the deficits associated with FXS (Bear et al., 2004; Dölen and Bear, 2008). Our laboratory has identified the upregulation of mGluR5 expression in autistic children (Fatemi and Folsom, 2011; Fatemi et al. 2011a) and this finding has been replicated by another group in patients with FXS (Lohith et al., 2013). In schizophrenia and mood disorders, we have determined reduced mGluR5 protein expression (Fatemi et al., 2013a), while other groups have found either no change (Corti et al., 2011; Gupta et al., 2005; Matosin et al., 2013, 2014) or only a reduction of this receptor in major depression (Deschwanden et al., 2011); details of which will be further discussed in section 4.

Taken together, changes in GABA-FMRP-mGluR5 signaling may have profound effects and contribute significantly to the etiology of major psychiatric disorders. These changes also offer the potential for therapeutic intervention. In particular, modulators of GABA receptor and mGluR5 activity are discussed in section 7 as potential means of ameliorating symptoms of schizophrenia, mood disorders, and autism.

2. GABA

2.1. GABA Receptors

GABA is the main inhibitory neurotransmitter in the brain. In the central nervous system approximately 20% of all neurons are GABAergic (Charych et al., 2009). There are two classes of GABA receptors: GABA_A and GABA_B, which differ in terms of composition, pharmacology, and action. There are 19 different GABA_A receptor subunits: alpha 1–6 (α1–6), beta 1–3 (β1–3), gamma1–3 (γ1–3), delta (δ), epsilon (ε), pi (π), theta (θ), and rho 1–3 (ρ1–3) that combine to form the various GABA_A receptors (Brandon et al., 2000; Ma et al., 2005). These receptors are either mostly synaptic/postsynaptic (α1–4, β1–3, γ1–3, δ, ε, θ, π, ρ1–3), predominantly extrasynaptic (α4, α5, α6, and δ), or both (α5, ρ1) (Errington, 2014; Wafford, 2014). GABA_A receptors are binding sites for benzodiazepines, steroids, and anesthetics, thus mediating their effects. While synaptic receptors, because of the presence of γ subunits, are sensitive to the action of benzodiazepines; extrasynaptic receptors the presence of the δ subunit are insensitive to benzodiazepines but are sensitive to the agonist 4,5,6,7-tetrahydroisoxazolo[5,4-c]pyridin-3-ol (THiP or Gaboxadol; Wafford and Ebert, 2006) as well as neurosteroids (Stell et al., 2003; Belelli and Lambert, 2005). Interestingly, extrasynaptic GABA_A receptors can be modulated both directly and indirectly by GABA_B, serotonin, dopamine, noradrenalin, and metabotropic glutamate receptors (Errington, 2014). Additionally, synaptic GABA_A receptors are composed of α1, α2, and α3 subunits along with β1–3, and γ2 subunits, while extrasynaptic receptors are composed of α1, α4, α5, β, γ2 and δ subunits (Farrar et al., 1999; Wafford, 2014). GABA_B receptors are heterodimeric receptors consisting of one GABA_B receptor 1 subunit (GABBR1) and one GABA_B receptor 2 subunit (GABBR2). GABA_B receptors are coupled via G-proteins to potassium and calcium channels as well as adenylate cyclase (Bowery, 2000). Additionally, the action of GABA_B receptors is slower than GABA_A receptors. Presynaptically, GABA_B receptors facilitate neurotransmitter release including glutamate and GABA, while postsynaptically, they generate inhibitory potentials (Bowery, 2000; Kuriyama et al., 2000). GABA_A receptors, upon binding GABA, lead to increased membrane permeability to both chloride and bicarbonate ions (Errington, 2014; Chebib and Johnston, 2000). There are essentially two types of GABAergic inhibition: 1) phasic inhibition which results from a transient rise in GABA in synapse and activation of postsynaptic GABA receptors and production of inhibitory postsynaptic potentials (IPSP); and 2) tonic inhibition which results from very low nanomolar (nM) concentrations of GABA and activation of extrasynaptic GABA receptors located in several brain sites such as cerebellar granule cells, layer II/III and V cortical pyramidal cells, medium spiny neurons of striatum, CA1 and CA3 pyramidal neurons, dentate granule cells, and thalamocortical neurons (Errington, 2014).

Altered GABAergic synaptic transmission has been associated with panic, anxiety, impaired learning and memory, and sleep abnormalities (Heulens et al., 2010; Pinna et al., 2006, 2009), and thus contributes to deficits associated with schizophrenia, bipolar disorder, major depression, and autism (Gonzalez-Burgos et al., 2011; Luscher et al., 2011; Möhler et al., 2012).

2.2. Changes in GABA Binding Site Density in Major Psychiatric Disorders

Initial pioneering studies by Benes and colleagues have demonstrated increased GABA_A receptor binding in brains of subjects with schizophrenia (Benes et al., 1992, 1996a,b, 1997). An initial study identified an increase in ³H-muscimol-labeled GABA_A receptor binding sites in layers II and III, but not IV or V of the anterior cingulate cortex of subjects with schizophrenia (Benes et al., 1992). Further experiments revealed increased GABA_A receptor binding sites in layers II, III, V, and VI in prefrontal cortex (PFC) of subjects with schizophrenia (Benes et al., 1996a) and in the area dentata (granule cell layer), CA4, CA3, subiculum, and presubiculum in the hippocampus of subjects with schizophrenia (Benes et al., 1996b). ³H-flunitrazepam-labeled benzodiazepine binding sites were also increased in the striatum oriens of CA3, subiculum, and presubiculum in the hippocampus of subjects with schizophrenia (Benes et al., 1997).

Quantitation of GABA receptor binding sites has also been performed in brains of subjects with autism (Blatt et al. 2001; Oblak et al. 2010, 2011). Reduced ³H-muscimol-labeled GABA_A receptor binding sites and ³H-flunitrazepam-labeled benzodiazepine binding sites have been found in hippocampus of subjects with autism (Blatt et al., 2001). Finally reduced ³H-muscimol-labeled GABA_A receptor binding sites have been observed in the posterior cingulate cortex and fusiform gyrus of subjects with autism (Oblak et al., 2011). Additionally, reduced GABA_B receptor binding sites, as measured using ³H-CGP54626-labeling, have also been identified in the posterior cingulate cortex and fusiform gyrus of subjects with autism (Oblak et al., 2010).

Taken together, these studies have demonstrated widespread changes in GABA binding sites in brains of subjects with schizophrenia and autism. Additional biochemical studies have identified changes in expression of mRNA and protein for individual GABA_A and GABA_B receptor subunits, primarily by Fatemi and collaborators in four psychiatric disorders spanning 17 GABA_{A and B} receptors in cerebellum, BA9, and parietal cortices of subjects with autism and primarily in cerebellum of subjects with schizophrenia, bipolar disorder, and major depression which are described below. Results from these studies will be contrasted with those reported by other investigators in the ensuring sections (vide infra).

2.3. Expression of Individual GABA Receptor Subunits in Brains of Subjects with Major Psychiatric Disorders

The gene that codes for GABRα1 (GABRA1) is located at 5q34-q35 (Buckle et al., 1989). In rat brain, mRNA for GABRα1 is distributed throughout the brain including the neocortex, hippocampus, globus pallidus, medial septum, thalamus, and cerebellum (Wisden et al., 1992). The GABRα1 subunit is expressed in most GABA_A receptors. Our laboratory has found that GABRα1 protein level was significantly increased in lateral cerebellum of subjects with major depression while there were no changes in subjects with schizophrenia or bipolar disorder (Fatemi et al., 2013b) (Table 1). Moreover, we found that mRNA for GABRα1 is significantly reduced in lateral cerebella of subjects with schizophrenia and major depression (Fatemi et al., 2013b; Table 1). Three studies have identified reduced expression of GABRα1 mRNA in prefrontal cortex (PFC) and dorsolateral prefrontal cortex (DLPC) of subjects with schizophrenia (Beneyto et al., 2011; Glausier and Lewis, 2011; Hashimoto et al., 2008), while two other studies found no change in GABRα1 mRNA in the same regions (Akbarian et al., 1995; Duncan et al., 2010) (Table 1). In subjects with autism, we determined that GABRα1 protein is also significantly reduced in superior frontal cortex [Brodmann Area 9 (BA9)], parietal cortex (BA40), and cerebellum without any change in mRNA levels (Fatemi et al., 2009a) (Table 1).

Table 1.

Summary of GABA_A and GABA_B receptor expression changes in major psychiatric disorders

Receptor Subunit	Autism	Schizophrenia	Bipolar Disorder	Major Depression	Reference
GABRα1	— mRNA, ↓ protein in BA9, BA40, Cer			↑ protein in Cer	Fatemi et al., 2009a, 2013b
		No change in protein in Cer; ↓ mRNA in Cer	No change in mRNA or protein in Cer	↓ mRNA in Cer	Fatemi et al., 2013b
		↓ mRNA in DLPFC			Beneyto et al., 2011; Hashimoto et al., 2008
		↓ mRNA in PFC pyramidal cells			Glausier and Lewis, 2011
		No change in mRNA in PFC			Akbarian et al., 1995
		No change in mRNA in DLPFC			Duncan et al., 2010
GABRα2	↓ protein in BA40; no change in Cer or BA9	↑ protein in Cer	↑ protein in Cer	↑ protein in Cer	Fatemi et al., 2009a, 2013b
	↓ mRNA in BA9; ↑ mRNA in Cer; no change in BA40	↓ mRNA in Cer	No change in mRNA in Cer	No change in mRNA in Cer	Fatemi et al., 2013b, 2014
		↑ mRNA in DLPFC			Beneyto et al., 2011
		No change in mRNA in DLPFC			Duncan et al., 2010
		No change in mRNA in PFC			Akbarian et al., 1995
GABRα3	↓ protein in BA40; no change in BA9 or Cer				Fatemi et al., 2009a
	↓ mRNA in BA9 and BA40; ↑ mRNA in Cer				Fatemi et al., 2014
		No change in mRNA or protein in Cer	No change in mRNA or protein in Cer	No change in mRNA or protein in Cer	Fatemi et al., 2013b
		No change in mRNA in DLPFC			Beneyto et al., 2011; Duncan et al., 2010
GABRα4	↓ protein in BA9; no change in BA40 or Cer				Fatemi et al., 2010b
	↓ mRNA in BA9; ↑ mRNA in Cer; no change in BA40				Fatemi et al., 2010b
		ND	ND	ND	Fatemi et al., 2013b
		↓ gene transcript in DLPFC;			Hashimoto et al., 2008
		medication-dependent ↓ mRNA in in BA9			Maldonado-Avilés et al., 2009
		No change in mRNA in DLPFC			Duncan et al., 2010
GABRα5	↓ protein in BA9 and BA40; no change in Cer				Fatemi et al., 2010b
	↓ mRNA in BA9; ↑ mRNA in Cer; no change in BA40				Fatemi et al., 2010b
		No change in mRNA or protein in Cer	No change in mRNA or protein in Cer	No change in mRNA or protein in Cer	Fatemi et al., 2013b
		↓ mRNA in DLPFC			Beneyto et al., 2011; Duncan et al., 2010
		No change in mRNA in PFC			Akbarian et al., 1995
GABRα6	↓ protein in BA9; no change in BA40 or Cer			↑ protein in Cer	Fatemi et al., 2013b, 2014
		↑ mRNA in Cer			Bullock et al., 2008
	↑ mRNA in BA9; ↓ mRNA in Cer	No change in mRNA or protein in Cer	No change in mRNA or protein in Cer	No change in mRNA or protein in Cer	Fatemi et al., 2013b, 2014
GABRβ1	↓ protein in BA9; no change in BA40 or Cer	↓ protein in Cer	↓ protein in Cer	↓ protein in Cer	Fatemi et al., 2010b
	↓ mRNA in BA9; ↑ mRNA in Cer; no change in BA40	No change in mRNA in Cer	No change in mRNA in Cer	No change in mRNA in Cer	Fatemi et al., 2010b, 2013b
		No change in mRNA in DLPFC			Beneyto et al., 2011
		No change in mRNA in PFC			Akbarian et al., 1995
GABRβ2	↓ protein in BA9; no change in BA40 or Cer	No change in mRNA or protein in Cer	↓ protein in Cer; no change in mRNA	No change in mRNA or protein in Cer	Fatemi et al., 2013b, 2014
	↓ mRNA in Cer; no change in BA9 or BA40				Fatemi et al., 2014
		↓ mRNA in DLPFC			Beneyto et al., 2011
		No change in mRNA in PFC			Akbarian et al., 1995
GABRβ3	↓ protein in BA40 and Cer; no change in BA9	↑ mRNA and no change in protein in Cer	No change in mRNA or protein in Cer	No change in mRNA or protein in Cer	Fatemi et al., 2009a, 2013b
		No change in mRNA in Cer			Bullock et al., 2008
	↓ protein in BA9				Samaco et al., 2005
	↓ mRNA in BA9; ↑ mRNA in Cer; no change in BA40	↑ mRNA in Cer			Fatemi et al., 2013b, 2014
		↓ gene transcript in DLPFC			Hashimoto et al., 2008
		No change in mRNA in DLPFC			Beneyto et al., 2011
GABRδ	↓ protein in BA9; no change in BA40 or Cer; No change in mRNA in BA9, BA40, Cer	No change in mRNA or protein in Cer	No change in mRNA or protein in Cer	No change in mRNA or protein in Cer	Fatemi et al., 2013b, 2014
		↑ mRNA in Cer			Bullock et al., 2008
		↓ mRNA in DLPFC			Hashimoto et al., 2008
		↓ mRNA in BA9			Maldonado-Avilés et al., 2009
GABRε	↓ protein in BA9; no change in BA40 or Cer; No change in mRNA in BA9, BA40, Cer	↑ protein in Cer; no change in mRNA in Cer	↑ protein in Cer; no change in mRNA in Cer	↑ protein in Cer; no change in mRNA in Cer	Fatemi et al., 2013b, 2014
GABRγ2	↓ protein in BA9; No change in mRNA in BA9, BA40, Cer	No change in mRNA or protein in Cer	No change in mRNA or protein in Cer	No change in mRNA or protein in Cer	Fatemi et al., 2013b, 2014
	↑ mRNA in Cer; no change in BA9 and BA40				Fatemi et al., 2014
		↓ gene transcript in DLPFC			Hashimoto et al., 2008
		No change in mRNA in PFC			Akbarian et al., 1995
GABRγ3	No change in protein in BA9, BA40, Cer	No change in mRNA or protein in Cer	No change in mRNA or protein in Cer	↑ protein in Cer; no change in mRNA in Cer	Fatemi et al., 2013b, 2014
	↓ mRNA in BA9; ↑ mRNA in BA40 and Cer				Fatemi et al., 2014
GABRπ	No change in mRNA or protein in BA9, BA40, and Cer	No change in mRNA or protein in Cer	No change in mRNA or protein in Cer	No change in mRNA or protein in Cer	Fatemi et al., 2013b, 2014
GABRθ	No change in protein in BA9, BA40, Cer	↓ protein in BA9 and Cer	↓ protein in BA9 and Cer	↓ protein in Cer	Fatemi et al., 2013a, 2014
	↓ mRNA in BA9; ↑ mRNA in Cer; no change in BA40	↑ mRNA in BA9; ↓ mRNA in Cer	No change in mRNA in Cer	No change in mRNA in Cer	Fatemi et al., 2013a, 2014
GABRρ2	↓ protein in BA9; No change in mRNA in BA9, BA40, Cer	No change in mRNA or protein in BA9 or Cer	↑ protein in Cer; No change in protein in BA9	↑ protein in Cer	Fatemi et al., 2013a, 2014
			↑ mRNA in BA9 and Cer	No change in mRNA in Cer	Fatemi et al., 2013a
		↓ mRNA in blood of subjects on risperidone			Ota et al., 2014
GABBR1	↓ protein in BA9, BA40, Cer	↓ protein in BA9 and Cer; no change in mRNA in Cer	↓ protein in BA9 and Cer; no change in mRNA in Cer	↓ protein in Cer; no change in mRNA in Cer	Fatemi et al., 2009b, 2011b
	↑ mRNA in BA40; ↓ mRNA in Cer				Fatemi et al., 2010b, 2014
GABBR2	↓ protein in Cer; No change in BA9 or BA40; no mRNA changes in BA9, BA40, Cer	↓ protein in Cer; no change in mRNA in Cer	↓ protein in Cer; no change in mRNA in Cer	↓ protein in Cer; no change in mRNA in Cer	Fatemi et al., 2009b, 2011b

↑, significant increase; ↓ significant decrease; —, no change; BA9, Brodmann Area 9; Cer, cerebellum; DLPFC, dorsolateral prefrontal cortex; ND, not determined; PFC, prefrontal cortex

The GABRα2 gene (GABRA2), is localized to 4q13-p12 (Buckle et al., 1989). In situ hybridization experiments have identified GABRα2 mRNA in multiple regions of rat brain including the neocortex, hippocampus, hypothalamus, and cerebellum (Laurie et al., 1992; Wisden et al., 1992). We observed that GABRα2 protein level was increased in lateral cerebella of subjects with schizophrenia, bipolar disorder, and major depression (Fatemi et al., 2013b) (Table 1). Our laboratory further identified reduced GABRα2 mRNA expression in lateral cerebella of subjects with schizophrenia. Beneyto et al (2011) also found increased mRNA expression for GABRα2 in DLPFC in subjects with schizophrenia (Table 1). Two other studies found no change in GABRα2 mRNA expression in PFC and DLPFC (Akbarian et al., 1995; Duncan et al., 2010) (Table 1). In subjects with autism, we found that GABRα2 protein was significantly reduced in BA40 (Fatemi et al., 2009a) while its mRNA expression was significantly reduced in BA9 and significantly increased in cerebellum (Fatemi et al., 2014) (Table 1). The GABRα2 subunit has been implicated in depression as GABRα2 KO mice display depressive behavior (i.e., increased immobility) in response to the forced swim test (FST) and the tail suspension test (TST) (Vollenweider et al., 2011).

GABRA3, the gene that codes for GABRα3 is located at Xq28 (Buckle et al., 1989). In rat brain, mRNA for GABRα3 has been found in multiple regions including septum, thalamus, cerebellum, and neocortex (Laurie et al., 1992; Wisden et al., 1992). We observed that GABRα3 protein and mRNA were not altered in lateral cerebellum of subjects with schizophrenia and mood disorders (Table 1). Similarly, other groups found no difference in mRNA for GABRα3 in DLPFC of subjects with schizophrenia vs. controls (Beneyto et al., 2011; Duncan et al., 2010) (Table 1). In BA40 of subjects with autism, we determined that both mRNA and protein levels for GABRα3 were significantly reduced (Fatemi et al., 2009a, 2014) (Table 1). We also observed that mRNA for GABRα3 was also significantly reduced in BA9 and increased in cerebellum of subjects with autism (Fatemi et al., 2014). Interestingly, GABRα3 KO mice display sensorimotor gating deficits as measured by prepulse inhibition (PPI) of the acoustic startle response (Yee et al., 2005), a deficit that is common to subjects with autism and schizophrenia.

The gene that codes for GABRα4 (GABRA4) is located at 4p14-q12 (McLean et al., 1995), and is a member of the novel group of extrasynaptic GABA_A receptors (Errington, 2014). mRNA for GABRα4 shows particularly strong expression in the hippocampus and thalamus of rat brain as well as caudate, putamen and neocortex (Laurie et al., 1992; Wisden et al., 1992). We have not yet determined if there are any changes in GABRα4 mRNA or protein in subjects with schizophrenia and mood disorders. Duncan et al. (2010) found no change in mRNA expression for GABRα4 in DLPFC of subjects with schizophrenia (Table 1). Hashimoto et al., (2008) determined a significant decrease in gene transcript for this receptor subunit in DLPFC of subjects with schizophrenia. However, this change was not verified at the level of mRNA by either qRT-PCR or another technique (Hashimoto et al., 2008). Maldonado-Avilés et al (2009) found reduced mRNA expression in BA9 of subjects with schizophrenia who were taking benzodiazepines, mood stabilizers, or antidepressants at the time of death, suggesting a medication-specific effect on GABRα4 expression (Table 1). Our laboratory observed that both mRNA and protein levels for GABRα4 were reduced in BA9 of subjects with autism (Fatemi et al., 2010a) (Table 1). In contrast, we have observed significantly increased mRNA for GABRα4 in cerebellum of subjects with autism (Fatemi et al., 2010a).

The GABRα5 gene (GABRA5) is localized to 15q11.2-q13 (Knoll et al., 1992) and is considered one of the extrasynaptic GABA_A receptors involved in tonic inhibitory neurotransmission (Wafford, 2014). mRNA for GABRα5 is expressed in the hippocampus, olfactory bulb, and hypothalamus in rat brain (Laurie et al., 1992; Wisden et al., 1992) as well as amygdala (Martin and Huntsman, 2014). We did not observe changes in GABRα5 mRNA or protein expression in lateral cerebellum of subjects with schizophrenia, bipolar disorder, or major depression (Table 1). However, previous reports have identified reduced GABRα5 mRNA in DLPFC of subjects with schizophrenia (Beneyto et al., 2011; Duncan et al., 2010), while Akbarian et al (1995) found no change in PFC (Table 1). In BA9 of subjects with autism, we observed that GABRα5 mRNA and protein were significantly reduced (Fatemi et al., 2010a) (Table 1). In contrast, GABRα5 mRNA was significantly increased in cerebellum of subjects with autism (Fatemi et al., 2010a). Improvements in cognitive behavior in a mouse model of Down Syndrome following the use of inverse agonists of GABRα5 (i.e., α5IA or RO4938581) suggest that this receptor subunit could be targeted to improve cognitive deficits associated with intellectual impairment (Braudeau et al., 2011; Martinez-Cue et al., 2013). Mice that display reduced GABRα5 in the hippocampus also show deficits in latent inhibition (Gerdjikov et al., 2008), further implicating this subunit in cognition. Additionally, because of the importance of this receptor in cognition, it has also been a focus of therapeutic intervention for both Alzheimer’s disease and schizophrenia (Wafford, 2014).

GABRA6, the gene that codes for GABRα6 subunit, is located at 5q31.1-q35 (Hicks et al., 1994) and is another member of the extrasynaptic GABA_A receptors (Wafford, 2014). GABRα6 mRNA has been found to localize exclusively to cerebellar granule neurons in rat brain (Laurie et al., 1992; Wisden et al., 1992). We observed a significant increase in GABRα6 protein in lateral cerebella of subjects with major depression while there were no changes observed in the lateral cerebella of subjects with schizophrenia and bipolar disorder (Fatemi et al., 2013b) (Table 1). GABRα6 mRNA has been shown to be increased in lateral cerebella of subjects with schizophrenia (Bullock et al., 2008). In BA9 of subjects with autism, we observed that GABRα6 protein level was significantly reduced while mRNA was significantly increased (Fatemi et al., 2014) (Table 1). In cerebellum of subjects with autism, we determined that GABRα6 mRNA level was significantly reduced (Fatemi et al., 2014) (Table 1).

The cytogenetic location of the gene which codes for GABRβ1 (GABRB1) is 4p12 (Kirkness et al., 1991). GABRβ1 mRNA localizes to numerous regions of rat brain with particularly strong expression in hippocampus, amygdala, and cerebellum (Wisden et al., 1992). Findings from our laboratory identified reduced protein expression of GABRβ1 in lateral cerebellum of subjects with schizophrenia, bipolar disorder, and major depression (Fatemi et al., 2013b) (Table 1) with no change in level of mRNA in cerebellum of subjects with schizophrenia and mood disorders (Table 1). Other laboratories have found no change in mRNA for GABRβ1 expression in PFC and DLPFC of subjects with schizophrenia, confirming our data in cerebellum (Akbarian et al., 1995; Beneyto et al., 2011) (Table 1). We observed that GABRβ1 mRNA and protein levels were significantly reduced in BA9 of subjects with autism (Fatemi et al., 2010a) (Table 1), while mRNA for GABRβ1 was significantly increased in cerebellum of subjects with autism (Fatemi et al., 2010a) (Table 1).

The gene that codes for GABRβ2 (GABRB2) is localized to 5q34-q35 (Russek and Farb, 1994). GABRβ2 mRNA has been found in olfactory bulb, neocortex, globus pallidus, thalamus and cerebellum of rat brain (Laurie et al., 1992; Wisden et al., 1992). In lateral cerebella of subjects with bipolar disorder, our laboratory observed a significant reduction in protein expression for GABRβ2 with no change in mRNA or protein for subjects with schizophrenia or major depression (Fatemi et al., 2013b) (Table 1). Similarly, mRNA for GABRβ2 has been found to be reduced in DLPFC of subjects with schizophrenia (Beneyto et al., 2011) while a separate study found no reduction in GABRβ2 mRNA in PFC of subjects with schizophrenia (Akbarian et al., 1995) similar to our observations (Table 1). We also observed that GABRβ2 protein was significantly reduced in BA9 of subjects with autism while in cerebellum, GABRβ2 mRNA level was significantly increased (Fatemi et al, 2014) (Table 1).

The GABRβ3 gene (GABRB3), is located at 15q11.2-q13. In rat brain, GABRβ3 mRNA has been localized to olfactory bulb, neocortex, hippocampus, hypothalamus, and cerebellum (Laurie et al., 1992; Wisden et al., 1992). While we determined that GABRβ3 protein level was unchanged in lateral cerebella of subjects with schizophrenia and mood disorders, its mRNA expression was significantly increased in subjects with schizophrenia (Fatemi et al., 2013b) (Table 1). There have been contradictory findings regarding mRNA for GABRβ3 in DLPFC of subjects with schizophrenia with one study reporting reduced expression in the transcript for β3 (Hashimoto et al., 2008) and a separate study finding no change in its mRNA (Beneyto et al., 2011). Bullock et al. (2008) found no change in GABRβ3 mRNA expression in lateral cerebella of subjects with schizophrenia (Bullock et al., 2008). We identified significant reductions in GABRβ3 protein in cerebellum and BA40 of subjects with autism (Fatemi et al., 2009a) (Table 1). Similarly, reduced GABRβ3 protein expression has also been reported in PFC of subjects with autism (Samaco et al., 2005). Our laboratory determined that mRNA for GABRβ3 was significantly reduced in BA9 of subjects with autism while significantly increased in cerebella of subjects with autism (Fatemi et al., 2014). GABRβ3 KO mice display deficits in learning and memory while exhibiting presence of epilepsy (DeLorey et al., 1998, 2008), suggesting that reduced expression of this subunit is relevant to the pathology of autism. Indeed, gene abnormalities in the 15q11-q13 locus, which includes GABRα5, GABRβ3, and GABRγ3 are found in 1–4% of subjects with autism (McCauley et al., 2004; Schroer et al., 1998). Epigenetic mechanisms and monoallelic expression of GABRβ3 may explain reduced expression of this subunit in autism (Hogart et al., 2007; Samaco et al., 2005).

GABRD, the gene that codes for the GABRδ subunit is located at 1p36.3 (Emberger et al. 2000), and exclusively localized extrasynaptically, exhibits higher sensitivity to GABA agonism, is activated at less than one micromolar (1 µM) GABA concentration, and shows lower degree of receptor desensitization (Wafford, 2014). In rat brain, GABRδ mRNA has been localized to the cerebellum, olfactory bulb, dorso-lateral geniculate nucleus, thalamus, striatum and the nucleus accumbens (Pirker et al. 2000; Schwarzer et al. 2001; Wisden et al. 1992). We did not observe any changes in GABRδ mRNA or protein expression in subjects with schizophrenia or mood disorders (Fatemi et al., 2013b). Other groups have demonstrated reduced expression of GABRδ mRNA in BA9 (Maldonado-Avilés et al., 2009) and DLPFC (Hashimoto et al., 2008) of subjects with schizophrenia (Table 1). Additionally, Bullock et al. (2008) found increased expression of GABRδ mRNA in lateral cerebella of subjects with schizophrenia. In BA9 of subjects with autism, we observed significantly reduced expression of the GABRδ subunit expression (Fatemi et al., 2014) (Table 1).

GABRE, the gene that codes for the GABRε subunit, clusters at Xq28 with the genes that code for GABRα3 and GABRθ (Korpi et al., 2002). GABRε mRNA has been localized to the septum, thalamus, hypothalamus, and amygdala in rat brain and is often coexpressed with mRNA for GABRθ (Moragues et al., 2002). We discovered that GABRε protein expression was significantly increased in lateral cerebella of subjects with schizophrenia, bipolar disorder, or major depression without any change in the level of its mRNA, indicating that protein changes observed were due to posttranslational processes (Fatemi et al., 2013b) (Table 1). Protein for GABRε however, was observed to be significantly reduced in BA9 of subjects with autism (Fatemi et al., 2014) (Table 1). Presence of the GABRε subunit (or its increased expression) is known to reduce GABA_A receptor sensitivity to benzodiazepines and anesthetics (Belujon et al., 2009; Kasparov et al., 2001; Thompson et al., 2002) and altered expression may impair responses to these pharmacologic agents.

GABRγ2 is coded by the GABRG2 gene, which is located at 15q31.1-q33.1 (Wilcox et al. 1992). There is widespread expression of GABRγ2 mRNA in both the developing and adult rat brain including the olfactory bulb, amygdala, septum, hypothalamus, and cerebellum (Laurie et al. 1992; Pirker et al. 2000; Schwarzer et al. 2001; Wisden et al. 1992). We observed no changes in protein or mRNA expression for GABRγ2 in lateral cerebellum of subjects with schizophrenia or mood disorders (Fatemi et al., 2013b) (Table 1). Akbarian et al (1995) similarly observed no change in GABRγ2 mRNA in PFC of subjects with schizophrenia, while Hashimoto et al (2008) found reduced expression of GABRγ2 transcript in DLPFC as measured by DNA microarray (Table 1). However, this finding was not verified by qRT-PCR (Hashimoto et al., 2008). We found that GABRγ2 protein was significantly reduced in BA9 of subjects with autism while its mRNA was significantly increased in cerebellum of subjects with autism (Fatemi et al., 2014) (Table 1). Similar to the α2 subunit, the γ2 subunit has been implicated in depression. Mice with selective heterozygous deletion of GABRγ2, display increased immobility following the forced swim test or the novelty-suppressed feeding test (Earnheart et al., 2007) and exhibit increased anxiety (Crestani et al., 1999).

The gene that codes for GABRγ3 (GABRG3) localizes to 15q11.2-q13 where it clusters with the genes for GABRα5 and GABRβ3 (Greger et al., 1995). mRNA for GABRγ3 has been identified in multiple areas of the rat brain including neocortex, caudate, putamen, nucleus accumbens, and cerebellum (Laurie et al., 1992; Wisden et al., 1992). We determined that GABRγ3 protein was significantly and specifically increased in lateral cerebella of subjects with major depression with no change observed in either schizophrenia or bipolar disorder (Fatemi et al., 2013b). mRNA for GABRγ3 was also observed to be significantly increased in cerebellum and BA40 of subjects with autism, while it was significantly reduced in BA9 of subjects with autism (Fatemi et al., 2014).

GABRπ is coded by the GABRP gene, which is located at 5q32-q33 (Whiting et al., 1999). GABRπ mRNA has been shown to be expressed in human hippocampus and temporal cortex (Hedblom and Kirkness, 1997). We did not identify significant alterations in GABRπ mRNA or protein levels in subjects with schizophrenia, bipolar disorder, major depression, or autism (Fatemi et al., 2013b, 2014) (Table 1).

GABRQ, the gene that codes for GABRθ, clusters with the gene for GABRε at Xq28 (Korpi et al. 2002). In human brain, GABRθ is expressed in multiple regions including the cerebral cortex (cingulate and piriform areas), hypothalamus, amygdala, hippocampus, substantia nigra, locus coeruleus, and hindbrain (Bonnert et al. 1999). A recent study by our laboratory discovered significant reductions in GABRθ protein in lateral cerebella of subjects with schizophrenia, bipolar disorder, and major depression and in BA9 of subjects with schizophrenia and bipolar disorder (Fatemi et al., 2013a) (Table 1). We also determined that GABRθ mRNA expression was reduced in lateral cerebella of subjects with schizophrenia and increased in BA9 of subjects with schizophrenia (Fatemi et al., 2013a). Our laboratory has observed that GABRθ mRNA was significantly reduced in BA9 and significantly increased in cerebellum of subjects with autism (Fatemi et al., 2014).

GABRR2, the gene that codes for GABRρ2, is located at 6q15 (Ma et al., 2005). In rat brain, mRNA for GABRρ2 has been identified in multiple regions including prefrontal cortex, hippocampus, and cerebellum (Alakuijala et al., 2005; Boue-Grabot et al., 1998). We observed significantly increased expression of GABRρ2 in lateral cerebella of subjects with bipolar disorder and major depression while there was no change in BA9 of subjects with schizophrenia and bipolar disorder (Fatemi et al., 2013a) (Table 1). mRNA expression for GABRρ2 was significantly increased in both lateral cerebella and BA9 of subjects with bipolar disorder (Fatemi et al., 2013a). GABA_A receptors that include GABRρ2 are known to be involved in mediating phasic inhibitory GABAergic transmission at interneuron-Purkinje cell synapses in cerebellum (Harvey et al., 2006). Altered expression of GABRρ2 is likely to change inhibitory neurotransmission. A recent report shows reduced expression of GABRρ2 mRNA in blood of subjects with schizophrenia who were taking risperidone (Ota et al., 2014). Moreover, Ota et al (2014) found that higher GABRρ2 mRNA levels were associated with more severe Positive and Negative Syndrome Scale (PANSS) negative, psychopathology and total score ratings, however, these comparisons did not remain significant following Bonferroni’s correction. The authors hypothesized that reduction in GABRρ2 mRNA levels resulted in a better response to antipsychotic treatment (Ota et al., 2014). Taken together, the GABRρ2 subunit may be a highly important biomarker in the pathology and treatment of schizophrenia, bipolar disorder, and major depression. We have also observed significantly reduced GABRρ2 protein level in BA9 of subjects with autism (Fatemi et al., 2014).

The cytogenetic location of GABBR1 is 6p21.3 (Goei et al., 1998) while the GABBR2 gene is located at 9q22.1 (Martin et al., 1999). Both GABBR1 and GABBR2 show strong expression in piriform cortex, hippocampus, olfactory bulb, and medial habenula with moderate expression in thalmic nuclei and cortex (Li et al., 2003). We have found that both GABBR1 and GABBR2 proteins were significantly reduced in lateral cerebella of subjects with schizophrenia, bipolar disorder, and major depression (Fatemi et al., 2011b) (Table 1). More recently, we have identified significant reductions in GABBR1 protein in BA9 of subjects with schizophrenia and bipolar disorder (Fatemi, unpublished observations). In subjects with autism, we observed that GABBR1 protein was significantly reduced in cerebellum, BA40, and BA9 while its mRNA expression was reduced in cerebellum, but increased in BA40 (Fatemi et al., 2009b, 2010a). We also determined that GABBR2 protein was significantly reduced in cerebellum of subjects with autism (Fatemi et al., 2009b) (Table 1). Altered expression of GABBR1 and GABBR2 have been associated with seizure disorder in animal models (Han et al., 2006; Princivalle et al., 2003; Straessle et al., 2003), which is relevant to subjects with autism where the presence of comorbid seizure disorder ranges from 4–44% (Tuchman and Rapin, 2002).

Collectively, postmortem studies by our laboratory have demonstrated detailed biochemical analysis of GABA_A and GABA_B receptor subunit distribution in cerebellum of subjects with schizophrenia, bipolar disorder, and major depression, showing significant alterations in several subunits as follows: 1) significant downregulation in protein levels for β1, θ, R1, and R2 and significant upregulation in protein for α2 and ε in subjects with schizophrenia (Table 2); 2) significant downregulation in protein levels for β1, β2, θ, R1, and R2 and significant upregulation in protein for α2, ε, and ρ2 in subjects with bipolar disorder (Table 3); 3) significant downregulation in protein levels for β1, θ, R1, and R2 and significant upregulation in protein for α1, α2, α6, ε, ρ2, and γ3 in subjects with major depression (Table 4); 4) significant downregulation in mRNA levels for α1, α2, and θ and significant upregulation in mRNA for β3 in subjects with schizophrenia (Table 2); 5) significant upregulation in mRNA for ρ2 in subjects with bipolar disorder (Table 3); and 6) significant downregulation in mRNA levels for α1 in subjects with major depression (Table 4) (Fatemi et al., 2013a,b). There are currently no other published studies correlating protein and mRNA levels for 17 GABA receptor subunits in cerebella of subjects with schizophrenia, bipolar disorder, or major depression with the exception of a study by Bullock et al (2008) who measured mRNA levels for three GABA receptor subunits. However, several groups have investigated receptor levels using different techniques, (i.e. in situ hybridization, microarray, qRT-PCR), and different brain areas (i.e., DLPFC, PFC) and demonstrated increases or decreases in mRNA levels for various GABA receptors limited to schizophrenic samples (Akbarian et al., 1995; Beneyto et al., 2011; Bullock et al., 2008; Duncan et al., 2010; Glausier and Lewis, 2011; Hashimoto et al., 2008; Maldonado-Aviles et al., 2009) (Tables 2–4). However, it is difficult to compare and contrast our data vs. others due to different brain sites and different techniques used. Experiments are currently underway in our laboratory to expand our studies using tissue from the prefrontal cortex in subjects with schizophrenia and bipolar disorder vs. healthy controls.

Table 2.

Summary of changes in GABA receptor mRNA and protein expression in subjects with schizophrenia

	Protein		mRNA
Receptor	Fatemi	Others	Fatemi	Othersa
GABRα1	— Cer	ND	↓ Cer	↓/— DLPFC, ↓/— PFC
GABRα2	↑ Cer	ND	↓ Cer	↑/— DLPFC, — PFC
GABRα3	— Cer	ND	— Cer	— DLPFC
GABRα4	ND	ND	ND	↓/— DLPFC, ↓ BA9
GABRα5	— Cer	ND	— Cer	↓ DLPFC, — PFC
GABRα6	— Cer	ND	— Cer	↑ Cer
GABRβ1	↓ Cer	ND	— Cer	— DLPFC, — PFC
GABRβ2	— Cer	ND	— Cer	↓ DLPFC, — PFC
GABRβ3	— Cer	ND	↑ Cer	↓/— DLPFC; — Cer
GABRδ	— Cer	ND	— Cer	↓DLPFC, ↓BA9; ↑Cer
GABRε	↑ Cer	ND	— Cer	ND
GABRγ2	— Cer	ND	— Cer	↓ DLPFC, — PFC
GABRγ3	— Cer	ND	— Cer	ND
GABRπ	— Cer	ND	— Cer	ND
GABRθ	↓ BA9 and Cer	ND	↑ BA9, ↓ Cer	ND
GABRρ2	— Cer	ND	— Cer and BA9	↓ Blood
GABRR1	↓ BA9 and Cer	ND	ND	ND
GABRR2	↓ Cer	ND	ND	ND

↑, increase; ↓, decrease; —, no change; BA9, Brodmann Area 9,; Cer, cerebellum; DLPFC, dorsolateral prefrontal cortex; ND, not determined; PFC, prefrontal cortex

^aPlease see Table 1 for specific references

Table 3.

Summary of changes in GABA receptor mRNA and protein expression in subjects with bipolar disorder

	Protein		mRNA
Receptor	Fatemi	Others	Fatemi	Others
GABRα1	— Cer	ND	— Cer	ND
GABRα2	↑ Cer	ND	— Cer	ND
GABRα3	— Cer	ND	— Cer	ND
GABRα4	ND	ND	ND	ND
GABRα5	— Cer	ND	— Cer	ND
GABRα6	— Cer	ND	— Cer	ND
GABRβ1	↓ Cer	ND	— Cer	ND
GABRβ2	↓ Cer	ND	— Cer	ND
GABRβ3	— Cer	ND	— Cer	ND
GABRδ	— Cer	ND	— Cer	ND
GABRε	↑ Cer	ND	— Cer	ND
GABRγ2	— Cer	ND	— Cer	ND
GABRγ3	— Cer	ND	— Cer	ND
GABRπ	— Cer	ND	— Cer	ND
GABRθ	↓ BA9 and Cer	ND	— BA9 and Cer	ND
GABRρ2	↑ Cer	ND	↑ BA9 and Cer	ND
GABRR1	↓ BA9 and Cer	ND	ND	ND
GABRR2	↓ Cer	ND	ND	ND

↑, increase; ↓, decrease; —, no change; BA9, Brodmann Area 9,; Cer, cerebellum; DLPFC, dorsolateral prefrontal cortex; ND, not determined; PFC, prefrontal cortex

Table 4.

Summary of changes in GABA receptor mRNA and protein expression in subjects with major depression

	Protein		mRNA
Receptor	Fatemi	Others	Fatemi	Others
GABRα1	↑ Cer	ND	↓ Cer	ND
GABRα2	↑ Cer	ND	— Cer	ND
GABRα3	— Cer	ND	— Cer	ND
GABRα4	ND	ND	ND	ND
GABRα5	— Cer	ND	— Cer	ND
GABRα6	↑ Cer	ND	— Cer	ND
GABRβ1	↓ Cer	ND	— Cer	ND
GABRβ2	— Cer	ND	— Cer	ND
GABRβ3	— Cer	ND	— Cer	ND
GABRδ	— Cer	ND	— Cer	ND
GABRε	↑ Cer	ND	— Cer	ND
GABRγ2	— Cer	ND	— Cer	ND
GABRγ3	↑ Cer	ND	— Cer	ND
GABRπ	— Cer	ND	— Cer	ND
GABRθ	↓ Cer	ND	— Cer	ND
GABRρ2	↑ Cer	ND	— Cer	ND
GABRR1	↓ Cer	ND	ND	ND
GABRR2	↓ Cer	ND	ND	ND

↑, increase; ↓, decrease; —, no change; BA9, Brodmann Area 9,; Cer, cerebellum; DLPFC, dorsolateral prefrontal cortex; ND, not determined; PFC, prefrontal cortex

To summarize our findings in autism with regard to GABA_A and GABA_B receptor protein expression, we identified: 1) significant downregulation for protein levels of α1, β3, R1, and R2 in cerebellum; 2) significant downregulation of α1, α2, α3, α5, β3, and R1 proteins in BA40; and 3) significant downregulation of α1, α4, α5, α6, β1, β2, δ, ε, γ2, ρ2, and R1 proteins in BA9 (Table 5). With regard to mRNA expression we discovered: 1) significant upregulation of α2, α3, α4, α5, β1, β3, γ2, γ3, and θ mRNA and significant downregulation of α6 and R1 mRNA in cerebellum; 2) significant upregulation of γ3 and R1 mRNAs and significant downregulation of α3 mRNA in BA40; and 3) significant uprgulation of α6 mRNA and significant downregulation of α2, α3, α4, α5, β1, β3, γ3, and θ mRNAs in BA9 (Table 5). Ultimately, it is the protein expression that has functional consequences and our findings of consistent, global reductions in the expression of GABA receptor subunits in brains of subjects with autism has the potential to alter the excitatory/inhibitory balance in the brain and help to explain learning deficits and presence of seizure disorder in this devastating neurodevelopmental disorder (Blatt and Fatemi, 2011).

Table 5.

Summary of changes in GABA receptor mRNA and protein expression in subjects with autism

	Protein		mRNA
Receptor	Fatemi	Othersa	Fatemi	Others
GABRα1	↓ BA9, BA40, Cer	ND	— BA9, BA40, Cer	ND
GABRα2	↓ BA40; — BA9, Cer	ND	↓BA9, ↑Cer, —BA40	ND
GABRα3	↓ BA40; — BA9, Cer	ND	↓BA9, BA40; ↑Cer	ND
GABRα4	↓ BA9; — BA40, Cer	ND	↓BA9, ↑Cer, —BA40	ND
GABRα5	↓ BA9, BA40; — Cer	ND	↓BA9, ↑Cer, —BA40	ND
GABRα6	↓ BA9; — BA40, Cer	ND	↑BA9, ↓Cer, —BA40	ND
GABRβ1	↓ BA9; — BA40, Cer	ND	↓BA9, ↑Cer, —BA40	ND
GABRβ2	↓ BA9; — BA40, Cer	ND	↓Cer; —BA9, BA40	ND
GABRβ3	↓ BA40, Cer; — BA9	↓ BA9	↓BA9, ↑Cer, —BA40	ND
GABRδ	↓ BA9; — BA40, Cer	ND	— BA9, BA40, Cer	ND
GABRε	↓ BA9; — BA40, Cer	ND	— BA9, BA40, Cer	ND
GABRγ2	↓ BA9; — BA40, Cer	ND	↑Cer; —BA9, BA40	ND
GABRγ3	— BA9, BA40, Cer	ND	↓BA9; ↑BA40, Cer	ND
GABRπ	— BA9, BA40, Cer	ND	— BA9, BA40, Cer	ND
GABRθ	— BA9, BA40, Cer	ND	↓BA9, ↑Cer, —BA40	ND
GABRρ2	↓ BA9; — BA40, Cer	ND	— BA9, BA40, Cer	ND
GABRR1	↓ BA9, BA40, Cer	ND	↑BA40; ↓Cer; —BA9	ND
GABRR2	↓ Cer; — BA9, BA40	ND	— BA9, BA40, Cer	ND

↑, increase; ↓, decrease; —, no change; BA9, Brodmann Area 9; BA40, Brodmann Area 40; Cer, cerebellum; DLPFC, dorsolateral prefrontal cortex; ND, not determined; PFC, prefrontal cortex

^aPlease see Table 1 for specific references

3. FMRP

FMRP is an RNA binding protein that primarily functions as a regulator of translation. FMRP targets 842 mRNAs - approximately 5% of all mRNAs - expressed in brain (Darnell and Klann, 2013). FMRP is highly expressed in neurons (Devys et al., 1993) where it is involved in mediating neurotransmitter release, synaptic transmission, and plasticity (Deng et al., 2011, 2013). In individuals with fragile X syndrome (FXS), the gene that codes for FMRP, fragile X mental retardation 1 (FMR1) is transcriptionally silenced due to hypermethylation of multiple (>200) CGG repeats in the 5’ untranslated portion of the gene including the CpG islands of the FMR1’s promoter region (Pieretti et al., 1991). Absence of FMRP results in loss of inhibition of protein synthesis mediated by mGluR5. As a result there is increased synthesis of synaptic proteins (Weiler and Greenough, 1999) which results in increased epileptiform discharges (Chuang et al., 2005), altered long-term depression (Hou et al., 2006), and an increase in long, immature dendritic spines (Irwin et al., 2002). The mGluR5 theory of FXS posits that these changes underlie the multiple pathologies of FXS including intellectual impairment and presence of seizure disorder (Bear et al., 2004; Dölen and Bear, 2008). However, as discussed below, underproduction of FMRP may also be associated with additional abnormalities seen in several psychiatric disorders which do not exhibit evidence for mutations in the FMR1 gene (Fatemi and Folsom, 2011; Fatemi et al., 2010b, 2011a, 2013a; Kelemen et al., 2013; Kovács et al., 2013; Lohith et al., 2013).

3.1. Reduction of FMRP Expression in Schizophrenia, Mood Disorders, and Autism

Significant reductions in FMRP were first identified in lateral cerebella of subjects with schizophrenia, bipolar disorder, and major depression (Fatemi et al., 2010b, 2013a) (Table 6). A follow-up study also identified reductions in FMRP in BA9 of subjects with schizophrenia and bipolar disorder (Fatemi et al., 2013a) (Table 6). In contrast, we found no change in FMRP mRNA in schizophrenia, bipolar disorder, or major depression, indicating that the observed changes were most likely secondary to posttranscriptional mechanisms. Subsequently, FMRP expression has been found to be reduced in peripheral blood lymphocytes of subjects with schizophrenia, confirming our initial observations (Kelemen et al., 2013; Kóvacs et al., 2013) (Table 6). This reduction was associated with lower IQ and earlier onset of illness (Kóvacs et al., 2013) as well as with deficits in visual perception (Kelemen et al., 2013) in subjects with schizophrenia. Thus, the observed reductions in FMRP protein levels result in functional consequences that could contribute to schizophrenic deficits. We similarly discovered reduced expression of FMRP in cerebellar vermis and BA9 of adults with autism (Fatemi and Folsom, 2011; Fatemi et al., 2011a) (Table 6). Moreover, we observed reduced expression of FMRP phosphorylated on serine 499 in cerebellar vermis of children and adults with autism and in BA9 of adults with autism (Rustan et al., 2013). Dephosphorylation of FMRP is known to lead to its ubiquitination and degradation (Nalavadi et al., 2012). In contrast, we discovered that there was no change in FMRP protein expression in children with autism and no change in mRNA levels in either age group (Fatemi and Folsom, 2011; Fatemi et al., 2011a) (Table 6). These findings are highly relevant as none of the individuals in these studies displayed the FMR1 mutation (Fatemi and Folsom, 2011; Fatemi et al., 2010b, 2011a, 2013a; Kóvacs et al., 2013). The phosphorylation status of FMRP in schizophrenia and mood disorders is currently unknown.

Table 6.

Summary of changes in expression of FMRP-mGluR5 signaling molecules in major psychiatric disorders

Protein of Interest	Autism	Schizophreniaa	Bipolar Disordera	Major Depressiona	Reference
FMRPa	↓ protein in BA9 and Cer; No change in mRNA in BA9 and Cer	↓ protein in BA9 and Cer; No change in mRNA in BA9 and Cer	↓ protein in BA9 and Cer; No change in mRNA in BA9 and Cer	↓ protein in Cer; No change in mRNA in Cer	Fatemi and Folsom, 2011; Fatemi et al., 2010a, 2011a, 2013a
		↓ protein in peripheral lymphocytes			Kelemen et al., 2013; Kóvacs et al., 2013
mGluR5	↑ dimerized protein in BA9 and Cer of children; no change in monomer in BA9 or Cer of children or adults.	↓ monomeric protein in BA9 and Cer of adults	↓ monomeric protein in BA9 and Cer of adults	↓ monomeric protein in Cer of adults	Fatemi and Folsom, 2011; Fatemi et al., 2011a, 2013a
		No change in protein in BA10			Corti et al., 2011
		No change in protein in PFC or striatum			Gupta et al., 2005
		No change in protein in DLPFC or ACC	No change in protein in ACC	No change in protein in ACC	Matosin et al., 2013, 2014
				↓ protein in PFC, CC, insula, thalamus, hippocampus	Deschwanden et al., 2011
		↓ mRNA in Cer; no change in BA9	↓ mRNA in BA9; no change in Cer	↓ mRNA in Cer	Fatemi et al., 2013a
		No change in mRNA in hippocampus			Ohnuma et al., 2000
		No change in mRNA in thalamus			Richardson-Burns et al., 2000
		No change in mRNA in BA9 or BA42			Volk et al., 2010
APP 120 kDa	↑ protein in BA9 of children and ↓ protein in Cer of adults	↓ protein in BA9; no change in Cer	↓ protein in BA9; no change in Cer	No change in protein in Cer	Fatemi et al., 2013c, unpublished observations
APP 88 kDa	↑ protein in BA9 of children	↓ protein in BA9; no change in Cer	↓ protein in BA9; no change in Cer	No change in protein in Cer	Fatemi et al., 2013c, unpublished observations
RAC1	↑ protein in BA9 of adults and children and Cer of adults	↑ protein in Cer; no change in BA9	↑ protein in Cer; no change in BA9	↑ protein in Cer	Fatemi et al., 2013c, unpublished observations
Homer 1	↓ protein in BA9 of adults	No change in protein in BA9 or Cer	↓ protein in BA9	No change in protein in Cer	Fatemi et al., 2013c, unpublished observations
STEP 61 kDa	↓ in BA9 of children	↓ protein in BA9	No change in protein in BA9 or Cer	No change in protein in Cer	Fatemi et al., 2013c, unpublished observations
		↑ protein in DLPFC and ACC			Carty et al., 2012
STEP 33 kDa	↓ in Cer of adults	↓ protein in BA9; no change in Cer	No change in protein in BA9 or Cer	No change in protein in Cer	Fatemi et al., 2013c, unpublished observations

^aadults;

↓, significant decrease; ↑, significant increase; ACC, anterior cingulate cortex; CC, cingulate cortex; Cer, cerebellum; DLPFC, dorsolateral prefrontal cortex; PFC, prefrontal cortex

Recent work has identified de novo mutations are overrepresented among FMRP target proteins in subjects with schizophrenia including protein UNC-13 homolog A (UNC13A), kinesin-like protein KIF1A (KIF1A), glutamate receptor, ionotropic, N-methyl-D-aspartate subunit 2A (GRIN2A), discs large, Drosophila homolog 2 (DLG2), G-protein coupled receptor kinase-interacting protein 1 (GIT1), protein tyrosine phosphatase, and receptor type gamma (PTPRG) (Fromer et al., 2014). Fromer et al. (2014) suggested that these findings may indicate pathogenic disruption of synaptic plasticity in subjects with schizophrenia. An analysis of the exome sequences of 2,536 subjects with schizophrenia and 2,543 controls found that FMRP targets including calcium/calmodulin-dependent protein kinase II alpha and beta (CAMK2A and CAMK2B), aconitase, mitochondrial (ACO2), and calcium channel, voltage-dependent, N-type, alpha 1B subunit (CACNA1B) were identified as particularly enriched for mutations (Purcell et al., 2014). The overall conclusions of these two studies were that overlap existed between genes that were targets of FMRP and mutations affecting these genes impacted cognition both in schizophrenia and autism, further confirming our initial observations. While Fromer et al (2014) identified direct evidence for specific FMRP target genes, Purcell et al.’s (2014) study supported a broad and indirect evidence for FMRP target gene involvement in schizophrenia.

4. mGluR5 Expression in Major Psychiatric Disorders

mGluR5 exists in monomeric and dimeric forms with dimerized mGluR5 assumed to be the active form of the receptor (Goudet et al., 2005; Romano et al., 2001; Schwendt and McGinty, 2007). mGluR5 has multiple functions including involvement in activity-dependent synaptic plasticity both developmentally and during adulthood as well as regulation of proliferation, differentiation, and survival of neurons (reviewed by Catania et al., 2007). As mentioned previously, in the absence of FMRP, unregulated mGluR5 signaling is believed to lead to the intellectual impairment associated with FXS (Bear et al., 2004; Dölen and Bear, 2008). Because of its normal physiological roles as well as its pathological role in FXS, mGluR5 has emerged as a target of therapeutic intervention in FXS, autism, and schizophrenia (Gürkan and Hagerman 2012; Hashimoto et al., 2013).

In subjects with schizophrenia, major depression, and bipolar disorder, our laboratory has identified reduced expression of the monomeric form of mGluR5 in lateral cerebellum (Fatemi et al., 2013a) (Table 6). qRT-PCR experiments also found that there was reduced mRNA expression for mGluR5 in the same brain region in schizophrenia and major depression (Fatemi et al., 2013a) (Table 6). We have also identified reduced expression of monomeric mGluR5 in BA9 of subjects with schizophrenia and bipolar disorder with mRNA values similarly reduced in BA9 of subjects with bipolar disorder (Fatemi et al., 2013a) (Table 6).

There are currently no other postmortem studies of mGluR5 in cerebellum of subjects with schizophrenia or mood disorders. However, results by other groups (Corti et al., 2011; Gupta et al. 2005; Matosin et al., 2013, 2014) found no change in levels of mGluR5 protein in other regions (anterior cingulate cortex, BA10, DLPFC, PFC, or striatum) in subjects with schizophrenia. A study by Deschwanden et al (2011), however, has identified a reduction in mGluR5 by western blotting in PFC in subjects with major depression, supporting our results. The above contradicting results can be due to several factors: 1) brain site differences; 2) different techniques used; and more importantly 3) lack of measurement of both monomeric and dimeric forms of mGluR5 by SDS-PAGE and western blotting. For example, as shown in Table 6, none of the above mentioned studies in schizophrenia discussed whether a monomer or dimer of mGluR5 was measured. The studies by Corti et al (2011) and Deschwanden et al (2011) showed a single band at 150 kDa, Gupta et al (2005) a band at 130 kDa, and Matosin et al (2013) a band at 130 kDa. Furthermore, Matosin et al. (2014) measured mGluR5 using receptor autoradiography which would not be able to identify either forms of mGluR5. Similarly, a recent study by Pretto et al. (2014) used immunocytochemistry to measure levels of mGluR5 whereby molecular species of mGluR5 could not be identified or quantified and reported a reduction in mGluR5 in FXS-ataxia syndrome. Similarly, three groups measured mRNA levels of mGluR5 in schizophrenia and did not report any changes in hippocampus, thalamus, BA9, or BA42 (Ohnuma et al., 2000; Richardson-Burns et al., 2000; Volk et al., 2010). Thus as recently discussed (Fatemi and Folsom, 2014), measurements of both monomer and dimer forms of mGluR5 by western blotting is a requirement for accurate measurement of any change in postmortem studies. In contrast, a positron emission tomography (PET) study found reduced expression in mGluR5 in the PFC, cingulate cortex (CC), insula, thalamus, and hippocampus of subjects with depression while western blotting found reduction in PFC (Deschwanden et al., 2011) (Table 6). Moreover, Volk et al (2010) found that subjects with schizoaffective disorder displayed significantly reduced mGluR5 mRNA than subjects with schizophrenia, further confirming our results in both schizophrenia and bipolar subjects who exhibit symptoms similar to subjects with a diagnosis of schizoaffective disorder.

In children with autism, we observed increased expression of the dimeric form of mGluR5 in BA9 and cerebellar vermis when compared to controls (Fatemi and Folsom, 2011; Fatemi et al., 2011a) (Table 6). The observed increased expression of mGluR5 in brains of children with autism appeared to be both age and disease specific (Fatemi and Folsom, 2014). Upregulation in the dimer form of mGluR5 in autistic children may be due to increased dimerization as a result of oxidative stress, ischemia, or via auto-induction (Copani et al., 2000). Altered expression of mGluR5 has also been observed in a number of disorders including Alzheimer’s disease (Tsamis et al., 2013), Huntington’s disease (Gulyás et al., 2014) and Down’s syndrome (Oka and Takashima, 1999), as well as in animal models of Parkinson’s disease (Sanchez-Pernaute et al., 2008). These changes may reflect greater excitotoxicity which contributes to the pathology of these disorders.

5. Expression of FMRP-mGluR5 Molecular Targets in Schizophrenia, Mood Disorders and Autism

In addition to some of the GABA receptors, a number of important molecules are targets of FMRP-mGluR5 signaling and synaptic transmission (Darnell et al., 2011; Darnell and Klann, 2013). We identified changes in four such targets – homer 1, amyloid beta A4 precursor protein (APP), ras-related C3 botulinum toxin substrate 1 (RAC1), and striatal-enriched protein tyrosine phosphatase (STEP) - in cerebellar vermis and BA9 of adults and children with autism when compared to controls (Fatemi et al., 2013c) (Table 6). APP is involved in synapse formation and neural plasticity (Priller et al., 2006; Turner et al., 2003). STEP is present in the postsynaptic density (PSD), where it is an important regulator of N-methyl-D-aspartate (NMDA) receptor function (Goebel-Goody et al., 2012). Homer proteins are components of the PSD and have multiple roles in synaptogenesis, receptor trafficking, and dopaminergic and glutamatergic signaling (Szumlinski et al., 2006). RAC1 is involved in functions that are relevant to neurotransmission, specifically the modulation of dendritic spine morphology and density (Luo et al., 1996; Nakayama et al., 2000; Threadgill et al., 1997).

More recently, we have investigated the protein expression of the same molecules in subjects with schizophrenia and mood disorders. In BA9 of subjects with schizophrenia, we determined reduced protein expression for APP and STEP (Folsom and Fatemi, unpublished observations). In BA9 of subjects with bipolar disorder, we determined significant reductions in APP and homer 1 (Folsom and Fatemi, unpublished observations). In lateral cerebella, we identified increased expression of RAC1 in subjects with schizophrenia, bipolar disorder, and major depression (Folsom and Fatemi, unpublished observations). When combined with previous observations of reduced expression of FMRP and mGluR5 in the same areas (Fatemi et al., 2010b, 2013a), our results provide further evidence that this signaling pathway is altered in schizophrenia and mood disorders.

6. FMRP Regulon Impacts Expression of GABA Receptor Subunits and mGluR5

A number of studies involving animal models of FXS have established that there is altered expression of GABA_A receptor subunits in brain as a result of reduced or absent FMRP expression (Adusei et al., 2010; El Idrissi et al., 2005; D’Hulst et al., 2006; Gantois et al., 2006; Hong et al., 2012). An initial study found that there was reduced protein expression of the GABRβ subunit in the cortex, hippocampus, diencephalon, and brainstem of adult FMR1 knockout (KO) mice (El Idrissi et al., 2005). A follow up study by the same group found reduced mRNA for GABRβ1 in cerebellum, reduced mRNA for GABRβ2 in hippocampus and cerebellum, and reduced mRNA for GABRβ3 in cortex of male FMR1 KO mice (Hong et al., 2012). However, mRNA for GABRβ3 was significantly upregulated in hippocampus of male FMR1 KO mice (Hong et al., 2012). A separate study found that mRNA for GABRδ was reduced in multiple brain regions of FMRP KO mice including hippocampus, neocortex, cerebellum, caudate, putamen, thalamus and olfactory bulb (Gantois et al., 2006). A comprehensive study of 17 GABA_A subunits found reduced expression of mRNA for GABRα1, GABRα3, GABRα4, GABRβ1, GABRβ2, GABRγ1, and GABRγ2 in cortex of FMR1 KO mice (D’Hulst et al., 2006).

Complex developmental patterns of GABA_A receptor subunits have also been identified in forebrain of FMR1 KO mice - i.e., downregulation of GABRα1 protein during early postnatal development while displaying similar expression as wild type during adulthood vs. GABRβ2 which displayed similar expression as wild type at postnatal day five (PN5), but showed significantly reduced expression at PN12 and adulthood (Adusei et al., 2010). Importantly, treatment with allosteric modulators of GABA_A receptors (ganaxolone) or GABA_B receptors (arbaclofen) rescued many FXS-related deficits in FMR1 KO mice including reducing audiogenic seizures, spine density, protein synthesis, and internalization of AMPA receptors (Henderson et al., 2012; Heulens et al., 2012).

It is well known that FMRP is involved in the normal maturation and function of GABAergic interneurons (Olmos-Serrano et al., 2010; Martin and Huntsman, 2014) with its loss or reduction leading to abnormal GABA release presynaptically, abnormal production of GAD65/67 enzymes, decreases in GABA transporter 1 expression and excitatory drive of parvalbumin positive GABAergic interneurons (Martin and Huntsman, 2014). As FMRP binds to the GABA_A delta subunit; reduction in δ subunit mRNA, protein, or both (D’Hulst et al., 2006; Curia et al., 2009) can result in defective tonic GABAergic conductance (Martin and Huntsman, 2014). Quite interestingly, reports by two groups indicate that absence of FMRP in amygdala of Fmr1 KO mice results in decreased tonic inhibitory tone, most likely due to several reasons: 1) limited GABA production and release, leading to reduced availability of GABA ligand for tonic receptors; 2) decreased capacity for tonically active currents; 3) lack of co-expression of α3 subunits with postsynaptic protein gephyrin; 4) small α5 subunit-mediated tonic currents; 5) reduced inhibitory synaptic efficacy; and 6) ineffective glutamate-related plasticity (Martin and Huntsman, 2014; Martin et al., 2014; Olmos-Serrano et al., 2010). Analogously, observation of significant decreases in protein expression for α3 (both mRNA and protein), α5, and δ subunits in brains of subjects with autism (Fatemi et al., 2009a, 2010a, 2014; Tables 1 and 5) demonstrate that similar mechanisms may be operational in idiopathic autism, probably due to reduction in levels of FMRP protein. A further mechanism whereby reduction in FMRP may lead to alterations in several GABA receptor subunits is potentially through failure of translation of one subunit, i.e., δ subunit causing decreased translation of other coassembled subunits (i.e., α4 and α5) (Sur et al., 2001; Martin and Huntsman, 2014).

Taken together, these studies show that FMRP impacts multiple GABA receptor subunit mRNAs, resulting in altered expression and, consequently leading to abnormalities associated with FXS. The exact mechanism by which FMRP alters GABA receptor subunit expression is currently unknown although FMRP effects may be at the transcriptional or translational levels. The reduction of mRNA for GABA receptors in FMR1 KO mice suggests some level of transcriptional control (D’Hulst et al., 2006; Gantois et al., 2006; Hong et al., 2012). Reduced protein expression in FMR1 KO mice also suggests that translational control may be abnormal (Adusei et al., 2010; El Idrissi et al., 2005). FMRP has been shown to directly bind GABBR1, GABBR2, and GABRδ mRNAs (Darnell et al., 2011; Miyashiro et al., 2003) while no direct binding of FMRP to mRNAs for GABRα1-α6, GABRβ1-β3, or GABRγ1-γ3 has been demonstrated (Brown et al., 2001; Chen et al., 2003; Darnell et al., 2011). Further experiments are needed to elucidate the exact molecular mechanism by which GABA receptor subunit expression is altered by FMRP regulon.

The recent findings of reduced expression of FMRP in multiple brain regions of subjects with schizophrenia, bipolar disorder, major depression, and autism (Fatemi and Folsom, 2011; Fatemi et al., 2010b, 2011a, 2013a) and in peripheral blood lymphocytes of subjects with schizophrenia (Kelemen et al., 2013; Kóvacs et al., 2013) are novel. They represent the first findings of altered FMRP expression outside of subjects with FXS, in individuals who do not carry the FMR1 mutation. The cause of reduced FMRP expression in major psychiatric disorders is currently unknown. In FXS, it is well-established that hypermethylation of the CGG-trinucleotide expansion at the FMR1 promoter inhibits production of FMRP (Colak et al., 2014; Pieretti et al., 1991). Hypomethylation of the FMR1 promoter has also been observed in an individual with the FMR1 mutation who did not display intellectual impairment (Burman et al., 1999). It is unclear whether similar epigenetic mechanisms are at work in schizophrenia, mood disorders, and autism, especially as none of the individuals studied thus far carried the FMR1 mutation (Fatemi and Folsom, 2011; Fatemi et al., 2010b, 2011a, 2013a; Keleman et al., 2013; Kóvacs et al., 2013). Moreover, none of these studies indicate any abnormalities in mRNA levels for FMRP, clearly pointing to translational abnormalities in these studies. Lastly, recent positive associations between FMRP target mutations in schizophrenia and autism (Fromer et al., 2014; Iossifov et al., 2012; Purcell et al., 2014) clearly expand the pathological role of FMRP in these diseases.

Our laboratory’s findings of reduced FMRP, increased mGluR5, and reduced expression of GABA receptors in autism are consistent with the mGluR5 theory of FXS. Due to the extensive phenotypic overlap between autism and FXS, these findings are not unexpected. As mentioned above, this pattern is similar to what has been observed in FMR1 KO mice (D’Hulst et al., 2006; Dölen and Bear et al., 2008; El Idrissi et al., 2005; Gantois et al., 2006). The theory is further strengthened by the findings of Lohith et al (2013) who found increased expression of mGluR5 protein in BA9 of subjects with FXS. Our results suggest increased mGluR5 activity in children with autism. Even in adults with autism, where mGluR5 levels were not significantly different than controls, ratios of mGluR5 to neuronal specific enolase (NSE) for both dimeric and total mGluR5 were increased by 96% in BA9 (Fatemi and Folsom, 2011; Fatemi et al., 2011a). In contrast, a recent article by Pretto et al. (2014) demonstrated reduced expression of both FMRP and mGluR5 in cerebella of subjects with the FMR1 premutation who had been diagnosed with fragile X tremor/ataxia syndrome (FXTAS). The discrepancy between these findings and ours may be due to brain regional differences as we found non-significant reductions in mGluR5 in cerebellar vermis of adults with autism (Fatemi et al., 2011a). Additionally, while Lohith et al (2013) found marginally increased mGluR5 expression in carriers and subjects with FXS based on receptor binding assays (p<0.056) and western blotting (p<0.048), when carriers were removed from analysis, both mGluR5 expression and density were higher although significance was not attained (Lohith et al., 2013), possibly due to reduced sample size. Thus, the status of the individuals tested (carrier vs. FXS/autism diagnosis) or presence/absence of FXTAS or brain regional differences may impact mGluR5 levels.

In schizophrenia and mood disorders, which have no obvious phenotypic overlap with FXS, our findings of reduced levels of both FMRP and monomeric mGluR5, suggest that an as yet unknown mechanism might be at work (Fatemi et al., 2013a). The discrepancy between our results and those of other laboratories (Corti et al., 2011; Gupta et al., 2005; Matosin et al., 2013, 2014; Ohnuma et al., 2000; Pretto et al., 2014; Richardson-Burns et al., 2000; Volk et al., 2010) may reflect region-specific differences in mGluR5 expression (BA9 vs. BA10 vs. DLPFC vs. ACC vs. cerebellum); the different analytic techniques that were used (western blotting vs. receptor binding assays vs. qRT-PCR vs. immunocytochemistry); more importantly, the lack of measurement of monomer and dimer forms of mGluR5 in three studies (Corti et al., 2011; Gupta et al., 2005; Matosin et al., 2013) and the non-quantitative technique of measuring mGluR5 by immunocytochemistry in one study (Pretto et al., 2014); and lastly, differences between levels of mRNA and protein for mGluR5. The hypothesis that brain regional expression of mGluR5 may also account for differences between studies has recently been advocated by other laboratories (Matosin et al., 2014).

7. Potential for Treatment

Altered expression of GABAergic and FMRP-mGluR5 signaling system components provide opportunities for targeted therapeutic intervention. Several review articles have been published regarding targeting the FMRP-mGluR5 signaling system, including the use of allosteric modulators of mGluR5, for the treatment of autism and FXS (Gürkan and Hagerman, 2012; Hagerman et al., 2014; Wang et al., 2010). Minocycline is a semisynthetic second-generation tetracycline that prevents glutamate-induced excitotoxicity (Maier et al., 2007; Pi et al., 2004). Treatment with minocycline has been shown to improve behavior in subjects with FXS (Paribello et al., 2010). Preliminary trials of minocycline in subjects with schizophrenia have demonstrated a beneficial effect on the negative symptoms, and cognitive functioning (Levkovitz et al., 2010; Miyaoka et al., 2008). mGluR5 negative allosteric modulator 2-Methyl-6-(phenylethynyl)pyridine (MPEP) has shown promise in animal studies (de Vrij et al., 2008) and may prove effective in humans. Fenobam, an antagonist of mGluR5 was shown to improve PPI deficits in six out of 12 subjects with FXS in a pilot study (Berry-Kravis et al., 2009). Norbin is a positive endogenous mediator of mGluR5 which increases cell surface expression of mGluR5 (Wang et al., 2009), and may be useful where there are deficits in mGluR5 expression. Less is known about mGluR5 modulators in schizophrenia although preclinical studies of glutamate receptor modulators, including mGluR5 modulators are currently underway (Hashimoto et al., 2013; Herman et al., 2012). In contrast, there have been multiple clinical trials of drugs that target GABA_A and GABA_B receptors in subjects with schizophrenia and mood disorders.

Drugs that modulate GABA function may prove to be beneficial for the treatment of schizophrenia. Pregnenolone and its metabolite allopregnanolone are positive allosteric modulators of GABA_A receptors. Pregnenolone has recently shown promise in ameliorating positive, negative, and cognitive deficits in schizophrenia (Kreinin et al., 2014; Marx et al., 2009; Ritsner et al., 2010, 2014). A proof-of-concept trial demonstrated that subjects with schizophrenia receiving pregnenolone showed improvement in negative symptoms as measured by the Scale for the Assessment of Negative Symptoms (SANS) and increased serum pregnenolone predicted scores on the Brief Assessment of Cognition in schizophrenia (BACS) (Marx et al., 2009). More recently, in a double-blind, randomized add-on trial, pregnenolone showed efficacy in improving negative symptoms, especially with respect to blunted affect, anhedonia, and avolition (Ritsner et al., 2014). Pregnenolone has also been shown to reduce deficits in visual attention, sustained attention, and executive function when compared against placebo (Kreinin et al., 2014). Additionally, pregnenolone has been shown to improve positive symptoms of schizophrenia as well as neuroleptic-induced extrapyramidal side-effects (Ritsner et al., 2010). In an animal model of schizophrenia, pregnenolone has been shown to correct deficits in PPI as well as hyperlocomotion (Wong et al., 2012). Antipsychotic drugs clozapine and olanzapine are known to increase levels of pregnenolone as well as allopregnanolone, in a dose dependent manner (Marx et al., 2003, 2006a,b), while haloperidol, risperidone, aripiprazole, quetiapine, and ziprasidone have no such effect (Marx et al., 2003, 2006b). Importantly, pregnenolone was increased in rat hippocampus and cerebral cortex by clozapine and olanzapine (Marx et al., 2006a,b), two brain areas implicated in the pathology of schizophrenia. The superior efficacy of clozapine in the treatment of refractory schizophrenia may be partly accounted for by its ability to increase pregnenolone and subsequent allosteric modulation of GABA_A receptors (Marx et al., 2006b).

Treatment with pregnenolone or allopregnanolone may be beneficial for mood disorders. Reduced levels of allopregnanolone in blood and cerebrospinal fluid have been associated with depression and anxiety disorders (reviewed by Schüle et al., 2014). Rats treated with fluoxetine display increased expression of pregnenolone and allopregnanolone in hippocampus (Marx et al., 2006a), suggesting that this may be part of its efficacy as an antidepressant. A recent double-blind, placebo-controlled trial of pregnenolone in subjects with bipolar disorder found that it was effective in treating depressive symptoms in this population (Brown et al., 2014).

Modulators of GABA_B receptors have also shown promise in reversing deficits in animal models of schizophrenia. rac-BHFF [(R, S)-5,7-di-tert-butyl-3-hydroxy-3-trifluoromethyl-3H-benzofuran-2-one], which acts as a positive allosteric modulator of GABA_B receptors has been shown to correct for sensorimotor gating deficits, as measured by PPI, in a dose-dependent manner (Frau et al., 2014). Baclofen, another GABA_B receptor agonist, has been shown to reverse deficits in social interaction and spatial working memory (Gandal et al., 2012). Baclofen can also reverse deficits in PPI and object recognition memory (Mizoguchi and Yamada, 2011). A small number of studies have investigated baclofen for the treatment of schizophrenia with mixed results. There have been some reports of clinical improvement when baclofen is used on its own (Gulmann et al., 1976) or as an adjunctive treatment (Frederiksen, 1975), while another study found no improvement (Bigelow et al., 1977). However, in light of recent animal experiments and the availability of additional GABA_B modulators, use of these drugs may prove beneficial in the treatment of schizophrenia.

8. Conclusions

The GABAergic signaling system has been hypothesized to contribute to the pathology of major psychiatric disorders including schizophrenia, bipolar disorder, major depression, and autism. Our laboratory has demonstrated comprehensive changes in GABA receptor expression in cerebellum in all four disorders. A potential mechanism to explain altered GABA receptor expression is through reduced expression of FMRP and altered FMRP-mGluR5 signaling as shown by changes in expression for mGluR5 and downstream targets including RAC1, APP, STEP, and homer 1. Alternatively, other mechanisms such as epigenetics or monoallelic expression, as seen in the case of GABRβ3, may also play important roles in GABA receptor expression. Drugs that affect: 1) the GABAergic signaling system such as pregnenolone, allopregnanolone, rac-BHFF, or baclofen; 2) allosteric modulators of mGluR5 activity such as MPEP, fenobam, or norbin; or 3) minocycline’s action preventing glutamate-induced excitotoxicity, are emerging as potential treatments for schizophrenia and mood disorders that may succeed where traditional treatments that target dopaminergic or serotonergic systems have not been as efficacious. Future studies should focus on these signaling systems as potential novel treatments for major psychiatric disorders.