Structural and functional characterization of dendritic arbors and ...

J Neurophysiol 114: 942–957, 2015. First published June 3, 2015; doi:10.1152/jn.00824.2014.

Structural and functional characterization of dendritic arbors and GABAergic synaptic inputs on interneurons and principal cells in the rat basolateral amygdala Paul M. Klenowski,1* Matthew J. Fogarty,2* Arnauld Belmer,1 Peter G. Noakes,2,3 Mark C. Bellingham,2 and Selena E. Bartlett1 1

Translational Research Institute and Institute for Health and Biomedical Innovation, Queensland University of Technology, Brisbane, Queensland, Australia; 2School of Biomedical Sciences, The University of Queensland, Brisbane, Queensland, Australia; and 3Queensland Brain Institute, The University of Queensland, Brisbane, Queensland, Australia Submitted 17 October 2014; accepted in final form 2 June 2015

Klenowski PM, Fogarty MJ, Belmer A, Noakes PG, Bellingham MC, Bartlett SE. Structural and functional characterization of dendritic arbors and GABAergic synaptic inputs on interneurons and principal cells in the rat basolateral amygdala. J Neurophysiol 114: 942–957, 2015. First published June 3, 2015; doi:10.1152/jn.00824.2014.—The basolateral amygdala (BLA) is a complex brain region associated with processing emotional states, such as fear, anxiety, and stress. Some aspects of these emotional states are driven by the network activity of synaptic connections, derived from both local circuitry and projections to the BLA from other regions. Although the synaptic physiology and general morphological characteristics are known for many individual cell types within the BLA, the combination of morphological, electrophysiological, and distribution of neurochemical GABAergic synapses in a three-dimensional neuronal arbor has not been reported for single neurons from this region. The aim of this study was to assess differences in morphological characteristics of BLA principal cells and interneurons, quantify the distribution of GABAergic neurochemical synapses within the entire neuronal arbor of each cell type, and determine whether GABAergic synaptic density correlates with electrophysiological recordings of inhibitory postsynaptic currents. We show that BLA principal neurons form complex dendritic arborizations, with proximal dendrites having fewer spines but higher densities of neurochemical GABAergic synapses compared with distal dendrites. Furthermore, we found that BLA interneurons exhibited reduced dendritic arbor lengths and spine densities but had significantly higher densities of putative GABAergic synapses compared with principal cells, which was correlated with an increased frequency of spontaneous inhibitory postsynaptic currents. The quantification of GABAergic connectivity, in combination with morphological and electrophysiological measurements of the BLA cell types, is the first step toward a greater understanding of how fear and stress lead to changes in morphology, local connectivity, and/or synaptic reorganization of the BLA. basolateral amygdala; dendritic spines; GABA; synaptic currents

a complex neural circuit that encodes emotional salience of the sensory information it receives. The processing of information by neurons within the amygdala guides and instructs the formation and consolidation of memories that determine behavioral responses to emotional stimuli (reviewed in Phelps and LeDoux 2005; Sah et al. 2003).

THE AMYGDALA FORMS

* P. M. Klenowski and M. J. Fogarty contributed equally to this work. Address for reprint requests and other correspondence: S. E. Bartlett, Translational Research Inst., Inst. for Health and Biomedical Innovation, Faculty of Health, Queensland Univ. of Technology, Brisbane, Queensland, Australia (e-mail: [email protected]). 942

Furthermore, strong evidence suggests that the amygdala plays a role in incentive learning and motivational behaviors associated with the rewarding effects of addictive substances (Everitt et al. 1999; Robbins and Everitt 2002). Abnormalities in amygdala function have recently been shown to contribute to the pathophysiology of disease states including posttraumatic stress disorder and drug addiction (Koob et al. 2014; Weston 2014). The basolateral amygdala (BLA) is the main input center of the amygdala and consists of lateral, basolateral, and basomedial subdivisions (Sah et al. 2003). Sensory information is relayed to the BLA via glutamatergic afferent connections from the cortex and the thalamus (McDonald 1998; Pitkänen 2000; Sah et al. 2003). This glutamatergic input enters via the internal and external capsules (Mascagni et al. 1993). Glutamatergic synapses are formed at the distal dendritic processes and spines of BLA principal neurons (Brinley-Reed et al. 1995; Farb and Ledoux 1997, 1999; LeDoux et al. 1991; Mascagni et al. 1993), which then send projections to the central nucleus of the amygdala (Tye et al. 2011), bed nucleus of the stria terminalis (Weller and Smith 1982), nucleus accumbens (Stuber et al. 2011), and medial prefrontal cortex (Senn et al. 2014) to facilitate the coordination of emotional responses. Principal neurons, which constitute ⬃80% of the BLA cell population, are characterized by large piriform/triangle-shaped soma, dendritic spines, and the presence of a thicker apical dendrite (McDonald 1982, 1984, 1992). Studies have also provided subclassification of the BLA principal cells based on characteristics including semipyramidal and stellatelike morphology and distinct electrophysiological properties (Rainnie et al. 1993; Sah et al. 2003). The BLA also contains GABAergic interneurons, which make local connections that inhibit BLA output as well as providing feedback inhibition (Carlsen 1988; Ehrlich et al. 2009; McDonald et al. 2002; Rainnie et al. 1991; Spampanato et al. 2011). Multiple interneuron subtypes within the BLA have been identified on the basis of their morphology and expression of neuropeptides and calcium binding proteins (Kemppainen and Pitkanen 2000; McDonald and Mascagni 2001, 2002). Previous studies have reported that BLA interneurons generally have smaller cell bodies compared with principal cells, lack spines, and have no apparent apical dendrite (McDonald 1982). These cells receive cortical inputs that are regionally similar to principal cells (Smith et al. 2000); however, their efferent connections appear to be distributed

0022-3077/15 Copyright © 2015 the American Physiological Society

www.jn.org

GABAERGIC INPUTS ON BLA PRINCIPAL CELLS AND INTERNEURONS

extensively within the local BLA network (Carlsen 1988; Ehrlich et al. 2009; McDonald et al. 2002; Rainnie et al. 1991). The integration of sensory inputs and subsequent generation of emotional responses is largely determined by the proportion of excitatory and inhibitory input to BLA principal cells (Davis et al. 1994), in addition to local inhibition provided by GABAergic interneurons (Ehrlich et al. 2009; Spampanato et al. 2011). Furthermore, it has been proposed that the pattern of local innervation within the BLA promotes synchronous firing of principal cells that facilitates rapid oscillations and synaptic plasticity (Muller et al. 2006). Despite the availability of general morphology and information regarding neuronal organization within the BLA (Faber et al. 2001; McDonald 1982, 1984; Rainnie et al. 1993; Washburn and Moises 1992), limited quantitative data are available on BLA principal cell and interneuron morphology and the relative amount of inhibitory innervation these cells receive. We therefore decided to quantify the morphology, inhibitory spontaneous synaptic currents, and distribution of GABAergic neurochemical synapses within the entire dendritic arbor of BLA interneurons and principal cells. Specifically, we assessed the morphological properties of individual Neurobiotin (NB)filled BLA neurons from ex vivo rat brain slices and divided the cell population into principal cells, spiny interneurons, and aspiny interneurons on the basis of measurements of soma volume, dendritic arbor length, and spine density. We also analyzed differences in spontaneous inhibitory postsynaptic currents (sIPSCs) and quantified the localization and density of neurochemical GABAergic synapses between the cell types. Our analysis shows that marked differences in spontaneous inhibitory synaptic activity, morphology, and synaptic density are observed between principal cells and interneurons in the BLA. Furthermore, we have also identified some important morphological properties within the neuronal subtypes that may give rise to the observed level of cell subtype heterogeneity within the BLA. MATERIALS AND METHODS

Ethics statement. All experimental procedures were approved by The University of Queensland and the Queensland University of Technology Animal Ethics Committees and complied with the policies and regulations regarding animal experimentation and other ethical matters (Drummond 2009). They were conducted in accordance with the Queensland Government Animal Research Act 2001, associated Animal Care and Protection Regulations (2002 and 2008), as well as the Australian Code for the Care and Use of Animals for Scientific Purposes, 8th Edition (National Health and Medical Research Council, 2013). Surgical and experimental procedures. A total of forty 8-wk-old male Wistar rats were used in this study. Rats were deeply anesthetized by intraperitoneal injection of pentobarbital sodium (60 – 80 mg/kg; Vetcare, Brisbane, Australia) and then decapitated. The brain was quickly removed, and a sagittal cut was made along the midline of the forebrain. The tissue block was fixed rostral end down with cyanoacrylate glue onto a metal bath, and the lateral side was rested on an agar block previously superglued to the metal bath. During the dissection, the tissue was bathed in ice-cold high-Mg2⫹ Ringer solution that contained (in mM) 130 NaCl, 3 KCl, 26 NaHCO3, 1.25 NaH2PO4, 5 MgCl2, 1 CaCl2, and 10 D-glucose (osmolarity ⬃300 mosM; Vapro 5520 osmometer, Wescor, South Logan, UT) (Bellingham and Berger 1996). Ringer solution was continuously bubbled with 95% O2-5% CO2 to maintain pH at 7.4.

943

The metal bath containing the fixed tissue block was then filled with the same ice-cold Ringer solution used in dissection. Coronal sections (300 ␮m thick) were cut with a vibratome (DSK 1000, Ted Pella), which had a large water bath surrounding the metal tissue bath, also kept ice cold. The dorsal hippocampus, caudate putamen, and external capsule were used as landmarks to locate the amygdala. During sectioning, the BLA was identified with a dissection microscope, as characteristic bifurcations appeared in the lateral hemisphere. Usually four serial sections containing BLA were obtained per hemisphere. We targeted healthy-looking neurons from the lateral and basolateral divisions within the BLA complex in a randomized manner (see Fig. 1). Cut slices were then moved from ice-cold bath solution (high Mg2⫹) and incubated for 60 min in the same solution kept at 34°C in a water bath. The sections were then moved to a normal Ringer solution in a similar incubation chamber and kept at room temperature (22–24°C) for 45 min prior to the start of recording and labeling. Normal Ringer solution used for recording contained (in mM) 130 NaCl, 3 KCl, 26 NaHCO3, 1.25 NaH2PO4, 1 MgCl2, 2 CaCl2, and 10 D-glucose, continuously bubbled with 95% O2-5% CO2 to maintain pH at 7.4. The electrodes were pulled from borosilicate glass capillaries (Vitrex Modulohm, Edwards Medical) and were filled with pipette solution (giving a tip resistance of ⬃3– 4 M⍀). The pipette solution contained 2% NB (Vector Labs, Burlingame, CA) in an artificial intracellular solution containing (in mM) 135 Cs⫹-methanesulfonate (Cs⫹MeSO4), 6 KCl, 1 ethylene glycol-bis(␤-aminoethyl ether)N,N,N=,N=-tetraacetic acid (EGTA), 2 MgCl2, 5 Na-4-(2-hydroxyethyl)piperazine-1-ethanesulfonic acid (Na-HEPES), 3 ATP-Mg2⫹, and 0.3 GTP-Tris (pH 7.25 with KOH, osmolarity 305 ⫾ 5 mosM) (Kanjhan and Vaney 2008). This NB-containing solution was placed in the pipette tip with a Hamilton syringe (1–2 ␮l), and the pipette was then backfilled with the same solution without NB. Visualizing and recording neurons in the BLA. A single brain slice was placed in a tissue chamber, stabilized with the aid of a metal mesh on a Nikon E600FN (Tokyo, Japan) microscope fitted with IR-DIC optics, and continuously superfused (1–2 ml/min) with the normal Ringer solution at room temperature (22–24°C). The tissue and recording electrodes were viewed on a Sony monitor through a ⫻60 water-immersion objective and infrared video camera (Hamamatsu) mounted above the camera port of the microscope. Neurobiotin electroporation. Recordings and voltage pulse protocols were made with an Axopatch 1D amplifier (Axon Instruments, Foster City, CA). Data were acquired at a sampling rate of 10 kHz, low-pass filtered at 2 kHz, and stored on a Windows computer with pCLAMP 10.2 software and a Digidata 1332A digitizer (Axon Instruments). Individual neurons within the BLA were targeted visually, and a patch electrode was advanced toward the cell body with the aid of a manipulator (MPC-200, Sutter Instrument). The electrode tip was maneuvered against the neuronal soma, and then continuous gentle suction was applied until a stable seal of preferably ⬎300 M⍀ was obtained. The semiloose seal NB electroporation procedure was modified from that described for retinal ganglion cells and brain stem motor neurons (Kanjhan and Bellingham 2013; Kanjhan and Vaney 2008). Square-wave voltage steps (0.5 s, 1 Hz) of 5–25 mV that generated current pulses of 300 –500 pA were applied for 6 –9 min. In the semiloose configuration, the cell membrane usually breaks down easily in the presence of NB in the pipette solution (Kanjhan and Vaney 2008). If a tight seal formed during suction through the pipette, then the cell was stimulated with 50- to 100-mV square-wave pulses of 500-ms duration at 1 Hz until these caused a significant drop in the membrane input resistance (e.g., ⬃80 M⍀), which usually occurred within 10 –30 s. As soon as the membrane broke, pulses were stopped, and sIPSCs were recorded at a holding membrane potential of 0 mV. Immunocytochemistry. After filling, the sections were left to rest in the bath for a minimum of 5 min to allow diffusion of NB. Sections

J Neurophysiol • doi:10.1152/jn.00824.2014 • www.jn.org

944


were then fixed in 4% paraformaldehyde in 0.1 M phosphate buffer (pH 7.4) for 20 –30 min at room temperature and subsequently washed three or four times for 30 min in 0.1 M phosphate-buffered saline pH 7.4 (PBS) at 4°C. Sections were incubated for 2 h in PBS containing 4% bovine serum albumin and 0.5% Triton X-100 at 4°C to block nonspecific background labeling. Sections were then incubated for 4 h at 4°C in Cy3-streptavidin (Sigma; 1:500 in blocking solution) to visualize NB. The labeling quality of cellular morphology was checked with a basic fluorescence microscope (Olympus BX50, Olympus) with sections bathed in 0.1 M PBS in small petri dishes. Only neurons observed to have exceptionally good filling and no vacuolations or blebbing in the dendritic processes were included in this study. To identify neurochemical GABAergic synapses on BLA neurons, we used antibodies against key pre- and postsynaptic components of GABAergic synapses and digital filtering in a manner identical to previous studies (Fogarty et al. 2013). GABAergic presynaptic terminals were labeled with a primary antibody against VGAT (vesicular inhibitory amino acid transporter), a GABA-glycine cotransporter [mouse anti-VGAT, Synaptic Systems, Goettingen, Germany, catalog no. 131011 (Witkovsky et al. 2008)] prevalent in both GABAergic and glycinergic inhibitory neurons, localizing to the synaptic vesicles within the terminal ends of these neurons (Chaudhry et al. 1998; Dumoulin et al. 1999; McIntire et al. 1997). For inhibitory postsynaptic specializations, we used a primary antibody against the GABAA receptor ␣1-subunit [rabbit anti-GABAA ␣1-subunit, Synaptic Systems, catalog no. 224203 (Mendu et al. 2012; Muir and Kittler 2014)], a postsynaptic receptor component of inhibitory GABAA synapses (Vicini et al. 2001). Antibodies were tested for specificity with primary antibody omission as controls. All primary antibodies were diluted at 1:200 in 0.1 M PBS-2% BSA-0.5% Triton X-100 and incubated for 48 h at 4°C. Primary antibodies were detected with a secondary antibody cocktail consisting of 1:400 anti-rabbit Alexa 488 (Invitrogen, Mulgrave, Australia) plus 1:400 anti-mouse Cy5 (Millipore, Billerica, MA), incubated for 48 h at 4°C. Between primary and secondary antibody incubations, slices were washed three times in 0.1 M PBS. After secondary antibody incubation, all slices were washed in 0.1 M PBS and mounted on slides (Menzel-Glaser, Braunschweig, Germany) in a standard glycerol-based p-phenylenediamine mounting medium and coverslipped (Menzel-Glaser). Sections were imaged with a Olympus FV1000 confocal microscope (Tokyo, Japan) using sequential laser scanning. Imaging and morphological quantification. Morphological properties of the BLA neurons filled with NB were analyzed from stacks of confocal images obtained at low and high magnifications (from ⫻20 to ⫻60) in a manner similar to previous studies using the same semiloose electroporation technique (Kanjhan and Bellingham 2013; Kanjhan and Vaney 2008). Images containing total dendritic arbors were obtained with a 0.75 NA ⫻20 Olympus air objective with a voxel size of 0.6199 ⫻ 0.6199 ⫻ 1.1 ␮m/pixel (x-, y-, and z-dimensions). Image z stacks of between 40 and 80 images were acquired at a z separation of 1.16 ␮m. For spine and neurochemical synapse quantification we obtained images with a 1.35 NA ⫻60 Olympus oil immersion objective at 1.7 zoom. The voxel size was 0.121 ⫻ 0.121 ⫻ 0.33 ␮m/pixel for spine images and 0.082 ⫻ 0.082 ⫻ 0.33 ␮m/pixel for neurochemical synapses (x-, y-, and z-dimensions). Image z stacks of dendrites contained between 15 and 35 images acquired at a step size of 0.3 ␮m. Morphological properties (dendritic branching, length, and dendritic spines) of filled cells were analyzed with Neurolucida software (MBF Bioscience, Williston, VT) in a manner identical to previous reports (Fogarty et al. 2015). An entire arbor consists of the entirety of the length of the dendritic trees emanating from the neuronal soma. A dendritic tree consists of all of the branches emanating from a single primary (1st order) branch extending from the neuronal soma (Larkman and Mason 1990). Dendritic processes were classified as spines only if length was no greater than 6 ␮m (Harris and Kater 1994). Criteria for classifying

cells as principal neurons, as opposed to interneurons, were as follows: 1) cell body must be larger than 25 ␮m in long-axis diameter (Washburn and Moises 1992) and 2) the cell must have an apical dendritic tree (McDonald 1982, 1984) of no less than 400 ␮m in length. Interneurons lacked an apical dendrite and displayed reduced dendrite length compared with principal dendrites, with the largest mean tree length of any labeled interneuron being 230 ␮m and the longest single tree in any interneuron measuring 310 ␮m in length. For principal neurons, the apical dendritic tree was considered to be the tree with the longest length, the smallest of which was 449 ␮m, smaller than only 9 (5.7%) of 156 basal dendritic trees. These longer basal dendritic trees may come from a population of principal neurons within the amygdala that consist of two or more large apical dendritic trees (McDonald 1982). In accordance with previous morphological studies in the BLA (Faber et al. 2001; McDonald 1982; Washburn and Moises 1992), these neurons were pooled with all principal neurons. Spiny and aspiny interneurons were classified on the basis of total dendritic arbor spine density, where all aspiny interneurons had spine densities of ⬍10 spines per 100 ␮m. Furthermore, in the subset of interneurons that were labeled with VGAT we found that five of eight spiny and three of four aspiny interneurons displayed positive immunolabeling, while we did not detect any immunolabeling for VGAT within principal cells (data not shown). Neurochemical synapse quantification. Robust density estimates of pre- and postsynaptic components of synapses on NB-filled neurons with IMARIS software (Bitplane, South Windsor, CT) have been established previously (Fogarty et al. 2013). Briefly, the neuronal surface was determined with Cy3-streptavidin conjugate-labeled NB. Presynaptic and postsynaptic fluorescence channels were then filtered based on the neuronal surface, so that only the fluorescence relevant to pre- and postsynaptic elements of the filled neuron remained. This required removal of all presynaptic (VGAT) fluorescence signals inside the filled cell and all postsynaptic (GABAA ␣1) fluorescence signals outside the filled cell. Semiautomated spot analysis allowed for proper quantification of the punctuate fluorescence of the presynaptic components localized to within 1 ␮m of the postsynaptic components of the filled cell as outlined previously (Fogarty et al. 2013), here expressed as neurochemical synapses per square micrometer of neuronal surface area. Electrophysiology analysis. Electrophysiological recording of sIPSCs was carried out after whole cell recording conditions were established. pCLAMP 10.2 (Axon Instruments) and a Windows computer were used to record epochs of data at a holding voltage of 0 mV, via a Digidata 1322 digitizer (sampling frequency 5–10 kHz, low-pass filter 2 kHz). Synaptic currents were detected and analyzed off-line with the optimally scaled sliding template algorithm in pCLAMP 10.2. Statistical methods. Means and SE were calculated for each data set. Where indicated, unpaired two-tailed t-tests, one-way ANOVAs with Tukey posttests, or two-way ANOVAs with Bonferroni posttests were conducted for all analysis involving the comparison of group means with Prism 5. Statistical significance was accepted at P ⬍ 0.05. All data in RESULTS are presented as means ⫾ SE. Correlations were performed with Pearson coefficients. RESULTS

Principal cell somas have a greater soma volume and greater dendritic arbor length compared with spiny and aspiny interneurons within the BLA. Morphological analysis was based on principal cells (n ⫽ 29) and interneurons (n ⫽ 31) sampled from the lateral and basolateral subdivisions of the BLA (Fig. 1). Principal cells were only included for analysis if they exhibited the morphological characteristics outlined above. Additionally, comparison of lateral and basolateral principal cells did not reveal any differences in soma volume,



945

Fig. 1. Neuronal location map showing locations of principal cells and spiny and aspiny interneurons sampled within the basolateral amygdala (BLA). We labeled and recorded from a range of topographic positions of the BLA at bregma ⫺2.54 (top left), ⫺2.76 (top right), ⫺3.00 (bottom left), and ⫺3.24 (bottom right) mm. This included the basolateral amygdala nucleus, anterior (BLA), basolateral amygdala nucleus, posterior (BLP), basomedial amygdala nucleus, posterior (BMP), lateral amygdala nucleus, dorsolateral (LaDL), lateral amygdala nucleus, ventrolateral (LaVL), and lateral amygdala nucleus, ventromedial (LaVM).

total arbor length, or spine density. Consequently, these cells were pooled together for analysis. Cell soma volumes (as calculated using volume of an ellipsoid) were greater in principal cells than in either spiny or aspiny interneurons within the BLA, while interneuron soma volumes did not differ [principal: 2,431 ⫾ 636 ␮m3 (n ⫽ 29), spiny: 1,185 ⫾ 156 ␮m3 (n ⫽ 23), aspiny: 949 ⫾ 194 ␮m3 (n ⫽ 8); principal vs. spiny, P ⫽ 0.0026; principal vs. aspiny, P ⫽ 0.0141; spiny vs. aspiny, P ⫽ 0.8944; 1-way ANOVA; Fig. 2, A–D]. Further examination of these BLA neurons revealed that the total dendritic length (the summed length of the total neuronal arbor) was greater for principal cells compared with either spiny or aspiny interneurons [principal: 2,034 ⫾ 127 ␮m (n ⫽ 29), spiny: 713 ⫾ 79 ␮m (n ⫽ 23), aspiny: 651 ⫾ 118 ␮m (n ⫽ 8); principal vs. spiny, P ⬍ 0.0001; principal vs. aspiny, P ⬍ 0.0001; spiny vs. aspiny; P ⫽ 0.9592; 1-way ANOVA; Fig. 2, A–C and E]. The mean tree length (the mean length of each dendritic tree per cell) was greater in principal apical dendrites than in both spiny and aspiny interneurons [principal apical: 793 ⫾ 101 ␮m (n ⫽ 29), spiny: 126 ⫾ 10 ␮m (n ⫽ 23), aspiny: 124 ⫾ 26 ␮m (n ⫽ 8); apical vs. spiny, P ⬍ 0.0001; apical vs. aspiny, P ⬍ 0.0001; 1-way ANOVA, Fig. 2, A–C and F]. Principal apical dendritic tree mean lengths were also significantly longer than principal basal mean tree lengths

[principal apical: 793 ⫾ 101 ␮m (n ⫽ 29), principal basal: 227 ⫾ 33 ␮m (n ⫽ 29); basal vs. apical, P ⬍ 0.0001]. There was no difference between principal basal, spiny, and aspiny interneuron dendritic tree mean lengths (basal vs. spiny, P ⫽ 0.1277; basal vs. aspiny, P ⫽ 0.3973; spiny vs. aspiny, P ⬎ 0.9999). The average reach (distance from the soma to the end of the dendrite) of principal apical dendrites was 224 ⫾ 14 ␮m (n ⫽ 27). All basic dendritic morphometric properties are presented for ease of comparison in Table 1. Comparison of spine densities between principal cell dendrites and spiny and aspiny interneuron dendrites in the BLA. The total spine density per 100 ␮m of principal neurons was greater than that of both spiny and aspiny interneurons, with spiny interneurons more spine dense than aspiny interneurons [principal: 43.6 ⫾ 2.5 (n ⫽ 26), spiny: 24.9 ⫾ 2.1 (n ⫽ 23), aspiny: 8.0 ⫾ 2.0 (n ⫽ 8); principal vs. spiny, P ⬍ 0.0001; principal vs. aspiny, P ⬍ 0.0001; spiny vs. aspiny, P ⫽ 0.0015]. The proximal dendrite spine density (spines per 100 ␮m of 1st- and 2nd-order dendrites) was greatest in principal cell apical dendrites (Fig. 3), which had significantly more spines compared with both spiny and aspiny interneuron dendrites within the BLA [principal apical 42 ⫾ 3 (n ⫽ 25), spiny: 28 ⫾ 4 ␮m (n ⫽ 23), aspiny: 12 ⫾ 3 ␮m (n ⫽ 8); apical vs. spiny, P ⫽ 0.0329; apical vs. aspiny, P ⬍ 0.0004; 1-way


946


Fig. 2. Increased somatic volume and dendrite arbor length in principal cells (A) compared with spiny (B) and aspiny (C) interneurons in the BLA. A–C show Neurobiotin-filled principal cell, spiny interneuron, and aspiny interneuron, respectively. The principal cell apical dendrite is indicated by the arrow in A. D shows that the somatic volume (␮m3) in principal cells (triangles, n ⫽ 29) is higher compared with spiny (diamonds, n ⫽ 23) and aspiny (circles, n ⫽ 8) interneurons. E shows that the total dendritic length (␮m) in principal cells (triangles) is higher compared with spiny (diamonds) and aspiny (circles) interneurons. F shows that the mean tree dendritic length (␮m) in principal cell apical dendritic trees (filled triangles) is higher compared with principal cell basal dendritic trees (open triangles) and spiny (diamonds) and aspiny (circles) interneurons. *P ⬍ 0.05, **P ⬍ 0.01, ****P ⬍ 0.0001; 1-way ANOVA of means ⫾ SE. Scale bar, 75 ␮m.

ANOVA; Fig. 3]. There was no difference in spine density between proximal apical dendrites and proximal basal dendrites of BLA principal cells [principal apical 42 ⫾ 3 (n ⫽ 25), principal basal: 31 ⫾ 3 (n ⫽ 26); apical vs. basal, P ⫽ 0.1306]. Basal proximal dendrites had increased spine densities compared with aspiny interneuron proximal dendrites (P ⫽ 0.0383); however, there was no difference in proximal dendritic spine densities between principal basal dendrites and spiny interneuron dendrites (P ⫽ 0.9130) or between spiny and aspiny interneuron dendrites (P ⫽ 0.1237). The distal dendrite spine density (spines per 100 ␮m of 3rd-order and later dendrites) was the same for principal cell apical dendrites and basal dendrites [principal apical: 57 ⫾ 4 (n ⫽ 22), principal basal: 52 ⫾ 6 (n ⫽ 21); apical vs. basal, P ⫽ 0.8923]. Spiny and aspiny interneuron distal dendrites within the BLA had significantly fewer spines compared with principal cell apical dendrites, with principal cell basal dendrites also having greater spine densities compared with aspiny but not spiny interneurons [spiny: 36 ⫾ 5 ␮m (n ⫽ 23), aspiny:

4 ⫾ 2 ␮m (n ⫽ 8); basal vs. spiny, P ⫽ 0.0568; basal vs. aspiny, P ⬍ 0.0001; apical vs. spiny, P ⫽ 0.0072; apical vs. aspiny, P ⬍ 0.0001; 1-way ANOVA; Fig. 4]. Spiny interneuron distal dendrite spine density was higher compared with aspiny interneurons (P ⫽ 0.0026). Comparison of proximal and distal spine densities revealed that both proximal principal apical and basal dendrites contained significantly fewer spines compared with distal apical and basal dendrites (proximal apical vs. distal apical, P ⫽ 0.0064; proximal basal vs. distal basal, P ⫽ 0.0012; Student’s t-test). There was no difference between the proximal and distal spine densities of spiny interneurons (P ⫽ 0.2427), but aspiny interneuron proximal dendritic spines were significantly more dense than distal spines (P ⫽ 0.0284). All basic dendritic spine density properties are presented for ease of comparison in Table 1. Branch order analysis of dendritic branch number, dendritic segment length, and dendritic spines in principal neurons, spiny interneurons, and aspiny interneurons. After the characterization of the gross morphologies of principal neurons and



947

Table 1. General morphological parameters of principal neurons, spiny interneurons, and aspiny interneurons Parameter

Principal Neuron

Spiny Interneuron

Aspiny Interneuron

P Values

Soma volume, ␮m3

2,431 ⫾ 311

1,185 ⫾ 156

949 ⫾ 194

Total dendrite length, ␮m

2,034 ⫾ 127

713 ⫾ 79

651 ⫾ 118

Mean tree length, ␮m

Basal: 227 ⫾ 33 Apical:793 ⫾ 101

126 ⫾ 10

124 ⫾ 26

Proximal spines per 100 ␮m

Basal: 31 ⫾ 3 Apical: 42 ⫾ 3

28 ⫾ 4

12 ⫾ 3

Distal spines per 100 ␮m

Basal: 52 ⫾ 6 Apical: 57 ⫾ 4

36 ⫾ 5

4⫾2

P vs. S: **0.0026 P vs. A: *0.0141 S vs. A: 0.6944 P vs. S: ****0.0001 P vs. A: ****0.0001 S vs. A: 0.9592 PB vs. PA: ****0.0001 PB vs. S: 0.1277 PB vs. A: 0.3973 PA vs. S: ****0.0001 PA vs. A: ****0.0001 S vs. A: 0.9999 P vs. S: *0.0414 P vs. A: ***0.0006 S vs. A: 0.1237 P vs. S: **0.0092 P vs. A: ****0.0001 S vs. A: **0.0026

All data presented as means ⫾ SE for principal neurons [P; n ⫽ 29], spiny interneurons (S; n ⫽ 23), and aspiny interneurons (A; n ⫽ 8). PA, principal apical; PB, principal basal. *P ⬍ 0.05, **P ⬍ 0.01, ***P ⬍ 0.001, ****P ⬍ 0.0001, 1-way ANOVAs with Tukey’s posttest.

interneurons, we analyzed the dendritic arbors and spine densities with regard to their branch order characteristics. Our comprehensive assessment of the dendritic trees quantified the number of dendritic segments per branch order, the mean length of dendritic segments per branch order, and the total number of spines per branch order of each neuronal tree type (Fig. 5). A summary of branch order data and analysis is presented in Table 2. The mean dendritic branch segment number per branch order was significantly altered between neuronal tree types (P ⬍ 0.002; Fig. 5A) and branch order (P ⬍ 0.0001; Fig. 5A). Mean number of dendritic segments per branch order was significantly lower in apical dendrites compared with all three other types at first-order branches [principal apical: 1.0 ⫾ 0.0 (n ⫽ 29), principal basal: 5.4 ⫾ 0.4 (n ⫽ 29), spiny: 6.0 ⫾ 0.3 (n ⫽ 23), aspiny: 6.0 ⫾ 0.4 (n ⫽ 8); apical vs. basal, P ⬍ 0.0001; apical vs. spiny, P ⬍ 0.0001; apical vs. aspiny, P ⬍ 0.0009; 2-way ANOVA; Fig. 5A]. At second-order branches, apical dendrites also had reduced numbers of dendritic segments compared with basal and spiny interneuron trees [principal apical: 2.1 ⫾ 0.1 (n ⫽ 29), principal basal: 7.1 ⫾ 0.6 (n ⫽ 29), spiny: 6.5 ⫾ 0.6 (n ⫽ 23), aspiny: 4.0 ⫾ 0.8 (n ⫽ 8); apical vs. basal, P ⬍ 0.0001; apical vs. spiny, P ⬍ 0.0001; 2-way ANOVA; Fig. 5A]. These results were primarily due to principal neurons in our criteria only having a single apical dendrite and thus only having a solitary first-order branch and few second-order branches. At third-order branches, principal neuron basal dendrites had significantly increased numbers of dendritic segments compared with all three other tree types [principal apical: 3.2 ⫾ 0.3 (n ⫽ 29), principal basal: 6.7 ⫾ 0.6 (n ⫽ 29), spiny: 4.2 ⫾ 0.6 (n ⫽ 23), aspiny: 1.5 ⫾ 0.5 (n ⫽ 8); apical vs. basal, P ⬍ 0.0002; basal vs. spiny, P ⬍ 0.0346; basal vs. aspiny, P ⬍ 0.0003; 2-way ANOVA; Fig. 5A]. The number of dendritic segments in basal dendrites remained higher than those in spiny interneurons at fourth-order branches [principal apical: 3.0 ⫾ 0.3 (n ⫽ 29), principal basal: 4.7 ⫾ 0.6 (n ⫽ 29), spiny: 2.1 ⫾ 0.4 (n ⫽ 23), aspiny: 1.5 ⫾ 0.5 (n ⫽ 8); basal vs spiny, P ⬍ 0.0153; 1-way ANOVA; Fig. 5A]. At fifth-order branches and beyond, the mean number of dendritic segments

per branch order was significantly higher in apical dendrites compared with all other types and in basal dendrites compared with spiny and aspiny interneurons [principal apical: 7.4 ⫾ 1.4 (n ⫽ 29), principal basal: 4.0 ⫾ 1.0 (n ⫽ 29), spiny: 1.1 ⫾ 0.4 (n ⫽ 23), aspiny: 0.5 ⫾ 0.5 (n ⫽ 8); apical vs. basal, P ⫽ 0.0004; apical vs. spiny, P ⬍ 0.0001; apical vs aspiny, P ⬍ 0.0001; basal vs. spiny, P ⫽ 0.0064; basal vs. aspiny, P ⫽ 0.0349; 2-way ANOVA; Fig. 5A]. The mean length of dendritic segments per branch order was significantly different for each tree type (P ⫽ 0.0002; Fig. 5B). The mean length of dendritic segments per branch order was unchanged at first-, second-, and third-order branches (Fig. 5B). At fourth-order branches the mean length of dendritic segments for principal neuron apical dendrites was significantly greater than that of spiny interneurons [principal apical: 57.4 ⫾ 7.7 ␮m (n ⫽ 29), principal basal: 40.6 ⫾ 5.0 ␮m (n ⫽ 29), spiny: 27.1 ⫾ 5.5 ␮m (n ⫽ 23), aspiny: 44.1 ⫾ 19.5 ␮m (n ⫽ 8); apical vs. spiny, P ⫽ 0.0046; 2-way ANOVA; Fig. 5B], and at fifth order and beyond the mean length of dendritic segments for principal neuron apical dendrites was significantly greater than that of all other neuronal tree types [principal apical: 87.8 ⫾ 13.9 ␮m (n ⫽ 29), principal basal: 30.5 ⫾ 5.2 ␮m (n ⫽ 29), spiny: 15.6 ⫾ 5.2 ␮m (n ⫽ 23), aspiny: 1.4 ⫾ 1.4 ␮m (n ⫽ 8); apical vs. basal, P ⬍ 0.0001; apical vs. spiny, P ⬍ 0.0001; apical vs. aspiny, P ⬍ 0.0001; 2-way ANOVA; Fig. 5B]. The total of number dendritic spines per branch order was significantly different depending on the branch order (P ⬍ 0.0001; Fig. 5C) and tree type (P ⫽ 0.0279; Fig. 5C). The total number of dendritic spines per branch order was unchanged at first- and second-order branches (Fig. 5C). At third-order branches total number of dendritic spines for principal neuron basal dendrites was significantly increased compared with aspiny interneurons [principal apical: 76.1 ⫾ 11.5 (n ⫽ 29), principal basal: 137.5 ⫾ 24.8 (n ⫽ 29), spiny: 88.8 ⫾ 25.2 (n ⫽ 23), aspiny: 3.3 ⫾ 1.7 (n ⫽ 8); basal vs. aspiny, P ⫽ 0.0076; 2-way ANOVA; Fig. 5C]. The total number of dendritic spines per branch order was unchanged at fourth-order branches (Fig. 5C). At fifth-order branches and beyond, the


948


rons have a lower spine density compared with principal basal (P ⫽ 0.0019) and spiny (P ⫽ 0.0024) interneurons. These data show that principal apical dendrites display increased dendritic complexity (branches), length, and distal synaptic substrates (spines) at distal branch ramifications compared with other tree types. Furthermore, there is no morphological difference between spiny and aspiny interneurons in either branch number or dendritic length when analyzed in this manner. Decreased GABAergic neurochemical synapse density on principal cell somas and dendritic arbors compared with interneuron somas and dendritic arbors. In addition to the detailed morphological examinations of individually filled

Fig. 3. Proximal dendrite spine density is highest in apical principal cell dendrites (A) followed by basal principal cell dendrites (B), spiny interneurons (C), and aspiny interneurons (D) in the BLA. E shows the proximal spine density (spines per 100 ␮m) in principal cell apical dendrites (filled triangles, n ⫽ 25) compared with principal cell basal dendrites (open triangles, n ⫽ 26) and spiny (diamond, n ⫽ 23) and aspiny (circles, n ⫽ 8) interneuron dendrites. *P ⬍ 0.05, ***P ⬍ 0.001, 1-way ANOVA of means ⫾ SE. Scale bar, 5 ␮m.

total number of dendritic spines was increased in principal neuron apical dendritic trees compared with all other types, while that in principal neuron basal trees was increased compared with spiny interneurons [principal apical: 222.2 ⫾ 35.0 (n ⫽ 29), principal basal: 105.8 ⫾ 41.0 (n ⫽ 29), spiny: 14.8 ⫾ 5.8 (n ⫽ 23), aspiny: 0.3 ⫾ 0.3 (n ⫽ 8); apical vs. basal, P ⫽ 0.0001; apical vs. spiny, P ⬍ 0.0001; apical vs. spiny, P ⬍ 0.0001; basal vs. spiny, P ⫽ 0.0106; 2-way ANOVA; Fig. 5C]. When dendritic length was taken into account for the above analysis (i.e., a spine density per 100 ␮m), branch order (P ⬍ 0.0001) and tree type (P ⬍ 0.003) were significantly different, with the spine density of aspiny interneurons significantly less than that of principal apical dendrites at fourth-order branches (P ⫽ 0.0286). At fifth-order branches, principal apical dendrites had significantly fewer spines than principal basal (P ⫽ 0.001) and spiny (P ⫽ 0.003) interneurons. Aspiny interneu-

Fig. 4. Distal dendrite spine density is increased in apical (A) and basal (B) principal cell dendrites compared with spiny (C) and aspiny (D) interneurons. Aspiny interneuron distal spine density is less than all other groups, including basal principal cell dendrites (B). E shows the distal spine density (spines per 100 ␮m) in principal cell apical dendrites (filled triangles, n ⫽ 22) compared with principal cell basal dendrites (open triangles, n ⫽ 21) and spiny (diamonds, n ⫽ 23) and aspiny (circles, n ⫽ 8) interneuron dendrites. **P ⬍ 0.01, ****P ⬍ 0.0001, 1-way ANOVA of means ⫾ SE. Scale bar, 5 ␮m.



Fig. 5. Branch order analysis of dendritic segment branch number per branch order (A), dendritic segment length per branch order (B), and dendritic spines per branch order (C) in principal neuron apical (n ⫽ 29) and basal (n ⫽ 29) dendrites, spiny interneurons (n ⫽ 23), and aspiny interneurons (n ⫽ 8). These data show no differences between spiny interneurons and aspiny interneurons in any of the analyses performed, supporting the pooling of their respective data. *P ⬍ 0.05, **P ⬍ 0.01, ***P ⬍ 0.001, ****P ⬍ 0.0001, 2-way ANOVA of means ⫾ SE with Bonferroni posttests.

principal cells and interneurons, we used immunolabeling, confocal microscopy, and imaging processing techniques (see MATERIALS AND METHODS and Fogarty et al. 2013) to locate the colocalized pre- and postsynaptic components of putative GABAergic synapses on NB-filled BLA cells. To check for antibody penetration into brain slices under our processing conditions, we examined raw confocal z-slice images at z intervals of 4 ␮m for imaging channels of the NB-filled cell (Cy3), the presynaptic GABA component (VGAT), and the postsynaptic (GABAA␣1) component and found that antibody penetration was present up to 32 ␮m (the entirety of the soma and proximal dendritic tree), as shown in image series for a representative principal neuron and interneuron (Fig. 6). Because neurochemical GABAergic synapse density [spiny: 0.086 ⫾ 0.013 spots/␮m2 (n ⫽ 8), aspiny: 0.063 ⫾ 0.012

949

spots/␮m2 (n ⫽ 4); P ⫽ 0.295, Student’s t-test], IPSC frequency [spiny: 1.46 ⫾ 0.11 Hz (n ⫽ 5), aspiny: 1.24 ⫾ 0.18 Hz (n ⫽ 3); P ⫽ 0.3, Student’s t-test], and branch order morphometry were not different between spiny and aspiny interneurons, the interneuron subtypes were pooled for neurochemical synapse density and inhibitory synaptic physiology analysis. Neurochemical GABAergic synapse density (spots/␮m2) was lower on principal cell somas and proximal dendrites compared with interneurons [principal: 0.0554 ⫾ 0.005 spots/ ␮m2 (n ⫽ 13), interneuron: 0.0780 ⫾ 0.010 spots/␮m2 (n ⫽ 12); P ⫽ 0.0451, Student’s t-test; Fig. 7, A, B, and E]. The density of neurochemical GABAergic synapses on the distal dendrites of principal cells was also lower compared with interneurons [principal: 0.0891 ⫾ 0.013 spots/␮m2 (n ⫽ 7), interneuron: 0.1234 ⫾ 0.010 spots/␮m2 (n ⫽ 8); P ⫽ 0.0493, Student’s t-test, Fig. 7, C, D, and F]. Figure 8 shows two three-dimensional (3D)-rendered neurons (Cy3) with GABAergic neurochemical synaptic “spots” (Fig. 8, A and B). Sholl analysis of the dendritic arbor shows that interneurons have significantly fewer interactions (P ⬍ 0.0001), consistent with interneurons having shorter branches that are predominantly within 80 ␮m from the soma (10-␮m bins; Fig. 8C). GABAergic neurochemical synapses on the soma and dendritic arbors (50-␮m bins; Fig. 8D) are significantly increased for interneurons compared with principal neurons (P ⫽ 0.0419) and diminish rapidly with increasing radial distance from the soma in both neuronal types (P ⬍ 0.0001). However, interneurons have significantly higher densities of putative GABAergic synapses than principal neurons on dendrites between 0 and 50 ␮m away from the soma (P ⫽ 0.0419). As a percentage, 67% of principal neuron putative GABAergic synapse density and 64% of interneuron neurochemical GABAergic synapse density were located on the soma. On dendrites up to 50 ␮m radially extending from the soma, principal neurons exhibited 17% and interneurons 28% of their GABAergic neurochemical synapse density. Dendrites between 50 and 100 ␮m contained 9% of principal neuronal and 8% of interneuronal GABAergic neurochemical synapses, while 3% of principal neuronal GABAergic neurochemical synapses were between 100 and 150 ␮m and the final 3% beyond 150 ␮m. Less than 1% of the total putative GABAergic synapse density of interneurons was found beyond 100 ␮m away from the soma. Figure 8 illustrates that the vast majority of the interneuron dendritic arbor is perisomatic, along with most of the putative GABAergic synapses on both principal neurons and interneurons, although interneurons are more likely to have a greater amount of inhibitory synapses on their dendrites compared with principal neurons. BLA principal cells exhibit reduced inhibitory postsynaptic currents compared with interneurons. To determine whether morphological differences in putative GABAergic synapse density were accompanied by functional differences in synaptic activity, we analyzed sIPSCs recorded from BLA principal cells (n ⫽ 12) and interneurons (n ⫽ 8) (Table 3). While there was no difference in IPSC amplitude, half-width, or decay time, we did observe significant reduction in the IPSC rise time and increase in the IPSC frequency for interneurons compared with principal cells (Fig. 9 and Table 3).


950


Table 2. Branch order characteristics for different tree types of BLA cells Principal Apical

Principal Basal

Spiny Interneuron

Aspiny Interneuron

1.0 ⫾ 0.0

5.4 ⫾ 0.4

6.0 ⫾ 0.3

6.0 ⫾ 0.4

1st-order mean branch segment length, ␮m 1st-order branch mean total spines 2nd-order branch segments

15.7 ⫾ 3.4 7.3 ⫾ 2.0 2.1 ⫾ 0.1

27.6 ⫾ 3.0 50.8 ⫾ 7.8 7.1 ⫾ 0.6

31.4 ⫾ 6.2 57.3 ⫾ 12.8 6.5 ⫾ 0.6

33.8 ⫾ 7.6 10.3 ⫾ 2.2 4.0 ⫾ 0.8

2nd-order mean branch segment length, ␮m 2nd-order branch mean total spines 3rd-order branch segments

33.3 ⫾ 5.9 30.8 ⫾ 6.4 3.2 ⫾ 0.3

35.7 ⫾ 2.2 86.6 ⫾ 11.4 6.7 ⫾ 0.6

39.0 ⫾ 6.1 93.0 ⫾ 29.0 4.2 ⫾ 0.6

37.5 ⫾ 7.6 14.5 ⫾ 4.6 1.5 ⫾ 0.5

3rd-order mean branch segment length, ␮m 3rd-order branch mean total spines 4th-order branch segments 4th-order mean branch segment length, ␮m 4th-order branch mean total spines 5th-order and above branch segments

47.3 ⫾ 5.1 76.1 ⫾ 11.5 3.0 ⫾ 0.3 57.4 ⫾ 7.7 96.5 ⫾ 17.5 7.4 ⫾ 1.4

40.0 ⫾ 3.3 137.5 ⫾ 24.8 4.7 ⫾ 0.6 40.6 ⫾ 5.0 103.6 ⫾ 18.3 4.0 ⫾ 1.0

44.9 ⫾ 4.3 88.8 ⫾ 25.2 2.1 ⫾ 0.4 27.1 ⫾ 5.5 31.0 ⫾ 6.4 1.1 ⫾ 0.4

20.6 ⫾ 8.9 3.3 ⫾ 1.7 1.5 ⫾ 0.5 44.1 ⫾ 19.5 5.2 ⫾ 3.2 0.5 ⫾ 0.5

5th-order and above mean branch segment length, ␮m

87.8 ⫾ 13.9

30.5 ⫾ 5.2

15.6 ⫾ 5.2

1.4 ⫾ 1.4

222.2 ⫾ 35.0

105.8 ⫾ 41.1

14.8 ⫾ 5.8

0.29 ⫾ 0.29

Branch Order Properties 1st-order branch segments

5th-order branch mean total spines

P Values PA vs. PB: ****P ⬍ 0.0001 PA vs. S: ****P ⬍ 0.0001 PA vs. A: ***P ⬍ 0.0009 Not significant Not significant PA vs. PB: ****P ⬍ 0.0001 PA vs. S: ****P ⬍ 0.0001 Not significant Not significant PA vs. PB: ****P ⫽ 0.0002 PB vs. S: *P ⫽ 0.0346 PB vs. A: ***P ⫽ 0.0003 Not significant PB vs. A: **P ⫽ 0.0076 PB vs. S: *P ⫽ 0.0153 PA vs. S: **P ⫽ 0.0045 Not significant PA vs. PB: ***P ⫽ 0.0004 PB vs. S: **P ⫽ 0.0064 PB vs. A: *P ⫽ 0.0349 PA vs. S: ****P ⬍ 0.0001 PA vs. A: ****P ⬍ 0.0001 PA vs. PB: ****P ⬍ 0.0001 PA vs. S: ****P ⬍ 0.0001 PA vs. A: ****P ⬍ 0.0001 PA vs. PB: ***P ⫽ 0.0001 PB vs. S: *P ⫽ 0.0106 PA vs. S: ****P ⬍ 0.0001 PA vs. A: ****P ⬍ 0.0001

All data presented as means ⫾ SE for principal neurons (n ⫽ 29), spiny interneurons (n ⫽ 23), and aspiny interneurons (n ⫽ 8). BLA, basolateral amygdala. *P ⬍ 0.05, **P ⬍ 0.01, ***P ⬍ 0.001, ****P ⬍ 0.0001, 2-way ANOVAs with Bonferroni’s posttest.

In the subset of cells with both GABAergic currents recorded and successful neuronal morphology recovery, we performed correlations to determine the relationship between IPSC frequency and GABAergic neurochemical synapse density (per ␮m2) in the same cell (Fig. 9D). The Pearson correlation coefficient was 0.6127, with a P value (P ⫽ 0.0342) indicating that this association is unlikely to be random. DISCUSSION

In this study we have used a combination of electrophysiology, immunolabeling, and imaging techniques to provide a detailed morphological analysis of principal cells and interneurons, the two main neuron types within the BLA complex. Furthermore, we have determined the density of putative GABAergic synaptic inputs along the entire dendritic arbor of each cell type and, in a subset of the NB-filled BLA neurons, correlated this with neurophysiological properties, using electrophysiological recordings of sIPSCs. Previous investigations in rat BLA using Golgi staining (McDonald 1982, 1984, 1992) or intracellular labeling of cells with fluorescent tracers including Lucifer yellow (Washburn and Moises 1992) and biocytin (Faber et al. 2001; Rainnie et al. 1993) have provided important information regarding morphology of neurons within this region. We have extended these studies by generating z-stack confocal microscopy fluorescent images using NB-filled BLA neurons, to make measurements of cell body volume, total arbor length, mean tree lengths, branch order characteristics, and spine densities from young adult male Wistar rats at 8 wk of age. Our morphological

analysis also combines neurochemical identification of GABAergic synapses with functional analysis of inhibitory synaptic activity in single neurons from the rat BLA. There are a number of previous studies reporting dendritic arbor length in rat BLA principal cells, with total dendritic arbor lengths in 8-wk-old Wistar rats estimated at 1,822 ␮m by Golgi-Cox staining (Vyas et al. 2002). Our ex vivo slice NB filling and neuronal tracing gives a mean of 2,034 ␮m for principal cell total dendritic arbor, although estimates range from a low of 1,338-1,447 ␮m in 9- to 10-wk-old SpragueDawley rats (Padival et al. 2013) to a high of 2,900 ␮m in 5-wk-old Long-Evans rats (Koss et al. 2014). It is important to note that in 17- to 20-day-old Wistar rats principal neuron arbors from the lateral amygdala were reported to be 6,2996,722 ␮m by a dye-filling method with DAB postprocessing (Faber et al. 2001). A combination of animal age, processing artifact, and Pythagorean error correction of camera lucida tracings may explain the disparity of this estimate compared with other estimates. However, it highlights a limitation of our own study, notably, that parts of the dendritic tree may be cut during the slicing process and contribute to smaller arbor sizes, although this is a caveat common to all studies using sectioning (Faber et al. 2001) regardless of method. Previous morphological characterization of BLA principal cells has revealed that these neurons exhibit large, complex dendritic arbors and contain a more prominent apical dendrite similar to that of cortical neurons (McDonald 1982). We therefore used the presence of an apical dendrite in addition to soma diameter to rudimentarily distinguish principal cells from



951

Fig. 6. Consistent levels of punctate labeling of presynaptic (VGAT) and postsynaptic (GABAA ␣1-subunit) components during immunohistochemistry in brain slices show antibody penetration into tissue. A and B show z stacks of 3 pseudocolored channels (purple ⫽ VGAT; green ⫽ GABAA ␣1; red ⫽ Cy3-avidin-Neurobiotin) for representative principal neurons (A) and interneurons (B). Each row represents a single z plane of 4 ␮m apart. Scale bar, 25 ␮m.

interneurons. BLA principal apical dendrites displayed distinct characteristics compared with all other dendritic tree types. We found that these dendrites that had a mean length of 793 ␮m, which was 3.5-fold longer than the mean length of basal

dendrites, and projected the greatest distance from the cell body (Fig. 2). Furthermore, examination of branch segments and mean segment length per branch order showed that principal apical dendrites displayed increased complexity at distal


952


Fig. 7. Increased somatic/proximal and distal dendritic GABAergic neurochemical synaptic density in interneurons (B) compared with principal cells (A) in the BLA. A and B show the somatic and proximal presynaptic VGAT (purple dots) and postsynaptic GABAA ␣1 subunit (green dots) puncta identified with IMARIS algorithm overlaid on transparent gray (Cy3) principal cells (A) and interneurons (B). C and D show distal dendritic presynaptic VGAT (purple dots) and postsynaptic GABAA ␣1-subunit (green dots) puncta identified with IMARIS algorithm overlaid on transparent gray (Cy3) principal cells (C) and interneurons (D). E and F show increased mean GABAergic neurochemical synapse density (synapses per ␮m2) in interneurons (squares, n ⫽ 12) compared with principal cells (triangles, n ⫽ 13) for soma and proximal dendrites (E) and distal dendrites (F). *P ⬍ 0.05, unpaired 2-tailed t-test of means ⫾ SE. Scale bars: A and B, 20 ␮m; C and D, 10 ␮m.

(5th order and above) branches compared with basal dendrites, which formed more complex proximal branching patterns that tapered off at more distal branch orders (Fig. 5). Our data also show that principal cell dendrite spine density increased at distal regions compared with proximal regions (Figs. 3 and 4). In addition, our analysis showed that BLA principal cells display large variations in soma volume. The observed level of somatic variation, dendritic organization, and spine localization is likely to provide important morphological control over the integration of synaptic inputs, which would be expected to

influence cellular output and potentially contribute to differences in BLA principal cell physiology. Previous studies using anterograde and retrograde tracing have shown that excitatory input from cortical and thalamic regions is received primarily at distal dendritic processes and spines of BLA principal neurons (Brinley-Reed et al. 1995; Farb and Ledoux 1997, 1999). More recently, there has been a plethora of research demonstrating that these connections play a pivotal role in associative learning tasks such as fear conditioning and extinction (reviewed in Duvarci and Pare 2014;



Johansen et al. 2011; Pape and Pare 2010) and the formation of memories that encode incentive value to rewarding stimuli (Everitt et al. 1999; Robbins and Everitt 2002). Our data show that principal cell apical dendrite structure is organized to support the formation of high levels of distal synaptic inputs, which are likely to be excitatory (Fig. 5). Furthermore, reduced proximal spine density suggests that this region may receive a greater portion of inhibitory inputs compared with distal dendrites. Indeed, our neurochemical results show a higher number

953

Table 3. Electrophysiological properties of spontaneous inhibitory postsynaptic currents recorded from BLA principal cells and interneurons Amplitude, Rise time, pA ms

Decay time, ms

Principal 22.8 ⫾ 2.1 4.0 ⫾ 0.3* 19.3 ⫾ 1.7 Interneuron 29.9 ⫾ 2.6 2.8 ⫾ 0.4* 19.3 ⫾ 1.4

Half-width, ms 8.5 ⫾ 1.0 8.1 ⫾ 1.1

Frequency, Hz

n

0.8 ⫾ 0.1*** 12 1.4 ⫾ 0.1*** 8

Data are electrophysiological properties of spontaneous inhibitory postsynaptic currents recorded from BLA principal cells and interneurons, showing mean ⫾ SE peak amplitude, 10 –90% rise time, 90 –10% decay time, halfwidth, and frequency. *P ⬍ 0.05, ***P ⬍ 0.001, unpaired 2-sided t-test.

of GABAergic synapses on the soma and proximal dendrites compared with the distal dendrites of principal neurons (Fig. 8), potentially facilitating inhibitory control of principal neuron firing (Miles et al. 1996). Analysis of BLA spiny and aspiny interneuron morphology showed that these neurons have cell body volumes that are on average twofold smaller and also have dendritic arbors that are on average three times shorter than principal cells (Fig. 2). Analysis of the two subsets of interneurons sampled showed that their general morphology measurements were fairly well conserved, with the exception of distal dendrite spine density (Fig. 4). After we divided our interneuron data set into spiny and aspiny interneurons on this basis, examination of these subgroups did not reveal any significant morphological differences, nor did we find any obvious localization patterns, as each BLA subnucleus we sampled from contained both spiny and aspiny interneurons (Fig. 1). Interestingly, we found that interneurons classified as spiny had similar proximal and distal spine densities compared with principal cell basal dendrites. This finding is noteworthy, as previous studies have suggested that BLA interneurons are generally spine sparse or aspiny (McDonald 1982; Sah et al. 2003; Washburn and Moises 1992). Species and age differences, in combination with the use of intracellularly filled neurons and analysis of serial z stacks performed at high magnification, may have contributed to these higher estimates of spine densities on BLA interneurons compared with previous reports. Additionally, as we processed all slices in the same manner and recorded from all cells at room temperature, these variables are unlikely to have contributed to variations in spine densities between the cell types. Previous studies have shown that a large amount of BLA inhibitory activity is provided by locally projecting interneurons (Bienvenu et al. 2012; Muller et al. 2006; Rainnie et al. 2006; Woodruff and Sah 2007), and electron microscopic analysis has revealed that BLA interneuron subtypes display Fig. 8. Sholl analysis of principal neuron and interneuron dendritic arbors and GABAergic neurochemical synapse numbers. A and B show somatic and proximal presynaptic VGAT (purple dots) and postsynaptic GABAA ␣1subunit (green dots) puncta identified with IMARIS algorithm overlaid on transparent gray (Cy3) principal cell (A) and interneuron (B). C shows Sholl analysis of dendritic branching, demonstrating significant increases in the number of branches and their location further away from the soma in principal neurons (n ⫽ 29) compared with interneurons (n ⫽ 31) with 10-␮m bins. D shows Sholl analysis of the GABAergic neurochemical synapse density, demonstrating increased somatic and proximal GABAergic neurochemical synapse density in interneurons (n ⫽ 12) compared with principal neurons (n ⫽ 13) using somatic and 50-␮m bins. *P ⬍ 0.05, ****P ⬍ 0.0001, 2-way ANOVAs with Bonferroni posttests. Scale bar, 20 ␮m. J Neurophysiol • doi:10.1152/jn.00824.2014 • www.jn.org

954


Fig. 9. Increased spontaneous inhibitory postsynaptic current (IPSC) frequency in interneurons (B) compared with principal cells (A) in the BLA. A and B show representative whole cell voltage-clamp recordings from principal cells (A) and interneurons (B) at 0-mV holding potential. C shows the mean frequency of spontaneous IPSC events, which was increased in interneurons (squares, n ⫽ 12) compared with principal cells (triangles, n ⫽ 8). D shows mean IPSC frequency and mean GABAergic neurochemical synaptic density from the same cells plotted against each other. This correlation is indicative of a moderate association between IPSC frequency recorded with whole cell patch clamp and GABAergic synapses as labeled with immunohistochemical techniques. ****P ⬍ 0.0001, unpaired 2-tailed t-test of means ⫾ SE.

distinct patterns of axonal innervation (McDonald and Betette 2001; Muller et al. 2003, 2005, 2006, 2007). These analyses showed that BLA interneurons expressing calbindin (CB) and parvalbumin (PV), which constitute ⬃50% of the cell popula-

tion (McDonald and Mascagni 2001; Rainnie et al. 2006), formed axonal arborizations that target the soma, proximal dendrites, and the axon initial segment of principal cells (McDonald and Betette 2001; Muller et al. 2005, 2006). Dense interconnections are also formed between somas and axon terminals of PV-containing interneurons (Muller et al. 2005). Regional differences in the connectivity of interneuron subtypes are also observed for somatostatin-containing interneurons, which predominantly target distal dendrites and spines of principal cells (Muller et al. 2007). More recently, an elegant study performed by Bienvenu and colleagues identified two additional interneuron subtypes, axo-axonic and amygdalostriatal. Axo-axonic cells primarily innervated the initial axon segment of principal cells, while amygdalo-striatal cells formed perisomatic innervation of principal neurons and also medium spiny neurons in the amygdala-striatum border (Bienvenu et al. 2012). Further work is required to determine whether the spiny and aspiny interneurons characterized in this study correspond to BLA interneurons expressing specific markers, such as PV, CB, or somatostatin, or have different innervation targets. While the above studies have provided a comprehensive classification of individual BLA interneuron axonal targets, a quantification of GABAergic synaptic connections to different neuron types within the BLA has been lacking. Consequently, we used a recently published method that allows for mapping of the neurochemical synaptic density in a single reconstructed 3D neuron (Fogarty et al. 2013) to quantify and compare putative GABAergic synaptic density on principal cells and interneurons in the rat BLA. Our Sholl analysis revealed an increased number of neurochemical GABAergic synapses at the soma and proximal dendrites of principal cells, which was reduced on dendrites with increased radial distance away from the soma (Fig. 8). Coupled with observed differences in proximal and distal spine density, our results suggest that the ratio of inhibitory to excitatory innervation is increased along the distal to proximal axis of principal cells in the BLA. This arrangement was also observed for BLA interneurons; however, densities of neurochemical GABAergic synapses at this cell type were significantly increased compared with principal cells (Fig. 7). Given that inhibitory input to BLA cells (McDonald et al. 2011) and locally projecting interneurons predominantly innervates perisomatic targets and subtypes shown to selectively target dendrites at distal regions (CB and somatostatin expressing) primarily interact with principal cells (Bienvenu et al. 2012; Muller et al. 2007), the high density of neurochemical GABAergic synapses formed at distal dendrites of BLA interneurons may arise from nonlocal inhibitory input; however, the source of this input remains to be determined. Comparison of neurochemical GABAergic synapse labeling between cell types showed that BLA interneurons receive a significantly higher density of putative GABAergic synapses compared with principal cells (Figs. 7 and 8). This was also accompanied by significant increases in the frequency of sIPSCs (Fig. 9 and Table 3) compared with principal cells. Although our neurochemical immunolabeling is limited by antibody penetration (Fogarty et al. 2013), and is therefore confined to subsampling of cell bodies and neuronal arbors that lie within 32 ␮m of the surface of the slice, we found an association between sIPSC frequency and neurochemical GABAergic density for cells in which both measurements were



obtained (Fig. 9). Because spontaneous synaptic currents consist of a mixture of miniature (TTX insensitive) synaptic currents and synaptic currents caused by presynaptic neuron firing (TTX sensitive) (Bellingham and Berger 1996; Doze et al. 1991), sIPSC frequency reflects a combination of both basal presynaptic release probability and presynaptic activity, and both factors are permissively limited by the number of inhibitory synapses. A parsimonious explanation for increased sIPSC frequency in BLA interneurons is increased density of inhibitory synapses. However, our data do not rule out the possibility of either increased presynaptic activity or basal release probability playing a role. Furthermore, our postsynaptic neurochemical GABAergic density estimates were determined by using an antibody against the GABAA ␣1-subunit to mark postsynaptic sites. Although the high level of GABAA ␣1 mRNA expression observed in the BLA suggests that most neurons in this region contain this subunit (Persohn et al. 1992; Wisden et al. 1992), it is possible that ␣1 immunoreactivity may be confined to the distal dendrites of some principal cells (McDonald and Mascagni 2004). While it is conceivable that a proportion of the neurochemical GABAergic synapses we quantified arose via connections between neurons in and outside the BLA (presumably those formed at distal dendrites), we believe that the cellular organization observed within this region suggests that the majority are formed by BLA interneuron-BLA interneuron and BLA interneuron-BLA principal cell contacts. Hence, our data suggest that BLA interneurons not only provide strong regulation of principal cell output by targeting the soma and proximal dendrites but also form a densely interconnected inhibitory network consisting of a high number of interneuron-interneuron GABAergic contacts. This level of somatodendritic organization, coupled with reciprocal excitatory connections provided by principal cells (McDonald et al. 2005; Smith et al. 2000), which can fire synchronously (Pare et al. 1995; Pare and Gaudreau 1996), may facilitate the propagation of inhibitory signals within the BLA. Furthermore, the subsequent inhibitory feedback onto principal cells is thought to contribute to rhythmic cell firing and synchronization of BLA activity (Pare et al. 2002). This was recently demonstrated for rhythmic compound IPSCs of PV interneurons, which improved spike-timing precision and synchronized principal cell firing in the BLA (Ryan et al. 2012). We believe that our technique is well suited to investigate changes in BLA morphology and activity, which have been implicated in a number of disease states, including posttraumatic stress disorder and drug addiction (Koob et al. 2014; Mahan and Ressler 2012; Roozendaal et al. 2009). Posttraumatic stress disorder and other forms of chronic stress are often associated with increases in amygdala activity, resulting in the dysregulation of fear responses and causing sensitization to stressful stimuli (Mahan and Ressler 2012; Roozendaal et al. 2009). Previous studies have shown that emotional responses generated by the BLA may result from increased excitatory input and changes in glutamate receptor ratios (Christian et al. 2012; Hubert et al. 2014; LeDoux 2003) and repeated restraint stress can cause changes in BLA spine localization (Padival et al. 2013). It would be interesting to investigate whether these changes result from altered levels of not only excitatory input but also inhibitory synapses within the BLA and other nuclei within the amygdala. Because a large majority of the inhibitory

955

signaling within the BLA is provided by locally projecting interneurons, one would expect reductions in interneuron synaptic number and/or activity to have a profound effect on the excitatory output of this region. Inhibitory synaptic plasticity has previously been demonstrated within the BLA (Szinyei et al. 2007), and reduced GABAergic activity facilitates longterm potentiation of thalamic input to the lateral amygdala (Bissiere et al. 2003). The use of the methodology described here combined with rodent models of fear conditioning, chronic stress, and drug addiction will allow further investigations to uncover whether such phenomena are accompanied by changes in morphology and inhibitory synaptic connectivity. GRANTS This work was supported by National Health and Medical Research Council funding (to S. E. Bartlett and M. C. Bellingham) and an Australian Research Council Future Fellowship (to S. E. Bartlett). DISCLOSURES No conflicts of interest, financial or otherwise, are declared by the author(s). AUTHOR CONTRIBUTIONS Author contributions: P.M.K., M.J.F., P.G.N., M.C.B., and S.E.B. conception and design of research; P.M.K., M.J.F., and A.B. performed experiments; P.M.K., M.J.F., and A.B. analyzed data; P.M.K., M.J.F., A.B., P.G.N., M.C.B., and S.E.B. interpreted results of experiments; P.M.K. and M.J.F. prepared figures; P.M.K. and M.J.F. drafted manuscript; P.M.K., M.J.F., P.G.N., M.C.B., and S.E.B. edited and revised manuscript; P.M.K., M.J.F., P.G.N., M.C.B., and S.E.B. approved final version of manuscript. REFERENCES Bellingham MC, Berger AJ. Presynaptic depression of excitatory synaptic inputs to rat hypoglossal motoneurons by muscarinic M2 receptors. J Neurophysiol 76: 3758 –3770, 1996. Bienvenu TC, Busti D, Magill PJ, Ferraguti F, Capogna M. Cell-typespecific recruitment of amygdala interneurons to hippocampal theta rhythm and noxious stimuli in vivo. Neuron 74: 1059 –1074, 2012. Bissiere S, Humeau Y, Luthi A. Dopamine gates LTP induction in lateral amygdala by suppressing feedforward inhibition. Nat Neurosci 6: 587–592, 2003. Brinley-Reed M, Mascagni F, McDonald AJ. Synaptology of prefrontal cortical projections to the basolateral amygdala: an electron microscopic study in the rat. Neurosci Lett 202: 45– 48, 1995. Carlsen J. Immunocytochemical localization of glutamate decarboxylase in the rat basolateral amygdaloid nucleus, with special reference to GABAergic innervation of amygdalostriatal projection neurons. J Comp Neurol 273: 513–526, 1988. Chaudhry FA, Reimer RJ, Bellocchio EE, Danbolt NC, Osen KK, Edwards RH, Storm-Mathisen J. The vesicular GABA transporter, VGAT, localizes to synaptic vesicles in sets of glycinergic as well as GABAergic neurons. J Neurosci 18: 9733–9750, 1998. Christian DT, Alexander NJ, Diaz MR, Robinson S, McCool BA. Chronic intermittent ethanol and withdrawal differentially modulate basolateral amygdala AMPA-type glutamate receptor function and trafficking. Neuropharmacology 62: 2430 –2439, 2012. Davis M, Rainnie D, Cassell M. Neurotransmission in the rat amygdala related to fear and anxiety. Trends Neurosci 17: 208 –214, 1994. Doze VA, Cohen GA, Madison DV. Synaptic localization of adrenergic disinhibition in the rat hippocampus. Neuron 6: 889 –900, 1991. Drummond GB. Reporting ethical matters in The Journal of Physiology: standards and advice. J Physiol 587: 713–719, 2009. Dumoulin A, Rostaing P, Bedet C, Levi S, Isambert MF, Henry JP, Triller A, Gasnier B. Presence of the vesicular inhibitory amino acid transporter in GABAergic and glycinergic synaptic terminal boutons. J Cell Sci 112: 811– 823, 1999. Duvarci S, Pare D. Amygdala microcircuits controlling learned fear. Neuron 82: 966 –980, 2014.


956


Ehrlich I, Humeau Y, Grenier F, Ciocchi S, Herry C, Luthi A. Amygdala inhibitory circuits and the control of fear memory. Neuron 62: 757–771, 2009. Everitt BJ, Parkinson JA, Olmstead MC, Arroyo M, Robledo P, Robbins TW. Associative processes in addiction and reward. The role of amygdalaventral striatal subsystems. Ann NY Acad Sci 877: 412– 438, 1999. Faber ES, Callister RJ, Sah P. Morphological and electrophysiological properties of principal neurons in the rat lateral amygdala in vitro. J Neurophysiol 85: 714 –723, 2001. Farb CR, LeDoux JE. NMDA and AMPA receptors in the lateral nucleus of the amygdala are postsynaptic to auditory thalamic afferents. Synapse 27: 106 –121, 1997. Farb CR, Ledoux JE. Afferents from rat temporal cortex synapse on lateral amygdala neurons that express NMDA and AMPA receptors. Synapse 33: 218 –229, 1999. Fogarty MJ, Hammond LA, Kanjhan R, Bellingham MC, Noakes PG. A method for the three-dimensional reconstruction of Neurobiotin-filled neurons and the location of their synaptic inputs. Front Neural Circuits 7: 153, 2013. Fogarty MJ, Noakes PG, Bellingham MC. Motor cortex layer V pyramidal neurons exhibit dendritic regression, spine loss, and increased synaptic excitation in the presymptomatic hSOD1(G93A) mouse model of amyotrophic lateral sclerosis. J Neurosci 35: 643– 647, 2015. Harris KM, Kater SB. Dendritic spines: cellular specializations imparting both stability and flexibility to synaptic function. Annu Rev Neurosci 17: 341–371, 1994. Hubert GW, Li C, Rainnie DG, Muly EC. Effects of stress on AMPA receptor distribution and function in the basolateral amygdala. Brain Struct Funct 219: 1169 –1179, 2014. Johansen JP, Cain CK, Ostroff LE, LeDoux JE. Molecular mechanisms of fear learning and memory. Cell 147: 509 –524, 2011. Kanjhan R, Bellingham MC. Neurobiotin electroporation for combined structural and functional analysis of neurons in developing mouse brain slices. In: In Excitation and Inhibition of Neural Responses, edited by Pilowsky PM, Farnham MM, Fong AY. New York: Humana, 2013, p. 151–165. Kanjhan R, Vaney DI. Semi-loose seal Neurobiotin electroporation for combined structural and functional analysis of neurons. Pflügers Arch 457: 561–568, 2008. Kemppainen S, Pitkanen A. Distribution of parvalbumin, calretinin, and calbindin-D(28k) immunoreactivity in the rat amygdaloid complex and colocalization with gamma-aminobutyric acid. J Comp Neurol 426: 441– 467, 2000. Koob GF, Buck CL, Cohen A, Edwards S, Park PE, Schlosburg JE, Schmeichel B, Vendruscolo LF, Wade CL, Whitfield TW Jr, George O. Addiction as a stress surfeit disorder. Neuropharmacology 76B: 370 –382, 2014. Koss WA, Belden CE, Hristov AD, Juraska JM. Dendritic remodeling in the adolescent medial prefrontal cortex and the basolateral amygdala of male and female rats. Synapse 68: 61–72, 2014. Larkman A, Mason A. Correlations between morphology and electrophysiology of pyramidal neurons in slices of rat visual cortex. I. Establishment of cell classes. J Neurosci 10: 1407–1414, 1990. LeDoux J. The emotional brain, fear, and the amygdala. Cell Mol Neurobiol 23: 727–738, 2003. LeDoux JE, Farb CR, Milner TA. Ultrastructure and synaptic associations of auditory thalamo-amygdala projections in the rat. Exp Brain Res 85: 577– 586, 1991. Mahan AL, Ressler KJ. Fear conditioning, synaptic plasticity and the amygdala: implications for posttraumatic stress disorder. Trends Neurosci 35: 24 –35, 2012. Mascagni F, McDonald AJ, Coleman JR. Corticoamygdaloid and corticocortical projections of the rat temporal cortex: a Phaseolus vulgaris leucoagglutinin study. Neuroscience 57: 697–715, 1993. McDonald AJ. Neurons of the lateral and basolateral amygdaloid nuclei: a Golgi study in the rat. J Comp Neurol 212: 293–312, 1982. McDonald AJ. Neuronal organization of the lateral and basolateral amygdaloid nuclei in the rat. J Comp Neurol 222: 589 – 606, 1984. McDonald AJ. Projection neurons of the basolateral amygdala: a correlative Golgi and retrograde tract tracing study. Brain Res Bull 28: 179 –185, 1992. McDonald AJ. Cortical pathways to the mammalian amygdala. Prog Neurobiol 55: 257–332, 1998.

McDonald AJ, Betette RL. Parvalbumin-containing neurons in the rat basolateral amygdala: morphology and co-localization of Calbindin-D(28k). Neuroscience 102: 413– 425, 2001. McDonald AJ, Mascagni F. Colocalization of calcium-binding proteins and GABA in neurons of the rat basolateral amygdala. Neuroscience 105: 681– 693, 2001. McDonald AJ, Mascagni F. Immunohistochemical characterization of somatostatin containing interneurons in the rat basolateral amygdala. Brain Res 943: 237–244, 2002. McDonald AJ, Mascagni F. Parvalbumin-containing interneurons in the basolateral amygdala express high levels of the alpha1 subunit of the GABAA receptor. J Comp Neurol 473: 137–146, 2004. McDonald AJ, Mascagni F, Mania I, Rainnie DG. Evidence for a perisomatic innervation of parvalbumin-containing interneurons by individual pyramidal cells in the basolateral amygdala. Brain Res 1035: 32– 40, 2005. McDonald AJ, Muller JF, Mascagni F. GABAergic innervation of alpha type II calcium/calmodulin-dependent protein kinase immunoreactive pyramidal neurons in the rat basolateral amygdala. J Comp Neurol 446: 199 – 218, 2002. McDonald AJ, Muller JF, Mascagni F. Postsynaptic targets of GABAergic basal forebrain projections to the basolateral amygdala. Neuroscience 183: 144 –159, 2011. McIntire SL, Reimer RJ, Schuske K, Edwards RH, Jorgensen EM. Identification and characterization of the vesicular GABA transporter. Nature 389: 870 – 876, 1997. Mendu SK, Bhandage A, Jin Z, Birnir B. Different subtypes of GABA-A receptors are expressed in human, mouse and rat T lymphocytes. PLoS One 7: e42959, 2012. Miles R, Toth K, Gulyas AI, Hajos N, Freund TF. Differences between somatic and dendritic inhibition in the hippocampus. Neuron 16: 815– 823, 1996. Muir J, Kittler JT. Plasticity of GABAA receptor diffusion dynamics at the axon initial segment. Front Cell Neurosci 8: 151, 2014. Muller JF, Mascagni F, McDonald AJ. Synaptic connections of distinct interneuronal subpopulations in the rat basolateral amygdalar nucleus. J Comp Neurol 456: 217–236, 2003. Muller JF, Mascagni F, McDonald AJ. Coupled networks of parvalbuminimmunoreactive interneurons in the rat basolateral amygdala. J Neurosci 25: 7366 –7376, 2005. Muller JF, Mascagni F, McDonald AJ. Pyramidal cells of the rat basolateral amygdala: synaptology and innervation by parvalbumin-immunoreactive interneurons. J Comp Neurol 494: 635– 650, 2006. Muller JF, Mascagni F, McDonald AJ. Postsynaptic targets of somatostatincontaining interneurons in the rat basolateral amygdala. J Comp Neurol 500: 513–529, 2007. Padival MA, Blume SR, Rosenkranz JA. Repeated restraint stress exerts different impact on structure of neurons in the lateral and basal nuclei of the amygdala. Neuroscience 246: 230 –242, 2013. Pape HC, Pare D. Plastic synaptic networks of the amygdala for the acquisition, expression, and extinction of conditioned fear. Physiol Rev 90: 419 – 463, 2010. Pare D, Collins DR, Pelletier JG. Amygdala oscillations and the consolidation of emotional memories. Trends Cogn Sci 6: 306 –314, 2002. Pare D, Gaudreau H. Projection cells and interneurons of the lateral and basolateral amygdala: distinct firing patterns and differential relation to theta and delta rhythms in conscious cats. J Neurosci 16: 3334 –3350, 1996. Pare D, Pape HC, Dong J. Bursting and oscillating neurons of the cat basolateral amygdaloid complex in vivo: electrophysiological properties and morphological features. J Neurophysiol 74: 1179 –1191, 1995. Persohn E, Malherbe P, Richards JG. Comparative molecular neuroanatomy of cloned GABAA receptor subunits in the rat CNS. J Comp Neurol 326: 193–216, 1992. Phelps EA, LeDoux JE. Contributions of the amygdala to emotion processing: from animal models to human behavior. Neuron 48: 175–187, 2005. Pitkänen A. Connectivity of the Rat Amygdaloid Complex. New York: Oxford Univ. Press, 2000. Rainnie DG, Asprodini EK, Shinnick-Gallagher P. Inhibitory transmission in the basolateral amygdala. J Neurophysiol 66: 999 –1009, 1991. Rainnie DG, Asprodini EK, Shinnick-Gallagher P. Intracellular recordings from morphologically identified neurons of the basolateral amygdala. J Neurophysiol 69: 1350 –1362, 1993. Rainnie DG, Mania I, Mascagni F, McDonald AJ. Physiological and morphological characterization of parvalbumin-containing interneurons of the rat basolateral amygdala. J Comp Neurol 498: 142–161, 2006.


GABAERGIC INPUTS ON BLA PRINCIPAL CELLS AND INTERNEURONS Robbins TW, Everitt BJ. Limbic-striatal memory systems and drug addiction. Neurobiol Learn Mem 78: 625– 636, 2002. Roozendaal B, McEwen BS, Chattarji S. Stress, memory and the amygdala. Nat Rev Neurosci 10: 423– 433, 2009. Ryan SJ, Ehrlich DE, Jasnow AM, Daftary S, Madsen TE, Rainnie DG. Spike-timing precision and neuronal synchrony are enhanced by an interaction between synaptic inhibition and membrane oscillations in the amygdala. PLoS One 7: e35320, 2012. Sah P, Faber ES, Lopez De Armentia M, Power J. The amygdaloid complex: anatomy and physiology. Physiol Rev 83: 803– 834, 2003. Senn V, Wolff SB, Herry C, Grenier F, Ehrlich I, Grundemann J, Fadok JP, Muller C, Letzkus JJ, Luthi A. Long-range connectivity defines behavioral specificity of amygdala neurons. Neuron 81: 428 – 437, 2014. Smith Y, Pare JF, Pare D. Differential innervation of parvalbumin-immunoreactive interneurons of the basolateral amygdaloid complex by cortical and intrinsic inputs. J Comp Neurol 416: 496 –508, 2000. Spampanato J, Polepalli J, Sah P. Interneurons in the basolateral amygdala. Neuropharmacology 60: 765–773, 2011. Stuber GD, Sparta DR, Stamatakis AM, van Leeuwen WA, Hardjoprajitno JE, Cho S, Tye KM, Kempadoo KA, Zhang F, Deisseroth K, Bonci A. Excitatory transmission from the amygdala to nucleus accumbens facilitates reward seeking. Nature 475: 377–380, 2011. Szinyei C, Narayanan RT, Pape HC. Plasticity of inhibitory synaptic network interactions in the lateral amygdala upon fear conditioning in mice. Eur J Neurosci 25: 1205–1211, 2007.

957

Tye KM, Prakash R, Kim SY, Fenno LE, Grosenick L, Zarabi H, Thompson KR, Gradinaru V, Ramakrishnan C, Deisseroth K. Amygdala circuitry mediating reversible and bidirectional control of anxiety. Nature 471: 358 –362, 2011. Vicini S, Ferguson C, Prybylowski K, Kralic J, Morrow AL, Homanics GE. GABAA receptor alpha1 subunit deletion prevents developmental changes of inhibitory synaptic currents in cerebellar neurons. J Neurosci 21: 3009 –3016, 2001. Vyas A, Mitra R, Shankaranarayana Rao BS, Chattarji S. Chronic stress induces contrasting patterns of dendritic remodeling in hippocampal and amygdaloid neurons. J Neurosci 22: 6810 – 6818, 2002. Washburn MS, Moises HC. Electrophysiological and morphological properties of rat basolateral amygdaloid neurons in vitro. J Neurosci 12: 4066 – 4079, 1992. Weller KL, Smith DA. Afferent connections to the bed nucleus of the stria terminalis. Brain Res 232: 255–270, 1982. Weston CS. Posttraumatic stress disorder: a theoretical model of the hyperarousal subtype. Front Psychiatry 5: 37, 2014. Wisden W, Laurie DJ, Monyer H, Seeburg PH. The distribution of 13 GABAA receptor subunit mRNAs in the rat brain. I. Telencephalon, diencephalon, mesencephalon. J Neurosci 12: 1040 –1062, 1992. Witkovsky P, Gabriel R, Krizaj D. Anatomical and neurochemical characterization of dopaminergic interplexiform processes in mouse and rat retinas. J Comp Neurol 510: 158 –174, 2008. Woodruff AR, Sah P. Networks of parvalbumin-positive interneurons in the basolateral amygdala. J Neurosci 27: 553–563, 2007.
