«Distribution Agreement In presenting this thesis or dissertation as a partial fulfillment of the requirements for an advanced degree from Emory ...»
5.4.3 PS reduced excitability of BLA principal neurons in adulthood We also measured neuronal excitability in patch clamped neurons by eliciting trains of action potentials with direct current injection (Figure 5.8). By injecting depolarizing current into principal neurons at resting membrane potential, we were able to calculate rheobase, the minimum current required to elicit an action potential. We found an age-dependent increase in the rheobase. There was nearly a full order of magnitude increase in rheobase by P60 from initial values of 24.4 ± 4.3 pA in control pups and 28.8 ± 3.4 pA in PS pups at P10 (Figure 5.8A; Twoway ANOVA, main effect of age: F5,135 = 53.55, p 0.0001). There was a significant interaction of age and PS, such that neurons from PS animals had a rheobase nearly twice as large as those from controls at P60 but not earlier (323.3 ± 32.0 pA vs. 176.3 ± 17.7 pA for controls, Bonferonni post-test, p 0.01; Two-way ANOVA, interaction effect: F5,135 = 3.24, p 0.01).
By increasing the magnitude of the direct current injections, we determined the maximum steady-state action potential frequency for BLA principal neurons (Figure 5.8B). There was a significant increase with age but no effect of PS (Two-way ANOVA, main effect of age: F5,128 = 10.93, P 0.0001; main effect of stress: F1,128 = 0.02, p 0.05). However, neurons from PS rats required more current to elicit the maximum spike rate, consistent with the reduced excitability already observed; this effect of PS was most evident at P60, when PS neurons required 44% more current than controls (Figure 5.8C; 1582 ± 113 pA vs. 1100 ± 184 pA for controls; Two-way ANOVA, main effect of stress: F1,128 = 5.32, p 0.05). The reduced excitability in mature neurons from PS rats is clearly portrayed by their F-I curves, with right-shifted curves for PS neurons at P28 and P60 relative to controls (Figure 5.8D).
5.4.4 PS altered the developmental trajectory of GABAergic transmission in the BLA Considering the role of GABAergic transmission in coordinating neural development, we sought to determine if early deficits in this neurotransmitter system were caused in the BLA by PS (Figure 5.9). We began by characterizing the waveforms of GABAA receptor-mediated postsynaptic currents (PSCs) in developing BLA principal neurons (Figure 5.9A). Stimulationevoked GABAA PSCs from control animals had consistent kinetics across the first postnatal
month but became slower in adulthood (Figure 5.9B; Two-way ANOVA, main effect of age:
F5,142 = 3.9, p 0.01). Interestingly, GABAA PSCs were significantly slower in neurons from PS rats relative to those from control rats, specifically during a window early in the first postnatal month, but were faster than in neurons from control rats in adulthood. As shown in Figure 5.9C, at P14 the decay time constant of GABAA PSCs was 59% larger in neurons from PS animals (15.7 ± 2.3 ms) than those from controls (9.8 ± 1.8 ms, Bonferroni post-test, p 0.05; Two-way ANOVA, interaction effect: F5,142 = 2.79, p 0.05). Figure 5.9D depicts GABAA PSCs in neurons at P60, when the effect of PS is reversed compared to P14: decay time constants are 39% smaller in neurons from PS animals at P60 (10.7 ± 1.3 ms) than those from controls (17.4 ± 3.4 ms, Bonferroni post-test, p 0.05).
We also investigated the effect of PS on synaptic plasticity of GABAergic transmission in the developing BLA (Figure 5.10). We have previously shown a developmental change in short-term plasticity of GABAA receptor-mediated PSCs in the BLA, such that immature PSCs exhibit short-term depression that is lost with age (Ehrlich et al., 2013). As expected, in neurons from control animals we found short-term depression of the fifth pulse of a train of 20Hz GABAA PSCs at P10 (66% of pulse 1), P14 (72% of pulse 1), and P17 (94% of pulse 1) that was lost by P21 (102% of pulse 1; Figure 5.10A; Two-way ANOVA, main effect of age: F5,129 = 7.02, p 0.0001). However, short-term depression was also present at P60 (75% of pulse 1), creating an inverted-U of pulse ratio vs. age.
PS had no significant effect on short-term plasticity of GABAA receptor-mediated PSCs.
However, PS tended to promote short-term facilitation during a specific developmental window, increasing the amplitude of pulse 5 relative to pulse 1 starting at P21 (Figure 5.10A; Two-way ANOVA, main effect of PS: F1,129 = 2.59, P 0.05). As illustrated in Figure 5.10B, PSCs in a number of neurons from PS animals at P21 and P28 exhibited short-term facilitation, a phenomenon scarcely observed in neurons from control animals. At P21, only 1 of 9 control neurons exhibited PSCs with short-term facilitation greater than 15% (on average, 102 ± 18% of pulse 1) while 6 of 14 PS neurons exhibited such a facilitation (on average, 131 ± 27% of pulse 1). For trains of IPSCs elicited at 50Hz, a similar developmental trajectory to 20Hz trains was observed for both control and PS neurons, but short-term depression was present at all ages and in both groups, such that the curves were shifted lower on the y-axis (Figure 5.10C).
5.4.5 PS reduced BLA expression of the GABAA receptor α1 subunit during a development critical period The timing of developmental critical periods is regulated by GABAA receptors containing the α1 subunit (Fagiolini et al., 2004). Having observed the postnatal emergence of some effects of PS on emotional behavior and BLA neuron physiology, namely principal neuron excitability and kinetics and plasticity of GABAergic transmission, we hypothesized early changes to α1 subunit expression in the BLA could contribute to the altered developmental trajectory following PS. We characterized expression of mRNA for the α1 subunit in the BLA of PS and control rats across postnatal development using quantitative RT-PCR (Figure 5.11). PS caused a significant and robust reduction in α1 expression, beginning around P17 (Figure 5.11A; Two-way ANOVA, main effect of PS: F1,20 = 14.13, P 0.01). Expression of α1 subunit mRNA in the BLA at P10 and P14 was only reduced in PS animals by 1.17- and 1.27-fold, respectively; by P17, α1 expression was reduced by 2.47-fold in the BLA of PS animals relative to controls, and this reduction increased to 3.19- and 3.45-fold by P21 and P60, respectively (Figure 5.11B).
5.5 Discussion We have provided the first evidence that amygdala electrophysiology is altered by ELS.
By characterizing developmental trajectories, we identified age-specific changes to GABAergic transmission and receptor expression in the infant BLA due to PS. Considering the role of the GABA system in shaping circuit maturation, these effects of PS are likely to influence other aspects of amygdala development. In addition, we identified changes due to PS to neuronal intrinsic physiology as early as P10 and found reductions in BLA neuronal excitability that emerge in adolescence. Our PS paradigm reduced amygdala-dependent behavior, in agreement with reduced BLA principal neuron excitability, suggesting the effects of PS reported here may underlie the risk it confers for psychiatric disorders, including SZ and ASDs. The results of this study are summarized in Figure 5.12.
5.5.1 PS Reduces Anxiety-Like Behavior and May Reduce Sociability We found that PS significantly reduced anxiety-like behavior in offspring, manifested as increased time spent on the open arms of the EPM and more head dips. While a number of studies have described anxiogenic studies of PS (Zagron and Weinstock, 2006; Baker et al., 2008;
Darnaudery and Maccari, 2008), there is also a precedent for anxiolytic effects. For example, in one study PS throughout gestation had an anxiolytic effect on the EPM in male offspring, increasing the time spent in open arms (Estanislau and Morato, 2006). In another study, PS on the last 11 days of gestation significantly increased the number of head dips in the open arm of the plus maze (Mairesse et al., 2007). Reduced anxiety-like behavior in adulthood may be due to dampened amygdala activity, and we observed reduced excitability of BLA neurons in adult PS offspring. Interestingly, a study comparing predictable and unpredictable stress restricted to the last week of gestation found that the predictable stressor elicited the most robust increase in anxiety-like behavior and stress reactivity (Richardson et al., 2006), suggesting the effects we observed here may be caused by unpredictable stress application late in gestation. Future studies will be required to address whether the anxiolytic effect of PS constitutes dampened emotionality, or instead reflects stress inoculation and increased resilience or increased risk-taking behavior.
We also found a trend toward reduced anxiety-like behavior in PS offspring as measured by isolation-induced USVs, considered an expression of anxiety-like states (Hofer et al., 2002).
Although this data is preliminary, PS rats at P17 tended to exhibit fewer calls than age-matched controls upon separation from their litters. In contrast, at P7 very little difference was observed, raising the possibility that the anxiolytic effects of PS emerge between P7 and P17. Interestingly, alterations to GABAergic transmission in the BLA emerged around P17, which may underlie the effect on anxiety-like behavior. PS has been previously shown to suppress USVs in pups at P14, and was interpreted by Morgan and colleagues as an expression of behavioral inhibition (1999).
Isolation-induced USVs are an inherently social behavior (Hofer et al., 2002; Harmon et al.,
2008) and reduced USVs may alternatively reflect ASD-like deficits. Less frequent USVs have been previously reported in rodent models of ASDs (Umeda et al., 2010; Higashida et al., 2011).
Diminished USVs in rat pups may model the lack of crying observed in some children with ASDs (Crawley, 2007).
As another indication of reduced social behavior in our PS animals, we found a trend towards decreased sociability. PS rats spent less time interacting with conspecifics in the social choice test, and may have reduced preference for a novel conspecific over a novel, inanimate object. Additional studies are needed to support this preliminary finding. If confirmed, these data would further suggest our PS paradigm reduces social behavior and models some aspects of ASDs (Moy et al., 2004; Moy et al., 2009; Mines et al., 2010; Ryan et al., 2010). Interestingly, reduction of BLA inhibition and ablation of a subset of interneurons corresponded with reduced sociability (Truitt et al., 2007), suggesting amygdala inhibition contributes to the proper function of the amygdala in social processing. We found alterations to GABAergic transmission in the adult BLA and as early as P14, and identified potential deficits in social behavior at P17 and adulthood but not before P14.
5.5.2 PS Alters Electrophysiological Properties of Developing BLA Neurons We identified a number of changes to electrophysiological properties of BLA neurons due to PS, some detectable throughout development and some beginning at specific ages. RMP and AP threshold typically become more hyperpolarized in BLA principal neurons across the first few postnatal weeks (Ehrlich et al., 2012), and both were more hyperpolarized in PS animals than controls throughout development. The effect of PS on RMP and AP threshold was around 1-2 mV on average at each time point. Corroborating the effect on AP threshold, AP amplitude was also increased due to PS, possibly reflecting greater expression of voltage-gated sodium currents mediating the AP. Conversely, there was no effect of PS on input resistance or membrane time constant; both properties exhibited the normative decrease with age until around P21.
Shifting RMP more hyperpolarized should serve to make BLA principal neurons less excitable, while the same change to AP threshold should increase excitability. Therefore, these two effects may negate each other, with no net change in neuron excitability. Supporting this notion, rheobase, a measure of neuronal excitability, was unaffected by PS in neurons before P28, despite the early changes to RMP and AP threshold. However, the effect of PS on AP threshold may function to shift the operational voltage range of these neurons, thereby reducing the influence on action potential generation of currents that are normally active near threshold, like IT and IA. The lack of effect of PS on rheobase suggests no direct effect on neuronal excitability, but interfering with the propensity to activate IT and IA may reduce the propensity of a neuron to exhibit membrane potential oscillations, which are critical for the mature function of BLA principal neurons (see Chapter 6).
PS also caused a non-significant increase in the amplitude of the fAHP at P28 and P60.
The larger fAHP may be due to the larger AP amplitude, providing greater activation of voltagegated potassium channels that repolarize the AP and contribute to the AHP. However, the larger fAHP may also reflect direct alterations in the expression of those potassium channels, including the delayed rectifier, KV3 family. A larger fAHP would be expected to increase maximum spike rate and the expression of high-frequency doublet firing. However, while maximum spike rate nearly doubled from P10 to P21, as we have previously reported (Ehrlich et al., 2012), we observed no difference in maximum spike rates due to PS.
As mentioned above, PS increased the rheobase of BLA principal neurons at P28 and P60 without any effect on maximum spike rate. Clearly illustrated in the F-I curves for neurons at P28 and P60, PS right-shifted the curves, reflecting greater current required to elicit any given spike rate. Reduced excitability was found at P28 and P60 despite no reduction of input resistance, suggesting voltage-gated currents regulating inter-spike intervals, like IA, are altered in adulthood by PS. Diminished excitability of BLA output neurons at P60 may underlie the reduced anxiety observed in the EPM and trend towards reduced sociability, as amygdala activation is known to contribute to these behaviors. Importantly, SZ patients exhibit a deficit in amygdala activation to social and emotional stimuli (Schneider et al., 1998; Baas et al., 2008; Rasetti et al., 2009), and the effect of treatment to reduce the deficit in amygdala activation predicts the behavioral benefit for individual patients (Hooker et al., 2013). Amygdala hypo-activation in SZ may be recapitulated by the reduced BLA neuron excitability seen in our adult PS animals.