Immunohistochemical Identification of Interneuronal Subpopulations in the Basolateral Amygdala of the Rhesus Monkey (Macaca mulatta)

Abstract

Inhibitory circuits in the basolateral nuclear complex of the amygdala (BNC) critical for controlling the acquisition, expression, and extinction of emotional responses are mediated by GABAergic interneurons (INs). Studies in rodents have demonstrated that separate IN subpopulations, identified by their expression of calcium-binding proteins and neuropeptides, play discrete roles in the intrinsic circuitry of the BNC. Far less is known about IN subpopulations in primates. In order to fill in this gap in our understanding of primate INs, the present investigation used dual-labeling immunohistochemistry for IN markers to identify subpopulations expressing cholecystokinin (CCK), calbindin (CB), calretinin (CR), and somatostatin (SOM) in somata and axon terminals in the monkey BNC. In general, colocalization patterns seen in somata and axon terminals were similar. It was found that there was virtually no colocalization of CB and CR, the two calcium-binding proteins investigated. Three subtypes of CCK-immunoreactive (CCK+) INs were identified on the basis of their expression of CR or CB: (1) CCK+/CR+; (2) CCK+/CB+); and (3) CCK+/CR−/CB−. Almost no colocalization of CCK with SOM was observed, but there was extensive colocalization of SOM and CB. CCK+, CR+, and CCK+/CR+ double-labeled axon terminals were seen surrounding pyramidal cell somata in basket-like plexuses, as well as in the neuropil. CB+, SOM+, and CB+/SOM+ terminals did not form baskets, suggesting that these IN subpopulations are mainly dendrite-targeting neurons. In general, the IN subpopulations in the monkey are not dissimilar to those seen in rodents but, unlike rodents, CB+ INs in the monkey are not basket cells.

INTRODUCTION

Inhibitory circuits in the amygdala are critical for controlling the acquisition, expression, and extinction of emotional responses (Ehrlich et al., 2009). In the basolateral nuclear complex of the amygdala (BNC) these inhibitory mechanisms are provided by GABAergic interneurons (INs). Immunohistochemical studies in the rat suggest that the BNC contains at least four distinct subpopulations of INs that can be distinguished on the basis of their content of calcium-binding proteins and neuropeptides. These subpopulations are: (1) parvalbumin+/calbindin+ (PV+/CB+) neurons, (2) somatostatin+/− calbindin+ (SOM+/CB+) neurons, (3) large multipolar cholecystokinin+ (CCK+) neurons that are often CB+, and (4) small bipolar and bitufted neurons that exhibit extensive colocalization of calretinin (CR), CCK, and vasoactive intestinal peptide (VIP) (McDonald, 2020). Some of these IN classes exhibit several different subtypes. For example, a subtype of SOM+ neurons expresses neuropeptide Y (NPY) (McDonald, 1989; Urban, 2020), and subtypes of CCK+ INs express the CB1 cannabinoid receptor (CB1R) and/or the type 3 vesicular glutamate transporter protein (VGluT3) (McDonald and Mascagni, 2001b; Katona et al., 2001; Omiya et al., 2015). These separate IN subpopulations in rodents play discrete roles in the intrinsic circuitry of the BNC by innervating distinct compartments of pyramidal projection neurons (PNs) and/or other INs, and by having different inputs/receptors and electrophysiological properties (Ehrlich et al., 2009; Spampanato et al., 2011; Bienvenu et al., 2012; Capogna, 2014; Krabbe et al., 2018; Rovira-Esteban et al., 2017; Lucas and Clem, 2018; Beyeler and Dabrowska, 2020; McDonald, 2020).

Far less is known about IN subpopulations in the primate BNC. As in rodents virtually all INs in the monkey BNC are GABAergic (McDonald and Augustine, 1993; Pitkanen and Amaral, 1994). Golgi studies of the human and non-human primate BNC reported that INs were morphologically heterogeneous (Braak and Braak, 1983; McDonald and Augustine, 2020). Single-labeling immunohistochemical studies in primates indicate that the same calcium-binding proteins and neuropeptides expressed in rodent BNC INs are also expressed in primates. These include PV (Pitkänen and Amaral, 1993a; Sorvari et al., 1995), CB (Pitkänen and Amaral, 1993b; Sorvari et al., 1996a), CR (McDonald, 1994; Sorvari et al., 1996b), SOM (Amaral et al., 1989; Pantazopoulos et al., 2017), and CCK (McDonald and Mascagni, 2019). However, to the author’s knowledge there have been only four dual-labeling immunohistochemical studies examining coexpression of IN markers in the primate BNC; these investigations demonstrated coexpression of NPY in a subpopulation of SOM+ neurons in the monkey and human BNC (Schwartzberg et al., 1990; McDonald et al., 1995) and coexpression of PV and CB in a subpopulation of INs in the monkey and human BNC (Pantazopoulos et al., 2006; Mascagni et al., 2009). Because of the paucity of dual- or multi-labeling immunohistochemical studies of IN markers in the primate BNC, knowledge of BNC IN subpopulations is very limited in these species. In order to help fill in this gap in our understanding of primate IN subpopulations, the present study examined dual-localization of CCK, CB, CR, and SOM in somata and axon terminals in the monkey BNC. Knowledge of the structure and function of GABAergic INs in the primate BNC should lead to a better understanding of how disruption in GABAergic inhibition produces hyperexcitability and comcomitant increases in anxiety, emotional dysregulation, and seizure activity in humans (Prager et al., 2016).

EXPERIMENTAL PROCEDURES

Tissue preparation

These immunohistochemical experiments were performed on amygdalas obtained from three rhesus monkeys (Macaca mulatta) housed at Yerkes National Primate Research Center at Emory University, Atlanta, Georgia. Animals were deeply anesthetized with an overdose of sodium pentobarbital (100 mg/kg), and then perfused with one of two fixatives: (1) a mixture of 4% paraformaldehyde/0.2% glutaraldehyde/0.2% picric acid (Zamboni’s fixative) in phosphate buffer (0.1 M, pH 7.4; PB), or (2) 4% paraformaldehyde in PB (Table 1). Coronal blocks through the amygdala were cryoprotected in 35% sucrose in PB and frozen until processed for immunohistochemistry. The care of the animals and all anesthesia and euthanasia procedures in this study were performed in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals and were approved by the Institutional Animal Care and Use Committee of Emory University. All efforts were made to minimize animal suffering and to use the minimum number of animals necessary to produce reliable scientific data.

Table 1.

Macaques used in this study

Animal #	Species	Age, Sex	Fixative
M28	Rhesus	11.3 y.o., male	4% Paraformaldehyde
M102	Rhesus	2.8 y.o., male	Zamboni’s
M60	Rhesus	3.9, y.o., female	Zamboni’s

Blocks containing the amygdala were thawed and placed in descending concentrations of sucrose in PB (25%, 15%, 10%, 0%) over 2–3 days. The amygdalas were excised and 50 μm coronal sections were cut on a vibratome (Leica VT 1200; Nussloch, Germany) and collected in PB in wells of tissue culture chamber slides. Sections were incubated in 1% borohydride in PB for 30 min, and then rinsed thoroughly for 45–60 min in several changes of PB. A one-in-ten series of sections through each amygdala was processed for each dual-labeling immunohistochemical experiment (see below), which generated 6–7 equally spaced sections at 500 μm intervals through each amygdala. In addition, a one-in-ten series of sections through each amygdala was Nissl-stained (cresyl violet) so that borders of amygdalar nuclei could be recognized in neighboring sections processed for immunofluorescence, using blood vessels and fiber bundles as landmarks. Dual-labeling immunohistochemistry was performed in tissue culture chamber slides. Other amygdalar sections from these monkey brains were used for antibody controls (see below), antibody dilution tests, or for other studies.

Immunofluorescence experiments

Dual localization of CCK with CR, CB, or SOM.

The following antibodies were used for these studies: mouse anti-CCK (1:2000, antibody #9303 obtained from CURE-UCLA, Los Angeles, CA), rabbit anti-CR (1:4000, #7697, Swant Inc., Marly, Switzerland), rabbit anti-calbindin D-28K (1:8000, antiserum R-8701 obtained from Dr. Kenneth Baimbridge, University of British Columbia), and rabbit anti-somatostatin-28 (1:500, #T-4547, Peninsula Laboratories, San Carlos, CA) (Table 2). The diluent for all antibodies was a mixture of 3% normal goat serum, 1% bovine serum albumin, and 0.4% Triton X-100 in 0.05 M phosphate buffered saline (PBS, pH 7.4) Sections were incubated in a cocktail of two primary antibodies (either CCK/CR, CCK/CB or CCK/SOM) overnight at 4 °C, rinsed in three changes of PBS (10 min each), and then incubated in a cocktail of Alexa Fluor 546-labeled goat anti-mouse IgG (1:400, Thermo Fisher Scientific, West Columbia, SC) and Alexa Fluor 488-labeled goat anti-rabbit IgG (1:400; Thermo Fisher Scientific) for 3 h at room temperature. Secondary antibodies were highly cross-adsorbed by the manufacturer to insure specificity for primary antibodies raised in mouse or rabbit. Sections were then rinsed in PBS followed by two changes in PB (10 min each), mounted on glass slides, and coverslipped using Vectashield HardSet Antifade mounting medium (Vector Laboratories, Burlingame, CA).

Table 2.

Primary antibodies used in this study

Molecule	Species	Provider	Catalog #	RRID	Specificity	Dilution
CCK	Mouse	CURE, UCLA	9303	AB_2314185	McDonald and Mascagni, 2001a, 2001b	1:2000
CR	Rabbit	Swant	7697	AB_2619710	Manufacturer information	1:4000
CR	Rabbit	Chemicon	AB5054	AB_2068506	Manufacturer information	1:4000
CB	Rabbit	K. Baimbridge	R-8701	NA	Conde et al., 1994	1:8000
CB	Mouse	Swant	300	AB_10000347	Manufacturer information	1:4000
SOM	Rabbit	Peninsula	T-4547	AB_518618	Manufacturer information	1:500

Dual localization of CB with CR or SOM.

The following antibodies were used for these studies: mouse anti-CB (1:8000, #300, Swant, Inc.), rabbit anti-CR (1:4000, #7697, Swant Inc.), and rabbit anti-somatostatin-28 (1:500; #T-4547, Peninsula Laboratories) (Table 2). Sections were incubated in a cocktail of primary antibodies (either CB/CR or CB/SOM) overnight at 4 °C, then incubated in a cocktail of Alexa Fluor 546-labeled goat anti-mouse IgG (1:400, Thermo Fisher Scientific, West Columbia, SC) and Alexa Fluor 488-labeled goat anti-rabbit IgG (1:400; Thermo Fisher Scientific) for 3 h at room temperature. Sections were coverslipped as described above.

Immunoperoxidase staining for CR.

Also available for analysis were sections used in a previous study of CR+ structures in the monkey BNC performed in this laboratory. These sections were processed for immunoperoxidase staining using a rabbit anti-CR antibody (1:4000; Chemicon) with nickel-enhanced DAB as a chromogen (see McDonald, 1994 for details).

Analysis

Sections processed for immunofluorescence were examined with a Zeiss LSM 510 Meta confocal microscope. Fluorescence of Alexa Fluor-488 and Alexa Fluor-546 dyes was analyzed using filter configurations for sequential excitation/imaging via 488-nm and 543-nm channels. Nuclei of the macaque amygdala were identified based on the description of Price and coworkers (1987) (Fig. 1). Counts of single- and double-labeled somata were performed in the two younger animals (M102 and M60) in two major subdivisions of the BNC, the magnocellular subdivision of the basal nucleus (Bmg) and the dorsomedial subdivision of the lateral nucleus (Ldm), in the middle rostrocaudal three-fifths of the amygdala (i.e., at levels of Fig. 1C–E of Price et al., 1987) (Fig. 1). These two animals were chosen because the BNC of the oldest animal (M28) had light sporadic nonspecific fluorescence, perhaps caused by the presence of lipofuschin, that was also seen in sections that were not immunostained. The Bmg was easily identified because it is bounded medially and laterally by components of the longitudinal association fiber bundle (Price et al., 1987) (Fig. 1). The Ldm is separated from the Bmg by the lateral component of the longitudinal association fiber bundle. These fiber bundles, as well as myelinated zones dorsal to the BNC, contain neurons of the intercalated masses (IMs; Zikopoulos et al., 2016). Although the IMs contain CB+ and CR+ neurons (Zikopoulos et al., 2016), the pattern of staining for these calcium-binding proteins, as well as the staining of neuropeptide-containing axons, in the IMs was distinct from that seen in the BNC nuclei (see Figure 10C of Zikopoulos et al., 2016). As a result the borders of the BNC nuclei with the IMs were readily discernable. All fields sampled for cell counts in the Bmg and Ldm were located within the borders of the Ldm and Bmg, and did not overlap the IMs.

Fig. 1.

Nissl-stained coronal section through the amygdala of the macaque. Cross indicates orientation (D, dorsal; V, ventral; M, medial; L lateral). The level of this section is midway along the rostrocaudal extent of the amygdala and corresponds to section 201 of BrainMaps coronally-sectioned macaque brain (Data set ID 155; sections containing the amygdala are numbered 184–218 from rostral to caudal). See Crosby and Humphrey (1944) for a discussion of how the expansion of the temporal lobe in primates, which rotates the amygdala, reconfigures the spatial relationships of amygdalar nuclei in primates versus rodents and other non-primates. Asterisks indicate two components of the longitudinal association bundle. Note that the Ldm is separated from the Bmg by the lateral component of the longitudinal association fiber bundle. Nuclei of the monkey amygdala: ABmg, magnocellular accessary basal nucleus; ABpc, parvicellular accessary basal nucleus; ABs, superficial accessary basal nucleus; Bmg, magnocellular basal nucleus; Bpc, parvicellular basal nucleus; CL, lateral central nucleus; CM, medial central nucleus; Ldm, dorsomedial lateral nucleus; Lvl, ventrolateral lateral nucleus; M, medial nucleus; PAC, periamygdaloid cortex. Scale bar = 700 μm.

Counts of single- and double-labeled somata at 200× magnification were made from images of separate, randomly-selected 400 × 400 μm fields obtained using merged (red/green) channels (double-labeled cells appear yellow). Images of the non-merged red and green channels were also obtained. Some PNs in the monkey BNC exhibited low levels of CB or CR. However, as in rat (McDonald and Mascagni, 2001a, 2001b), PNs were easily distinguished from the intensely stained non-pyramidal INs at the antibody dilutions used in this study. Only somata of INs were counted. This analysis of dual-localization of IN markers in somata are considered semi-quantitative since stereological methods were not used.

Analyses of single- and double-labeled axon terminals at 630× magnification were made from images of 140 × 140 μm fields displayed for merged (red/green) channels (double-labeled cells appear yellow). Images of the non-merged red and green channels were also obtained.

Antibody specificity

All of the primary antibodies used in this study have been used in previous studies of the brain, and their specificity has been documented in control studies by investigators or manufacturers (Table 2).

To examine method specificity in the present study control sections were processed with one of the two primary antibodies omitted. In all cases only the color of the corresponding secondary fluorescent antibody was observed, and only on the appropriate channel. These results indicate that the secondary antibodies were specific for rabbit or mouse IgGs, and that there was no “crosstalk” between the red and green channels (Wouterlood et al., 1998).

RESULTS

General observations

As in previous studies in primates, each of the marker-labeled IN subpopulations was found to be fairly evenly distributed throughout the nuclei of the BNC (Amaral et al., 1989; Pitkänen and Amaral, 1993b; McDonald et al., 1995; Sorvari et al, 1996b; McDonald and Mascagni, 2019). Likewise, each of the double-labeled subpopulations in this study appeared to be evenly distributed. All portions of the macaque BNC in all three animals were examined for qualitative analysis of colocalization of interneuronal markers. The same marker colocalization combinations were seen in the BNC in all three animals. Counts of single- and double-labeled somata in Bmg and Ldm were performed in brains M102 and M60. The counts for each colocalization combination were similar in both brains in each of the two nuclei examined (see Tables 3–5). These analyses revealed that although similar patterns of colocalization were seen in all BNC nuclei, the percentages of interneuronal somata that exhibited colocalization could be different in different nuclei.

Table 3.

Colocalization of CCK and CR in the basal magnocellular (Bmg), and dorsomedial lateral (Ldm) nuclei

Nucleus	Brain	CCK Single labeled	CR Single labeled	CCK-CR Double labeled	% CCK Double labeled	% CR Double labeled
Bmg	M102	24	56	9	27.3% (9/33)	13.8% (9/65)
	M60	21	36	9	30.0% (9/30)	20.0% (9/45)
	Combined	45	92	18	28.6% (18/63)	16.4% (18/110)
Ldm	M102	16	88	19	54.3% (19/35)	17.8% (19/107)
	M60	16	90	17	51.5% (17/33)	15.9% (17/107)
	Combined	32	178	36	53.0% (36/68)	16.8% (36/214)

Table 5.

Colocalization of CB and SOM in the basal magnocellular (Bmg), and dorsomedial lateral (Ldm) nuclei

Nucleus	Brain	SOM Single labeled	CB Single labeled	SOM-CB Double labeled	% SOM Double labeled	% CB Double labeled
Bmg	M102	23	27	15	39.5% (15/38)	35.7% (15/42)
	M60	25	29	21	45.7% (21/46)	42.0% (21/50)
	Combined	48	56	36	42.9% (36/84)	39.1% (36/92)
Ldm	M102	17	22	15	46.9% (15/32)	40.5% (15/37)
	M60	19	20	19	50.0% (19/38)	48.7% (19/39)
	Combined	36	42	34	48.6% (34/70)	44.7% (34/76)

Dual localization of CCK and calcium-binding proteins in BNC somata

CCK+ INs were found in low density in the BNC and were outnumbered by CR+ INs. Most somata of CCK+ INs were about 13–15 μm in diameter (medium-sized), but smaller somata (10 μm in diameter) were also seen (Fig. 2). The somata of medium-sized INs were typically angular/multipolar whereas the somata of small neurons were spherical, ovoid or fusiform. Most somata of CR+ INs were small (10 μm in diameter) and spherical or ovoid (Fig. 2). However, medium-sized (13–15 μm in diameter) and large (20 μm in diameter) CR+ somata were also seen. Both small (Fig. 2A, B) and medium-sized INs were double-labeled (CCK+/CR+). In the Bmg CCK+/CR+ double-labeled neurons constituted 28.6% of CCK+ INs and 16.4 % of CR+ INs (Table 3). In the Ldm CCK+/CR+ double-labeled neurons constituted 53.0% of CCK+ INs and 16.8% of CR+ INs (Table 3).

Fig. 2.

Photomicrographs showing dual-localization of CCK, CR, and CB. (A, B) CCK+ neurons (red) and CR+ neurons (green) in the Ldm (A) and Bmg (B). Yellow indicates colocalization of CCK and CR. Arrows indicate CCK+/CR+ double-labeled INs. (C, D) CCK+ neurons (red) and CB+ neurons (green) in the Ldm (C) and Bmg (D). Unlabeled arrows indicate CCK+/CR+ double-labeled INs. In each field single-labeled CB+ INs (IN) are indicated by arrows. The remaining single-labeled CB+ neurons in each field are larger and more lightly-stained; they are presumptive pyramidal cells. (E, F) CB+ neurons (red) and CR+ neurons (green) in the Ldm (E) and Bmg (F). Note absence of colocalization. Scale bar in (A) = 20 μm. ((B-D) are at the same magnification). Scale bar in (F) = 20 μm. ((E) is at the same magnification).

Small, medium-sized, and large CB+ interneuronal somata were observed in the BNC; their morphologies resembled those of CR+ INs. In the Bmg CCK+/CB+ double-labeled neurons constituted 12.9% of CCK+ INs and 12.7% of CB+ INs (Table 4). In the Ldm CCK+/CB+ double-labeled neurons constituted 31.1% of CCK+ INs and 28.8% of CB+ INs (Table 4). Both small and medium-sized somata were double-labeled (Fig. 2C, D).

Table 4.

Colocalization of CCK and CB in the basal magnocellular (Bmg), and dorsomedial lateral (Ldm) nuclei

Nucleus	Brain	CCK Single labeled	CB Single labeled	CCK-CB Double labeled	% CCK Double labeled	% CB Double labeled
Bmg	M102	29	24	4	12.1% (4/33)	14.3% (4/28)
	M60	32	38	5	13.5% (5/37)	11.6% (5/43)
	Combined	61	62	9	12.9% (9/70)	12.7% (9/71)
Ldm	M102	21	27	9	30.0% (9/30)	25.0% (9/36)
	M60	21	20	10	32.2% (10/31)	33.3% (10/30)
	Combined	42	47	19	31.1% (19/61)	28.8% (19/66)

Dual localization of CR and CB revealed that there were almost no CR+/CB+ double-labeled INs in the BNC (Fig. 2E, F). Counts of somata in the Bmg in brains M102 and M60 found a total of 1 CR+/CB+ double-labeled IN, 124 CR+ single-labeled INs, and 61 CB+ single-labeled INs. Counts of somata in the Ldm in brains M102 and M60 found a total of 2 CR+/CB+ double-labeled INs, 114 CR+ single-labeled INs, and 71 CB+ single-labeled INs.

Colocalization of SOM with CB and CCK in BNC somata

Like CR+ and CB+ INs, SOM+ INs in the BNC were morphologically heterogeneous. Most had medium-sized fusiform or oval somata that averaged about 15 μm in diameter, although both smaller neurons (10 μm in diameter) and larger neurons (20 μm in diameter) were also seen. owever there was extThere was extensive expression of SOM in CB+ INs (Figs. 3, 4A, 4B). In the Bmg SOM+/CB+ double-labeled neurons constituted 42.9% of SOM+ neurons and 39.1% of CB+ INs (Table 5). In the Ldm SOM+/CB+ double-labeled neurons constituted 48.6% of SOM+ neurons and 44.7% of CB+ INs (Table 5). Small, medium and large somata were double-labeled.

Fig. 3.

Low power photomontage of CB+ neurons (red), SOM+ neurons (green), and double-labeled CB+/SOM+ neurons (yellow) in the Bmg. Double-labeled neurons are indicated by arrows. Lateral is left, medial is right, dorsal is up, and ventral is down. Scale bar = 50 μm.

Fig. 4.

Photomicrographs showing dual-localization of SOM with CB and CCK. (A, B) CB+ neurons (red) and SOM+ neurons (green) in the Ldm (A) and Bmg (B). Yellow indicates colocalization of CB and SOM. (C, D) CCK+ neurons (red) and SOM+ neurons (green) in the Ldm (C) and Bmg (D). Note lack of double-labeled neurons. Some SOM+ somata are contacted by several CCK+ punctae (arrows; see also Fig. 4). Scale bar in B = 20 μm. ((A) is at the same magnification). Scale bar in (C) = 50 μm. ((D) is at the same magnification).

Dual localization of CCK and SOM revealed that there were very few CCK+/SOM+ double-labeled INs in the BNC (Fig. 4C, D). Counts of somata in the Bmg in brains M102 and M60 found a total of 3 CCK+/SOM+ double-labeled INs, 61 CCK+ single-labeled INs, and 56 SOM+ single-labeled INs. Counts of somata in the Ldm in brains M102 and M60 found a total of 2 CCK+/SOM+ double-labeled INs, 28 CCK+ single-labeled INs, and 60 SOM+ single-labeled INs.

Colocalization of interneuronal markers in axon terminals

In sections stained for CCK, CR, or SOM the neuropil in all BNC nuclei contained many small round or oval punctae that were the size (0.5–2.0 μm) and shape of axonal terminals. A moderate number of double-labeled punctae was seen in the neuropil in sections stained for CCK and CR (Fig. 6A, B). CCK+, CR+, and CCK+/CR+ punctae were also observed contributing to basket-like plexuses in contact with presumptive pyramidal cell somata (Fig. 6E). CR+ punctae forming contacts with somata were also observed in immunoperoxidase preparations (Fig. 7). Consistent with the analysis of somata, very few double-labeled punctae were seen in sections stained for CCK and SOM (Fig. 6C, D). However, about 30% of SOM+ somata in the BNC were contacted by at least 3 CCK+ punctae (Figs. 4C, D, 5). Because no third marker specific for axon terminals was used in this study, no attempt to count single- and double-labeled axon terminals in CCK/CR and CCK/SOM preparations was attempted.

Fig. 6.

Photomicrographs of immunostained punctae in the BNC. (A and B) CCK+ punctae (red) and CR+ punctae (green) in the neuropil of Ldm (A) and Bmg (B). Yellow indicates colocalization of CCK and CR. Note that there are a moderate number of yellow CCK+/CR+ punctae that average about 1.0–2.0 μm in diameter (arrows). These resemble axon terminals. (C and D) CCK+ punctae (red) and SOM+ punctae (green) in the Ldm (C) and Bmg (D). Note that there are no yellow CCK+/SOM+ punctae in the Ldm field and very few (arrows) in the Bmg field. (E) CCK+ punctae (red) and CR+ punctae (green) in Bmg. Note three presumptive pyramidal neuronal somata (PN) receive basket-like contacts from several CCK+ punctae and a few CR+ punctae. All three PN somata are also contacted by some yellow CCK+/CR+ punctae (arrows). The upper two PN somata exhibit light CCK immunoreactivity consistent with previous studies demonstrating that PNs express CCK mRNA (Marsicano and Lutz, 1999) and very low levels of CCK peptide (Mascagni and McDonald, 2003). (F) A CB+ “bundle” in the parvicellular subdivision of the accessory basal nucleus. (G) A simpler CB+ “bundle” in the parvicellular subdivision of the accessory basal nucleus, consisting of a single strand of CB+ punctae, one of which also expresses SOM (arrow). Scale bar in (D) = 5 μm. ((A-C) are at the same magnification). Scale bar in (E) = 5 μm ((F), (G) are at the same magnification).

Fig. 7.

Photomicrographs of CR+ structures (black) in Ldm (A) and Bmg (B) in an immunoperoxidase-stained section using nickel-enhanced DAB as a chromogen. Somata are stained light pink using pyronin Y as a Nissl counterstain. Asterisks indicate Nissl-stained somata that appear to be contacted by basket-like plexuses of CR+ punctae. These CR− negative somata are larger than those of most CR+ neurons (see lower left corner of (B)) and probably represent pyramidal neurons. Note size difference of these presumptive pyramidal neurons in Ldm versus Bmg. Scale bar = 20 μm for (A), (B).

Fig. 5.

Photomicrographs of CCK+ punctae (red) contacting SOM+ somata (green) in the BNC. (A) A SOM+ soma is contacted by 4–5 CCK+ punctae. (B) One SOM+ soma (upper right) is not contacted by any CCK+ punctae, but the other SOM+ soma and proximal dendrite (lower left) is contacted by 5–6 CCK+ punctae. (C) A SOM+ soma and proximal dendrite is contacted by 7–8 CCK+ punctae. Scale bar = 10 μm for (A)-(C).

CB+ punctae were extremely scarce in the lateral nucleus and Bmg in sections stained with both the rabbit or mouse CB antibodies, but numerous in the accessory basal nucleus (AB) and parvicellular subdivision of the basal nucleus (Bpc) in sections stained with the mouse monoclonal CB antibody. Some of these CB+ punctae were in the form of linear or curvilinear clusters that appear to correspond to the CB+ “bundles” observed in the monkey and human AB and Bpc (Pitkänen and Amaral, 1993b; Sorvari et al., 1996a) (Fig. 6F, G). Some CB+ punctae in bundles were also SOM+ (Fig. 6G). Bundles were not seen in sections stained with the rabbit polyclonal CB antibody.

DISCUSSION

This is the first dual-labeling immunohistochemical study to identify interneuronal subpopulations in the monkey BNC on the basis of coexpression, or lack thereof, of CCK/CR, CCK/CB, CB/CR, SOM/CCK, and SOM/CB. Fig. 8 depicts the findings of the present study along with results of a previous study of the expression of PV in the monkey BNC (Mascagni et al., 2009). The same interneuronal subpopulations are present in both the Bmg and Ldm, but the relative number of INs in the separate subpopulations differ. CR+ and CB+ INs are basically two separate subpopulations. There are three types of CCK+ INs: (1) CCK+/CR+; (2) CCK+/CB+; and (3) CCK+/CR−/CB−. CCK+ INs are distinct from SOM+ and PV+ INs (Mascagni et al., 2009). Collectively, CB+ neurons that express CCK, SOM, or PV makeup the great majority of CB+ INs. All of these subpopulations exhibited morphological heterogeneity in regards to somatic size and shape. Because of variability in the size distributions of somata in each of the interneuronal populations, the percentages of colocalization obtained in this non-stereological study are approximations of the actual percentages that exist in each nucleus. For the most part, colocalization patterns seen in axon terminals match those seen in somata.

Fig. 8.

Venn diagrams illustrating the relative sizes of interneuronal subpopulations, and their degree of overlap, in the monkey Bmg and Ldm. The data on PV+ neurons, some of which exhibit coexpression of CB, but not CCK or SOM, is taken from a previous study (Mascagni et al., 2009). The present study did not test colocalization of SOM and CR; the separate subpopulations depicted are based on studies in rat, which have IN subpopulations in the BNC that are similar to those in monkey, but which exhibit no colocalization of CR and SOM (McDonald and Mascagni, 2002).

Dual-localization of CCK and calcium-binding proteins

CR+ and CB+ interneuronal somata exhibited almost no colocalization in the monkey BNC, indicating that these two calcium-binding proteins are in separate interneuronal subpopulations, just like in the rat BNC (McDonald and Mascagni, 2001a). Three subtypes of CCK+ INs were observed in the monkey BNC: (1) CCK+/CR+; (2) CCK+/CB+; and (3) CCK+/CR−/CB−. These three subtypes were also seen in the rat (Mascagni and McDonald, 2003), and in both species a greater degree of colocalization of CCK with CB was seen in the lateral nucleus versus the basal nucleus. There are two major subpopulations of CCK+ INs in the rat BNC that can be identified on the basis of morphological differences. Large CCK INs (CCK_L) have somata that average 15–20 μm in diameter, whereas small CCK INs (CCK_S) have somata that average 10 μm in diameter (Mascagni and McDonald, 2003). These two CCK+ IN subtypes in the rat express different calcium-binding proteins. Some CCK_L neurons express CB but none express CR, whereas some CCK_S neurons express CR but none express CB (Mascagni and McDonald, 2003). This size distinction was not seen in the monkey BNC in the present study; subpopulations of both small and medium-sized CCK+ INs expressed CR or CB.

Given colocalization of CCK and CR in a subpopulation of IN somata, it was not surprising to observe small CCK+/CR+ punctae, most likely representing axon terminals, in the BNC. These axon terminals, as well as single-labeled CCK+ and CR+ terminals, were seen in the neuropil as well as contributing to the formation of basket-like plexuses surrounding large somata that resemble those of PNs. This suggests that some CCK+ INs, CR+ INs, and CCK+/CR+ INs in the monkey BNC are basket cells. Baskets were also observed in a previous single-labeling study of CCK+ INs in the monkey BNC (McDonald and Mascagni, 2019), but were not described in previous studies of CR+ INs in the monkey (McDonald, 1994) or human BNC (Sorvari et al., 1996b). Our previous immunoperoxidase study of CR+ INs in the monkey BNC focused on their somatodendritic morphology (McDonald, 1994). As part of the present study we reexamined this material and found that PN somata in Nissl counter-stained sections of the monkey BNC were innervated by CR+ axons that form multiple axosomatic contacts, thus corroborating the immunofluorescence findings of the present study. However, no CR+ baskets were seen in the human BNC, which suggests that CR+ axons mainly innervate dendrites in humans (Sorvari et al., 1996b).

While it seems likely that most if not all of the CCK+, CCK+/CR+ terminals originate from interneurons, there could be other sources of the CR+ and CB+ terminals seen in the BNC. Some of the CR+ and CB+ terminals could originate from the lightly-labeled CR+ and CB+ BNC PNs seen in this study, or from the midline and intralaminar thalamic nuclei, which contain many CR+ and CB+ neurons in primates (Fortin et al., 1996; Münkle et al., 2000), and have robust projections to the monkey BNC (Mehler, 1980). The intercalated masses in the monkey also contain CR+, CB+, and CR+/CB+ neurons (Zikopoulos et al., 2016), and studies in rodents have shown that a neuronal subpopulation in the intercalated masses innervates interneurons in the BNC (Bienvenu et al., 2015).

Besides CCK+ and CCK+/CR+ basket cells, the other major basket cell subpopulation in the BNC of rodents, monkeys, and humans, are PV+ INs (Pitkänen and Amaral, 1993a; Sorvari et al., 1995, 1996c; Rainnie et al., 2006; Vereczki et al., 2016; McDonald, 2020). The placement of these basket cell inhibitory synapses close to the axon initial segment of PNs, the main projection neurons of the BNC, allows basket cells to tightly regulate the timing of spike firing of PN outputs to other brain regions (Veres et al., 2017). In the mouse BNC it has been found that both PV basket cells and CCK basket cells provide roughly equal regulation of PN firing (Veres et al., 2017). It is well established that networks of distinct subtypes of PV+ neurons in the rodent BNC, including basket cells, are interconnected by GABAergic synapses and electrical synapses (Muller et al., 2005; Woodruff and Sah, 2007a; Andrási et al., 2017). Moreover, a recent investigation found evidence that basket cells expressing CCK in the mouse BNC are also interconnected by both GABAergic synapses and electrical synapses, and that these CCK basket cell networks are distinct from PV+ basket cell networks (Andrási et al., 2017). These same networks may also exist in the monkey since synapses of PV+ axon terminals with PV+ INs (Pitkänen and Amaral, 1993a), as well as CCK+ axon terminals with CCK+ INs (McDonald and Mascagni, 2019), have been observed in the BNC. It remains to be determined the extent to which CR+ INs in the monkey BNC form IN networks. IN networks in the cortex and BNC are important for their contribution to rhythmic oscillations in these regions (Buzsáki and Chrobak, 1995; Whittington and Traub, 2003; Bartos et al., 2007; Andrási et al., 2017; Feng et al., 2019). These IN network rhythms in the BNC are critical for synaptic plasticity associated with the formation, retrieval, and extinction of emotional memories (Paré et al., 2002; Pape and Pare, 2010; Bocchio et al., 2017; Davis et al., 2017).

Almost no CB+ puncta were seen in the lateral nucleus and Bmg of the monkey in the present study. The failure to stain CB+ axon terminals in these nuclei in the present study suggests that either CB is not found in axon terminals in these nuclei, or is at levels that are below detection by immunofluorescence. However, many CB+ punctae were seen in the accessory basal nucleus (AB) and parvicellular basal nucleus (Bpc), where some were clustered in “bundles”. Similar findings were reported in the monkey BNC with a sensitive immunoperoxidase technique using the same monoclonal CB antibody used in this study (Pitkänen and Amaral, 1993b). Consistent with the results of the present study they reported that in most portions of the BNC “it was difficult to identify calbindin-immunoreactive fibers and terminals.” However CB+ “bundles” were seen in AB and Bpc in both the monkey and human BNC (Pitkänen and Amaral, 1993b; Sorvari et al., 1996a). In these previous light microscopic studies, as well as in the present study, it was not possible to determine if the bundles were composed of axons, thin beaded distal dendrites, or both. These BNC bundles closely resemble the longer bundles of CB+ processes in the cerebral cortex that are associated with a particular type of cortical IN termed a double-bouquet cell (DeFelipe et al., 1989). Electron microscopic analysis of cortical bundles revealed that they were composed of both CB+ beaded dendrites as well as CB+ axons that form symmetrical (inhibitory) synapses with dendritic shafts and spines of PNs (DeFelipe et al., 1989).

Previous dual-localization studies in monkey indicate that about 25–33% of PV+ INs in the BNC express CB, and none express SOM or CCK (Mascagni et al., 2009; Fig. 8). In both rodents and primates many PV+ INs are axo-axonic (chandelier) cells that do not co-express CB or neuropeptides (Pitkänen and Amaral, 1993a; Sorvari et al., 1995; Muller et al., 2006; Veres et al., 2014). Axo-axonic cells have also been seen in Golgi studies of the monkey BNC (McDonald and Augustine, 2020). As discussed above, a separate subpopulation of PV+ INs are basket cells in rodents and primates (Pitkänen and Amaral, 1993a; Sorvari et al., 1995; Muller et al., 2006; Rainnie et al., 2006; Vereczki et al., 2016). PV+ basket cells in rodents also express CB (Vereczki et al., 2016; Veres et al., 2017). However, the failure to see CB+ baskets in the BNC of the monkey and human amygdala suggests that PV+ basket cells in primates do not express CB (Pitkänen and Amaral, 1993a; Sorvari et al., 1995; present study). These data suggest that all CB+ INs in the primate BNC (including PV+/CB+, SOM+/CB+, and CCK/CB+ INs) are dendrite-targeting INs.

Somatostatin-immunoreactive neurons

The results of the present study indicate that SOM+ neurons in the monkey BNC are distinct from CCK+ INs and constitute almost half of CB+ INs in both the Bmg and Ldm. Collectively, SOM+, CCK+, and PV+ INs (Mascagni et al., 2009) comprise the great majority of CB+ INs in both nuclei, but are found in three separate subpopulations. Possible colocalization of SOM with CR was not tested in the present study, but studies in the rat have shown no coexpression (McDonald and Mascagni, 2002). As in a previous study in the monkey (Amaral et al., 1989), SOM+ punctae in the present investigation were found in the neuropil but did not form basket-like plexuses around PNs, which suggests that in addition to SOM+/CB+ neurons, single-labeled SOM+ interneurons also mainly innervate dendrites, as in rodents (Muller et al., 2007; Wolff et al., 2014). Many SOM+ INs, but not other IN subpopulations, coexpress NPY in both the rat, monkey, and human BNC (McDonald, 1989; Schwartzberg et al., 1990; McDonald et al., 1993; Urban, 2020).

Functional considerations

This investigation, and previous studies of the primate BNC, demonstrate that there are several separate IN subpopulations in the monkey BNC. As in rodents, these subpopulations undoubtedly play discrete roles in the intrinsic circuitry of the BNC by innervating distinct PN domains and/or other INs, and by having different inputs/receptors and electrophysiological properties (Ehrlich et al., 2009; Spampanato et al., 2011; Bienvenu et al., 2012; Capogna, 2014; Krabbe et al., 2018; Rovira-Esteban et al., 2017; Lucas and Clem, 2018; Beyeler and Dabrowska, 2020; McDonald, 2020). One of the differences among distinct IN subpopulations is their innervation of different PN compartments. Some innervate the perisomatic compartment of PNs, including somata and proximal dendrites (basket cells), or axon initial segments (axo-axonic cells), while others appear to mainly innervate dendrites. Differential targeting of perisomatic versus dendritic compartments of pyramidal cells by distinct subpopulations of primate BNC interneurons should have profound functional significance (Freund and Buzsaki, 1996; Miles et al., 1996). CCK+, CR+, and PV+ basket cells and PV+ axo-axonic cells in the monkey BNC are in a position to regulate pyramidal cell firing via their perisomatic innervation (Woodruff and Sah, 2007a, 2007b; Veres et al., 2014, 2017; Barsy et al., 2017; Andrási et al., 2017). This is critical for the timing of synchronous rhythmic oscillations involved in the acquisition, expression, and extinction of emotional responses in the BNC (Paré et al., 2002; Pape and Pare, 2010; Bocchio et al., 2017; Davis et al., 2017; Morozov and Ito, 2020). In contrast, dendritic inhibition provided by various subpopulations of CB+ and SOM+ INs in the monkey BNC could regulate synaptic plasticity by shunting excitatory inputs to dendrites, suppressing the generation of calcium-dependent action potentials in dendrites, and/or preventing back-propagation of action potentials from somatic to dendritic compartments (Miles et al., 1996; Stuart et al., 1997; Klausberger, 2009; Wolff et al., 2014; Müllner et al., 2015). The finding in the present study that many SOM+ somata are innervated by CCK+ axon terminals suggests that some CCK+ INs may be involved in a disynaptic disinhibition of PN dendrites. Similarly, fear conditioning in the mouse requires inhibition of BNC SOM+ neurons by PV+ INs (Wolff et al., 2014; Letzkus et al., 2015).