«Distribution Agreement In presenting this thesis or dissertation as a partial fulfillment of the requirements for an advanced degree from Emory ...»
Interestingly, the reduction of BLA principal neuron excitability following PS was not observed until P28, suggesting it may occur downstream of earlier deficits. Specifically, increased rheobase may occur downstream of the reduction of GABAA receptor α1 subunit expression, initially observed around P17 and persisting into adulthood. Reduced α1 expression may lead to BLA principal neuron hyperexcitability, because α1-containing receptors are preferentially found in perisomatic inhibitory synapses with PV+ terminals (Nusser et al., 1996;
Fritschy et al., 1998; Pawelzik et al., 1999; Thomson et al., 2000; Nyiri et al., 2001; Klausberger et al., 2002). The reduction in intrinsic excitability starting in late adolescence may be a homeostatic response to increased excitability due to loss of inhibitory synaptic transmission.
5.5.3 PS Alters GABAergic Transmission and Receptor Expression in the BLA We found that stimulation-evoked IPSCs exhibited a different developmental trajectory from spontaneous IPSCs, which we reported previously. While spontaneous IPSCs in BLA principal neurons become faster from P14 to P21, corresponding with emergence of the GABAA receptor α1 subunit (Ehrlich et al., 2013), evoked IPSCs in control animals have consistent kinetics across the first postnatal month and are slower at P60. These data in the control condition constitute a novel finding, and are consistent with findings from a similar study in the developing visual cortex, in which GABAA receptor-mediated IPSCs were also recorded with local stimulation and pharmacological isolation (Jang et al., 2010). Jang and colleagues found an approximately two-fold increase in IPSC decay time constant from P21 to P60, comparable to the 1.98-fold increase we found across the same window. In addition, a study of GABAergic transmission in the developing hippocampus reported the emergence during the first postnatal month of a slow GABAA receptor-mediated IPSC. Like our slow IPSCs, these could be elicited with electrical stimulation, and the kinetics of this IPSC became slower with age (Banks et al., 2002).
The stimulation paradigm employed here and by Jang and colleagues clearly measures distinct aspects of GABAergic transmission than recording spontaneous IPSCs, and the underlying cause for slower IPSCs in adulthood is as of yet unknown. Maturation of stimulationevoked IPSC kinetics is clearly not reflective of GABAA receptor subunit expression, but may rather reveal diminished efficiency of the GABA uptake system or increased activation of extrasynaptic GABA receptors (Roepstorff and Lambert, 1994). The slow IPSCs at P14 may reflect a shift of inhibitory synapses distally in the dendritic arbor, increasing filtering. A shift more distally could be caused by accelerated expansion of BLA principal neuron dendritic arbors or a relative decrease of perisomatic inhibition due to delayed emergence of PV+ interneurons.
The number of perisomatic synapses is typically reduced at critical period onset (Katagiri et al., 2007), and this process may be exaggerated in PS animals. The slow, evoked IPSC in adulthood may also be due to a shift in the presynaptic population recruited by stimulation towards distal dendrite-targeting interneurons, either due to emergence of interneuron subtypes, maturation of axon collateralization, or changes to interneuron excitability. The electrophysiological properties and axon collateralization of PV+ interneurons, for example, are known to change postnatally (Doischer et al., 2008; Kuhlman et al., 2010), which may influence their likelihood to be recruited by local stimulation or of synapsing on a given postsynaptic neuron. Interestingly, slow inhibition is mediated in the mature BLA by neurogliaform cells, a subset of interneurons (Manko et al., 2012). Neurogliaform cells may emerge between P28 and P60 to provide the slow inhibition we observed upon stimulation. Future studies will be required to measure spontaneous IPSCs, to determine whether a distinct population of slow IPSCs emerges after P28 or IPSCs become slower across the board.
The normative trajectory of development of IPSC kinetics was altered by PS. A transient expression of very slow IPSCs was found around P14 in BLA principal neurons from PS animals.
The early expression of slow IPSCs will likely influence the excitatory-inhibitory balance in the developing BLA and may therefore alter the subsequent maturation of GABAergic synapses (for review, see Hensch, 2005). The change of IPSC kinetics at P14 due to PS may act upstream of the profound reduction of GABAA receptor α1 expression first observed at P17. In addition, the slow IPSCs normally expressed in adulthood were not observed in PS animals, which exhibited consistent kinetics from P21 to P60. These changes will likely influence the ability of IPSCs to regulate principal neuron activity in the mature BLA (see Chapter 6). Considering the explanations proposed above for the slow IPSCs observed at P60, PS may prevent the emergence of neurogliaform cells in the BLA or reduce GABA spillover and activation of extrasynaptic GABAA receptors. Diminished extrasynaptic GABAA receptor activation may contribute to the pathophysiology of SZ, as activation of this receptor population reverses memory deficits in animal models (Damgaard et al., 2011). As stated above, the underlying causes of these changes to IPSC kinetics following PS are unknown. Future studies will be necessary to identify a structural correlate of the change in IPSC kinetics due to PS.
In addition to IPSC kinetics, PS tended to alter the development of short-term plasticity of IPSCs. We observed an increase in pulse ratio constituting a loss of short-term depression across the first postnatal month in control animals, as we reported previously (Ehrlich et al., 2013). In neurons from PS animals, IPSCs elicited at 20 Hz exhibited short-term facilitation at P21 and P28, which was not observed in control animals. In neurons from both PS and control animals, short-term depression re-emerged between P28 and P60, although the pulse ratio remained greater in the PS group. Previous studies did not proceed into adulthood to observe this novel re-emergence of synaptic depression. The same trajectory was observed for IPSCs elicited at 50 Hz, although pulse ratios were generally smaller, likely due to depletion of docked synaptic vesicles (Zucker and Regehr, 2002; Elfant et al., 2008). Short-term plasticity is known to influence information processing in neurons (Rothman et al., 2009), suggesting PS alters the function of GABA in the amygdala during adolescence. Short-term facilitation and depression occur simultaneously and act in direct opposition, and PS may act to either promote facilitation or block depression in adolescence.
The degree of synaptic facilitation is inversely correlated with the release probability of a synapse and generally also with neurotransmitter output (Dobrunz and Stevens, 1997; Zucker and Regehr, 2002), so GABAergic release in the BLA may be diminished starting at P21 in PS animals relative to controls. Sensory experience is known to trigger the loss of short-term depression of IPSCs during development (Jiang et al., 2010; Sanes and Kotak, 2011), and we suggest precocious activation of the BLA following ELS exposure may accelerate the typical developmental reduction of GABA release probability. Furthermore, parvalbumin is known to suppress short-term facilitation of GABAergic synapses (Caillard et al., 2000; Vreugdenhil et al., 2003; Collin et al., 2005; Orduz et al., 2013) and begins to emerge in the BLA around P17, when short-term depression is lost (Berdel and Morys, 2000; Davila et al., 2008). We suggest PS interferes with the emergence of parvalbumin expression in the BLA, which could contribute to the short-term facilitation in PS animals that begins following P17. We found expression of the GABAA receptor α1 subunit, which is preferentially enriched at perisomatic inputs from PV+ interneurons, is reduced starting at P17, which may also reflect the loss of parvalbumin (Nusser et al., 1996; Fritschy et al., 1998; Pawelzik et al., 1999; Thomson et al., 2000; Nyiri et al., 2001;
Klausberger et al., 2002).
PS altered the expression of GABAA receptors, causing a reduction in the expression of the α1 subunit greater than two fold starting at P17. As mentioned above, this decrease is first observed when expression of α1 and parvalbumin typically emerge. These findings corroborate a recent study by Laloux and colleagues investigating GABAA receptor α1 protein expression in the amygdala (2012). Interestingly, Laloux and colleagues reported no effect at P14 but a trend towards reduction at P22, supporting the appearance of a PS effect around P17. They were investigating the whole amygdala and found a non-significant decrease in α1 expression, suggesting this effect of PS may be specific to the BLA, where we focused. Early postnatal stress has been shown to reduce expression of the α1 subunit in the adult amygdala and hippocampus, and our findings suggest the impact of ELS may begin well before adulthood (Caldji et al., 2003;
Hsu et al., 2003). In transgenic mice lacking the α1 subunit, cortical circuits are impaired in their ability to organize gamma oscillations (Bosman et al., 2005a), which have been observed in the BLA during the expression of fear and are likely perturbed following PS (Madsen and Rainnie, 2009; Sangha et al., 2009; Pape and Pare, 2010). The reduction of α1 expression in the developing and adult BLA due to PS may therefore contribute to the reduced anxiety-like behavior we observed in those animals. Interestingly, human SZ patients exhibit reduced GABAA receptor α1 subunit expression (Glausier and Lewis, 2011) and a rodent model of Fragile X syndrome displays reduced α1 expression beginning around P17 (Vislay et al., 2013).
Furthermore, a rodent model of SZ including several aspects of disease pathophysiology is produced simply by GABAA receptor blockade in the BLA (Berretta and Benes, 2006). These findings suggest the effects of PS to reduce α1 expression in the developing BLA may influence emotional outcomes and contribute to ASD and SZ etiology.
The GABAA receptor α1 subunit plays a key role in regulating critical periods in development (Hensch, 2005). The reduction of α1 following PS may lead to delayed critical period closure in the BLA. Reducing excitability of PV+ interneurons by disrupting PNNs causes re-opening of a developmental critical period (Pizzorusso et al., 2002), an effect mimicked simply by GABAA receptor antagonists (Harauzov et al., 2010). These studies suggest the reduction in α1 caused by PS in the developing BLA could extend a critical period for amygdala development.
The closure of critical periods in the BLA, reflected by the emergence of PV+ interneurons and PNNs, is known to trigger developmental changes to extinction learning (Gogolla et al., 2009).
PS may therefore delay the emergence of mature forms of extinction learning, which has been shown following stress during infancy (Callaghan and Richardson, 2011, 2012). Changes to critical period timing may also influence plasticity in the adolescent BLA and have long-term impact on connectivity between the BLA and regions with late-developing inputs, like the PFC (Bouwmeester et al., 2002b; Cunningham et al., 2002, 2008). Finally, altered GABAergic transmission or critical period timing due to PS may perturb the excitability of the developing BLA, resulting in the reduction in BLA principal neuron excitability we observed starting at P28.
Future studies will need to address the emergence of parvalbumin and PNNs in the BLA following PS.
The effects observed here may be unique to our PS paradigm, and future studies are required to elucidate the specific aspects of this paradigm, be they timing, intensity, predictability, or sex of the offspring, that contribute to the deficits we observed. Together, our findings on the effects of PS on amygdala development and offspring emotional behavior suggest dysfunction of the developing BLA contributes to emotional dysfunction and the etiology of neurodevelopmental disorders like SZ and ASDs. PS is a known risk factor for a variety of other psychiatric disorders, including anxiety, depression, attention deficit hyperactivity disorder, many of which also involve amygdala dysfunction (Weinstock, 2001; O'Connor et al., 2003; Van den Bergh and Marcoen, 2004; Huizink et al., 2007; Ronald et al., 2010; Markham and Koenig, 2011). Together, these findings suggest early changes to BLA neuron excitability and GABAergic transmission contribute to later deficits in amygdala function and emotional behavior.
Looking ahead, hypotheses about how the changes due to PS influence the function of the amygdala must be considered in the context of the role of GABAergic transmission in shaping the function of the BLA network. As described in Chapter 2, the production of network oscillations in the BLA is critically involved in the generation of fear states and emotional learning. In Chapter 6 we show that, while GABAergic transmission can function to reduce firing rates, it can also facilitate output of the BLA by organizing the activity of groups of neurons into network oscillations. By altering network oscillations in the BLA, changes to GABAergic transmission and principal neuron physiology may influence the development of the BLA and emotional behavior.
Figure 5.1 Prenatal Unpredictable Shock Stress Paradigm Figure 5.
1 Prenatal Unpredictable Shock Stress Paradigm. Pregnant dams were placed in an operant conditioning chamber on embryonic days (E)17-20 to receive 16 pseudo-random footshocks, separated into two blocks as depicted. Non-stressed controls were placed in the chamber but received no shocks.
Figure 5.2 Increased body weight in PS rats.
Figure 5.2 Increased body weight in PS rats.
Body weight, depicted as a function of age, was significantly larger in PS rats than controls (Two-way ANOVA, main effect of PS: F1,69 = 4.09; p 0.05). At P60, there was a 10.2% increase in body weight due to PS (408.9 ± 40.7 g for PS vs.
371.1 ± 32.7 g for control). 14 ≤ N 16.
Figure 5.3 PS had an anxiolytic effect in adult rats in the elevated plus-maze (EPM).
Figure 5.3 PS had an anxiolytic effect in adult rats in the elevated plus-maze (EPM).