«Distribution Agreement In presenting this thesis or dissertation as a partial fulfillment of the requirements for an advanced degree from Emory ...»
Two specific, overlapping GABAergic circuits are implicated in regulating CP timing – GABAA receptors containing the α1 subunit as well as PV+ interneurons. GABAA receptors are pentameric channels composed of subunits from a variety of families, including α, β, γ, δ, ε, θ, π, and ρ. The most common subtype configuration includes 2 α, 2 β, and 1 γ subunit. The α subunits influence the kinetics, localization, and drug sensitivity of GABAA receptors (Nusser et al., 1996;
Hevers and Luddens, 2002). There are six known α-subunits, and the relative expression of these subunits changes throughout development. As the brain develops, expression shifts from the α2 subunit toward the α1 subunit, which confers faster kinetics (Dunning et al., 1999; Cohen et al., 2000; Bosman et al., 2002; Mohler et al., 2004; Bosman et al., 2005b; Eyre et al., 2012).
Emergence of the α1 subunit has specifically been linked to CP timing. Selective deletion of fast synaptic GABA activity prevents CP onset (Hensch et al., 1998), and activation of receptors containing the α1 subunit directly influences CP onset (Huntsman et al., 1994; Fagiolini et al., 2004). Further supporting a role for the α1 subunit in postnatal development, genetic deletion of this subunit preserves GABAergic terminals early in postnatal development, but GABAergic transmission is perturbed by P11 and lost by P18 (Fritschy et al., 2006).
The α1 subunit is the most highly expressed α-subunit in the mature brain, being included in 60% of adult GABAA receptors and having dense immunoreactivity and in-situ hybridization throughout the adult brain (Hornung and Fritschy, 1996; Davis et al., 2000) ( Laurie et al., 1992, 1331359; Hornung and Fritschy, 1996; Davis et al., 2000). The adult, mouse BLA contains high expression of the α1 subunit (Heldt and Ressler, 2007; Pirker et al., 2000, 11113332) and postsynaptic currents (PSCs) mediated by both α1- and α2-containing GABAA receptors (Marowsky et al., 2004). During the first three weeks after birth, α1 mRNA emerges and reaches full expression in the rat BLA (Zhang et al., 1992). Supporting a role for the α1 subunit in BLA development, transgenic mice lacking this subunit exhibit enhanced conditioned fear as adults (Wiltgen et al., 2009). Conversely, knockdown of α1 in the BLA selectively during adulthood has not been shown to impact emotional behavior (Heldt and Ressler, 2010), raising the possibility that changes in α1 expression in the BLA have the most potent effect during development, when inhibitory synaptic transmission is helping to organize amygdala maturation.
As described in Section 126.96.36.199, PV+ interneurons are a subtype of basket cells that play a fundamental role in the function of cortical-like circuits, including the BLA. These interneurons and their function primarily emerge postnatally and also influence circuit development, likely through their interaction with GABAA receptors containing α1 subunits. In the hippocampus, release sites of PV+ interneurons have a high abundance of GABAA receptor α1 subunits, relative to synapses of other interneuron subtypes (Nusser et al., 1996; Fritschy et al., 1998; Pawelzik et al., 1999; Thomson et al., 2000; Nyiri et al., 2001; Klausberger et al., 2002).
In the BLA, PV+ interneurons do not emerge until P17 and reach maturity at P30 (Berdel and Morys, 2000), during the same window when α1 expression emerges. In contrast, principal neurons are numerous in the BLA at birth (Berdel et al., 1997a) and interneurons containing the calcium binding proteins calbindin and calretinin can be found in the incipient BLA as early as E14-15 (Legaz et al., 2005). While maturation of PV+ interneurons specifically in the BLA has not been studied, a similar population in the mouse neocortex matures significantly from P7 to P21, showing increased high frequency spike discharge, alterations to intrinsic membrane properties (Goldberg et al., 2011) and faster action potentials (Lazarus and Huang, 2011). A similar developmental trajectory is observed in these interneurons in the rat striatum (Plotkin et al., 2005). The axonal arbors of basket cells expand markedly between P7 and P28 (Doischer et al., 2008) as their synapses on pyramidal cells increase in strength (Kuhlman et al., 2010).
PV+ interneurons associate with another marker of circuit maturity, specializations of the extracellular matrix known as perineuronal nets (PNNs). PNNs are found in the adult BLA in humans, rhesus macaques, and rodents (Hartig et al., 1995; Pantazopoulos et al., 2008; Gogolla et al., 2009). These structures are responsible for the stabilization of synapses and are preferentially found on fast-spiking basket cells that express parvalbumin, and degradation of PNNs increases the excitability of these interneurons (Dityatev et al., 2007). The formation of PNNs during development and the subsequent increase in activation of PV+ interneurons contribute to the closing of CP plasticity (Pizzorusso et al., 2002; Hensch, 2005; Nowicka et al., 2009).
Importantly, reducing excitability of PV+ interneurons by disrupting PNNs causes re-opening of these CPs (Pizzorusso et al., 2002), and disruption of PNNs and CP reopening are caused by GABAA receptor antagonists (Harauzov et al., 2010). As discussed in the following section, their emergence triggers developmental changes in emotional learning (Gogolla et al., 2009).
188.8.131.52 Changing Contributions of the Amygdala to Emotional Behavior Postnatal development also involves a number of changes to the expression of emotional behaviors as well as the contribution of the amygdala to those behaviors (for reviews, see Landers (for reviews, see Wiedenmayer, 2009; Landers and Sullivan, 2012; King et al., 2013). As one example, development of the amygdala is suggested to underlie changes to avoidance behavior (Ernst and Fudge, 2009). Rat pups exhibit an unlearned defensive behavior, freezing in the presence of an adult male, around P12 but not earlier (Takahashi, 1992). Also beginning at P12, exposure to an adult male elicits activation of the amygdala, suggesting changes in the function or connectivity of the amygdala contribute to this behavioral emergence (Moriceau et al., 2004).
The contribution of the developing BLA to emotional behavior has been studied in depth using classical fear conditioning (described in Section 1.2.3), the learned association of a benign, ‘conditioned’ stimulus with a noxious, ‘unconditioned’ stimulus (Fendt and Fanselow, 1999;
Davis, 2000; LeDoux, 2000). Fear conditioning studies in humans have shown that conditioned responses increase across childhood (Gao et al., 2010). Studies of fear conditioning and the well characterized, underlying neural circuit nicely illustrate the contribution of amygdala development to a CP for emotional learning in infancy. For example, while pairing an odor (conditioned stimulus, CS) with a shock (unconditioned stimulus, US) leads to avoidance of the odor in rats as young as P10, just two days earlier, at P8, this pairing leads to paradoxical approach to the CS (Sullivan et al., 2000). Furthermore, the paradigm at P10 but not P8 causes activation of the BLA, as measured by immediate early gene expression. This early appetitive conditioning to shocks likely follows from an inability to categorize sensory stimuli; rats at P3 and P6 exhibit the same responses to milk infusions and footshocks, and not until P12 do pups exhibit appropriate and distinct responses to these stimuli (Camp and Rudy, 1988). The deficit in assigning emotional valence to perceptual stimuli early in development suggests a lack of mature functionality of the amygdala.
The amygdala directly controls the CP for the approach behavior following fear conditioning, acting downstream of the stress response. Infusion of glucocorticoids into the amygdala during training leads to precocious avoidance behavior and amygdala activation at P8, while infusion of a glucocorticoid receptor antagonist into the amygdala during training reproduces the immature approach behavior after P10 (Moriceau et al., 2006). Similarly, administration of exogenous glucocorticoids causes precocious activation of the amygdala and defensive behavior following exposure to a predator odor at P8 (Moriceau et al., 2004), and andrenalectomy blocks the emergence of this response (Takahashi and Rubin, 1993). These findings support the notion that the CP is defined by the emergence of stress-induced glucocorticoid release in the amygdala, although changes to the amygdala may also act upstream of glucocorticoid release. The SHRP ends around P12 (described in Section 184.108.40.206), and one main function may be to suppress amygdala activation before this age (Moriceau et al., 2004).
Maturation of the dopamine system also plays a role in this fear learning CP during infancy. The paradoxical approach behavior at P8 corresponds with reduced dopamine efflux in the amygdala. Furthermore, glucocorticoid infusion into the amygdala at P8 increases dopamine efflux, and dopamine receptor antagonists block the glucocorticoid-induced, precocious aversion (Barr et al., 2009).
While aversive conditioning to odors is possible as early as P10, more complex behavioral responses emerge later, as the amygdala circuit continues to develop. Defensive behaviors including freezing responses and heart rate suppression emerge around 2 weeks after birth (Campbell and Ampuero, 1985; Hunt, 1999) and CS-induced potentiation of startle responses emerges between P18 and P23, depending on the modality of the conditioned stimulus (Hunt et al., 1994; Barnet and Hunt, 2006). Also, at two weeks-old, rats do not exhibit trace conditioning, when the CS and US are presented with a fixed delay, but develop this capacity gradually over the subsequent two weeks (Moye and Rudy, 1987; Barnet and Hunt, 2005).
Finally, the expression of contextual fear memories is temporarily suppressed in mice around P30, along with learning-induced synaptic changes in the BLA (Pattwell et al., 2011). The delayed development of the capacity to learn and express fearful associations is likely due to protracted maturation of amygdala circuitry.
There is a developmental profile for the contribution of the amygdala not only to fear conditioning, but also to the learned suppression of conditioned fear known as extinction (described in Section 1.2.3). The argument for extinction involving suppression, rather than erasure, of fear conditioning stems from the persistent fear-inducing capacity of the CS; after a period following extinction training, conditioned fear responses can recover due to re-exposure to the US without the CS, termed “reinstatement,” or following exposure to the environmental context of the CS-US pairing, termed “renewal” (Bouton et al., 2006; Myers and Davis, 2007).
However, early in development extinction seems to constitute permanent erasure of fear, as animals younger than P24 do not exhibit reinstatement or renewal of extinguished fear (Kim and Richardson, 2007, 2008). Developmental changes in synaptic plasticity in the BLA are implicated in this switch, as NMDA receptors are required for extinction at P24 but not the fear erasure exhibited at P17 (Langton et al., 2007). Interestingly, inactivation of the mPFC interferes with extinction at P24 but not at P17, suggesting the late development of connectivity of the BLA and mPFC also influences the mechanism of extinction (however, see Nair et al., 2001; Kim et al., 2009; Sotres-Bayon and Quirk, 2010).
A number of studies specifically implicate maturation of the amygdala GABA system in the development of fear learning. The emergence of conditioned avoidance at P10 corresponds with a switch in synaptic plasticity in the BLA; tetanic stimulation of the external capsule induces LTP in the BLA before P7 and LTD at P10, but application of a GABAA receptor antagonist in adulthood restores the immature form of synaptic plasticity (Thompson et al., 2008). Nearly every study of LTP in the adult BLA includes application of GABAA antagonists because stimulation of BLA inputs elicits feed-forward inhibition that shunts excitatory input and blocks LTP (Rainnie et al., 1991a; McKernan and Shinnick-Gallagher, 1997; Weisskopf et al., 1999; Li et al., 2011).
That GABAA receptor antagonists are not required to elicit LTP at P7 suggests feed-forward inhibition in the BLA may not be present at this young age. GABAergic inhibition not only suppresses LTP in the adult BLA, but also prevents generalization of conditioned fear to nonpresented stimuli, a process implicated in anxiety disorders (Shaban et al., 2006; Bergado-Acosta et al., 2008). Interestingly, juvenile mice exhibit generalization of conditioned fear (Ito et al.,
2009) which may be due to immature GABA circuits being unable to limit synaptic plasticity in the amygdala. Amygdala GABAA receptors are also required at P18 for infantile amnesia, the short-term suppression of learned fear associations during infancy (Kim et al., 2006b; Tang et al., 2007).
Developmental changes to GABAergic transmission in the amygdala are also implicated in the maturation of fear extinction, which is a GABA-receptor dependent phenomenon in adulthood (see Section 1.2.3). The switch in extinction mechanisms, from erasure to suppression, likely depends on maturation of amygdala inhibition, as activation of amygdala GABAA receptors is required for extinction at P24 but not the fear erasure exhibited at P17 (Kim and Richardson, 2007, 2010). Furthermore, PNNs (described in Section 220.127.116.11) emerge around P24 and their degradation in adult animals reverts extinction to the immature, erasure mechanism (Gogolla et al., 2009).