1. Neuroscience
Download icon

Cell-type-specific control of basolateral amygdala neuronal circuits via entorhinal cortex-driven feedforward inhibition

  1. E Mae Guthman
  2. Joshua D Garcia
  3. Ming Ma
  4. Philip Chu
  5. Serapio M Baca
  6. Katharine R Smith
  7. Diego Restrepo
  8. Molly M Huntsman  Is a corresponding author
  1. University of Colorado Anschutz Medical Campus, United States
Research Article
  • Cited 0
  • Views 1,827
  • Annotations
Cite this article as: eLife 2020;9:e50601 doi: 10.7554/eLife.50601

Abstract

The basolateral amygdala (BLA) plays a vital role in associating sensory stimuli with salient valence information. Excitatory principal neurons (PNs) undergo plastic changes to encode this association; however, local BLA inhibitory interneurons (INs) gate PN plasticity via feedforward inhibition (FFI). Despite literature implicating parvalbumin expressing (PV+) INs in FFI in cortex and hippocampus, prior anatomical experiments in BLA implicate somatostatin expressing (Sst+) INs. The lateral entorhinal cortex (LEC) projects to BLA where it drives FFI. In the present study, we explored the role of interneurons in this circuit. Using mice, we combined patch clamp electrophysiology, chemogenetics, unsupervised cluster analysis, and predictive modeling and found that a previously unreported subpopulation of fast-spiking Sst+ INs mediate LEC→BLA FFI.

Introduction

The ability of animals to learn to associate sensory stimuli with outcomes plays an important role in their survival. Specifically, animals learn to associate environmental stimuli with particular valences (rewarding or aversive outcomes). The BLA is key in the formation of these associations (Duvarci and Pare, 2014; Janak and Tye, 2015). Following a small number of presentations of a stimulus and valence information, such as a reward or threatening foot-shock, BLA excitatory PNs undergo plastic changes. For example, after a few trials BLA neurons fire selectively to novel odors that are informative about the outcome in an associative olfactory learning task (Schoenbaum et al., 1999). This plasticity is critically important as future presentations of the stimulus can then elicit robust firing in subsets of BLA PNs that output to downstream regions to guide goal-directed behavior (Beyeler et al., 2016; Janak and Tye, 2015; Quirk et al., 1995).

Accumulating evidence demonstrates that this learning induced-plasticity is input-specific (Kim and Cho, 2017; McKernan and Shinnick-Gallagher, 1997; Nabavi et al., 2014); however, the circuit mechanisms underlying this input specific plasticity remain unknown. Importantly, GABAergic inhibition regulates the ability of BLA PNs to undergo LTP with plasticity unable to occur in the presence of intact inhibition (Bissière et al., 2003; Krabbe et al., 2018). Further, local dendritic targeting GABAergic INs expressing Sst exert strong control over BLA learning processes in vivo: Sst+ INs are normally inhibited for the duration of the sensory stimulus-valence pairing and their activation impairs learning (Wolff et al., 2014). In acute slice preparations, FFI exerts this inhibitory control over BLA plasticity (Bazelot et al., 2015; Bissière et al., 2003; Tully et al., 2007). In cerebral cortex and hippocampus, FFI is mediated by perisomatic targeting, PV+ INs to ensure the temporal fidelity of PN action potential (AP) output (Glickfeld and Scanziani, 2006; Pouille and Scanziani, 2001; Tremblay et al., 2016). However, in BLA, FFI plays a dual role regulating PN AP output (Lang and Paré, 1997) and gating their plasticity (Bazelot et al., 2015; Bissière et al., 2003; Tully et al., 2007). Which IN population mediates FFI to control plasticity in BLA remains unclear: perisomatic targeting PV+ INs are candidates as they mediate FFI in cerebral cortex and hippocampus (Glickfeld and Scanziani, 2006; Pouille and Scanziani, 2001; Tremblay et al., 2016), but anatomical studies suggest that Sst+ INs may play this role in the BLA (Smith et al., 2000; Unal et al., 2014). Further, the role of Sst+ INs in regulating BLA learning in vivo (Wolff et al., 2014) suggests a role in regulating BLA plasticity via FFI.

The LEC receives extensive innervation from the olfactory bulb (Igarashi et al., 2012) and piriform cortex (Johnson et al., 2000), and neurons in this area respond to odorants (Leitner et al., 2016; Xu and Wilson, 2012). Restricted firing in LEC suggests that it may play a role in modulating odor-specific, experience- and state-dependent olfactory coding (Xu and Wilson, 2012). LEC stimulation exerts an inhibitory effect on the olfactory input from the olfactory bulb to BLA and, to a lesser extent, piriform cortex (Mouly and Di Scala, 2006). Given that BLA is involved in encoding the motivational significance of olfactory cues in associative learning (Schoenbaum et al., 1999) and that the LEC→BLA synapse is plastic (Yaniv et al., 2003), the LEC→BLA circuit could be involved in plasticity during olfactory learning. In addition to its role in olfactory processing, the LEC plays an important role in multimodal sensory processing (Keene et al., 2016; Tsao et al., 2013) and the BLA is a major target of LEC efferents (McDonald, 1998). Finally, the LEC projects along the perforant pathway to engage the hippocampal trisynaptic circuit and receives input to its deep layers from CA1 (Neves et al., 2008). In turn, it is these deep layer LEC neurons that innervate the BLA (McDonald, 1998; McDonald and Mascagni, 1997). Taken together, this raises the possibility that the LEC→BLA circuit could be involved in sensory-valence learning across sensory modalities or for more multimodal natural stimulus information under hippocampal influence. Here we address the circuit mechanisms underlying BLA dependent learning by studying the role of distinct populations of inhibitory interneurons in mediating LEC→BLA circuit motifs.

Due to the relative scarcity of cortical afferent synapses onto BLA PV+ INs (Smith et al., 2000), the role of BLA Sst+ INs in regulating learning in vivo (Wolff et al., 2014), and the importance of FFI circuits in gating BLA LTP (Bazelot et al., 2015; Bissière et al., 2003; Tully et al., 2007), we hypothesized that Sst+ INs provide FFI onto local BLA PNs. We tested this hypothesis by pairing patch clamp recordings in BLA with unsupervised cluster analysis, predictive modeling, and chemogenetic manipulations and found that a previously unreported subpopulation of fast spiking (FS) Sst+ INs mediate FFI in the LEC→BLA circuit.

Results

LEC afferents preferentially drive disynaptic FFI in BLA

To examine how LEC afferents, a major cortical source of polysynaptic inhibition to BLA (Lang and Paré, 1997), engaged BLA neuronal circuitry, we prepared acute horizontal slices of mouse brain containing both BLA and LEC and recorded synaptic responses of BLA PNs (Figure 1A, B). A single stimulation of LEC elicited evoked EPSCs (EPSCs) and IPSCs (IPSCs) in PNs (Figure 1B). EPSCs were blocked by glutamate receptor antagonists DNQX (20 μM) and D-APV (50 μM) but not by the GABAA receptor antagonist gabazine (gbz; 5 μM) (Figure 1C, D; p = 6.72 × 10−4, Kruskal-Wallis test). In contrast, IPSCs were blocked by either gbz or DNQX/APV (Figure 1C, D; p = 6.45 × 10-4, Kruskal-Wallis test), consistent with a monosynaptic glutamatergic nature of the EPSCs and a polysynaptic GABAergic nature of the IPSCs. Additionally, IPSC onset was delayed relative to EPSC onset by ~3 ms (Figure 1E; p = 0.023, paired t-test) consistent with prior reports of polysynaptic inhibitory circuits in BLA (Arruda-Carvalho and Clem, 2014; Hübner et al., 2014; Lucas et al., 2016).

LEC afferents preferentially drive disynaptic FFI in BLA.

(A) Experimental schematic. (B) Representative image of experimental preparation. LEC and BLA as well as neighboring hippocampus and piriform cortex are labeled. The external capsule can be seen running laterally of BLA and forming its border with piriform cortex. Note the stimulation electrode positioned in LEC and recording pipette in BLA. (C) Representative traces of EPSCs and IPSCs in a BLA PN in response to LEC stimulation (top: control; middle: gbz, 5 μM; bottom: DNQX, 20 μM, D-APV, 50 μM). Arrowheads: stimulation (artifacts truncated). Scale bars: 100 pA, 10 ms. (D) EPSCs blocked by DNQX/APV (Kruskal-Wallis test: p = 6.72 × 10−4; ncontrol = 8, 4, ngbz = 7, 4, nDNQX/APV = 7, 4). IPSCs blocked by gbz and DNQX/APV (Kruskal-Wallis test: p = 6.45 × 10−4; ncontrol = 8, 4, ngbz = 7, 4, nDNQX/APV = 7, 4). (E) Onset latency of IPSC is delayed relative to the EPSC (paired t-test: p=0.023, n = 8, 4). (F) 20 voltage traces from a representative BLA PN (top: control; bottom: gbz) in response to 5 × stimulation of LEC at 20 Hz. Arrowheads: stimulation. Scale bars: 40 mV, 20 ms. (G) GABAA receptor blockade increases AP firing in BLA PNs in response to LEC stimulation (Wilcoxon signed-rank test: p=0.031, n = 8, 3). (H) Baseline Vm did not differ between control and gbz (Wilcoxon signed-rank test: p=0.38, n = 8, 3). Summary statistics in E presented as mean ± standard error of the mean (s.e.m.). Summary statistics in D, G, H presented as median with interquartile range (IQR). Individual data points presented adjacent to the summary statistics. *p < 0.05, ***p < 0.001. See Figure 1—source data 1 for a table with full details on all statistical tests used in this figure. See Figure 1—source data 2 for a table of all individual data points displayed in Figure 1.

The two main archetypal circuit motifs that mediate polysynaptic inhibition are feedforward and feedback inhibition (Tremblay et al., 2016). In order to distinguish these motifs, we recorded from BLA PNs in response to a stimulus train of five pulses to the LEC at 20 Hz. If eIPSCs were the result of FFI alone, we would expect to see little to no AP activity in BLA PNs in response to stimulation. However, if the eIPSCs were a result of combined feedforward and feedback inhibition, we would expect to see the PNs fire in response to LEC stimulation. Recording from BLA PNs in current clamp (−60 mV), we found they rarely fired in response to stimulation (Figure 1F,G). These data suggest LEC stimulation preferentially drives FFI in BLA PNs.

To determine whether our experimental setup recapitulates LEC suppression of BLA PNs, as found in vivo (Lang and Paré, 1997), we tested the effect of gbz on the firing of BLA PNs in response to LEC stimulation (Figure 1F–H). However, GABAA receptor antagonism might increase the firing of BLA PNs by bringing them closer to AP threshold via membrane depolarization. To control for GABAA receptor blockade induced depolarization of the recorded PNs, we maintained the membrane voltage (Vm) of the PNs at −60 mV throughout these experiments (Figure 1H; p=0.38, Wilcoxon signed rank test). gbz mediated blockade of FFI led to a significant increase in PN firing in response to LEC stimulation (Figure 1F,G; p=0.031, Wilcoxon signed rank test). These findings confirm that the horizontal slice preparation contained a functionally complete LEC→BLA circuit supporting its use for the study of FFI.

Cell type specificity of the LEC→BLA circuit

We performed minimal stimulation experiments in the LEC→BLA circuit using whole cell patch clamp electrophysiology to examine putative unitary synaptic events between neurons (Gabernet et al., 2005; Kumar and Huguenard, 2001). Briefly, stimulation intensity was tuned to a threshold intensity where LEC stimulation elicits eEPSCs in BLA neurons with a ~ 50% success rate and an all-or-none amplitude (Figure 2—figure supplement 1); at lower stimulation intensities EPSCs do not occur. As these EPSCs likely represent the response of the neuron to the activation of a single LEC neuron, this method provides a reliable measure of the unitary EPSC (uEPSC) between the stimulated and recorded neurons (Gabernet et al., 2005; Kumar and Huguenard, 2001). To identify dendritic-targeting Sst+ and perisomatic-targeting PV+ INs in acute brain slices, we generated Sst-tdTomato and PV-tdTomato mouse lines by crossing Sst-ires-Cre and PV-ires-Cre mice to Ai9 tdTomato reporter line. Sst-tdTomato mice showed reliable labeling of Sst+ cells and low overlap with PV+ cells and PV-tdTomato mice showed reliable labeling of PV+ cells and low overlap with Sst+ cells (Figure 2—figure supplement 2). When minimally stimulating LEC we found that, whereas only a subset of BLA PNs (12/24 cells) and Sst+ INs (16/35 cells) responded with EPSCs, nearly every PV+ IN (15/16 cells) responded to the stimulation with an EPSC (Figure 2A,B). DNQX and D-APV abolished eEPSCs in all cell types, confirming the glutamatergic nature of the response (Figure 2—figure supplement 1).

Figure 2 with 3 supplements see all
Cell type specificity of the LEC→BLA circuit.

(A) Experimental schematic. (B) Proportion of neurons with detectable uEPSCs following LEC stimulation. (C) Mean traces of uEPSC successes from representative neurons. Left to right: PN (gray), Sst+ IN with a detectable uEPSC (purple), PV+ IN (green). Lower middle: mean trace of lack of response in a representative Sst+ IN without a detectable uEPSC (pink). Arrowhead: truncated stimulus artifact. Scale bars: 5 pA, 5 ms. Insets: maximal firing to a current injection; scale bars: 20 mV, 200 ms. (D) uEPSC amplitude is larger in Sst+/PV+ INs compared to PNs (Kruskal-Wallis test: p = 5.55 × 10−4). (E) Left: uEPSC 20%-80% risetime is faster in Sst+/PV+ INs compared to PNs (Kruskal-Wallis test: p = 1.66 × 10−5). Right: uEPSC τDecay is faster in Sst+/PV+ INs compared to PNs (Kruskal-Wallis test: p = 2.18 × 10−5). (F) uEPSC threshold intensity is lower for Sst+ INs (Kruskal-Wallis test: p = 3.33 × 10−4). Summary statistics in D, E, and F presented as median with IQR. **p<0.01, ***p<0.001. For all statistical tests: nSst = 16, 8, nPV = 15, 5, nPN = 12, 8. See Figure 2—figure supplement 1 for example minimal stimulation experiments and data showing the glutamatergic nature of the LEC→BLA synapse. See Figure 2—figure supplement 2 for validation of the IN reporter mouse lines. See Figure 2—figure supplement 3 for additional characterization of the LEC→BLA circuit. See Figure 2—source data 1 for a table with full details on all statistical tests used in this figure. See Figure 2—source data 2 for a table of all individual data points displayed in Figure 2 and corresponding figure supplements.

Next, we compared putative uEPSCs across cell types (Figure 2C-F). We found that the uEPSC amplitude was larger in Sst+ and PV+ INs compared to PNs (Figure 2D; p = 5.55 × 10−4, Kruskal Wallis test). Additionally, we found the uEPSC kinetics were faster in the INs compared to PNs (Figure 2E; 20%-80% risetime: p = 1.66 × 10−5, Kruskal Wallis test; τDecay: p = 2.18 × 10−5, Kruskal Wallis test). Finally, we found no differences between cell types in uEPSC latency or jitter (Figure 2—figure supplement 3). Together with our data demonstrating that BLA PNs rarely fire in response to LEC stimulation (Figure 1E, F), the statistically indistinguishable low jitters are consistent with a monosynaptic connection between LEC projection neurons and BLA Sst+ INs, PV+ INs, and PNs (Doyle and Andresen, 2001 ).

Since minimal stimulation likely reports the response of a BLA neuron to a single upstream LEC neuron (Gabernet et al., 2005; Kumar and Huguenard, 2001), we can use threshold stimulation intensity as an indirect measure of convergence of LEC projection neurons onto different BLA cell types. If threshold stimulation is lower for one BLA neuronal population compared to the others, it would follow that convergence of LEC inputs onto that cell type is likely greater relative to the other populations as it requires the activation of fewer LEC neurons. We found that threshold intensity was lower for the Sst+ INs compared to other cell types (Figure 2F; p = 3.33 x 10−4, Kruskal Wallis test).

Taken together, these data demonstrate that LEC stimulation leads to large and fast unitary currents in BLA Sst+ and PV+ INs and small and slow unitary currents in BLA PNs. Finally, although Sst+ and PV+ INs display equivalent uEPSCs, the finding that Sst+ INs have a lower threshold stimulation intensity compared to PV+ INs and PNs suggest that LEC afferents may have a greater functional convergence onto BLA Sst+ compared to PV+ INs.

A fast spiking phenotype distinguishes BLA Sst+ INs targeted by LEC afferents

Though little is known about BLA Sst+ INs, they appear to have diverse electrophysiological properties ex vivo (Krabbe et al., 2018; Sosulina et al., 2010) and responses to stimuli in vivo (Krabbe et al., 2018; Wolff et al., 2014), suggesting Sst may be expressed by a broad range of GABAergic IN subtypes, similar to cerebral cortex (Tremblay et al., 2016). Only a subset of Sst+ INs responded to cortical stimulation raising the question whether these cells represented a distinct cell type. To address this hypothesis, we ran an unsupervised cluster analysis using Ward’s method (Ward, 1963) based on 15 membrane properties from 105 Sst+ INs. Applying Thorndike’s procedure (Thorndike, 1953) suggested two distinct clusters (Groups I and II). A large majority of Sst+ INs that responded to cortical stimulation clustered into Group I (14/16 cells) whereas a large majority of non-responsive Sst+ INs clustered into Group II (17/19 cells) (Figure 3A). To better understand what membrane properties best distinguished Group I and II Sst+ INs, we used decision tree analysis (Breiman et al., 1984; Therneau and Atkinson, 1984) which returned maximum firing rate, hyperpolarization induced sag, and AHP latency as the most salient parameters for cluster separation (Figure 3—figure supplement 1). To reduce potential model overfitting with the decision tree analysis, we used an additional predictive modeling technique, the random forest method (Liaw and Wiener, 2002). Like decision tree analysis, the random forest method returned maximum firing rate and hyperpolarization induced sag in addition to AP halfwidth as the most salient parameters for defining BLA Sst+ INs (Figure 3B,C).

Figure 3 with 2 supplements see all
A fast spiking phenotype distinguishes BLA Sst+ INs targeted by LEC afferents.

(A) Unsupervised clustering analysis revealed two clusters of Sst+ INs. Majority of Sst+ INs with a uEPSC present following LEC stimulation cluster in Group I (14/16, Sst+ INs with uEPSC; 60/105, all Sst+ INs). Majority of Sst+ INs with no response to LEC stimulation cluster in Group II (17/19, Sst+ INs without uEPSC; 45/105, all Sst+ INs). (B) Results of Random Forest model. Plot shows mean decrease in Gini impurity for each of the 15 parameters included in the model (nSst = 105, 23). Gini impurity measures how often a random Sst+ IN would be clustered incorrectly if labeled randomly according to the distribution of labels in the set. (C) First two dimensions of the principal components analysis (PCA) of the random forest proximity matrix (nSst = 105, 23). (D) Left: maximum AP firing from representative PV+ (top, green), Group I (middle, purple) and II (bottom, pink) Sst+ INs. Scale bars, main traces: 5 mV, 50 ms; inset: 20 mV, 400 ms. Right: voltage responses of the same neurons. Darker traces show responses to −200 pA and rheobase current injections. Lighter traces show responses to intermediate current injections were used to determine membrane resistance (−100 pA to sweep immediately before rheobase, Δ10 pA each sweep, +100 pA maximum current injection). Note lack of sag, rebound AP in Group I Sst+ IN (sag: closed arrowheads; rebound AP: open arrowheads). Scale bars: 10 mV, 200 ms; 50 pA, 200 ms. (E) APs of the same representative PV+ (green), Group I (purple), and II (pink) Sst+ INs at rheobase. Halfwidth: bar through width of AP. Scale bars: 20 mV, 1 ms. (F) Phase plots of PV+, Group I, and Group II Sst+ INs shows rate of voltage change for Group I and II APs at rheobase. Data presented as mean ± s.e.m. nPV = 52, 19, nSstI = 60, 23, nSstII = 45, 18. (G) Left: reconstructed soma and dendrites of a representative PV+ and Group I Sst+ INs (scale bar: 200 μm). Sholl rings shown beneath reconstruction. Right: confocal images of the biocytin filled PV+ and Group I Sst+ INs (scale bar: 200 μm). (H) No significant differences between PV+ and Group I Sst+ INs for soma perimeter (left; unpaired t-test: p=0.26) or area (right; Mann-Whitney U test: p=0.28). (I) No significant differences between PV+ and Group I Sst+ INs for (left to right) dendrite branchpoints (unpaired t-test: p=0.40), endpoints (unpaired t-test: p=0.30), length (unpaired t-test: p=0.97), or surface area (Mann-Whitney U test: p=0.95). (J) Results of Sholl analysis showing dendrite length (left) and surface area (right) as a function of distance from the soma in PV+ and Group I Sst+ INs. Data presented as mean ± s.e.m. nPV = 6, 4, nSstI = 8, 6. Summary statistics in H (perimeter) and I (branchpoints, endpoints, and length) presented as mean ± s.e.m. Summary statistics in H (area) and I (area) presented as median and IQR. For all statistical tests in (H) and (I): nPV = 6, 4, nSstI = 8, 6. See Figure 3—figure supplement 1 for decision tree data. See Figure 3—figure supplement 2 for data on Group II Sst+ IN morphology. See Figure 3—source data 1 for a table of summary data and statistical comparisons on all membrane properties studied for Group I and Group II Sst+ INs and PV+ INs. See Figure 3—source data 2 for a table with full details on all statistical tests used in this figure. See Figure 3—source data 3 for a table of the results of the hierarchical cluster analysis and used by the Decision Tree and Random Forest models to determine which membrane properties best distinguished the clusters (related to Figure 3A-C and Figure 3—figure supplement 1). See Figure 3—source data 4 for all individual data points of BLA IN membrane properties (related to Figure 3—source data 1). See Figure 3—source data 5 for all individual data points for phase plot in Figure 3F. See Figure 3—source data 6 for all individual data points on BLA IN morphology (related to Figure 3H-J).

Having observed two subpopulations of Sst+ INs, we compared membrane properties across Group I and II Sst+ INs and PV+ INs (Figure 3D–F, Figure 3—source data 1). These data show that Group I Sst+ and PV+ INs differed from Group II Sst+ INs in a subset of membrane properties such as maximum firing rate and hyperpolarization induced sag. Further, all three IN subtypes differed in AP halfwidth with PV+ INs firing the fastest APs and Group I Sst+ INs firing faster APs than Group II Sst+ INs (see Figure 3—source data 1 for detailed statistics on membrane properties). The membrane properties of Group I Sst+ and PV+ INs are consistent with a FS phenotype typically seen in cortical, hippocampal, and BLA PV+ INs that is characterized by the ability to fire high frequency trains of brief APs (Tremblay et al., 2016; Woodruff and Sah, 2007a; however, see Large et al. (2016); Ma et al. (2006); Nigro et al. (2018) for examples of FS Sst+ INs in cerebral cortex).

In order to compare the morphology of the IN populations, we included biocytin in the patch pipette in a subset of recordings to allow for post hoc visualization of the different IN subtypes. Using these biocytin-filled neurons, we created reconstructions of the somatic and dendritic morphology of the INs (Figure 3G; see Figure 3—figure supplement 2 for Group II Sst+ IN data which were not included in formal analyses due to low number of recovered morphologies [nSstII = 3, 2]). When we compared the somatic and dendritic morphology of PV+ and Group I Sst+ INs, we found no significant differences in any measurement (Figure 3H–J, see Figure 3—source data 2 for statistical analysis).

Taken together, the membrane properties of BLA Sst+ and PV+ INs reveal two distinct subpopulations of Sst+ INs that are readily distinguished at the biophysical level by their FS phenotype and at the functional circuit level by synaptic responses to cortical stimulation. However, the IN subtypes do not appear to have any readily observable differences in their somatic and dendritic morphology. For clarity, we refer to the Group I and II Sst+ INs as FS and non-fast spiking (nFS) Sst+ INs, respectively.

Probing the LEC→BLA circuitry suggests distinct functional feedforward/feedback roles for IN subtypes

BLA Sst+ INs have lower threshold stimulation intensity compared to PV+ INs and PNs (Figure 2F). Since these data suggest a higher rate of convergence of LEC afferents onto Sst+ compared to PV+ INs, we wanted to test the hypothesis that LEC input to BLA may preferentially recruit Sst+ over PV+ INs. To do this, we recorded from BLA Sst+ and PV+ INs in current clamp at rest and stimulated LEC with five pulses at 20 Hz. The stimulation intensity was set to the empirically derived threshold stimulation for BLA PNs (defined as the median PN threshold stimulation; 273.00 nC; Figure 2F) allowing us to determine how BLA INs respond when LEC activity is sufficient to ensure that PNs receive input. We found that stimulation led to robust spiking in BLA Sst+ INs whereas PV+ INs rarely fired (Figure 4A–C; p=0.016, Mann-Whitney U test).

Figure 4 with 1 supplement see all
Distinct microcircuit functional roles for IN subtypes.

(A) Experimental schematic for B-E. (B) Representative Sst+ (top, purple) and PV+ (bottom, green) responses to LEC stimulation at 20 Hz (arrowhead: stimulation artifacts; scale bars: 25 mV, 100 ms). Insets show maximal firing frequency for the representative IN in response to a square current pulse (scale bars: 5 mV, 200 ms). (C) Sst+ INs fire more APs per stimulus compared to PV+ INs (Mann-Whitney U test: p=0.016; nSst = 8, 3; nPV = 9, 3). (D) Sst+ INs have a more depolarized Vrest compared to PV+ INs (unpaired t-test: p=0.034; nSst = 8, 3; nPV = 9, 3). (E) Cumulative probability distribution of all subthreshold EPSPs. Inset, middle: mean subthreshold EPSP from representative neurons in (A; Sst+: purple, PV+: green; scale bars: 1 mV, 5 ms; Kolmogorov-Smirnov test: p = 2.19 x 10−21, nSst = 161 events, nPV = 779 events; Mann-Whitney U test: p = 0.030, nSst = 5, 3, nPV = 8, 3). (F) Experimental schematic for G and H. (G) Representative paired recording experiments between BLA and INs (left to right: nFS Sst+, FS Sst+, PV+). Bottom: AP in BLA PN (scale bar: 40 mV). Top: overlaid current responses of BLA INs to PN AP (50 pA, 2 ms); successful trials shown in color. Insets show maximal firing frequency and response to a −200 pA current injection in the representative neurons (scale bars: 20 mV, 200 ms). (H) uEPSC amplitude is larger in PV+ compared to FS and nFS Sst+ INs (Kruskal-Wallis test test: p=0.025, nnFS-Sst = 5, 5, nFS-Sst = 8, 6, nPV = 8, 8) (I) Experimental schematic for J-L. (J) Representative sEPSC traces from BLA INs (top to bottom: nFS Sst+, FS Sst+, PV+; scale bars: 20 pA, 100 ms). Insets show maximal firing frequency and response to a −200 pA current injection in the representative neurons (scale bars: 20 mV, 300 ms). (K) Scatter plot of sEPSC frequency and amplitude. Right: nFS Sst+ INs have smaller amplitude sEPSCs compared to FS Sst+ and PV+ INs (one-way ANOVA: p=0.022). Bottom: PV+ INs have more frequent sEPSCs compared to nFS Sst+ INs (one-way ANOVA: p=0.022). (L) nFS Sst+ IN sEPSCs have slower decay kinetics compared to FS Sst+ and PV+ IN sEPSCs (Kruskal-Wallis test: p=0.0033). Summary statistics in D and K presented in color as mean ± s.e.m. and in C, E (inset), H, and L in color as median with IQR. *p < 0.05 or False Discovery Rate corrected significance threshold where applicable, ***p < 0.001. See Figure 4—figure supplement 1 for additional characterization of the BLA inhibitory microcircuitry. See Figure 4—source data 1 for a table with full details on all statistical tests used in this figure. See Figure 4—source data 2 for a table of all individual data points displayed in Figure 4 and corresponding figure supplements.

To probe the underlying mechanisms of the preferential spiking of Sst+ INs, we used two additional measures: resting Vm (Vrest) and subthreshold EPSP amplitude. We found that Vrest was more depolarized in Sst+ compared to PV+ INs (Figure 4D; p = 0.034, unpaired t-test). Because individual INs displayed variability in the number of APs/stimulus, each IN would have a different number of subthreshold EPSPs. Therefore, we compared both the distribution of all subthreshold EPSP amplitudes and the mean cellular amplitudes across Sst+ and PV+ INs. Regardless of the analysis method, we found that evoked subthreshold EPSPs were larger in Sst+ compared to PV+ INs (Figure 4E; all subthreshold EPSP events: p = 2.19 x 10−21, Kolmogorov-Smirnov test; cellular subthreshold EPSP: p = 0.030, Mann-Whitney U test). Thus, the data show that two distinct mechanisms underlie the recruitment of BLA Sst+ INs by LEC afferents: Sst+ INs were more depolarized at rest compared to PV+ INs, and, despite the concomitant decrease in the driving force through glutamate receptors, the magnitude of evoked subthreshold EPSPs was greater in Sst+ compared to PV+ INs. Finally, when we classified the Sst+ INs from this experiment (Figure 3A), we found that all but one (7/8 cells) were FS Sst+ INs, consistent with our finding that LEC afferents appear to selectively target FS cells among the Sst+ INs.

Taken together with the data in Figures 2 and 3, these data show that LEC stimulation leads to synaptic responses in BLA FS Sst+ INs, PV+ INs, and PNs but not in nFS Sst+ INs. Although FS Sst+ and PV INs have equivalent unitary events following LEC stimulation, the responses of these cell types to LEC input diverge in important ways. Specifically, FS Sst+ INs have a lower threshold stimulation intensity compared to PV+ INs, and FS Sst+ INs have greater subthreshold EPSP amplitudes and are more likely to fire in response to PN threshold stimulation of LEC compared to PV+ INs. These findings suggest that LEC afferents to BLA have a greater functional convergence onto FS Sst+ INs compared to PV+ INs and raise the question whether they are involved in BLA FFI.

In neocortical circuits, PV+ INs mediate FFI whereas Sst+ typically mediate feedback inhibition (Tremblay et al., 2016); nevertheless, our data point to a role for Sst+ INs in BLA FFI, consistent with prior hypotheses that BLA Sst+ INs mediate FFI and PV+ INs mediate feedback inhibition (Duvarci and Pare, 2014). Thus, we wanted to test for responses of Sst+ and PV+ INs to local BLA PNs to probe for potential roles in feedback inhibition. To do this, we recorded PN-IN pairs in BLA, drove AP firing in PNs and compared uEPSC responses to the AP across IN subtypes (Figure 4F–H). We found responses in all IN subtypes (8/10 FS Sst+, 5/5 nFS Sst+, 8/12 PV+ INs with uEPSCs in response to PN APs) and found that BLA PV+ INs responded to PNs with larger amplitude uEPSCs compared to either FS or nFS Sst+ INs (Figure 4H; p=0.025, Kruskal-Wallis test). Further, we found no difference in the latency or jitter of the uEPSCs across cell types (Figure 4—figure supplement 1). These data are consistent with a larger role for PV+ INs in BLA feedback inhibition relative to Sst+ INs.

Finally, we wanted to look at the level of spontaneous excitation onto the different IN subtypes as different levels of spontaneous glutamatergic activity could lead to different levels of basal excitability for the INs in the LEC→BLA circuit (Figure 4I,J). We recorded spontaneous EPSCs (sEPSCs) across BLA IN subtypes. We found that PV+ INs had larger amplitude sEPSCs than either Sst+ IN subtype and that PV+ and FS Sst+ INs had more frequent sEPSCs than nFS Sst+ INs (Figure 4K; amplitude: p=0.022, one-way ANOVA; frequency: p=0.022, one-way ANOVA). Additionally, sEPSCs in nFS Sst+ INs had slower decay kinetics compared to the other BLA IN subtypes (Figure 4L, Figure 4—figure supplement 1; τDecay: p=0.0033, Kruskal-Wallis test). These data suggest that although FS Sst+ and PV+ INs may have divergent functional roles with regards to feedforward and feedback inhibition, they both receive greater levels of spontaneous excitatory input compared to nFS Sst+ INs.

Sst+ INs mediate LEC→BLA FFI

To test whether Sst+ or PV+ INs provide cortically evoked FFI onto BLA PNs, we generated Sst-hM4Di and PV-hM4Di mice by crossing the Sst-ires-Cre or PV-ires-Cre mice to the ROSA-hM4Di-mCitrine mice. These mice selectively express hM4Di in Sst+ (Sst-hM4Di) or PV+ (PV-hM4Di) INs. hM4Di is an inhibitory chemogenetic receptor that, when bound to its ligand clozapine-N-oxide (CNO), activates the Gi/o signaling cascade to hyperpolarize and block neurotransmitter release from neurons (Armbruster et al., 2007; Stachniak et al., 2014). To validate the efficacy of the Sst-hM4Di and PV-hM4Di mouse lines, we patched onto BLA mCitrine+ neurons and recorded Vm in current clamp (Ihold = 0 pA). Application of 10 μM CNO reduced Vm of mCitrine+ neurons in both lines (Figure 5A; pSst = 0.017, pPV = 0.031, one-sample t-test). Thus, we used the Sst-hM4Di and PV-hM4Di mice to selectively hyperpolarize Sst+ or PV+ INs.

Figure 5 with 1 supplement see all
Sst+ INs mediate LEC→BLA FFI.

(A) Left: Effect of CNO on Vm. Scale bars: 5 mV, 3 min. Right: CNO hyperpolarizes mCitrine+ cells in both mouse lines (pSst = 0.017, pPV = 0.031, one-sample t-test, nSst = 7, 4, nPV = 6, 3). (B) IPSC is delayed relative to EPSC (paired t-tests: pSst-Ctrl = 0.0018, pSst-CNO = 8.80 x 10−4, pPV-CNO = 0.0042, pWT-CNO = 0.0012; Wilcoxon signed-rank tests: pPV-Ctrl = 2.44 x 10−4, pWT-Ctrl = 0.0024; nSst = 13, 6, nPV = 13, 7, nWT = 12, 6). (C) CNO- mediated Sst+ IN inactivation reduces FFI (Sst-hM4Di mouse line; Wilcoxon signed-rank tests: pIPSC = 0.040, pIE = 0.013, n = 13, 6). Top: Experimental schema, IPSC traces with and without CNO. Arrowheads: stimulation artifacts (truncated). Scale bars: 100 pA, 50 ms. (D) PV+ IN inactivation has no effect on FFI (paired t-tests: pIPSC = 0.35, pIE = 0.85, n = 13, 7). Display as for (C), but in PV-hM4Di mouse line. Scale bars: 25 pA, 50 ms. (E) CNO has no effect on FFI in wildtype littermates (Wilcoxon signed-rank tests: pIPSC = 0.73, pIE = 0.68, n = 12, 6). Display as for (C), but in hM4Di-/- littermates of Sst- and PV-hM4Di mice. Scale bars: 50 pA, 50 ms. Summary statistics in A, B (Sst-hM4Di both conditions, PV-hM4Di CNO condition, WT littermates CNO condition), and D in color as mean ± s.e.m. Summary statistics in B (PV-hM4Di and WT littermates control conditions), C, and E in color as median with IQR. *p < 0.05, **p < 0.01, ***p < 0.001. See Figure 5—figure supplement 1 for lack of effect of IN inactivation on EPSCs. See Figure 5—source data 1 for a table with full details on all statistical tests used in this figure. See Figure 5—source data 2 for a table of all individual data points displayed in Figure 5 and corresponding figure supplements.

To assess the role of Sst+ and PV INs in BLA FFI, we stimulated LEC and recorded evoked EPSCs and IPSCs in BLA PNs using Sst-hM4Di and PV-hM4Di mice before and after bath application of CNO (10 μM). To determine the effect of IN inactivation on FFI we quantified two measures: evoked IPSC amplitude and inhibition-excitation balance (I/E balance). We defined I/E balance as the evoked IPSC amplitude for each cell normalized to its evoked EPSC amplitude. We examined this measure in addition to evoked IPSC amplitude to control for potential differences across cells with regards to the number of excitatory inputs to BLA activated by LEC stimulation. Consistent with the disynaptic nature of LEC-driven FFI in BLA (Figure 1), LEC stimulation elicited an IPSC that was significantly delayed relative to the EPSC in BLA PNs in both Sst-hM4Di and PV-hM4Di mice in control conditions and in the presence of CNO (Figure 5B, see Figure 5—source data 1 for detailed statistical analysis). Importantly, these data indicate that hM4Di expression or CNO application does not alter the ability of LEC stimulation to drive FFI. When we perfused CNO to inactivate the different IN subtypes, CNO reduced IPSC amplitude and I/E balance in PNs from Sst-hM4Di mice by 30.2% and 40.2% respectively but had no effect on either measure in PNs from PV-hM4Di mice (Figure 5C,D; pSst-IPSC = 0.040, pSst-IE = 0.013, Wilcoxon signed rank tests; pPV-IPSC = 0.35, pPV-IE = 0.85, paired t-tests). Finally, demonstrating the effects of Sst+ and PV+ IN inactivation were specific to FFI, bath application of CNO had no effect on eEPSCs in either mouse line (Figure 5—figure supplement 1).

To control for off-target effects of CNO (Gomez et al., 2017), we repeated the experiments in wildtype (WT) littermates of Sst-hM4Di and PV-hM4Di mice. In WT littermates, LEC stimulation elicited the delayed EPSC-IPSC pairing consistent with FFI (Figure 5B, see Figure 5—source data 1 for detailed statistical analysis), and CNO application had no effect on LEC-driven excitation or FFI in BLA (Figure 5E, Figure 5—figure supplement 1, see Figure 5—source data 1 for detailed statistical analysis). Taken together with our data demonstrating that LEC afferents selectively synapse onto FS Sst+ INs among the Sst+ IN subtypes (Figures 23), these data show that FS Sst+, but not PV+ or nFS Sst+, INs mediate LEC-driven FFI in BLA.

Discussion

Our data show that a previously unreported subpopulation of fast spiking Sst+ INs mediates LEC-driven FFI in BLA (Figure 6). Although LEC afferents synapsed onto both PV+ and FS Sst+ INs (Figures 2 and 3), only FS Sst+ INs fired in response to LEC activity (Figure 4); conversely, PV+ INs received stronger synaptic input from local BLA PNs than both populations of Sst+ INs. Finally, inactivation of Sst+, but not PV+, INs led to a reduction in FFI onto BLA PNs (Figure 5). Our findings are consistent with prior experiments in BLA indicating that Sst+ INs form the anatomical circuit underlying FFI (Duvarci and Pare, 2014; Smith et al., 2000; Unal et al., 2014). In summary, these data illuminate a circuit mechanism for the feedforward inhibitory control mediated by entorhinal afferents onto BLA PNs (Lang and Paré, 1997). The role of BLA FFI in gating plasticity at afferent synapses onto PNs (Bazelot et al., 2015; Bissière et al., 2003; Tully et al., 2007) and of local Sst+ INs in gating BLA-dependent learning in vivo (Wolff et al., 2014) suggests a potential role for this circuit in olfactory and other forms of sensory-valence learning (Keene et al., 2016; Kitamura et al., 2017; McDonald, 1998; Mouly and Di Scala, 2006; Schoenbaum et al., 1999; Tsao et al., 2013; Xu and Wilson, 2012).

Schematic of findings.

(A) FS Sst+ INs (purple) fire APs in response to LEC input to drive FFI onto BLA PNs. nFS Sst+ INs (pink) do not respond to LEC input. PV+ INs (green) respond with a subthreshold excitatory response to LEC input. (B) Inactivation of FS Sst+ IN removes the feedforward inhibitory control over BLA PNs. PV+ INs receive large uEPSC input from local BLA PNs compared to both Sst+ IN subtypes.

BLA Sst+ INs are a heterogenous population of dendritic targeting INs with diverse firing properties (Krabbe et al., 2018; Muller et al., 2007; Sosulina et al., 2010; Wolff et al., 2014). This is similar to what is observed in cerebral cortex (Tremblay et al., 2016). Accordingly, our cluster analysis and predictive modeling techniques revealed two distinct subpopulations of Sst+ INs that could be distinguished based on membrane properties, most notably maximum firing rate, hyperpolarization induced sag, and AP halfwidth (Figure 3). These membrane properties suggest differential expression of HCN and Kv3 channels in the Sst+ IN subpopulations (Poolos et al., 2002; Rudy and McBain, 2001). Consistent with this, Kv3.2 is expressed in a large subpopulation of BLA Sst+ INs (McDonald and Mascagni, 2006).

How IN heterogeneity maps onto specific functions within the BLA is critical for understanding survival circuits. We observed that our cluster analysis based on intrinsic properties separated the populations of Sst+ INs based on whether they displayed a synaptic response to LEC stimulation: 87.5% of Sst+ INs with a synaptic response were classified as fast spiking and 89.5% of Sst+ without a synaptic response were non-fast spiking. Although canonically PV+ INs are thought to be synonymous with FS INs, our data indicate that, at least in BLA, this is not the case. In BLA, our data reveal two functionally distinct populations of FS INs can be separated based on PV or Sst expression. Further, taken together with recent descriptions of FS Sst+ INs in cortex (Large et al., 2016; Ma et al., 2006; Naka et al., 2019; Nigro et al., 2018), we argue that a fast spiking phenotype is insufficient to classify an IN solely as PV+. Indeed, despite similar physiological phenotypes, in neocortex, FS Sst+ INs have distinct functions from PV+ INs. For example, whereas a well characterized circuit role of PV+ INs is their control of thalamocortical driven FFI onto layer 4 PNs (Tremblay et al., 2016), a recent elegant study demonstrated that neocortical FS Sst+ receive their input from layer 4 PNs and provide strong feedback inhibition to these PNs that is capable of silencing sensory evoked responses in the PNs (Naka et al., 2019). Further, in piriform and neocortex, FS Sst+ INs show clear biophysical and circuit connectivity differences from other Sst+ IN subtypes (Large et al., 2016; Ma et al., 2006; Naka et al., 2019; Nigro et al., 2018). Recordings and manipulations of neocortical Sst+ IN subtypes in awake, behaving animals show that FS Sst+ and nFS Sst+ INs have disparate functioning in the neocortical circuit during sensory processing (Muñoz et al., 2017; Naka et al., 2019). Thus, our data are consistent with recent findings from cerebral cortex and demonstrate that FS Sst+ INs are biophysically and functionally distinct class of Sst+ INs. Further, we show that BLA FS Sst+ INs mediate feedforward inhibition in the LEC→BLA circuit.

In cerebral cortex and hippocampus, a major circuit role of perisomatic PV+ INs is to regulate AP output of PNs via FFI (Glickfeld and Scanziani, 2006; Pouille and Scanziani, 2001; Tremblay et al., 2016). Similarly in BLA, PV+ INs provide strong regulation of local PN activity (Andrási et al., 2017; Veres et al., 2017; Woodruff and Sah, 2007b). Nevertheless, anatomical data indicates that BLA PV+ INs receive more synaptic input from local PNs as compared to afferent sources (Smith et al., 2000). Our data demonstrating that PV+ INs receive larger uEPSCs from local PNs compared to Sst+ INs (Figure 4) and other recent work showing that PV+ INs fire APs more readily in response to local PN activation compared to cholecystokinin expressing INs (Andrási et al., 2017) provide a physiological compliment to the anatomy findings. Thus, PV+ INs are positioned to provide feedback and lateral inhibition to control BLA spiking output whereas FS Sst+ INs provide feedforward inhibition to regulate PN responses to sensory afferents to BLA. Together, these inhibitory networks would work in concert to fine-tune patterns of BLA activity to allow animals to learn about and react to stimuli. Our results illuminate an important point: archetypal circuit motifs that reappear across brain regions may be realized by different types of neurons in these brain areas depending on the functional role of the local circuit motif.

Although our data indicate that BLA PV+ INs do not fire in response to LEC stimulation, afferent activation does lead to postsynaptic responses in these INs (Figures 2 and 4). When considered with our finding that local PNs preferentially target this IN subtype (Figure 4), this finding suggests an intriguing possibility for PV+ IN functioning. Specifically, if afferents drive specific subsets of BLA PNs to fire, this afferent driven subthreshold depolarization of PV+ INs would prime them to be more likely to fire in response to the subsequent PN activity. As BLA PV+ IN firing is sufficient to prevent AP firing in postsynaptic PNs (Andrási et al., 2017; Veres et al., 2017; Woodruff and Sah, 2007b), this motif could provide an important source of lateral inhibition to select for activity in specific subsets of BLA PNs in response to sensory stimuli. Indeed, in vivo electrophysiological and Ca2+ imaging studies in behaving animals indicate that sensory stimuli drive activity in specific subpopulations of BLA PNs rather than widespread activity across BLA PNs (Beyeler et al., 2016; Grewe et al., 2017; Schoenbaum et al., 1999).

Do the findings of the LEC→BLA circuit extend to other afferents to BLA (e.g. thalamic and other cortical inputs)? A recent study (Lucas et al., 2016) shows that PV+ INs mediate polysynaptic inhibition in the lateral but not basal subnucleus of the BLA following stimulation of cortical (external capsule) or thalamic (internal capsule) fiber tracts to BLA. As stimulation of these fiber tracts leads to activation of local PNs and feedback inhibition (Szinyei et al., 2000), their results could be indicative of PV+ INs mediating feedback inhibition in the lateral subnucleus of the BLA, as has been suggested previously (Duvarci and Pare, 2014). The horizontal slice preparation does not allow for unambiguous discrimination between the lateral and basal subnuclei of the BLA; therefore, we cannot determine the subnucleus within BLA where we conducted our recordings. Since the majority of LEC afferents target the basal subnucleus of the BLA (McDonald, 1998), it is conceivable that we recorded predominantly from neurons in this BLA subnucleus. If this were the case, this would provide an alternative explanation for the apparent discrepancy between our data and the data reported by Lucas and colleagues (Lucas et al., 2016); namely, that PV+ INs provide FFI in the lateral subnucleus and Sst+ INs provide FFI in the basal subnucleus. However, BLA PV+ INs of both the lateral and basal subnuclei of the BLA receive relatively little input from a variety of cortical sources (Smith et al., 2000). Instead, cortical inputs target PV-/calbindin+ INs throughout the BLA (Unal et al., 2014). Interestingly, the major population of PV-/calbindin+ INs in BLA are Sst+ INs (McDonald and Mascagni, 2002). Taken together these data suggest that our findings that FS Sst+ INs mediate BLA FFI will likely hold across cortical inputs to the BLA. Nevertheless, future experiments will be necessary to test this premise.

What might be the functional role of BLA FS Sst+ IN mediated FFI? In BLA, FFI plays a critical role in exerting control over the plasticity of PNs (Bazelot et al., 2015; Bissière et al., 2003; Tully et al., 2007). BLA Sst+ INs primarily target the dendrites of PNs (Muller et al., 2007; Wolff et al., 2014). One function of dendritic targeting INs is to regulate the local Ca2+ flux in dendrites (Chiu et al., 2013; Miles et al., 1996; Müllner et al., 2015). Recent work in cerebral cortex and BLA demonstrates that dendritic disinhibition via Sst+ INs inhibition is critical in gating cortical pyramidal neuron plasticity ex vivo and in gating learning in vivo via regulation of learning dependent changes in pyramidal neuron activity (Adler et al., 2019; Williams and Holtmaat, 2019; Wolff et al., 2014). Specifically, sensory stimuli inhibit Sst+ IN firing in BLA (Wolff et al., 2014), and further decreasing Sst+ IN activity facilitates learning whereas increasing Sst+ IN activity impairs learning (Wolff et al., 2014). Taken in the context of these findings, our data suggest an intriguing potential circuit mechanism explaining how this pattern of activity in vivo may lead to learning. Sensory stimulus mediated inhibition of Sst+ INs would lead to a reduction in FFI and consequent dendritic disinhibition which could allow for plastic changes in the BLA PN to determine appropriate behavioral responses to the sensory stimulus in the future. Future experiments should test the hypothesis that Sst+ IN mediated FFI of BLA PNs gates BLA plasticity and learning.

Taken together, the data presented in this study demonstrate the existence of a previously unreported population of FS Sst+ INs in BLA. This population of INs is biophysically and functionally distinct from PV+ and other Sst+ INs. We show that these FS Sst+ INs provide FFI at the corticoamygdalar synapse in contrast to nFS Sst+ and PV+ INs that do not engage in this circuit motif. Our data highlight the importance of probing IN heterogeneity within the context of local circuits as a key aspect for understanding circuit functioning. As FFI gates synaptic plasticity in BLA (Bazelot et al., 2015; Bissière et al., 2003; Tully et al., 2007), our data raise the possibility that the FFI mediated by this novel BLA population of FS Sst+ INs serves as the underlying circuit mechanism for learning in the survival circuits of the amygdala.

Materials and methods

Key resources table
Reagent
type
Reagent or
resource
SourceIdentifierAdditional
information
AntibodyRat monoclonal anti-SstMilliporeCat#MAB354; RRID: AB_2255365(1:100)
AntibodyGuinea pig polyclonal anti-PVSynaptic SystemsCat#195004; RRID:AB_2156476(1:500)
AntibodyDonkey polyclonal anti-rat Alexa Fluor 488Jackson ImmunoResearchCat#712-545-150(1:500)
AntibodyDonkey polyclonal anti-guinea pig Alexa Fluor 488Jackson ImmunoResearchCat#706-545-148(1:500)
AntibodyStreptavidin Conjugated Alexa Fluor 488ThermoFisher ScientificCat#S11223(1:500)
Commercial assay, kitiFX-enhancerInvitrogenCat#I36933
Chemical compound, drugPicric acidElectron Microscopy SciencesCat#I9556
Chemical compound, drugGlutaraldehydeSigmaCat#G-7651
Chemical compound, drugSucroseSigma-AldrichCat#S7903
Chemical compound, drugGlucoseSigma-AldrichCat#G8270
Chemical compound, drugSodium chlorideSigma-AldrichCat#S9888
Chemical compound, drugPotassium chlorideSigma-AldrichCat#P9333
Chemical compound, drugSodium phosphate monobasic dihydrateSigma-AldrichCat#71505
Chemical compound, drugSodium bicarbonateSigma-AldrichCat#S6297
Chemical compound, drugCalcium chlorideSigma-AldrichCat#21115
Chemical compound, drugMagnesium chlorideSigma-AldrichCat#68475
Chemical compound, drugCesium methanesulfonateAcros OrganicsCAS: 2550-61-0
Chemical compound, drugHEPESSigma-AldrichCat#H3375
Chemical compound, drugEGTASigma-AldrichCat#E3889
Chemical compound, drugSodium phosphocreatineSigma-AldrichCat#P7936
Chemical compound, drugQX-314 (Lidocaine N-ethyl bromide)Sigma-AldrichCat#L5783
Chemical compound, drugMagnesium-ATPSigma-AldrichCat#A9187
Chemical compound, drugSodium-GTPSigma-AldrichCat#G8877
Chemical compound, drugPotassium gluconateSigma-AldrichCat#G4500
Chemical compound, drugDNQXTocrisCat#0189; CAS: 2379-57-9
Chemical compound, drugD-APVTocrisCat#0106;
CAS: 79055-68-8
Chemical compound, drugGabazine (SR 95531 hydrobromide)TocrisCat#1262; CAS: 104104-50-9
Chemical compound, drugClozapine-N-oxideEnzo Life SciencesCat#BML-NS105
Chemical compound, drugPhosphate Buffered Saline (PBS) Tablets, 100 mLVWRCat#E404-200TABS
Chemical compound, drugTriton X-100VWRCat#0694–1L
Chemical compound, drugNormal Donkey SerumJackson Immuno-research Labs IncCat#017-000
-121
Chemical compound, drugBovine Serum AlbuminSigma-AldrichCat#A9647-100G
Commercial assay, kitProlong Gold antifade reagentInvitrogenCat#P36934
OtherNail PolishElectron Microscope SciencesCat#72180
Chemical compound, drugBiocytinThermoFisherCat#280220.2–0.5%
OtherRaw and analyzed dataThis paperAvailable on github and by requesthttps://github.com/emguthman/Manuscript-Codes (copy archived at https://github.com/elifesciences-publications/Manuscript-Codes
Strain, strain background (Mus musculus)Mouse: C57BL/6JJackson LabRRID: IMSR_JAX:000664All sexes used
Strain, strain background (Mus musculus)Mouse: Ssttm2.1(cre)Zjh/JJackson LabRRID: IMSR_JAX:013044All sexes used
Strain, strain background (Mus musculus)Mouse: B6;129P2-Pvalbtm1(cre)Arbr/JJackson LabRRID: IMSR_JAX:008069All sexes used
Strain, strain background (Mus musculus)Mouse: B6;129S6-Gt(ROSA)26Sortm9(CAG-tdTomato)Hze/JJackson LabRRID: IMSR_JAX:007905All sexes used
Strain, strain background (Mus musculus)Mouse: B6N.129-Gt(ROSA)26Sortm1(CAG-CHRM4*,-mCitrine)Ute/JJackson LabRRID: IMSR_JAX:026219All sexes used
Software, algorithmpClamp 10.6Molecular DevicesRRID: SCR_011323
Software, algorithmMATLAB_R2018aMathworksRRID: SCR_001622
Software, algorithmRhttp://www.r-project.org/RRID: SCR_001905
Software, algorithmIllustratorAdobeRRID: SCR_010279
Software, algorithmPhotoshopAdobeRRID: SCR_014199
Software, algorithmSlideBook 6.0Intelligent Imaging Innovations (3i)RRID: SCR_014300
Software, algorithmFijiFijiRRID: SCR_002285
Software, algorithmGrid/Collection StitchingFijiRRID: SCR_016568
Software, algorithmNeurolucidaMBF BiosciencesRRID: SCR_001775

Contact for reagent and resource sharing

Request a detailed protocol

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Molly M Huntsman (molly.huntsman@ucdenver.edu).

Experimental model and subject details

Request a detailed protocol

All experiments were conducted in accordance with protocols approved by the Institutional Animal Care and Use Committee at the University of Colorado Anschutz Medical Campus. Slice electrophysiology experiments were conducted on mice aged postnatal days 35–70. Immunohistochemistry experiments were conducted on mice aged postnatal days 60–120. Experiments were conducted regardless of the observed external genitalia of the mice at weaning. To track estrous, vaginal swabs were collected from mice with vaginas that were used in experiments. The following mouse lines were used in the experiments: C57Bl/6J (Jackson Lab #000664), Sst-tdTomato, PV-tdTomato, Sst-hM4Di, PV-hM4Di, and wildtype littermates of the Sst-hM4Di and PV-hM4Di mice. Sst-tdTomato, PV-tdTomato, Sst-hM4Di, and PV-hM4Di mouse lines were generated by crosses of Sst-ires-Cre (Jackson Lab #013044) and PV-ires-Cre (Jackson Lab #008069) lines with either the Rosa-CAG-LSL-tdTomato-WPRE (Ai9; Jackson Lab #007905) or the R26-hM4Di/mCitrine (Jackson Lab #026219) lines. Please see Key Resources Table for more details on strain information.

Method details

Acute slice preparation for electrophysiology

Request a detailed protocol

Animals were first anesthetized with CO2 and decapitated. Brains were quickly dissected and placed in an ice-cold, oxygenated (95% O2-5% CO2) sucrose-based slicing solution (in mM: sucrose, 45; glucose, 25; NaCl, 85; KCl, 2.5; NaH2PO4, 1.25; NaHCO3, 25; CaCl2, 0.5; MgCl2, 7; osmolality, 290–300 mOsm/kg). 300–400 µm horizontal slices were obtained using a vibratome (Leica Biosystems, Buffalo Grove, IL, USA). Slices were incubated in oxygenated (95% O2-5% CO2) artificial cerebral spinal fluid (ACSF; in mM: glucose, 10; NaCl, 124; KCl, 2.5; NaH2PO4, 1.25; NaHCO3, 25; CaCl2, 2; MgCl2, 2; osmolality 290–300 mOsm/kg) at 36 °C for at least 30 min. All reagents were purchased from Sigma-Aldrich (St. Louis, MO, USA).

Electrophysiology

Request a detailed protocol

Slices were placed in a submerged slice chamber and perfused with ACSF heated to 32-37°C. Slices were visualized using a moving stage microscope (Scientifica: Uckfield, UK; Olympus: Tokyo, Japan) equipped with 4× (0.10 NA) and 40× (0.80 NA) objectives, differential interference contrast (DIC) optics, infrared illumination, LED illumination (CoolLED, Andover, UK), a CoolSNAP EZ camera (Photometrics, Tuscon, AZ, USA), and Micro-Manager 1.4 (Open Imaging, San Francisco, CA, USA). Whole cell patch clamp recordings were made using borosilicate glass pipettes (2.5-5.0 MΩ; King Precision Glass, Claremont, CA, USA) filled with intracellular recording solution. For voltage clamp experiments on FFI a cesium methanesulfonate (CsMe) based intracellular solution was used (in mM: CsMe, 120; HEPES, 10; EGTA, 0.5; NaCl, 8; Na-phosphocreatine, 10; QX-314, 1; MgATP, 4; Na2GTP, 0.4; pH to 7.3 with CsOH; osmolality adjusted to approximately 290 mOsm/kg). For all remaining voltage clamp experiments and for all current clamp experiments, a potassium gluconate based intracellular solution was used (in mM: potassium gluconate, 135; HEPES, 10; KCl, 20; EGTA, 0.1; MgATP, 2; Na2GTP, 0.3; pH to 7.3 with KOH; osmolality adjusted to approximately 295 mOsm/kg). A subset of the potassium gluconate recordings were supplemented with 0.2-0.5 % biocytin to allow for post-hoc morphological analysis. Access resistance was monitored throughout the experiments and data were discarded if access resistance exceeded 25 MΩ or varied by more than ± 20%. No junction potential compensation was performed. Data were acquired with a Multiclamp 700B amplifier and were converted to a digital signal with the Digidata 1440 digitizer using pCLAMP 10.6 software (Molecular Devices, Sunnyvale, CA). Data were sampled at 10 kHz and lowpass filtered at 4 kHz. Offline, current data were filtered using a 3rd order Savistky-Golay filter with a ± 0.5 ms window after access resistance was assessed. Mean traces were created by first aligning all events by their point of maximal rise (postsynaptic currents) or by threshold (APs) and then obtaining the mean of all events; mean subthreshold EPSPs were not aligned prior to averaging.

Cell-type identification

Request a detailed protocol
Principal Neurons (PNs)
Request a detailed protocol

PNs were targeted based on their large, pyramidal-like soma. Recordings were terminated if the physiology of the neuron was inconsistent with BLA PNs (e.g. high membrane resistance, narrow AP halfwidth, large and fast spontaneous EPSCs).

Interneurons (INs)
Request a detailed protocol

INs were targeted based on fluorescence in the Sst-tdTomato, PV-tdTomato, SST-hM4Di, and PV-hM4Di mouse lines. A 470 nm LED was used to identify mCitrine+ INs in SST-hM4Di and PV-hM4Di mouse lines, and a 535 nm LED was used to identify tdTomato+ INs in the SST-tdTomato and PV-tdTomato mouse lines (CoolLED, Andover, UK).

Pharmacology

Request a detailed protocol

DNQX, D-APV, and gbz were purchased from Tocris Biosciences (Bristol, UK) and CNO was purchased from Enzo Life Sciences (Farmingdale, NY). DNQX stock was made at 40 mM and diluted to a final concentration of 20 μM in ACSF; D-APV stock was made at 50 mM and diluted to a final concentration of 50 μM in ACSF; GBZ stock was made at 25 mM and diluted to a final concentration of 5 μM in ACSF; and, CNO stock was made at 10 mM and diluted to a final concentration of 10 μM in ACSF. All stocks were stored at −20°C and CNO was used within one month of making the stock solution.

Electrophysiology experimental design

Request a detailed protocol
FFI, voltage clamp
Request a detailed protocol

The LEC was stimulated using a bipolar stimulating electrode (FHC, Inc, Bowdoin, ME, USA). Evoked EPSCs (Vhold = −70 mV) and IPSCs (Vhold = 0 mV) were recorded from BLA PNs in response to LEC stimulation. To assess the effects of different drugs on the EPSCs and IPSCs, ACSF containing DNQX, D-APV, gbz, and/or CNO was perfused onto the slice for five minutes prior to and continuously during the experiment. Effects of DNQX/APV and gbz on EPSCs and IPSCs were recorded using in an unpaired design where some PNs were recorded under control conditions and in the presence of DNQX/APV and gbz (given sequentially with time for washout) whereas others were recorded under control conditions in the presence of DNQX/APV or gbz. Effects of CNO on EPSCs and IPSCs were examined with a paired design where all PNs were recorded in both control conditions and in the presence of CNO.

FFI, current clamp
Request a detailed protocol

Membrane voltage of BLA PNs was recorded in response to 5 stimulations of the LEC at 20 Hz. Ihold was adjusted such that Vrest of the PNs was approximately −60 mV. To assess the role of GABAA receptor mediated inhibition on PN AP firing, gbz was perfused onto the slice for five minutes prior to and continuously during the experiment.

Current injections
Request a detailed protocol

Membrane voltage of BLA neurons was recorded in current clamp in response to a series of square hyperpolarizing and depolarizing current injections. Prior to initiation of the series of current injections, Vm of the BLA neurons was adjusted to approximately −60 mV. Each cell was subjected to two series of 600 ms square current injections: −100 pA to +100 pA at 10 pA intervals and −200 pA to +400 pA at 25 pA intervals. The data collected in these experiments were used to determine active and passive membrane properties of the neurons.

Minimal stimulation
Request a detailed protocol

EPSCs (Vhold = −70 mV) were recorded in voltage clamp in BLA PNs, Sst+ INs, and PV+ INs in response to LEC stimulation. Stimulation intensity was adjusted such that LEC stimulation resulted in recorded EPSCs having a success rate of approximately 50% and an all-or-none amplitude response.

Recruitment of BLA INs by LEC afferents
Request a detailed protocol

The median stimulation intensity necessary to observe putative uEPSCs in BLA PNs (273 μA × ms) was used as the empirically derived PN threshold stimulation intensity. Membrane voltage responses of BLA Sst+ and PV+ INs were recorded in current clamp in response to 5 stimulations of the LEC at 20 Hz at the empirically derived PN threshold stimulation intensity.

Paired PN-IN recordings
Request a detailed protocol

Paired recordings were made between BLA PNs and nearby Sst+ or PV+ INs. We used a 2.5 nA, 2 ms current injection to drive a single AP in the PN (Ihold adjusted such that Vm≈ -60 mV. BLA PN APs were repeated at 0.25 Hz and the response of the IN (Vhold = -70 mV) was recorded.

Spontaneous EPSC recordings
Request a detailed protocol

sEPSCs were recorded for 5 min in INs (Vhold = −70 mV) with no drugs in the bath.

Pharmacological effects of CNO on membrane potential
Request a detailed protocol

Membrane voltage of mCitrine+ neurons in BLA was recorded in the presence of 20 μM DNQX, 50 μM D-APV, and 5 μM GBZ. To assess the effects of CNO on membrane potential, baseline Vm was allowed to stabilize and was recorded for 3 min in the absence of CNO. Following recording of baseline Vm, ACSF containing 10 μM CNO was perfused onto the slice and Vm was recorded for an additional 10 min. Vm was separated into 30 s bins. ΔVm was defined as the difference in Vm between the mean Vm during the 3 min of baseline recordings and the mean Vm during the last 3 min of CNO application.

Definitions of electrophysiological parameters

Request a detailed protocol

Evoked EPSC/IPSC detection and amplitude

Request a detailed protocol

EPSCs (Vhold = -70 mV) were defined as negatively deflecting postsynaptic events that exceeded the mean baseline current (500 ms before stimulation) by 6 × the median absolute deviation of the baseline current and that occurred within 20 ms of the end of the electrical stimulus artifact. IPSCs (Vhold = 0 mV) were defined as positively deflecting postsynaptic events that exceeded the mean baseline current by 6 × the median absolute deviation of the baseline current and that occurred within 20 ms of the end of the electrical stimulus artifact. To ensure that the detected EPSCs/IPSCs were related to the stimulus, we subsampled 25% of the sweeps in the experiment (or 5 sweeps if the experiment consisted of <20 sweeps) and found the maximal peak negative (EPSC detection) or positive (IPSC detection) deflection from baseline in the 20 ms after the stimulus artifact in those sweeps. Then, we found the median peak time for those sweeps and repeated the analysis over all sweeps in the experiment with a detection threshold of 6 × the median absolute deviation of the baseline current and with a window set to ± 5 ms (EPSC detection) or ± 7.5 ms (IPSC detection) around the median peak time. The amplitude of each EPSC and IPSC was defined as the difference between the peak amplitude of the detected EPSC or IPSC and the mean baseline current for that sweep. The EPSC or IPSC amplitude for each cell in an experiment was defined as the mean of the amplitudes recorded from that cell (EPSC/IPSC successes only; failures were not included in EPSC/IPSC amplitude calculation). Where EPSC failure amplitude is reported on a per sweep basis, it was defined as the maximal negative deflection from the current trace within ± 5 ms of the mean current peak time for that cell. If all sweeps in an experiment were EPSC/IPSC failures, the mean EPSC/IPSC amplitude was defined as the maximal negative (EPSC) or positive (IPSC) deflection in the mean current trace that occurred within 20 ms of the end of the electrical stimulus artifact.

Success rate

Request a detailed protocol

Success rate was defined as 100% × nsuccessfulsweepsntotalsweeps where successful sweeps were defined as sweeps where an EPSC was detected.

I/E balance

Request a detailed protocol

I/E balance was defined as the ratio of IPSC to EPSC amplitude recorded in the same neuron.

EPSC/IPSC 20–80% risetime

Request a detailed protocol

20–80% risetime was defined as the time it took an EPSC or IPSC to reach 80% of its peak amplitude from 20% of its peak amplitude. 20–80% risetime was calculated for each sweep unless obscured by a spontaneous event. The risetime for each cell was defined as the mean of all risetimes recorded from that cell.

EPSC/IPSC latency and jitter

Request a detailed protocol

Latency of EPSCs and IPSCs was defined as the time between the end of the electrical stimulation or the peak of the PN action potential (in paired recordings) and the point of 20% rise for an EPSC or IPSC as calculated for the 20–80% risetime. The EPSC or IPSC latency for each cell was defined as the mean of the latencies recorded from that cell. The EPSC or IPSC jitter for each cell was defined as the standard deviation of the latencies recorded from that cell.

EPSC/IPSC τDecay

Request a detailed protocol

EPSC τDecay was determined using a single exponential fit, f(t)= Aet/τ. IPSC τDecay was defined as the weighted time-constant of IPSC decay. Briefly, a double exponential fit, f(t)= A1et/τ1+A2et/τ2, was used to obtain the parameters to determine the weighted time-constant where τweighted= (τ1A1+τ2A2)/(A1+A2). τDecay was calculated using the mean EPSC or IPSC trace for a cell.

uEPSC detection and amplitude

Request a detailed protocol

uEPSCs were defined as negative current deflections recorded in the IN that exceeded a detection threshold of 6x the median absolute deviation of the baseline current and occurred within 3ms of the PN AP peak during paired recordings. The amplitude of each uEPSC was defined as the difference between the peak amplitude of the detected uEPSC and the mean baseline current for that sweep. The uEPSC amplitude for each cell in an experiment was defined as the mean of the amplitudes recorded from that cell (successes only).

sEPSC detection and amplitude

Request a detailed protocol

sEPSCs were detected by a combined template and threshold method. Briefly, a template was made by subsampling 10% of local negative peaks exceeding at least 5 × the median absolute deviation of a rolling baseline current (50ms prior to the peak). The template current was then truncated from its 20% rise point through the end of the decay time constant for the template current. Next, all local negative peaks exceeding 5 × the median absolute deviation of a rolling baseline current (50ms prior to the peak) were collected. The template current was then scaled to each individual putative sEPSC peak and each sEPSC peak was assigned a normalized charge integral relative to the template. Finally, a normalized charge integral cutoff was chosen to exclude obvious noise/non-physiological events below a certain normalized charge integral. sEPSC amplitude was defined as the difference between the peak amplitude of each detected current and its corresponding baseline current. sEPSC for each cell was defined as the median peak amplitude for that cell.

sEPSC frequency

Request a detailed protocol

sEPSC frequency for each sESPC was defined as the inverse of the interevent intervals of the sEPSCs. The frequency measure for each neuron was defined as the median of the sEPSC frequencies for that cell.

Membrane resistance

Request a detailed protocol

Membrane resistance was defined as the slope of the best fit line of the I-V plot using the −100 pA to +100 pA (10 pA steps) series of current injections. Mean voltage response to each current injection step was defined as the difference between baseline mean membrane voltage (100 ms prior to current injection) and the mean membrane voltage during the 100 ms period from 50 ms after the start of the injection to 150 ms after the start of the current injection. This 100 ms window was chosen to allow for measurement of the change in Vm after the membrane had charged and prior to any potential HCN channel activation. The I-V plot was constructed using all current steps below rheobase.

Maximum firing rate

Request a detailed protocol

Maximum firing rate was defined as the inverse of the inter-spike interval (ISI) during the first 200 ms of the most depolarizing current injection step before attenuation of AP firing was observed. Max FR was calculated using the −200 pA to +400 pA (25 pA steps) series of current injections.

AP threshold

Request a detailed protocol

AP threshold was defined as the voltage at which dVdt exceeded 20 V/s. AP threshold was calculated at the rheobase sweep of the -200 pA to +400 pA (25 pA steps) series of current injections.

AP amplitude

Request a detailed protocol

Amplitude of the AP was defined as the voltage difference between the peak of the AP and its threshold potential. AP amplitude was calculated at the rheobase sweep of the −200 pA to +400 pA (25 pA steps) series of current injections.

AP halfwidth

Request a detailed protocol

AP halfwidth was defined as the time between the half-amplitude point on the upslope of the AP waveform to the half-amplitude point on the downslope of the AP waveform. AP halfwidth was calculated at the rheobase sweep of the −200 pA to +400 pA (25 pA steps) series of current injections.

After-hyperpolarization potential (AHP) magnitude

Request a detailed protocol

AHP magnitude was defined as the difference between the most hyperpolarized membrane voltage of the AHP (occurring within 100 ms after AP threshold) and AP threshold. AHP magnitude and latency data were calculated at the rheobase sweep of the −200 pA to +400 pA (25 pA steps) series of current injections. ΔAHP data were calculated at the rheobase + 50 pA sweep of the −200 pA to +400 pA (25 pA steps) series of current injections.

AHP latency

Request a detailed protocol

AHP latency was defined as the time from AP threshold and the peak of the AHP.

ΔAHP

Request a detailed protocol

ΔAHP was defined as the difference between the first and last AHP (ΔAHP=AHPlastAHPfirst).

AP phase plot

Request a detailed protocol

The AP phase plot was obtained by plotting the rate of change of the mean AP for each cell from the rheobase sweep of the −200 pA to +400 pA (25 pA steps) series of current injections as a function of the corresponding membrane voltage.

Latency to first AP

Request a detailed protocol

AP latency was defined as the time from the initiation of the current injection to the peak of the first AP. AP latency was calculated at the rheobase sweep of the −200 pA to +400 pA (25 pA steps) series of current injections.

Firing rate adaptation ratio (FR adaptation)

Request a detailed protocol

Firing rate adaptation was defined as the ratio of the first and the average of the last two ISIs, such that Firingrateadaptation=ISIfirstmeanISIlast 2ISI. Firing rate adaptation was calculated at the rheobase +50 pA sweep of the -200 pA to +400 pA (25 pA steps) series of current injections.

AP broadening

Request a detailed protocol

AP broadening was defined as the ratio of the AP halfwidths of the first two APs (Broadening=halfwidthsecondhalfwidthfirst). AP broadening was calculated at the rheobase +50 pA sweep of the -200 pA to +400 pA (25 pA steps) series of current injections.

AP amplitude adaptation

Request a detailed protocol

AP amplitude adaptation was defined as the ratio of the AP amplitude of the average of the last three APs and the first AP, such that APamplitudeadaptation=meanAmplitudelast 3APsAmplitudefirst. AP amplitude adaptation was calculated at the rheobase +50 pA sweep of the -200 pA to +400 pA (25 pA steps) series of current injections.

Membrane decay τ

Request a detailed protocol

Membrane decay τ was determined by using a single exponential fit, f(t)= Aet/τ, to fit the change in Vm induced by a −100 pA sweep in the −100 pA to +100 pA (25 pA steps) series of current injections.

Hyperpolarization-induced sag

Request a detailed protocol

Hyperpolarization-induced sag was calculated using the equation, Vmin-VssVmin-Vbl×100%, where Vmin was defined as the most hyperpolarized membrane voltage during the current injection, Vss was defined as the mean steady-state membrane voltage (last 200 ms of the current injection), and Vbl was defined as the mean baseline membrane voltage (100 ms prior to current injection). Hyperpolarization-induced sag was measured from the -200 pA current injection.

Rebound spikes

Request a detailed protocol

Rebound spikes were defined as the number of APs in the 500 ms following the −200 pA current injection.

APs per stimulus

Request a detailed protocol

The number of APs per stimulus was defined as the number of APs occurring within 50 ms of the stimulus.

Vrest

Request a detailed protocol

Vrest was defined as Vm (Ihold = 0 pA) during a 500 ms baseline prior to LEC stimulation during the experiments testing the recruitment of BLA INs by LEC afferents.

Subthreshold EPSP amplitude

Request a detailed protocol

Subthreshold EPSP amplitude was defined as the maximal, non-stimulus artifact, voltage deflection within 40 ms after LEC stimulation.

Immunohistochemistry

Request a detailed protocol

Biocytin filled neurons

Request a detailed protocol

To perform immunostaining of biocytin filled neurons, slices containing biocytin filled neurons were fixed in 4% PFA overnight at 4°C. After fixation, slices were transferred to PBS. Biocytin filled INs (n = 8 Sst+ Group 1; 3 Sst+ Group 2; 6 PV+) were blocked (1X PBS, 0.3% triton, 5% BSA, 5% Normal Donkey Serum) for 4 hours before 24-hour incubation with streptavidin conjugated Alexa Fluor 488 (1:500, ThermoFisher Scientific, Waltham, MA, USA) at 4°C. Slices were mounted with Prolong Gold and sealed for long-term storage. Slices were imaged using an Axio Observer microscope (Carl Zeiss, Okerkochen, Germany); equipped with a CSU-X1 spinning disc unit (Yokogawa, Musashino, Tokyo, Japan); 488 nm/40 mW laser; Plan-NeoFluar 40X (0.75 NA) air objective lens; and Evolve 512 EM-CCD camera (Photometrics, Tucson, AZ, USA). SlideBook 6.0 software (3i, Denver, CO, USA) enabled instrument control and data acquisition. Images were acquired in sections by following branched points from the cell soma. Images were stitched in Fiji software using Grid/Collection stitching (Preibisch et al., 2009) with an unknown position type.

Histological validation of PV+ and Sst+ IN identity

Request a detailed protocol

To perform the PV and Sst immunostaining, mice (n = 3 Sst-tdTomato mice, 3 PV-tdTomato mice) were sacrificed and transcardially perfused with ice cold 4% PFA (with 1.5% picric acid and 0.05% glutaraldehyde) followed by 30% sucrose protection. After the brain sank, coronal BLA slices of 30 μm thickness were obtained. Before application of blocking solution, slices were incubated for 30 minutes at room temperature with iFX-enhancer (Invitrogen, ThermoFisher Scientific, Waltham, MA, USA). After blocking, the slices were incubated with either rat anti-Sst antibody (1:100, MAB354, Millipore, Burlington, MA, USA) or guinea pig anti-PV antibody (1:500, 195004, Synaptic Systems, Göttingen, Germany) for at least 48 hours at 4°C. Then, a secondary antibody of either donkey anti-rat or donkey anti-guinea pig Alexa Fluor 488 (1:500, Jackson ImmunoResearch) was applied overnight at 4°C. The slices were imaged using a confocal laser scanning microscope (TCS SP5II, Leica Application Suite, Leica Biosystems, Buffalo Grove, IL, USA) with 10 × 0.40 NA and 20 × 0.70 NA dry objectives to determine the neuron identity.

Morphological analysis

Request a detailed protocol

Analysis of somatic and dendritic morphology

Request a detailed protocol

To determine morphological characteristics of biocytin filled neurons, stitched images were imported to Neurolucida (MBF Bioscience, Williston, VT, USA) to perform tracing. All analysis including sholl analysis (50 μM rings), dendrite branching, etc. was performed from traces using Neurolucida Explorer (MBF Bioscience).

Quantification and statistical analysis

Request a detailed protocol

Statistical analyses

Request a detailed protocol

All data analysis (except decision tree and random forest analyses) were performed offline using custom written MATLAB code. Normality of the data was assessed using the Anderson-Darling test. For assessment of whether a single group differed from a normal distribution centered around zero, a one-sample t-test was used. For a test between two groups, a paired or unpaired t-test was used where appropriate. For tests between two groups of non-normal data, a Mann-Whitney U or Wilcoxon signed-rank test was used where appropriate. For tests between three or more groups of normal data with one independent variable, a one-way ANOVA was used with Tukey’s post-hoc test to examine differences between groups. A Kruskal-Wallis test was used to examine differences between three or more groups of non-normal data with one independent variable. A Mann-Whitney U test was used as a post-hoc test following a significant result in a Kruskal-Wallis test and was corrected for multiple comparisons using the FDR method (Curran-Everett, 2000). The critical significance value was set to α = 0.05 or was set to a FDR-corrected value (αFDR) for multiple comparisons. All statistical tests were two-tailed. Unless otherwise stated, experimental numbers are reported as n = x, y where x is the number of neurons and y is the number of mice. Statistical parameters are reported in the Results section and figure legends display p values and sample sizes.

Unsupervised cluster analysis

Request a detailed protocol

Unsupervised cluster analysis using Ward’s method (Ward, 1963) was used to classify Sst+ INs. Briefly, this method involves plotting each neuron in multidimensional space where each dimension corresponds to a given parameter. For our data, we plotted each Sst+ IN in 15-dimensional space (where each dimension corresponds to a z-score transformation of one of the 15 membrane properties obtained from the current injection experiments; values were z-score transformed so that parameters with large values, for example membrane resistance, would not influence the cluster analysis more than those with small values, for example halfwidth). From here, the analysis proceeded along n – one stages where n is the number of Sst+ INs. At stage n = 1, the two closest cells in the 15-dimensional space are grouped together. At subsequent stages, the closest cells are grouped together until only one group of all objects remains. We determined the final number of clusters by using the Thorndike procedure (Thorndike, 1953) where large distances between group centroids at one cluster stage relative to other stages are indicative of significant differences between groups (see Figure 3A, inset).

Decision tree analysis

Request a detailed protocol

Recursive partitioning analyses (Breiman et al., 1984) were conducted in R (1.01.136). The R package rpart (Therneau and Atkinson, 1984) was used for recursive partitioning for classification. Decision trees were plotted using rpart.plot package, and random seeds were set using the rattle package. Predicted classes were determined from the unsupervised cluster analysis and were used to determine important discriminating parameters. The model was internally cross validated with a nested set of subtrees and the final model was selected as the subtree with the least misclassification error. Individual decision trees can suffer from overfitting. Therefore, the input parameters of the model were independently cross validated by bootstrapping 500 subsamples (without replacement and using random seeds). Each tree was pruned by choosing the complexity parameter with the lowest cross validated error. The mean correct prediction of the test set classifications (n = 20% of sample) by the pruned tree generated by modeling the training sets (n = 80% of sample) was 82.9%.

Random forest analysis

Request a detailed protocol

Supervised classification random forest was employed to reduce potential model overfitting. Random forest classification was conducted using the randomForest package in R (Liaw and Wiener, 2002). Random forest classification employs random subsampling of both parameters and bootstrapped subsamples of the dataset (with replacement). 10,000 decision trees were generated, and the modal tree was used as the final classification model. The out of bag estimate of the error rate of the model was 11.43%. An additional cross validation step was conducted by bootstrapping 80% of the initial sample to create the random forest model and tested against a subsample (n = 20% of sample; bootstrapped without replacement). The mean accuracy after 20 random forest runs was 90.7%. Gini impurity was calculated to determine parameter importance. Visualization of the frequency of individual samples falling within the same node across one run of the classification algorithm was obtained by calculating a proximity matrix for the samples and plotting it against the first two principle components.

Data display

Request a detailed protocol

Data visualizations were created in MATLAB and Adobe Illustrator. After analysis was completed, Neurolucida traces of soma and dendrites were thickened by 7 pixels (Figure 3) or 3-7 pixels (Figure 3—figure supplement 2) in Adobe Photoshop to improve visibility in figures. Normal data are presented as the mean ± s.e.m. Non-normal data are presented as the median with error bars extending along the interquartile range.

References

  1. 1
  2. 2
  3. 3
  4. 4
  5. 5
  6. 6
  7. 7
  8. 8
    Classification and Regression Trees
    1. L Breiman
    2. J Friedman
    3. CJ Stone
    4. RA Olshen
    (1984)
    Wadsworth: Belmont.
  9. 9
  10. 10
    Multiple comparisons: philosophies and illustrations
    1. D Curran-Everett
    (2000)
    American Journal of Physiology-Regulatory, Integrative and Comparative Physiology 279:R1–R8.
    https://doi.org/10.1152/ajpregu.2000.279.1.R1
  11. 11
  12. 12
  13. 13
  14. 14
  15. 15
  16. 16
  17. 17
  18. 18
  19. 19
  20. 20
  21. 21
  22. 22
  23. 23
  24. 24
  25. 25
  26. 26
  27. 27
  28. 28
  29. 29
    Classification and regression by randomForest
    1. A Liaw
    2. M Wiener
    (2002)
    R News 2:18–22.
  30. 30
  31. 31
  32. 32
  33. 33
  34. 34
  35. 35
  36. 36
  37. 37
  38. 38
  39. 39
  40. 40
  41. 41
  42. 42
  43. 43
  44. 44
  45. 45
  46. 46
  47. 47
  48. 48
  49. 49
  50. 50
  51. 51
  52. 52
  53. 53
  54. 54
  55. 55
  56. 56
    Rpart: Recursive Partitioning and Regression Trees
    1. T Therneau
    2. B Atkinson
    (1984)
    R Package Version.
  57. 57
  58. 58
  59. 59
  60. 60
  61. 61
  62. 62
  63. 63
  64. 64
  65. 65
  66. 66
  67. 67
  68. 68
  69. 69

Decision letter

  1. John Huguenard
    Reviewing Editor; Stanford University School of Medicine, United States
  2. Laura L Colgin
    Senior Editor; University of Texas at Austin, United States
  3. Jamie Maguire
    Reviewer; Tufts University School of Medicine, United States

In the interests of transparency, eLife publishes the most substantive revision requests and the accompanying author responses.

Acceptance summary:

This well written manuscript simplifies a very complex topic and clearly presents complicated electrophysiological studies. The study is conceptually sound and carefully conducted. This manuscript sheds light on an important, yet poorly understood, circuit impinging on the basolateral amygdala (BLA). This is a complex circuit and elucidating the functional dynamics of the local and extended circuitry of the BLA is challenging. This elegant electrophysiological study is a tour de force attempting to disentangle the function of the lateral entorhinal cortex (LEC) to BLA circuit that is implicated in olfactory learning. The data describe a novel fast-spiking SST+ interneuron in the BLA which is specifically recruited by LEC input and a suggests a unique mechanism that might gate plasticity in the BLA.

Decision letter after peer review:

Thank you for submitting your article "Cell type specific control of basolateral amygdala plasticity via entorhinal cortex-1 driven feedforward inhibition" for consideration by eLife. Your article has been reviewed by three peer reviewers, and the evaluation has been overseen by a Reviewing Editor and Laura Colgin as the Senior Editor. The following individual involved in review of your submission has agreed to reveal their identity: Jamie Maguire (Reviewer #2).

The reviewers have discussed the reviews with one another and the Reviewing Editor has drafted this decision to help you prepare a revised submission.

Essential revisions:

Overall, while the comprehensive approach is impressive and carefully teases out the circuits recruited by LEC input, the experiments on plasticity are underpowered and seem to rely on novel, i.e. non-standard, definitions of plasticity that might not be fully justified. The paper would be much stronger with additional experiments that would adequately power the plasticity results. However, given the extensive and important findings of the paper, the reviewers felt that even without the results on plasticity the paper would stand on its own. Accordingly, elimination of these results along with corresponding changes in title, Abstract, etc, would also be acceptable.

Detailed critique:

1) There are far too few cells in the analysis or the role of specific interneurons subtypes in plasticity induction. The investigators divide cells into "no plasticity" and LTP or LTD neurons, but there is no clear rationale for how those groups are created or why results are so heterogeneous. With such varied outcomes (and given the ambiguity in where exactly the recordings were located), it is important to increase the size of the dataset to determine whether there is a specific trend with inhibition blocked. It was unclear if the CNO application experiments were done blind – at least a portion of these should be carried out without investigator knowledge of either CNO presence or animal genotype. Also, Figure 1H-I would be better off in a separate figure or put together with Figure 6.

2) In neocortical areas, some SST neurons also express PV (see Hu, Cavendish, Agmon 2013 and others). Is this the case in the BLA? Are the authors suggesting that the FS-SST neurons are a novel subset of SST, or are PV cells that show off-target transgene expression? There are a number of studies that have investigated FS SST neurons; these studies should better cited. Can the authors stain the tissue with PV to see whether some SST neurons also show PV expression?

3) It is of concern that these data are incongruent with a prior study (Lucas et al., 2016). The authors suggest that they cannot accurately detect the boundaries between the BLA and LA in their horizontal prep. However, they indicate that their study was confined to the BLA throughout the paper. How was it verified? To better understand where they were recording from in their novel slice prep, they should target some injections to the BLA and then prepare their slice prep, to figure out their recording location. In addition, they should show a photo of the prep in their first figure, and indicate some anatomical landmarks that can be used by others in the field to take advantage of this slice preparation and replicate the results.

4) The Abstract talks about olfactory learning, but there is nothing in the paper that is related to this. Since it is rather speculative, it should be taken out.

5) The terminology and methodology around bidirectional plasticity should be improved and rationalized. If the authors are correct, then it is not so much bidirectional as heterogeneous, with some cells experiencing LTD, others LTP and yet others little effect at all. I think a very rigorous statistical approach is required to clearly demonstrate the existence of these distinct subpopulations over the more parsimonious explanation than on average there is little plasticity in this pathway. The experiments appear to be underpowered to really make this point.

https://doi.org/10.7554/eLife.50601.sa1

Author response

Essential revisions:

Overall, while the comprehensive approach is impressive and carefully teases out the circuits recruited by LEC input, the experiments on plasticity are underpowered and seem to rely on novel, i.e. non-standard, definitions of plasticity that might not be fully justified. The paper would be much stronger with additional experiments that would adequately power the plasticity results. However, given the extensive and important findings of the paper, the reviewers felt that even without the results on plasticity the paper would stand on its own. Accordingly, elimination of these results along with corresponding changes in title, Abstract, etc, would also be acceptable.

We would like to thank the reviewers for their critiques and suggested revisions to our manuscript. We agree that additional experiments would better power the plasticity results. We have made the requested revisions and removed these results with corresponding changes to the rest of the paper.

Detailed critique:

1) There are far too few cells in the analysis or the role of specific interneurons subtypes in plasticity induction. The investigators divide cells into "no plasticity" and LTP or LTD neurons, but there is no clear rationale for how those groups are created or why results are so heterogeneous. With such varied outcomes (and given the ambiguity in where exactly the recordings were located), it is important to increase the size of the dataset to determine whether there is a specific trend with inhibition blocked. It was unclear if the CNO application experiments were done blind – at least a portion of these should be carried out without investigator knowledge of either CNO presence or animal genotype. Also, Figure 1H-I would be better off in a separate figure or put together with Figure 6.

We thank the reviewers for this critique. As mentioned above, these data (old Figure 1H-I and old Figure 6) have been removed from the current manuscript.

2) In neocortical areas, some SST neurons also express PV (see Hu, Cavendish, Agmon 2013 and others). Is this the case in the BLA? Are the authors suggesting that the FS-SST neurons are a novel subset of SST, or are PV cells that show off-target transgene expression? There are a number of studies that have investigated FS SST neurons; these studies should better cited. Can the authors stain the tissue with PV to see whether some SST neurons also show PV expression?

Yes, there is some, albeit small (~16%), overlap between PV and Sst expression in the BLA. We stained tissue for PV expression in the Sst-tdTomato and for Sst expression in the PV-tdTomato mouse lines. These data were presented as Figure 2—figure supplement 2 and show 13.28 ± 1.11% overlap between PV and tdTomato expression in the Sst-tdTomato mouse line and 18.34 ± 2.24% overlap between Sst and tdTomato expression in the PV-tdTomato mouse line. These data were discussed the original submission and in the first paragraph of the subsection “Cell type specificity of the LEC→BLA circuit” in the new submission.

We propose that the FS Sst INs are a novel subset of Sst INs rather than a population of PV cells with off-target transgene expression. Our major arguments for this are the following:

1) FS Sst INs accounted for 57% of all Sst INs (Figure 2—source data 2). The PV/Sst overlap of ~16% in our transgenic lines (Figure 2—figure supplement 2) cannot account for the fraction of FS Sst INs.

2) If these FS Sst neurons were actually PV neurons with off-target cre expression in the Sst-cre mouse lines, we would expect to see a decrease in FFI when we inactivate PV INs. However, in Figure 5, we show that inactivation of Sst INs but not PV INs led to a decrease in FFI.

We have updated the Discussion to better cite other studies that have investigated the FS Sst INs. This discussion can be found in the third paragraph of the Discussion.

3) It is of concern that these data are incongruent with a prior study (Lucas et al., 2016). The authors suggest that they cannot accurately detect the boundaries between the BLA and LA in their horizontal prep. However, they indicate that their study was confined to the BLA throughout the paper. How was it verified? To better understand where they were recording from in their novel slice prep, they should target some injections to the BLA and then prepare their slice prep, to figure out their recording location. In addition, they should show a photo of the prep in their first figure, and indicate some anatomical landmarks that can be used by others in the field to take advantage of this slice preparation and replicate the results.

We think there has been a miscommunication on our end. The LA and BA are the two subnuclei that make up the BLA as a whole. This miscommunication likely arises from older literature calling the basal nucleus the basolateral nucleus of the basolateral amygdala. In the literature, the basal/basolateral nucleus of the basolateral amygdala has been abbreviated as BA when the basal amygdala nomenclature is used or BL when it is called the basolateral nucleus. Some authors have further subdivided the BA/BL into anterior and posterior regions, naming them BLA and BLP (not to be confused with the basolateral amygdala as a whole, this BLA is the anterior basolateral nucleus of the basolateral amygdala, a subnucleus of a subnucleus of the totality of the basolateral amygdala). For an example of this nomenclature, please see Author response image 1, an image from Paxinos and Franklin’s the Mouse Brain in Stereotaxic Coordinates, 4th ed. where the BA is called the BL with subsequent subdivision into BLA and BLP.

This nomenclature is confusing and can lead to a misunderstanding of which regions are being recorded from. For that reason, we referred to the two major subnuclei as LA and BA and the whole basolateral amygdala as BLA as this has become the standard nomenclature in the field. For examples of the LA/BA nomenclature (including in use in the horizontal slice) please see the above reference paper by Lucas and colleagues, the below paper by Andrasi and colleagues, and a recent review from Janak and Tye (Andrasi et al., 2017; Lucas et al., 2016; Janak and Tye, 2015).

With regards to our experimental preparation, we cannot accurately detect the boundaries between LA and BA in our horizontal prep without posthoc staining for VAChT, but we can unambiguously detect the boundaries of the BLA as a whole. (The only way to accurately detect LA and BA boundaries in the horizontal slice is with posthoc VAChT staining: Andrasi et al., 2017).

Similar to what is found in the coronal slice, the external capsule branches in two and creates medial and lateral boundaries around the BLA proper. In author response image 1, we have included images of our slice prep that shows the BLA in the horizontal slice surrounded by these branches of the external capsule. The anatomical landmarks we used to define the BLA were taken from Paxinos and Franklin’s the Mouse Brain in Stereotaxic Coordinates, 4th ed. A page from Paxinos showing the horizontal section of the mouse brain containing both the lateral and basal nuclei of the BLA is presented under the representative images of our recording and stimulation preparation to show that our recording sites correspond to the anatomical definitions of the BLA in the horizontal section.

All our recordings conform to the anatomically defined boundaries of the BLA. As should be clear from the slice images, no anatomical landmarks distinguish which subnuclei we recorded from, but we state with confidence that every recording was conducted within the BLA as a whole structure.

Author response image 1

4) The Abstract talks about olfactory learning, but there is nothing in the paper that is related to this. Since it is rather speculative, it should be taken out.

We agree with the reviewers and have taken out our mention of olfactory learning in the Abstract.

5) The terminology and methodology around bidirectional plasticity should be improved and rationalized. If the authors are correct, then it is not so much bidirectional as heterogeneous, with some cells experiencing LTD, others LTP and yet others little effect at all. I think a very rigorous statistical approach is required to clearly demonstrate the existence of these distinct subpopulations over the more parsimonious explanation than on average there is little plasticity in this pathway. The experiments appear to be underpowered to really make this point.

We agree with the reviewers that the plasticity experiments could be improved. As mentioned above, at the reviewers’ suggestion, we have removed these experiments from the manuscript.

https://doi.org/10.7554/eLife.50601.sa2

Article and author information

Author details

  1. E Mae Guthman

    1. Neuroscience Graduate Program, University of Colorado Anschutz Medical Campus, Aurora, United States
    2. Department of Pharmaceutical Sciences, University of Colorado Anschutz Medical Campus, Aurora, United States
    Contribution
    Conceptualization, Data curation, Formal analysis, Funding acquisition, Validation, Investigation, Visualization, Methodology
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-2190-7520
  2. Joshua D Garcia

    Department of Pharmacology, University of Colorado Anschutz Medical Campus, Aurora, United States
    Contribution
    Data curation, Methodology
    Competing interests
    No competing interests declared
  3. Ming Ma

    Department of Cell and Developmental Biology, University of Colorado Anschutz Medical Campus, Aurora, United States
    Contribution
    Data curation, Methodology
    Competing interests
    No competing interests declared
  4. Philip Chu

    1. Department of Pharmaceutical Sciences, University of Colorado Anschutz Medical Campus, Aurora, United States
    2. Department of Neurosurgery, University of Colorado Anschutz Medical Campus, Aurora, United States
    Contribution
    Data curation, Software, Formal analysis, Methodology
    Competing interests
    No competing interests declared
  5. Serapio M Baca

    1. Department of Pharmaceutical Sciences, University of Colorado Anschutz Medical Campus, Aurora, United States
    2. Department of Neurology, University of Colorado Anschutz Medical Campus, Aurora, United States
    Contribution
    Software
    Competing interests
    No competing interests declared
  6. Katharine R Smith

    Department of Pharmacology, University of Colorado Anschutz Medical Campus, Aurora, United States
    Contribution
    Data curation, Formal analysis, Supervision
    Competing interests
    No competing interests declared
  7. Diego Restrepo

    Neuroscience Graduate Program, University of Colorado Anschutz Medical Campus, Aurora, United States
    Contribution
    Conceptualization, Supervision, Investigation
    Contributed equally with
    Molly M Huntsman
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-4972-446X
  8. Molly M Huntsman

    1. Neuroscience Graduate Program, University of Colorado Anschutz Medical Campus, Aurora, United States
    2. Department of Pharmaceutical Sciences, University of Colorado Anschutz Medical Campus, Aurora, United States
    3. Department of Pediatrics, University of Colorado Anschutz Medical Campus, Aurora, United States
    Contribution
    Conceptualization, Resources, Supervision, Funding acquisition, Investigation, Methodology
    Contributed equally with
    Diego Restrepo
    For correspondence
    molly.huntsman@CUAnschutz.edu
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-5954-0023

Funding

National Institutes of Health (R01 NS095311)

  • Molly M Huntsman

National Science Foundation (DGE-1553798)

  • E Mae Guthman

National Institutes of Health (R01 DC000566)

  • Diego Restrepo

National Institutes of Health (T32 NS099042)

  • E Mae Guthman

National Institutes of Health (T32 GM763540)

  • Joshua D Garcia

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Acknowledgements

We would like to thank Nicole Arevalo and the University of Colorado Anschutz Medical Campus Breeding Core for their assistance with animal care. We would also like to thank members of the Huntsman and Restrepo labs and Alicia M Purkey for discussions on the data. This work was supported by US National Institutes of Health grants (R01 DC000566 to DR, R01 NS095311 to MMH, T32 NS099042 to EMG, and T32 GM763540 to JDG) and US National Science Foundation Graduate Research Fellowship (DGE-1553798 to EMG).

Ethics

Animal experimentation: This study was performed in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. All of the animals were handled according to approved institutional animal care and use committee (IACUC) protocols (#00039) of the University of Colorado Denver | Anschutz Medical Campus.

Senior Editor

  1. Laura L Colgin, University of Texas at Austin, United States

Reviewing Editor

  1. John Huguenard, Stanford University School of Medicine, United States

Reviewer

  1. Jamie Maguire, Tufts University School of Medicine, United States

Publication history

  1. Received: July 27, 2019
  2. Accepted: January 8, 2020
  3. Accepted Manuscript published: January 9, 2020 (version 1)
  4. Version of Record published: January 27, 2020 (version 2)

Copyright

© 2020, Guthman et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.

Metrics

  • 1,827
    Page views
  • 325
    Downloads
  • 0
    Citations

Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Download citations (links to download the citations from this article in formats compatible with various reference manager tools)

Open citations (links to open the citations from this article in various online reference manager services)