The Basolateral Amygdala GABAergic System in Health and Disease

Abstract

The brain comprises an excitatory/inhibitory neuronal network that maintains a finely tuned balance of activity critical for normal functioning. Excitatory activity in the basolateral amygdala (BLA), a brain region that plays a central role in emotion and motivational processing, is tightly regulated by a relatively small population of gamma-aminobutyric acid (GABA) inhibitory neurons. Disruption in GABAergic inhibition in the BLA can occur when there is a loss of local GABAergic interneurons, alterations in GABA_A receptor activation, or dysregulation of mechanisms that modulate BLA GABAergic inhibition. Disruptions in GABAergic control of the BLA emerge during development, in aging populations, or after a trauma, ultimately resulting in hyperexcitability. BLA hyperexcitability manifests behaviorally as an increase in anxiety, emotional dysregulation, or the development of seizure activity. This article reviews the anatomy, development, and physiology of the GABAergic system in the BLA, and circuits that modulate GABAergic inhibition, including the dopaminergic, serotonergic, noradrenergic, and cholinergic systems. We highlight how alterations in various neurotransmitter receptors, including the acid sensing ion channel 1a (ASIC1a), cannabinoid receptor 1 (CB1), and glutamate receptor subtypes, expressed on BLA interneurons, modulate GABAergic transmission and how defects of these systems affects inhibitory tonus within the BLA. Finally, we discuss alterations in the BLA GABAergic system in neurodevelopmental (autism/Fragile X syndrome) and neurodegenerative (Alzheimer’s disease) diseases, and after the development of epilepsy, anxiety, and traumatic brain injury. A more complete understanding of the intrinsic excitatory/inhibitory circuit balance of the amygdala and how imbalances in inhibitory control contribute to excessive BLA excitability will guide the development of novel therapeutic approaches in neuropsychiatric diseases.

Introduction

The brain is made up of a highly complex network of excitatory and inhibitory circuits that maintains exquisite balance in network activity. Hyperexcitability arises when there is an imbalance between excitation and inhibition (E/I) often as a result of deficiencies or disruption in gamma-aminobutyric acid (GABA) inhibitory system control. Hyperexcitability of the amygdala, in particular, can be strongly associated with anxiety, hypervigilance, and an inability to regulate emotions. Acquired deficiencies in the GABAergic inhibitory system have been observed after traumatic brain injury (TBI; Almeida-Suhett et al. 2014; Depue et al. 2014; Guerriero et al. 2015; Reger et al. 2012) and status epilepticus (SE; Fritsch et al. 2009; Gean et al. 1989; Prager et al. 2014b). In addition, amygdala hyperexcitability resulting in anxiety has been observed in neuropsychiatric disorders, such as posttraumatic stress disorder (PTSD; (Nuss 2015; Truitt et al. 2009), as well as in neurodevelopmental disorders, including autism/Fragile X syndrome (El-Ansary and Al-Ayadhi 2014; Martin et al. 2014; Olmos-Serrano et al. 2010), and neurodegenerative disorders, such as Alzheimer’s disease (AD; Klein et al. 2014; Palop and Mucke 2010).

Hyperexcitability of the basolateral nucleus of the amygdala (BLA) is associated with increased anxiety and often occurs in parallel with various neurodevelopmental, neurodegenerative, and neuropsychiatric disorders. The GABAergic inhibitory system is one target of therapeutic treatments to reduce anxiety and maintain homeostasis. For example, benzodiazepines, which allosterically enhance postsynaptic actions of GABA at the inhibitory type A GABA receptor (GABA_A receptor), are one first line treatment for anxiety (Farb and Ratner 2014) and seizure disorders. However, in many cases, benzodiazepines are ineffective and/or exacerbate symptoms, as has been observed in seizure models, for example, where administration of diazepam initially suppress seizures, but lead to rebound seizures that are similar to or longer in duration than that of animals that do not receive the anticonvulsant (Apland et al. 2014). Thus, the efficaciousness of current treatments targeting the GABAergic system has been called into question and new therapeutic targets merit preclinical investigation.

The purpose of this article is to review the BLA GABAergic system in health and disease. First, we review the anatomy and development of the GABAergic system in the BLA and the different ways that GABAergic inhibitory synaptic transmission is modulated. Second, we review how local GABAergic inhibitory neurotransmission in the BLA is altered in disease. This review focuses on five diseases: Autism/Fragile X, AD, epilepsy, TBI, and anxiety and trauma or stressor related disorders (such as PTSD), as these disorders are prime examples of acquired amygdala dysfunction that occur during development, aging, or after injury. By studying how deficiencies in the GABAergic inhibitory system in the amygdala contribute to disease outcomes, future research may be directed at developing new therapies to reduce excitability or increase inhibition.

The GABAergic System in the Basolateral Amygdala

GABAergic Interneurons

The amygdala is located in the medial temporal lobe and is comprised of 13 subnuclei (for a comprehensive review of the anatomical connections of the rat and human amygdala, see Pitkanen 2000; Sah et al. 2003; Whalen and Phelps 2009). The BLA makes up a large component of this network, receiving input from cortical and subcortical structures. The BLA, which is generally comprised of the lateral and basal portions, contains two main types of neurons: glutamatergic (pyramidal) principal neurons and GABAergic interneurons (McDonald 1992; Pare and Smith 1998). Principal neurons constitute the majority of the neurons in the BLA (80–85%), whereas GABAergic interneurons form approximately 15–20% of the neuronal population (Sah et al. 2003; Spampanato et al. 2011). GABAergic interneurons can be subdivided into those that express calbindin (CB) or calretinin (CR) and can be further subdivided into groups by neuropeptide expression (i.e., vasoactive intestinal peptide; VIP and/or cholecystokinin; CCK) or by the expression of the calcium binding protein parvalbumin (PV; Table 1) (Davila et al. 2008; Kemppainen and Pitkanen 2000; Mascagni and McDonald 2003; McDonald and Mascagni 2001a; McDonald and Mascagni 2002). Importantly, PV-immunopositive neurons make up about 40% of GABAergic interneurons and are the main source of the perisomatic innervation of principal cells, suggesting that their primary role is in feedback inhibition. CR interneurons make up about 25–30% of BLA GABAergic interneurons and primarily innervate other interneurons (Capogna 2014; McDonald and Mascagni 2001a; Muller et al. 2003; Muller et al. 2006).

Table 1.

Summary of Systems Modulating BLA GABAergic Inhibition. Note that no study has differentiated receptor localization to VIP +/−. Therefore, we have placed the receptor modulating VIP +/− in each category. For citations, see text.

Immunohistochemistry		Firing Patterns	Innervation of GABAergic Interneurons	Receptor Subtypes and role in Modulating GABAergic Transmission
General	Specific
CB	PV	Fast spiking, stuttering, non- adapting, adapting	~50% cortical, <1% thalamic to CB interneurons, VTA, SN, dorsal raphé nucleus, substantia innominate, ventral pallidum of basal forebrain (cholinergic and GABAergic)	D1 - ↑ firing, induce rhythmic oscillations D2 - ↓ presynaptic GABA release 5-HT2A - ↑ excitability GABAB - ↓ excitability
	CCK (VIP-)	Non-adapting, burst adapting		5-HT3A - ↑ excitability; but rapidly desensitizing α1-AR, α2-AR - ↑ AP firing and IPSCs CB1 - ↓ excitability GABAB - ↓ excitability
	SOM/NPY/NK1r		dorsal raphé nucleus	5-HT1A (NPY, NK1r) - ↓ presynaptic GABA release 5-HT2C (NPY) - ↑ excitability α1-AR, α2-AR - ↑ AP firing and IPSCs GABAB - ↓ excitability
CR	CCK (VIP+)	Adapting	VTA and SN (< PV interneurons)	5-HT3A - ↑ excitability; but rapidly desensitizing α1-AR, α2-AR - ↑ AP firing and IPSCs GABAB - ↓ excitability
Not localized to Specific GABAer gic Interneuronal Subpopulations			LC, NTS	M1-mAChR –↑ excitability M2-mAChR - ↓ excitability α7-, α4β2-nAChR ↑ excitability ASICs 1A - ↑ excitability AMPA lacking GluR2, NMDA - ↑ excitability GluK1 - ↑ presynaptic GABA release (dose dependent)

GABA_A Receptor Structure and Function

GABAergic inhibitory synaptic transmission plays a central role in the regulation of amygdala excitability. Pathological disruption of GABA_A receptors causes a disruption of the E/I balance and has been increasingly implicated in neurological and neurodegenerative diseases (Deidda et al. 2014). Fast inhibitory synaptic transmission within the central nervous system is mediated by the GABA_A receptor, a heteropentameric chloride permeable, GABA gated member of the cys-loop superfamily of ligand-gated ion channels. GABA_A receptors are formed from limited combinations of subunits that have diverse structural and functional properties (α1–6, β1–3, γ1–3, δ, ε, θ, and π; Olsen and Sieghart 2009).

Proper maturation of the GABAergic system in the BLA is essential in neurodevelopment. Dysfunction in the development of the GABAergic inhibitory system within the BLA may be associated with neurodevelopmental diseases, such as autism or Fragile X. In rat, the development of the mature GABAergic system in the BLA takes place between postnatal (P) day P14 and P30 with the emergence of PV-interneurons (Berdel and Morys 2000; Davila et al. 2008), an increase in the density of GABAergic fibers, and a decrease in the density of GABAergic cell bodies (Brummelte et al. 2007). Concurrently, GABA_A receptor mediated inhibitory postsynaptic currents (IPSCs) reach maturity between P21 and P28. Simultaneously, the reversal potential of GABA_A receptors expressed in principal neurons shifts from −55 mV at P7 to −70 mV by P21. This increase in hyperpolarization may be due, in part, to a switch from a greater expression of sodium-potassium-chloride cotransporter 1, which accumulates intracellular chloride and renders GABA_A receptors excitatory, to an increase in the potassium-chloride cotransporter 2, which extrudes chloride from the cell, rendering GABA_A receptors inhibitory (Ben-Ari et al. 2012; Ehrlich et al. 2013). In addition, a decrease in rise time and decay time constant occurs due to a change in the GABA_A receptor subunit composition (from primarily α2- to the α1-subunit; Ehrlich et al. 2013). This shift results in a GABAergic shunt that limits the extent of BLA activation (see below; Rainnie et al. 1991b).

The composition of GABA_A receptors have been found to be quite diverse because of their subunit assembly making their roles significantly different depending on the timing of activation and subcellular localization (Marowsky et al. 2004; Pouille and Scanziani 2001). The BLA of mature animals contains α1- and α2-subunits of the GABA_A receptor; α1-subunit containing GABA_A receptors are primarily expressed at the somal level of PV-GABAergic interneurons, but also exhibit co-immunoreactivity with the β2/3 subunits (McDonald and Mascagni 2004). Alternatively, GABA_A receptors on principal neurons primarily contain the α2-subunit, which is predominately responsible for the benzodiazepine allosteric potentiation of inhibitory currents (Marowsky et al. 2004). In addition, principal neurons in the BLA contain γ2 subunits, which likely contribute to the formation of α2βxγ2 pentameric GABA_A receptors that contribute to fast inhibitory synaptic transmission (Esmaeili et al. 2009). Extrasynaptically, the GABA_A receptor in the BLA is primarily made up of the α3 subunit, which strongly mediates tonic GABAergic currents (Marowsky et al. 2012). However, the α5 subunit, which is diazepam-sensitive and shapes the decay phase of the inhibitory postsynaptic currents (Marowsky et al. 2004) as well as the δ subunit, both of which are hallmark subunits that contribute to tonic inhibition (Farrant and Nusser 2005) are also expressed in the BLA, though not as strongly as the α3 (Marowsky et al. 2012).

Temporal dynamics and intra-amygdala regulation of excitatory activity

GABAergic interneurons can be differentiated by their firing properties. PV interneurons primarily fire short duration, non-adapting action potentials (Rainnie et al. 2006; Woodruff and Sah 2007b), whereas CB-expressing GABAergic interneurons fire broad action potentials, display firing adaptation, and primarily synapse with somata (see Table 1; Jasnow et al. 2009). Other interneurons expressing somatostatin (SOM), VIP, CR, and CCK also target dendrites or somata (Mascagni and McDonald 2003; Muller et al. 2007a). Though GABAergic interneurons only constitute a fraction of the total neuronal population, they tightly regulate network excitability, and lead to a low resting firing rate of principal neurons (Lang and Pare 1997; Pare and Gaudreau 1996; Woodruff and Sah 2007a).

The regulation of excitatory activity by local GABAergic interneurons is influenced by its firing properties (Lang and Pare 1997; Rainnie et al. 1991b). Most BLA GABAergic interneurons fire short-duration action potentials with little spike frequency adaptation in response to prolong depolarization, though specific subpopulations of GABAergic interneurons have different firing patterns (see Table 1). Principal neurons, by comparison, show spike frequency adaptation and prolonged after hyperpolarization in response to prolonged depolarizing currents (Pare et al. 1995; Rainnie et al. 1991a; Rainnie et al. 1991b; Sah et al. 2003). The axonal morphology of BLA GABAergic interneurons also allows for tight inhibitory control over principal neurons. GABAergic interneuron axons branch on average two to six times forming relatively dense terminal and collateral fields with principal neurons (Millhouse and DeOlmos 1983; Smith et al. 1998). BLA GABAergic projections participate in either feedback inhibition or transient dis-inhibition of principal neurons. Indeed, PV interneurons receiving strong excitatory local inputs from BLA projection neurons appear to be involved in feedback inhibition (Smith et al. 2000; Unal et al. 2014), whereas intercalated interneurons, which have recently been found to project to the BLA (Manko et al. 2011), appear to target PV- and CB-immunoreactive GABAergic interneurons and are likely to transiently disinhibit principal cells (Bienvenu et al. 2015).

The regulation of the firing rate by GABAergic interneurons controls the flow of information from the BLA, and evidence indicates that local inhibitory circuits in the amygdala mediate its functioning. Activation of the GABAergic system appears to play a central role in the synchronization of spiking activity, which can coordinate and enhance the effects of input signals, which precisely enables the activation of glutamatergic activity to drive behavioral responses (Courtin et al. 2014; Herry and Johansen 2014). Indeed, the initiation and expression of fear, for example, requires synchronization of amygdala activity, among other regions (Stujenske et al. 2014). Ongoing research has revealed that theta and the faster gamma oscillations may be fundamental to circuits underlying sensory processing and cognitive functions and that changes in emotional states may be mediated by alterations in BLA gamma coupling to theta frequency inputs (Fries 2009; Stujenske et al. 2014). Importantly, the activity of PV interneurons has been implicated in theta synchrony within the medial prefrontal cortex (PFC) (Courtin et al. 2014), as suppression of PV interneuronal activity in the PFC is necessary to disinhibit prefrontal projection neurons to the BLA, thereby synchronizing their firing by resetting local theta oscillations. While this work has not yet also been confirmed in the amygdala, it has been hypothesized that inhibiting PV interneurons in the BLA may also synchronize activity and enhance fear responses (Stujenske et al. 2014).

Modulation of GABA_A Receptor Mediated Inhibitory Synaptic Transmission in the BLA

GABAergic inhibition in the BLA is modulated by afferents from both cortical and subcortical brain regions (Figure 1A). In most cases afferents from these regions project to both principal neurons and GABAergic interneurons. In some cases, it has been determined that projections are directed to particular subpopulations of neurons. This section reviews the afferent projections that modulate and facilitate GABAergic inhibitory synaptic transmission in the BLA. More specifically, we discuss how activation of different receptor types modulate the release of GABA from the presynaptic terminal or alter the excitation of GABAergic interneurons (Figure 1B).

Figure 1. Modulation of GABAergic Inhibitory Synaptic Transmission in the BLA.

(A) Schematic representation of GABAergic projections from the prefrontal cortex (PFC) and glutamatergic projections from the thalamus. In addition, GABAergic interneurons in the BLA receive cholinergic projections from the substantia innominate (SN) and ventral pallidum (VP), dopaminergic projections from the ventral tegmental area (VTN) and substantia nigra (SN), noradrenergic projections from the locus coeruleus (LC) and nucleus of the solitary tract (NTS), and serotonergic projections from the dorsal raphé nucleus (DRN). (B) Schematic representation of receptors modulating GABAergic inhibitory synaptic transmission in the BLA. Postsynaptic (1A) M2-mAChRs and (2A) GABA_B receptors hyperpolarize GABAergic interneurons by reducing voltage gated Ca²⁺ channels and increasing K⁺ channel conductance. (1B) M1-mAChRs increase excitability (though primarily expressed on principal neurons) by suppressing several K⁺ currents and increasing voltage gated Ca²⁺ conductance. (2B) Activation of presynaptic GABA_B receptors inhibit neurotransmitter release on both excitatory and inhibitory synapses by inhibiting voltage-gated Ca²⁺ channels and possibly by interacting with vesicular release machinery.(3A) Activation of postsynaptic NMDA or AMPA receptors on interneurons increases excitability. (3B) Activation of GluK1-containing kainate receptors depolarize interneurons by increasing the presynaptic release of GABA or increasing excitability via activation of postsynaptic GluK1 receptors. (4) Activation of α₇-nAChRs and/or α₄β₂ nAChRs presynaptically modulate GABA release or regulate neuronal activity by their position on interneurons. (5A) Dopaminergic projections activate postsynaptic D1 receptors, which increase excitability by reducing slowly inactivating K⁺ currents while D2 (5B) receptors reduce presynaptic release of GABA. (6) Activation of ASIC1A receptors increase interneuronal excitability. Postsynaptically, activation of 5-HT₂(7A) and 5-HT_3A(7B) receptors increase interneuronal excitability via an increase in intracellular Ca²⁺ concentrations or increasing the interneuronal excitability, respectively. (7C) Activation of presynaptic 5-HT_1A receptors reduce quantal release and increases hyperpolarization. (8) Activation of α1 and α2 receptors depolarize interneurons, subsequently increasing action potential firing and enhancing inhibitory synaptic transmission. (9) Activation of CB1 receptors on CCK interneurons reduces presynaptic release by inhibiting voltage gated Ca²⁺ channels and activating voltage gated K⁺ channel.

Cortical and thalamic regulation of BLA GABAergic interneurons

The BLA receives extensive cortical and thalamic projections, which synapse onto both principal neurons and GABAergic interneurons. Stimulating afferents from either the cortical or thalamic pathways has been found to monosynaptically activate BLA and lateral amygdala (LA) GABAergic interneurons, primarily in a feedforward manner (Lang and Pare 1998; Rainnie et al. 1991b; Szinyei et al. 2000; Washburn and Moises 1992). Recent studies have identified to what type of interneuron cortical and thalamic inputs project. Unal and colleagues (Unal et al. 2014) found that BLA interneurons expressing CB receive about half of the cortical inputs to local-circuit cells of the BLA and constitute a major source of feedforward inhibition, while thalamic inputs form less than one percent of synapses on interneurons (Carlsen and Heimer 1988; LeDoux et al. 1991). Importantly, there appears to be a possible discrepancy in the regulation of GABAergic interneurons in the LA versus the BLA. Cortical inputs to the BLA primarily regulate CB-expressing interneurons, while GABAergic interneurons in the LA equally respond to both cortical and thalamic pathways (Szinyei et al. 2000; Unal et al. 2014).

Dopaminergic Afferents

The BLA receives dense dopaminergic innervation from the ventral tegmental area (VTA) and substantia nigra (SN; Asan 1997; Fallon and Ciofi 1992). VTA and SN projections synapse on BLA principal (projection) neurons and PV- and CR-immunopositive GABAergic interneurons (Brinley-Reed and McDonald 1999; Pinard et al. 2008). However, compared to CR-immunopositive interneurons, PV-interneurons appear to be the preferential target of dopaminergic synapses in the BLA (Pinard et al. 2008). By projecting to principal neurons and GABAergic interneurons, dopamine (DA) influences the activity of both excitatory and inhibitory cell types within the BLA (Kroner et al. 2005; Rosenkranz and Grace 1999). Via activation of D1 receptors, DA increases excitability and evoked firing of principal neurons by reducing slowly inactivating K⁺ currents while activation of D2 receptors increases input resistance. Moreover, D1 receptor activation increases evoked firing in fast-spiking BLA interneurons and the frequency of spontaneous IPSCs (sIPSCs) (Kroner et al. 2005). Activation of DA receptors has also been found to induce rhythmic inhibitory oscillations (Loretan et al. 2004; Ohshiro et al. 2011), though increases in excitatory transmission is required to precede GABAergic interneuronal burst firing (Ohshiro et al. 2011). Though DA fibers synapse onto both GABAergic interneurons and principal neurons, there appears to be a net increase in excitatory activity within the BLA in response to DA application. This increase in excitatory activity may be the result of: 1) reduced activation of GABAergic interneurons, which occurs when activation of D2 receptors on GABAergic interneurons causes a reduction in the probability of presynaptic quantal release (Seamans et al. 2001); 2) amygdala disinhibition and the subsequent increase in excitatory activity may occur when DA suppresses GABA release from PV-interneurons onto principal neurons but not interneurons (Chu et al. 2012); or 3) DA might also increase the excitatory drive onto disinhibitory interneurons, which would subsequently increase excitatory activity (Bissiere et al. 2003; Kemppainen and Pitkanen 2000).

Serotonergic Afferents

Serotonergic projections originating from the dorsal raphé nucleus (DRN) primarily innervate BLA principal neurons, PV interneurons and interneurons containing neuropeptide Y (NPY), a subgroup of CB and SOM interneurons (Ma et al. 1991; Muller et al. 2007b). Postsynaptically, serotonin (5-HT) neurotransmission leads to the activation of 5-HT receptors, which are grouped into seven families (5-HT₁ to 5-HT₇). 5-HT_1A receptors, which are G_i/o- protein coupled, have been localized to principal neurons (Stein et al. 2000) and the presynaptic terminal of GABAergic interneurons within the BLA (Kishimoto et al. 2000). 5-HT_1A receptor activation inhibits the discharge rate of principal neurons by inducing hyperpolarization and reduce GABA release from presynaptic terminals by activating potassium channels (Kishimoto et al. 2000; Stein et al. 2000). Alternatively, 5-HT₂ receptors, and more specifically the 5-HT_2A, 5-HT_2C receptors, are G_q/11-coupled membrane receptors that increase intracellular Ca²⁺ levels and increase interneuronal excitability (Bonn et al. 2012; Jiang et al. 2009). The 5-HT_3A receptor is a ligand-gated sodium, potassium, and calcium channel that also increases interneuronal excitability but leads to a rapidly desensitizing depolarization (Gharedaghi et al. 2014; Mascagni and McDonald 2007; Rainnie 1999).

While all 5-HT receptors have been documented in the amygdala, the majority (~65–70%) of GABA-immunoreactive neurons in the BLA exhibit 5-HT_2A immunoreactivity; a smaller percentage of 5-HT_2A receptors are present on principal neurons (Bombardi 2011). 5-HT_2A and 5-HT_3A receptors have been localized to specific interneuronal types in the BLA. 5-HT_2A receptors are found primarily on PV interneurons within the BLA (Bombardi 2011; McDonald and Mascagni 2007), and tightly control glutamatergic output by perisomatic inhibition (Holmes 2008; Muller et al. 2005), whereas 5-HT_3A receptors are primarily expressed on the CCK interneuronal subpopulation (Mascagni and McDonald 2007), which constitute only a small proportion of GABAergic interneurons in the BLA (Mascagni and McDonald 2003). In contrast, 5-HT_1A, which is expressed in low to moderate concentrations in the BLA (Asan et al. 2013 737), co-express with approximately 50% of NPY interneurons (Bonn et al. 2013), and about a third of neurokinin-1 receptor (NK_1r) interneurons (Hafizi et al. 2012), while approximately 30–40% of NPY interneurons express the excitatory 5-HT_2C receptor subtype (Bonn et al. 2012; Bonn et al. 2013). Importantly, 5-HT_1A and 5-HT₃ receptors have been localized on GABAergic nerve terminals in the BLA (Koyama et al. 1999; Koyama et al. 2000), where activation of these receptors inhibits or facilitates mIPSC frequency without effects on miniature IPSC (mIPSC) amplitude, respectively (Koyama et al. 2002).

Noradrenergic afferents

The BLA receives extensive noradrenergic (NA; norepinephrine, NE) innervation from the locus coeruleus (LC) and nucleus of the solitary tract (NTS) (Pitkanen 2000; Williams et al. 2000), which synapse onto GABAergic interneurons (Li et al. 2002). NA released from LC terminals activate three distinct classes of adrenoreceptors (AR; α1-, α2-, and β-AR), which have multiple subtypes and appear to be a more potent modulator of GABAergic inhibitory synaptic transmission than DA (Miyajima et al. 2010). While no study has yet anatomically identified to what type of interneuron the receptor subunits are localized in the BLA, electrophysiological evidence indicates that α1- and α2-AR activation depolarize SOM- and CCK-positive interneurons, resulting in action potential firing and enhanced inhibitory synaptic transmission (Braga et al. 2004b; Buffalari and Grace 2007; Kaneko et al. 2008). In addition to direct enhancement of inhibitory activity by LC afferents, NA enhancement of inhibitory activity in the BLA occurs indirectly. Indeed, activation of β1- and β3-AR in lateral paracapsular (LPCS) interneurons, which are a distinct class of GABAergic interneurons that border the BLA and external capsule and are thought to provide cortical feed-forward inhibition to the BLA (Marowsky et al. 2005), enhance LPCS GABAergic inhibition of the BLA (Silberman et al. 2010; Silberman et al. 2012).

Cholinergic afferents

The BLA is extensively innervated by fibers from the substantia innominate (SI) (nucleus basalis magnocellularis) and ventral pallidum (VP) of the basal forebrain (Carlsen et al. 1985; Emson et al. 1979; Woolf et al. 1984; Zaborszky et al. 1986). The extensive innervation leads to some of the highest levels of choline acetyltransferase (ChAT), the synthesizing enzyme for acetylcholine (ACh), and acetylcholinesterase (AChE), the hydrolyzing enzyme for ACh, in the BLA compared to other brain regions (Ben-Ari et al. 1977; Prager et al. 2013). The basal forebrain projects both cholinergic and non-cholinergic neurons to the BLA. Recent evidence indicates that approximately 10–15% of basal forebrain neurons projecting to the BLA are PV-immunopositive GABAergic interneurons (Mascagni and McDonald 2009), which primarily target PV interneurons in the BLA, but also target principal neurons (McDonald et al. 2011). By comparison, approximately 75–80% of the basal forebrain projection neurons are cholinergic (Carlsen et al. 1985; Mascagni and McDonald 2009; Zaborszky et al. 1986) and project primarily to dendritic shafts and spines of BLA principal neurons, but also innervate, to a small extent (approximately 7% of postsynaptic targets) PV GABAergic interneurons (Muller et al. 2011).

Stimulation of afferents from basal forebrain cholinergic neurons lead to the release of ACh, which extensively regulates neuronal excitability by acting on muscarinic (mAChR) and nicotinic (nAChR) acetylcholine receptors, both of which are abundantly present in the BLA (Hill et al. 1993; Mash and Potter 1986; Pidoplichko et al. 2013; Swanson et al. 1987; Zhu et al. 2005). mAChRs are G-protein coupled receptors that have five subtypes, designated M1-M5. M1, M3, and M5 receptors preferentially couple to G_q/11 proteins, which subsequently initiate signaling cascades that mobilize intracellular Ca²⁺, whereas M2 and M4 receptors couple to G_i/o proteins, which subsequently hyperpolarize neurons or inhibit neurotransmitter release (for a review, see Alger et al. 2014). While the BLA express M1-M4 mAChRs (Cortes et al. 1987; Mash and Potter 1986; McDonald and Mascagni 2010), M1 mAChR appears to be the predominant mAChR subtype in the amygdala. M1 mAChRs are primarily localized to principal neurons, and appear to increase excitability of BLA principal cells due to the suppression of several potassium currents, but M1 immunoreactivity has been observed in low levels on GABAergic interneurons (McDonald and Mascagni 2010; Muller et al. 2013). In contrast, M2 mAChRs are expressed on interneurons within the BLA, and predominately lead to hyperpolarization of GABAergic interneurons (McDonald and Mascagni 2011).

In the brain nAChRs are ligand-gated ion channels permeable to cations, including Ca²⁺, that produce membrane depolarization and post-synaptic excitation or the stimulation of neurotransmitter release (Dani and Bertrand 2007). nAChRs are comprised of nine different subunits (α_2–7 and β_2–4), which combine as either homomeric or heteromeric pentameric receptors (Dani and Bertrand 2007; Yakel 2013). The homomeric α₇ and the heteromeric α₄β₂ are the two major subtypes of nAChRs found in the mammalian brain (Gotti et al. 2009), have previously been found to be expressed in the BLA (Hill et al. 1993; Seguela et al. 1993), and appear to regulate neuronal excitability by presynaptically modulating neurotransmitter release, or directly regulate neuronal activity by their position on somatodendritic sites of interneurons or principal neurons (Klein and Yakel 2006; Pidoplichko et al. 2013). In addition, α₃β₄-nAChRs are also present on GABAergic interneurons and enhance GABAergic inhibitory synaptic transmission (Zhu et al. 2005).

The functional activity and subsequent modulation of either inhibition or excitation by mAChRs and nAChRs in the BLA appears to diverge from anatomical evidence. Indeed, while α₇-nAChRs are present on GABAergic interneurons and principal neurons and enhance both excitatory and inhibitory synaptic transmission, their activation powerfully modulates GABAergic inhibition, resulting in a net reduction in BLA excitability (Pidoplichko et al. 2013). What’s more, optogenetic activation of basal forebrain inputs to the BLA during periods of neuronal quiescence does not trigger excitatory responses; rather, muscarinic activation increased the inhibitory response, which may be a result of the contrasting spatiotemporal profile of cholinergic receptor activation (Unal et al. 2015). Importantly, light-induced activation of basal forebrain inputs transiently silence cells, which was followed by a longer duration inhibitory postsynaptic potential (IPSP) (Unal et al. 2015). This suggests the early IPSP is due to activation of nicotinic receptors and, based upon the results by Pidoplichko and colleagues (Pidoplichko et al. 2013), presumptively α₇-nAChR activation, though the subunit configuration was not tested by Unal and colleagues. By contrast, the late IPSP was mediated by M1 and not M2 mAChR activation. It is important to emphasize that this inhibitory effect only occurred in quiescent principal neurons. During periods of strong activation, mAChR inhibition appeared overwhelmed and M1 mediated excitation predominated (Unal et al. 2015).

Glutamate Receptors

GABAergic interneuronal excitability in the BLA is regulated, in part, by principal neurons within the BLA. Glutamatergic inputs make dual component synapses with both fast α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptors (AMPAR) and slower N-methyl-D-aspartate receptors (NMDAR), which are present on the postsynaptic membrane of interneurons and principal neurons (Mahanty and Sah 1998; McDonald 1992; Sah et al. 2003; Smith et al. 1998; Weisskopf and LeDoux 1999). Glutamatergic inputs to interneurons express AMPARs that have rapid kinetics and strong inward rectification, indicating calcium permeability and a lack of the GluA2 subunit (Mahanty and Sah 1998; Polepalli et al. 2010). Interneurons also express a heterogeneous population of NMDARs. These cells can be separated into groups that lack NR2B NMDAR subunits (Mahanty and Sah 1998; Polepalli et al. 2010), express NMDARs that contain mostly NR2B subunits and have fast decay kinetics (Polepalli et al. 2010; Williams 1993), or express NMDARs that have slow kinetics and largely contain GluN1/GluN2B heterodimers (Polepalli et al. 2010; Spampanato et al. 2011; Szinyei et al. 2003).

The heterogeneity of the subunits of AMPA and NMDA receptors within specific populations of interneurons is essential to the regulation of feedforward inhibition to principal cells. Polepalli and colleagues demonstrated, for example, that long-term potentiation (LTP) to interneurons is restricted to interneurons that contain GluR2-lacking AMPAR at the postsynaptic membrane. Importantly, it was only these neurons that provided feedforward inhibition to principal cells (Polepalli et al. 2010) and while it has not been specifically tested, it is likely that the CB-immunopositive interneurons may be the subpopulation of interneurons that provide the feedforward inhibition to principal cells (Unal et al. 2014).

In addition to AMPA and NMDA receptors, kainate receptors represent a distinct class of ionotropic glutamate receptors that are preferentially activated by kainic acid. Kainate receptors consist of five different subunits, namely GluK_1–3 and KA_1–2 (Chittajallu et al. 1999); the GluK_1–3 subunits form functional homomeric and heteromeric channels, whereas the KA subunits only generate functional receptors with distinct physiological properties when combined with the GluK_1–3 subunits (Herb et al. 1992; Schiffer et al. 1997). While kainate receptors are not widely distributed in the brain, the BLA has a markedly high expression of GluK1R subunit (Braga et al. 2003). Kainate receptors, and in particular the GluK1 subunit containing kainate receptor, have been found to enhance excitatory synaptic transmission in the BLA (Li and Rogawski 1998) by modulating pre- and postsynaptic release of glutamate (Braga et al. 2004a; Jiang et al. 2001). Moreover, presynaptic GluK1 containing kainate receptors have also been found to modulate GABA release in the BLA in a bidirectional manner. At low concentrations, activation of high affinity GluK1-containing kainate receptors depolarize both principal neurons and GABAergic interneurons, which leads to a substantial increase in GABA release. However, high concentrations of agonists activate low-affinity presynaptic GluK1-containing kainate receptors, which again depolarize both GABAergic interneurons and principal neurons, but suppress evoked GABA release, leading to an enhancement in BLA network excitability (Aroniadou-Anderjaska et al. 2012; Aroniadou-Anderjaska et al. 2007; Braga et al. 2003).

GABA_B Receptors

The late or slow component of inhibitory synaptic transmission is mediated by activation of GABA_B receptors, which are comprised of the GABA_B1 and GABA_B2 subunits (Craig and McBain 2014). Both GABA_B1 and GABA_B2 mRNA are expressed in the BLA (Bischoff et al. 1999; Clark et al. 2000; Durkin et al. 1999). Postsynaptically, the G_i/o-protein-coupled GABA_B receptors (primarily the GABA_B1B isoform) mediate hyperpolarization of postsynaptic membranes by inhibiting voltage-gated Ca²⁺ activation and activating inwardly rectifying potassium channels (Couve et al. 2000; Rainnie et al. 1991b; Sugita et al. 1993). Presynaptic GABA_B receptors, and primarily the GABA_B1A isoform, on the other hand, has been found to inhibit neurotransmitter release on both excitatory and inhibitory synapses by inhibiting voltage-gated Ca²⁺ channels and possibly by interacting with vesicular release machinery (Gassmann and Bettler 2012; Szinyei et al. 2000; Yamada et al. 1999).

Anatomically, GABA_B1 receptors are found in many amygdala nuclei. However, in the BLA, GABA_B receptor immunoreactivity is primarily found on GABAergic interneurons; very few principal neuronal somata in the BLA exhibit immunoreactivity for GABA_B1 receptors (McDonald et al. 2004). However, the light staining found in the neuropil is likely due to the staining of dendritic shafts and spines belonging to pyramidal cells (McDonald et al. 2004). Of the subpopulations of GABAergic interneurons, GABA_B1 receptor immunoreactivity is found primarily on large CCK+ neurons, but is also found, to a lesser extent, on small CCK+ interneurons as well. In addition, GABA_B1 receptor immunoreactivity is also found on the remaining subpopulations of GABAergic interneurons in the BLA (e.g., SOM, PV, and VIP neurons) (McDonald et al. 2004). More recently, GABA_B1 immunoreactivity has been examined at the synapse in the LA. Evidence indicates that GABA_B1 receptors are localized to the extrasynaptic terminal of both interneurons and principal neurons. Though the majority of receptors are found in the sensory afferent terminals of principal neurons (Pan et al. 2009) where they act to reduce glutamate release, GABA_B receptors are also present on inhibitory inputs to principal neurons where they act as autoreceptors (Szinyei et al. 2000).

Alterations in the modulation of inhibitory and excitatory synaptic transmission during development might be due, in part, to the activation of GABA_B receptors, which are functionally expressed early in development (Bosch and Ehrlich 2015). Indeed, the GABA_B receptor, which mediates paired-pulse depression (PPD) of sensory evoked IPSCs (Szinyei et al. 2000), is differentially regulated during development; intra-BLA inhibitory synapses show pronounced PPD in the first two weeks of development, but is reduced by the third week (Ehrlich et al. 2013). In other words, the contribution of PPD by GABA_B receptors may be primarily due to activation of GABA_B receptors in infancy, while PPD in old animals is only partially controlled by GABA_B receptors. Moreover, Bosch & Ehrlich (Bosch and Ehrlich 2015) found that presynaptic GABA_B receptors, which are present on sensory inputs to LA principal neurons, are activated as early as P8. Finally, tonic presynaptic control of ISPCs and EPSCs in the LA appears to be mediated by GABA_B receptors, and is likely enabled by ambient GABA that also emerged in adolescence (Bosch and Ehrlich 2015).

Acid Sensing Ion Channels (ASICs)

ASICs channels, and in particular the ASIC1a splice variant, is highly expressed in the BLA (Biagini et al. 2001; Waldmann et al. 1997). ASICs are heterotrimeric or homotrimeric proton-gated channels activated by extracellular acidosis, intracellular pH, and other factors (Wemmie et al. 2013). Until recently the precise role of ASIC1a activation in the BLA remained unknown. ASIC1a was thought to promote hyperexcitability as it was found to reduce fear and have antidepressant and anxiolytic effects (Coryell et al. 2009; Coryell et al. 2007; Dwyer et al. 2009; Ziemann et al. 2009). Indeed, electrophysiological evidence indicates that ASIC1a channels are present on principal neurons within the BLA and are activated by ammonium or by lowering extracellular pH, which increases glutamatergic activity (Pidoplichko et al. 2014). However, ASIC1a are also found on GABAergic interneurons within the BLA and their activation increase GABAergic activity. The increase in GABA_A receptor-mediated inhibitory synaptic transmission seems to predominate, suppressing overall excitability (Pidoplichko et al. 2014), which may be a result of the intrinsic organization of the BLA. Much of the excitatory input within the BLA is directed onto interneurons, which subsequently project back onto principal cells (Lang and Pare 1997; Smith et al. 1998).

Cannabinoid Receptors

Cannabinoids exert their effects by the activation of two known cannabinoid receptor subtypes: cannabinoid-type 1 (CB1) receptor, and the cannabinoid-type 2 receptor (CB2). The CB1 receptor, which is primarily expressed presynaptically and activated by retrograde transmission of endogenous cannabinoids, is coupled to G_i/o proteins. Activation of CB1 receptors decreases excitability of the presynaptic terminal by closing calcium (n- and P/Q type) channels, increasing G-protein coupled inwardly rectifying potassium channels, and decreasing cyclic adenosine monophosphate-dependent sodium conductance (Pertwee 1997; Schlicker and Kathmann 2001). The CB2 receptor is also coupled to G_i/o proteins, but its expression is primarily restricted to immunological tissues peripherally and is implicated in immunological functions (Schatz et al. 1997).

CB1 receptors are widely distributed in the brain, but are present in relatively high concentrations in the BLA and, in particular, are present on the presynaptic terminal of CCK interneurons (Katona et al. 2001; McDonald and Mascagni 2001b), which densely innervate principal neurons (McDonald and Pearson 1989). Activation of the presynaptic CB1 receptors on CCK GABAergic interneurons reduce the amplitude of sIPSCs, but do not affect mIPSCs (Katona et al. 2001), as CB1 receptors reduce GABA release via blockade of presynaptic N-type Ca²⁺ channels (Wilson et al. 2001). Activation of CB1 receptors has also been found to reduce excitatory synaptic transmission in the LA and decrease basal synaptic transmission, indicating that in the LA, CB1 modulation of neuronal activity was determined by CB1 receptors expressed on principal neurons (Azad et al. 2003). In addition to regulating GABA release, CB1 receptor activation appears essential for the expression of postsynaptic GABA_A receptors. The expression of α1 and α2 subunits of the GABA_A receptor are reduced in the amygdala of CB1^−/− mice. This reduction in subunit expression may be the result of a developmental neuroadaptation that compensates for the overstimulation of postsynaptic receptors due to the lack of inhibitory presynaptic activity exerted by CB1 receptors (Diana and Bregestovski 2005; Uriguen et al. 2011).

GABAergic Circuit Dysfunction within the BLA

A functional BLA GABAergic system is essential throughout one’s life. Deficiencies in GABAergic inhibitory synaptic transmission are associated with neurodevelopmental diseases, such as autism or Fragile X syndrome, but also appear in neurodegenerative diseases such as Alzheimer’s disease. In addition, deficiencies in the GABAergic system can appear as a result of brain trauma, such as after TBI, or may be acquired after SE. In this section, we will first provide an overview as to how a reduction in GABAergic inhibitory synaptic transmission within the BLA is associated with the development of anxiety. We will then provide an example of a neurodevelopmental and neurodegenerative disorder that result in deficiencies in the GABAergic system within the BLA (see Table 2). In addition, we will provide two examples of acquired GABAergic deficiencies. It must be noted that genetic variations may be an underlying factor in deficiencies of GABAergic inhibitory synaptic transmission. Unless genetic variations are directly involved in changing GABAergic function within the amygdala, we will not address these variations.

Table 2.

Summary of Alterations in the BLA GABAergic System in Disease. For citations, see text.

	Anxiety/Trauma	Autism/Fragile X	Alzheimer’s Disease	Epilepsy	TBI
Behavioral Symptoms	Anxiety, hypervigilance, stress, fear	Impaired social interaction, language/communication deficits, repetitive/restricted interest, aggression, anxiety, epilepsy	Memory impairment, impaired fear conditioning, anxiety, epilepsy	Seizures, status epilepticus	Cognitive and emotional deficits, PTE, anxiety, increased fear conditioning, PTSD
GABA Interneurons	↓ SOM/NK1r	↓ synaptic number but no interneuronal loss	↓ in total # interneurons	↓ total # interneurons ↓ density of SOM interneurons	↓ total # interneurons
GAD/GABA	↓ tonic/phasic IPSC ↓ GABA release	↓ GABA metabolism ↓ GAD65/67 ↓ tonic/phasic IPSC	No change in IPSC of 1- or 7- m/o apoE4 mice (in LA) ↑ IPSC in 18–20 m/o apoE4 mice (in LA) but < ↑ in EPSC	↓ sIPSC frequency and amplitude ↓ mIPSC amplitude	↓ sIPSC frequency and amplitude ↓ mIPSC frequency and amplitude
GABAA Receptors	↑ α2 in highly anxious rats	Delayed maturation of α1 and α2 ↓ α5, 8		↑ α1	↓ α1, β2, γ2

Anxiety and Posttraumatic Stress Disorders

Anxiety and/or stress related disorders, such as PTSD, develop when individuals are exposed to situations eliciting extreme stress or fear (Heim and Nemeroff 2001; van der Kolk 2003). These disorders are commonly associated with amygdala hyperactivity (Nuss 2015; Terburg et al. 2012) and are often treated by administering benzodiazepines, which mediate their actions via GABA_A receptors (Smith 2001). However, in many cases the treatment of anxiety disorders with benzodiazepine-like compounds may be ineffective, potentially due to deficits in GABA release, GABAergic interneuronal loss in the BLA, or alterations in GABA_A receptor functionality (Farb and Ratner 2014).

The loss of GABAergic interneurons or reductions in glutamate decarboxylase (GAD), an enzyme that catalyzes the decarboxylation of glutamate into GABA, may lead to deficits in the presynaptic release of GABA and contribute to increased anxiety and associated BLA hyperexcitability. Indeed, excess reductions in GAD, such as occurring when knocking out one of the GAD isoforms (GAD65), leads to reduced phasic and tonic inhibition, and subsequently results in BLA hyperexcitability, increased anxiety, and pathological fear memory reminiscent of PTSD (Lange et al. 2014; Muller et al. 2015; Walls et al. 2010).

Selectively targeting GABAergic interneurons in the BLA has been recently investigated for the development of anxiety-like behavior as well as in fear learning. Lesions to GABAergic interneurons that contain NK_1r, which colocalize with interneurons containing NPY, SOM, and CB, has been found to increase anxiety-like behaviors in rats (Truitt et al. 2009; Truitt et al. 2007). Notably, NK_1r containing interneurons account for only approximately 3% of the total population of GABAergic interneurons in the BLA. The loss of NK_1r containing interneurons does not result in a significant reduction in the total number of interneurons (Truitt et al. 2009). However, selective ablation of these interneurons and in particular the SOM-GABAergic interneurons, which include approximately half of the NK_1r-immunoreactive interneurons, likely disinhibit the synchronizing activity of projection neurons and may impair feedforward inhibition (Truitt et al. 2009). By comparison, increasing the number of GABAergic interneurons in the BLA will reduce anxiety and animals that had increased inhibitory neurons are less sensitive to unlearned fear, though they still could acquire conditioned fear responses (Cunningham et al. 2009). While anxiety appears to be regulated in part by NK_1r-containing interneurons, PV- and SOM-GABAergic interneurons bidirectionally regulate the acquisition of a fear memory through two distinct mechanisms. During an auditory cue, PV-interneurons are excited through direct sensory input from the auditory thalamus and cortex, and indirectly disinhibit principal neurons via inhibition of SOM neurons. However, during an aversive footshock, both PV and SOM interneurons are inhibited, most likely via the activation of other interneuron subtypes that contact both PV and SOM interneurons, suggesting that the interneurons exhibit distinct temporal dynamics that correlate with specific behavioral differences (Wolff et al. 2014).

While impaired GABA release or disinhibition of GABAergic interneurons may result in anxiety-like behavior and increased fear responses, so too will deficiencies in the activation of postsynaptic GABA_A receptors. Pharmacological alterations of GABA_A receptor activity by microinjection of GABA_A receptor agonists or antagonists induce anxiolytic or anxiogenic like effects, respectively (Barbalho et al. 2009; Da Cunha et al. 1992; Sanders and Shekhar 1995; Zangrossi et al. 1999). Moreover, highly anxious rats exhibit an increase in the expression of the α2 subunit of the GABA_A receptor in the BLA (Lehner et al. 2010; Skorzewska et al. 2014). However, it remains unknown whether alterations in the expression of other GABA_A subunits also contribute to increased anxiety. In all, these results indicate that deficiencies in GABAergic inhibitory synaptic transmission within the BLA contribute to BLA hyperexcitability and the subsequent development of anxiety and trauma-related disorders.

Autism Spectrum Disorders and Fragile X Syndrome

Autism spectrum disorders (ASDs), which include Fragile X syndrome, are a group of neurodevelopmental syndromes that are often associated with aggression, anxiety, and epilepsy (Parikh et al. 2008; Tuchman and Cuccaro 2011). Emerging evidence indicates a glutamatergic/GABAergic imbalance in multiple brain regions, including the amygdala, with greater levels of glutamatergic and reduced GABAergic activity, which results in the manifestation of symptoms associated with ASD (Coghlan et al. 2012; El-Ansary and Al-Ayadhi 2014). The amygdala has recently been implicated in ASD, including Fragile X (Suvrathan and Chattarji 2011), as increased activation of the left amygdala has been reported in functional magnetic resonance imaging studies of Fragile X patients (Watson et al. 2008).

Environmental models of autism or Fragile X knockout (KO) mice revealed deficiencies in the GABAergic system within the BLA. The reduced GABAergic inhibition appears to be a result of deficits in synaptic transmission and GABA metabolism, and not due to a loss of GABAergic interneurons (Kim et al. 2014). Indeed, in a Fragile X mental retardation 1 (FMR1) gene KO model of Fragile X syndrome, the total number of neurons, including GABA-immunopositive interneurons in the BLA is unaffected. Similarly, human studies of autistic children show little morphological alterations in the BLA as compared to developmentally typical children (see Blatt 2012 for a review). However, there appears to be a significant decrease in the total number of BLA inhibitory synaptic connections, indicating aberrant circuit development (Olmos-Serrano et al. 2010). Moreover, in the BLA of FMR1 KO mice, reductions in GAD1 mRNA and protein expression for GAD65/67 were observed and are associated with reduced presynaptic GABA release (Olmos-Serrano et al. 2010).

Though the overall number of amygdala GABAergic interneurons remains stable, mechanisms that modulate the activation of GABAergic interneurons may be impaired in ASD and Fragile X models. For example, FMR1 KO mice have reduced activation of SOM-expressing low-threshold-spiking interneurons in layer II/III of the somatosensory cortex causing reduced spike synchronization of BLA principal neurons (Paluszkiewicz et al. 2011b); reductions in spike synchronization from the cortex could subsequently impair neuronal activity in the BLA required in the expression of fear (Courtin et al. 2014) but also lead to hyperresponsivity within the amygdala (Rauch et al. 2006). In addition, in a rat model of ASD, reductions in dendritic morphology, including spine density, or distal connectivity between the PFC and BLA may lead to impaired cortical-BLA regulation (Bringas et al. 2013).

Deficits in GABAergic inhibitory synaptic transmission in ASD and Fragile X appear to also be a result of genetic variations in genes coding for particular subunits of the GABA_A receptor, as has been documented throughout multiple brain regions in ASD (Coghlan et al. 2012; Fatemi et al. 2010) and in Fragile X (Deidda et al. 2014). In the amygdala of ASD and Fragile X models, delays in the maturation of postsynaptic GABA_A receptors (Vislay et al. 2013) may lead to reductions in phasic and tonic GABAergic inhibitory synaptic transmission (D'Hulst et al. 2006; Olmos-Serrano et al. 2010). In the Fragile X model, the timing of the developmental expression of the α1 and α2 GABA_A receptor subunits is delayed, which, in turn, may have impaired the switch in GABA polarity (He et al. 2014) and proper neuronal connections in wiring up local neuron networks in the BLA (Cossart 2011; Paluszkiewicz et al. 2011a).

In addition to the deficits in phasic (synaptic) inhibition, tonic inhibition, which is mediated by extrasynaptic GABA_A receptors containing either the α5 or the δ subunits, is also compromised in the BLA of FMR1 KO mice (Martin et al. 2014) and in related ASDs (Zhang et al. 2008). Tonic inhibition, maintained by low levels of ambient GABA in the extrasynaptic space (Farrant and Nusser 2005), provides a persistent inhibitory conductance that regulates the E/I balance (Semyanov et al. 2004). The reduction in the expression of the α5 subunit of the GABA_A receptor narrows the integration window necessary for feedforward inhibition. Moreover, because of the reduced GABA release, more synchronized afferent inputs must be generated to result in an action potential and modulate the integration of postsynaptic excitatory and inhibitory potentials (Gabernet et al. 2005; Pouille and Scanziani 2001).

Alzheimer’s Disease

AD is associated with severe neuronal loss, with a predilection for brain regions within the medial temporal lobe, including the amygdala (Arnold et al. 1991; Braak and Braak 1991). Recent studies have found that in symptomatic AD patients, the amygdala, and in particular the basomedial and lateral nuclei, display between 14% and 60% volumetric loss, compared to controls, as well as non-uniform shape changes (Cavedo et al. 2011; Cavedo et al. 2014; Miller et al. 2015; Poulin et al. 2011). In addition, postmortem studies reveal that although there is damage throughout the amygdala, the degree of atrophy and neurofibrillary tangles of amygdala nuclei is greater in the corticomedial group than in the BLA, suggesting that there perhaps is a selective loss of neurons that contribute to the loss in amygdala volume (Tsuchiya and Kosaka 1990).

While the overall loss of neurons, which contributes to volumetric changes, has been observed in the amygdala, it remains unknown whether GABAergic interneuronal subpopulations or specific subunits of the GABA_A receptor are targeted in the BLA and contribute to the observed deficits in fear learning and increased anxiety. Indeed, a loss of GABAergic interneurons or alterations in the expression of GABA_A receptor subunits in the amygdala is possible, as it has been observed that in the canine prefrontal cortex PV- or CR-immunoreactive interneurons are resistant to neuronal death, whereas CB-positive interneurons are depleted (Pugliese et al. 2004), and in the mouse dentate gyrus, significantly fewer SOM-immunopositive neurons are observed, whereas in the CA1 hippocampal region, there is a significant loss of PV and SOM interneurons (Levenga et al. 2013). In addition, as observed in the hippocampus (see Armstrong et al. 2003; Iwakiri et al. 2009; Mizukami et al. 1998), GABA_A receptor subunit expression might also be reduced in the BLA, which could contribute to amygdala hyperactivity.

Though it may be hypothesized that there are alterations to the expression of GABA_A receptors and the interneuronal population in the BLA, only one study has examined alterations to the E/I balance in an AD model, and focused this examination on the LA. Using the apolipoprotein E4 (apoE4) targeted replacement (TR) mouse to model AD (Wang et al. 2005), 1-or 7- month old mice expressing apoE4 display reduced excitatory synaptic transmission in the LA but no changes in inhibitory synaptic transmission (Klein et al. 2010). However, aged (18–20 months) apoE4 animals display significant increases in both inhibitory and excitatory synaptic transmission and an increased seizure phenotype (Hunter et al. 2012; Klein et al. 2014), suggesting that increased excitatory synaptic transmission predominates. Importantly, the results indicate that it is unlikely that in the LA there are alterations in the subunit composition of the GABA_A receptor, but rather increased excitatory transmission may be the result of alterations in the presynaptic release of GABA. While it remains unknown why animals display increased excitatory activity in addition to the increased inhibitory activity, one hypothesis may be that the increased excitatory activity seen in apoE4 mice may be the result of a loss of inhibition from extrinsic afferent cortical inputs (Klein et al. 2014; Swanson and Petrovich 1998) or deficiencies in neuromodulatory mechanisms such as the loss of GABAergic interneurons but not cholinergic neurons in the basal forebrain (Loreth et al. 2012).

Epilepsy and Seizures

As a principal circuit projecting to many brain regions, hyperexcitability within the amygdala may be one source of seizure generation (Prager et al. 2013). For instance, spontaneous bursting activity has been found to first appear in the BLA of kindled animals (Smith and Dudek 1997; White and Price 1993) and seizure generation after a nerve agent exposure only occurs if AChE activity is significantly impaired in the amygdala compared to other seizurogenic brain regions (McDonough et al. 1987; Prager et al. 2013). The amygdala receives monosynaptic inputs from many frontal and temporal cortical areas that are known to generate and propagate seizure activity (Pitkanen et al. 1998). The convergence of input onto specific nuclei can then recruit a large number of neurons from interdivisional network connections, which may contribute to ictal-like activity within different amygdala nuclei. Efferents from amygdala nuclei may subsequently provide routes by which the amygdala can recruit other brain regions and lead to seizure propagation (Hirsch et al. 1997; Pitkanen 2000; Pitkanen et al. 1998).

The loss of neurons in the amygdala and subsequent reductions in amygdalar volume further implicates its role in the generation and propagation of seizures. The amygdala has previously been found to be severely damaged in patients with temporal lobe epilepsy and in both adults and children who experience SE (Pitkanen et al. 1998). While often occurring in combination with hippocampal damage, neuronal loss has been observed in the amygdale without any apparent damage to the hippocampus (Hudson et al. 1993; Miller et al. 1994; Pitkanen et al. 1998). The loss of neurons in the amygdala has also been observed in animal models with the BLA being among the highest damaged nuclei (Apland et al. 2010; Figueiredo et al. 2011; Prager et al. 2013; Prager et al. 2014a; Tuunanen et al. 1996). The loss of GABAergic interneurons in the amygdala in different seizure models has also been examined. Seizures or SE cause between 37% and 64% of GABAergic interneurons in the BLA to die, irrespective of the seizure model, though the loss of GABAergic interneurons was delayed by 7 days in animals that developed SE after a nerve agent exposure (Callahan et al. 1991; Figueiredo et al. 2011; Prager et al. 2014a); Tunnanen and colleagues found a 35% decrease in the density of SOM-immunoreactive neurons in a kindling model (Tuunanen et al. 1997).

Kindling or nerve agent induced SE causes long lasting changes in synaptic transmission in the BLA, including impaired feed-forward GABAergic inhibition and disinhibition of excitatory circuits (Rainnie et al. 1991a; Rainnie et al. 1992), and network reorganization resulting in BLA hyperexcitability (Prager et al. 2014a; Smith and Dudek 1997). The loss of feedforward inhibition has been observed indirectly as a significant increase in paired pulse facilitation beginning 24 h after SE and continuing up to 30 days after nerve agent exposure (Prager et al. 2014a; Zinebi et al. 2001), and directly as a prolonged reduction in GABA_A receptor mediated inhibitory synaptic transmission, which was likely due to the loss of GABAergic interneurons in the BLA (Prager et al. 2014b). The loss of inhibitory synaptic transmission has been found to cause a concomitant increase in excitatory synaptic transmission (Prager et al. 2014b), which is associated with an increase in both NMDA and non-NMDA receptor mediated glutamatergic transmission (Gean et al. 1989; Rainnie et al. 1992; Shoji et al. 1998).

Although alterations in GABAergic synaptic transmission have been observed in the amygdala after nerve agent induced SE, reductions in GABA_A receptor mediated IPSCs appears to be model specific. After nerve agent induced SE, there was a significant reduction in the frequency but not amplitude of GABA_A receptor mediated mIPSCs (Prager et al. 2014b), indicating that the deficits in inhibitory synaptic transmission resulted from the loss of GABAergic interneurons. However, seven to 10 days after kainate acid induced SE, there was an increase in α1 subunit and GAD expression, but a reduction in tonic inhibition (Fritsch et al. 2009). While few studies have addressed how the stoichiometry of GABA_A receptor subunits change in the BLA after prolonged SE, it is well known that the expression of GABA_A subunits are altered in other brain regions such as the hippocampus (Ferando and Mody 2012; Mohler 2006). Thus, perhaps in some cases of epilepsy, alterations in the stoichiometry of GABA_A receptor subunits may contribute to impaired inhibition in the BLA, whereas in other cases the loss of GABAergic inhibition may be the result of the death of interneurons.

Traumatic Brain Injury

Like many other disorders, TBI can affect many brain regions including the amygdala, and the disruption in neuronal excitability in surrounding regions may ultimately alter the homeostasis of the amygdala. The disruption in the E/I balance stems from an initial rise in glutamate release, which is responsible for excitotoxicity, but also a delayed disruption of excitatory glutamate circuits, which may underlie the cognitive and motor deficits observed after TBI (Guerriero et al. 2015). Alterations in both glutamatergic and GABAergic synaptic transmission and the expression of their corresponding receptors have been observed after TBI in many brain regions, including the BLA (Almeida-Suhett et al. 2014; Guerriero et al. 2015), though this work is in its infant stages. An increase in the NR1 subunit of the NMDA receptor has been observed in the amygdala two weeks after injury (Reger et al. 2012), while reductions in the α1, β2, and γ2 subunits of the GABA_A receptor were observed seven days after a mild TBI (Almeida-Suhett et al. 2014). Moreover, even when there is no overt neuronal death in the BLA, a delayed loss of GABAergic interneurons is observed after a mild TBI, which may contribute to increased anxiety-like behavior (Almeida-Suhett et al. 2014) and enhanced fear conditioning (Reger et al. 2012).

Conclusions and Future Directions

While alterations in GABAergic inhibitory synaptic transmission in different diseases were reviewed separately, it cannot go unstated that in many cases, comorbidity occurs within many of these diseases. For example, estimates of comorbidity between PTSD and some types of TBI, including combat related TBI, are as high as 73% (Hoge et al. 2008; Taylor et al. 2012). Moreover, epilepsy is often found to occur with diseases including autism/Fragile X (Berry-Kravis et al. 2010; Khetrapal 2010), schizophrenia (Kandratavicius et al. 2012), AD (Chan et al. 2015; Palop et al. 2007), and anxiety disorders (Trimble and Van Elst 2003; Vazquez and Devinsky 2003). Indeed, many of these disorders have comorbidities and often these comorbidities involve deficiencies within the BLA GABAergic system.

We are not arguing in this review that amygdala hyperactivity results in the development of symptoms associated with the diseases discussed above. Rather, the purpose of this review is to provide evidence that reduced GABAergic inhibition and alterations in the mechanisms that modulate GABAergic inhibition, contribute, in part, to amygdalar hyperexcitability; BLA hyperexcitability is common amongst these disorders and may lead to comorbid behavioral deficits. The extensive innervation of the amygdala by multiple brain regions has revealed that specific pathways modulate GABAergic inhibitory synaptic transmission and that these pathways may be disrupted in different diseases.

GABAergic activity in the BLA is modulated by dopaminergic, serotonergic, noradrenergic, and cholinergic activation, as well as the activation of various glutamate receptor subtypes, and the CB1 and ASIC1a receptor. In the diseases discussed, deficiencies in the release of monoamines or ACh, or alterations in glutamatergic receptor activity, can lead to reduced modulation of GABAergic inhibition, and more locally, greater excitation via deficiencies in either feed-forward or feedback inhibitory mechanisms. What’s more, because many of these systems are interconnected, deficiencies in one system may result in a cascading effect, which could contribute to disinhibition of excitatory neurons in the BLA, and subsequently increased anxiety-like behavior or increased seizure activity. Yet, the data in many cases is not conclusive. Much remains unknown about how alterations in neurotransmitter release, receptor activation, and stoichiometry contributes to the behavioral deficits and increased anxiety that is often associated with these disorders.

E/I balance in the amygdala is dependent upon functional neuromodulatory mechanisms and local interneuronal regulation. Neuromodulation is ineffective when there is a substantial loss of GABAergic interneurons, as has been observed in the amygdala in various neurological and neuropsychiatric disorders. In some cases it is known that a specific class of interneurons are differentially affected, while in most cases it remains unknown what type of interneuron is most susceptible to cell death. Moreover, the loss of GABAergic interneurons may be delayed as compared to the death of principal neurons. The immediate death may be due to the excitotoxic effects that occur with glutamatergic excitotoxicity (Zhou et al. 2013); however, one hypothesis may be that the delayed loss of GABAergic interneurons is due to an upregulation of D-serine, an endogenous co-agonist for NMDA receptors (Liu et al. 2009). Alternatively, even if there is no interneuronal degeneration, deficits in GAD may reduce the synthesis of GABA, which could subsequently reduce the concentration of GABA released in the synapse and impair inhibitory synaptic transmission. Alternatively, in many of the diseases discussed, alterations in expression of GABA_A receptor subunits have been observed in the BLA. While it is unknown whether these changes are transient or permanent, it can be assumed that alterations in the subunit stoichiometry may lead to reduced tonic and phasic inhibitory synaptic transmission.

This review comprised two major themes. First, we summarized the neuromodulatory systems that modulate GABA_A receptor mediated inhibitory synaptic transmission. Second we discussed how reduced GABAergic inhibition in the BLA throughout the lifespan can contribute to the behavioral manifestation of symptoms associated with autism and Fragile X, AD, epilepsy, TBI, and anxiety or stress related disorders. In each case, results indicate that BLA hyperexcitability is associated with deficits in mechanisms that modulate GABAergic inhibitory synaptic transmission, the loss of GABAergic interneurons, or alterations in GABA_A receptor subunit expression. However, in many of the diseases discussed above, much remains unknown about why the amygdala is hyperexcitable. By understanding how the GABAergic system is impaired, future research can target the functional aspects of the GABA_A receptor for potential therapeutic options. Future research might also develop new therapies that induce the growth of interneurons in specific brain regions or target and reduce excitation of the glutamatergic system. The latter option has been implemented after a nerve agent induced seizure, for example, where administering a GluK1-kainate receptor antagonist prevented neurodegeneration and associated increases in anxiety or seizure activity (Figueiredo et al. 2011; Prager et al. 2015). In all, identifying the alterations to the inhibitory system and the mechanisms that modulate inhibitory synaptic transmission is a fundamental prerequisite in the design of effective and well-tolerated therapeutic treatments for these and other neurological and neuropsychiatric disorders.

Significance Statement.

Deficits in the brain inhibitory systems can occur at any stage of life. The resulting hyperexcitability leads to the development of neurological and/or neuropsychiatric diseases. We assessed how the loss of inhibitory synaptic transmission and mechanisms that modulate inhibition in the basolateral amygdala lead to increased anxiety. In addition, we examine how different diseases including autism/Fragile X syndrome, Alzheimer’s disease, traumatic brain injury and epilepsy result in amygdalar hyperexcitability. By evaluating how deficiencies in inhibition within the amygdala contributes to these diseases, future research may be directed at developing new therapies to reduce excitability, which may alleviate the behavioral symptomology of neurologic diseases.

Abbreviations

5-HT

serotonin

ACh

acetylcholine

AChE

acetylcholinesterase

AD

Alzheimer’s disease

AMPA

α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid

apoE

apolipoprotein E

AR

adrenoreceptor

ASD

autism spectrum disorder

ASIC

acid sensing ion channel

Aβ

amyloid β protein

BLA

basolateral nucleus of the amygdala

CA1

cornus ammonis 1

CB

calbindin

CB1

cannabinoid receptor 1

CB2

cannabinoid receptor 2

CCK

cholecystokinin

ChAT

choline acetyltransferase

CR

calretinin

DA

dopamine

DRN

dorsal raphé nucleus

E/I

excitation and inhibition

FMR1

fragile X mental retardation 1 gene

GABA

gamma-aminobutyric acid

GAD

glutamate decarboxylase

GluK

kainate receptor

IPSC

inhibitory postsynaptic current

IPSP

inhibitory postsynaptic potential

KO

knockout

LA

lateral amygdala

LC

locus coeruleus

LTP

long-term potentiation

LPCS

lateral paracapsular interneuron

mAChR

muscarinic acetylcholine receptor

mIPSC

miniature inhibitory postsynaptic current

NA

noradrenaline

nAChR

neuronal nicotinic acetylcholine receptor

NE

norepinephrine

NK_ir

neurokinin 1 receptor

NMDA

N-methyl-D-aspartate receptor

NPY

neuropeptide Y

NTS

nucleus of the solitary tract

P

postnatal day

PD

Parkinson’s disease

PFC

prefrontal cortex

PPD

paired pulse depression

PTSD

posttraumatic stress disorder

PV

parvalbumin

SE

status epilepticus

SI

substantia innominate

sIPSC

spontaneous inhibitory postsynaptic current

SN

substantia nigra

SOM

somatostatin

TBI

traumatic brain injury

TR

targeted replacement

VIP

vasoactive intestinal peptide

VP

ventral pallidum

VTA

ventral tegmental area