Immunofluorescently labeling glutamic acid decarboxylase 65 coupled with confocal imaging for identifying GABAergic somata in the rat dentate gyrus—A comparison with labeling glutamic acid decarboxylase 67

Journal of Chemical Neuroanatomy 61–62 (2014) 51–63

Contents lists available at ScienceDirect

Journal of Chemical Neuroanatomy journal homepage: www.elsevier.com/locate/jchemneu

Immunofluorescently labeling glutamic acid decarboxylase 65 coupled with confocal imaging for identifying GABAergic somata in the rat dentate gyrus—A comparison with labeling glutamic acid decarboxylase 67 Xiaochen Wang a, Fei Gao a, Jianchun Zhu a, Enpu Guo b, Xueying Song a, Shuanglian Wang a, Ren-Zhi Zhan a,* a b

Department of Physiology, Shandong University School of Medicine, Jinan, China Division of General Surgery, The Second Affiliated Hospital, Shandong University of Traditional Chinese Medicine, China

A R T I C L E I N F O

A B S T R A C T

Article history: Received 22 January 2014 Received in revised form 11 July 2014 Accepted 12 July 2014 Available online 21 July 2014

As g-aminobutyric acid (GABA) is synthesized by two isoforms of glutamic acid decarboxylase (GAD), namely, GAD65 and GAD67, immunohistochemically targeting either isoform of GAD is theoretically useful for identifying GABAergic cell bodies. In practice, targeting GAD67 remains to be a popular choice. However, identifying GABAergic cell bodies with GAD67 immunoreactivity in the hippocampal dentate gyrus, especially in the hilus, is not without pitfalls. In the present study, we compared the characteristics of GAD65 immunoreactivity to GAD67 immunoreactivity in the rat dentate gyrus and examined perikaryal expression of GAD65 in four neurochemically prevalent subgroups of interneurons in the hilus. Experiments were done in normal adult Sprague-Dawley rats and GAD67-GFP knock-in mice. Horizontal hippocampal slices cut from the ventral portion of hippocampi were immunofluorescently stained and scanned using a confocal microscope. Immunoreactivity for both GAD67 and GAD65 was visible throughout the dentate gyrus. Perikaryal GAD67 immunoreactivity was denser but variable in terms of distribution pattern and intensity among cells whereas perikaryal GAD65 immunoreactivity displayed similar distribution pattern and staining intensity. Among different layers of the dentate gyrus, GAD67 immunoreactivity was densest in the hilus despite GAD65 immunoreactivity being more intense in the granule cell layer. Co-localization experiments showed that GAD65, but not GAD67, was expressed in all hilar calretinin (CR)-, neuronal nitric oxide synthase (nNOS)-, parvalbumin (PV)- or somatostatin (SOM)positive somata. Labeling CR, nNOS, PV, and SOM in sections obtained from GAD67-GFP knock-in mice revealed that a large portion of SOM-positive cells had weak GFP expression. In addition, double labeling of GAD65/GABA and GAD67/GABA showed that nearly all of GABA-immunoreactive cells had perikaryal GAD65 expression whereas more than one-tenth of GABA-immunoreactive cells lacked perikaryal GAD67 immunoreactivity. Inhibition of axonal transport with colchicine dramatically improved perikaryal GAD65 immunoreactivity in GABAergic cells without significant augmentation to be seen in granule cells. Double labeling GAD65 and GAD67 in the sections obtained from colchicine-pretreated animals confirmed that a portion of GAD65-immunoreactive cells had weak or even no GAD67 immunoreactivity. We conclude that for confocal imaging, immunofluorescently labeling GAD65 for identifying GABAergic somata in the hilus of the dentate gyrus has advantages over labeling GAD67 in terms of easier recognition of perikaryal labeling and more consistent expression in GABAergic somata. Inhibition of axonal transport with colchicine further improves perikaryal GAD65 labeling, making GABAergic cells more distinguishable. ß 2014 Elsevier B.V. All rights reserved.

Keywords: g-aminobutyric acid Calcium-binding proteins GAD67-GFP knock-in Hilus Hippocampus Neuronal nitric oxide synthase Neuropeptides

Abbreviations: GABA, g-aminobutyric acid; CR, calretinin; GAD, glutamic acid decarboxylase; nNOS, neuronal nitric oxide synthase; NPY, neuropeptide Y; PV, parvalbumin; SOM, somatostatin. * Corresponding author at: Department of Physiology, Shandong University School of Medicine, 44 Wenhua Xi Road, Jinan, Shandong 250012, China. Tel.: +1 86 15066152368. E-mail address: [email protected] (R.-Z. Zhan). http://dx.doi.org/10.1016/j.jchemneu.2014.07.002 0891-0618/ß 2014 Elsevier B.V. All rights reserved.

52

X. Wang et al. / Journal of Chemical Neuroanatomy 61–62 (2014) 51–63

Introduction The hippocampal dentate gyrus acts as a main gateway in processing information from the entorhinal cortex to the hippocampus (Freund and Buzsa´ki, 1996; Amaral et al., 2007). In addition to granule cells, the dentate gyrus also contains mossy cells and interneurons. Though interneurons only constitute a minority of the neuronal population, they function to prevent runaway excitation and contribute to hippocampus-dependent learning and memory (Andrews-Zwilling et al., 2012; Spiegel et al., 2013). Specifically, interneurons in the dentate gyrus may play important roles in sparse coding (Ewell and Jones, 2010; Morellini et al., 2010; Li et al., 2012; Marı´n-Burgin et al., 2012; Piatti et al., 2013; Yu et al., 2013). Most interneurons in the dentate gyrus are GABAergic (Houser, 2007). GABA is synthesized from glutamate by two isoforms of glutamic acid decarboxylase (GAD), namely glutamic acid decarboxylase 65 (GAD65) and glutamic acid decarboxylase 67 (GAD67). GAD67 is present throughout the cytosol, whereas GAD65 preferentially localizes to axonal terminals (Erlander et al., 1991; Esclapez et al., 1994). In addition to providing GABA for metabolic needs, GABA catalyzed by GAD67 also enters into synaptic vesicles, contributing to inhibitory neurotransmission, mainly under resting conditions. On the other hand, GABA catalyzed by GAD65 becomes more critical under intense neuronal activity (Tian et al., 1999; Patel et al., 2006) and may be responsible for tonic inhibition (Walls et al., 2010). Among the two isoforms of GAD, most studies have chosen to target GAD67 instead of GAD65 for identifying GABAergic cell bodies in different brain regions (Mu¨ller et al., 2001; Shetty and Turner, 2001; Dı´az-Cintra et al., 2007; Shetty et al., 2009). However, multiple problems exist in using GAD67 immunoreactivity to identify GABAergic somata in the dentate gyrus. First, due to dense perisomatic distribution of GAD67 in most nerve cells, it is hard to distinguish cytoplasmic staining from perisomatic staining in a portion of these cells. The overly robust GAD67 immunoreactivity in the hilus further makes isolation of GAD67-immunoreactive somata difficult (Liang et al., 2013). Additionally, it appears that the mossy fiber expression of GAD67 is both development- and activity-dependent (Schwarzer and Sperk, 1995; Maqueda et al., 2003; Sperk et al., 2012). Finally, some GABAergic somata, for example, a portion of somatostatinpositive cells, may lack GAD67 expression (Fukuda et al., 1998). Although in situ hybridization studies have shown that the signals of GAD65 mRNA in most hilar GABAergic somata are strong (Houser and Esclapez, 1994, 1996; Cze´h et al., 2013), immunohistochemically labeling GAD65 has rarely been used for determining the GABAergicity of cells. Since the use of confocal laser scanning microscope coupled with bright fluorophores has become routine in many laboratories, it is thus reasonable to examine the distribution pattern and the intensity of GAD65 immunoreactivity and to test if immunofluorescently labeling GAD65 would be useful in identifying GABAergic interneurons in terms of easy recognition and consistency. In the present study, we compared the characteristics of GAD65 immunoreactivity to these of GAD67 immunoreactivity in the rat dentate gyrus by examining perikaryal expression of GAD65 in four neurochemically prevalent subgroups of interneurons in the hilus and assessing whether all GABAergic cell bodies have visible GAD65 immunoreactivity. Experimental procedures Animals Animal experiments were approved by the Animal Ethics Committee of Shandong University School of Medicine and performed according to the guidelines for the care and use of laboratory animals set by the National Institutes of Health (Committee for the update of the guide for the care and use of laboratory animals, 2011). Experiments were mainly performed in 17 normal adult male Sprague-Dawley rats (400  50 g,

12–13 weeks old) which were purchased from Shandong University of Traditional Chinese Medicine (Jinan, Shandong, China) and left in a university animal facility for at least one week with a 12 h light/dark cycle and free access to food and water before the start of experiments. In addition to rats, 3 adult heterozygous glutamic acid decarboxylase 67-green fluorescent protein (GAD67-GFP) knock-in mice originally developed by Tamamaki et al. (2003) were used to verify some results obtained from rats. Primary antibody information Detailed information of primary antibodies used in the present study is shown in Table 1. The specificities of a mouse anti-GAD67 (EMD Millipore, Billerica, MA, USA) and a mouse anti-GAD65 (BD Biosciences, San Jose, CA, USA) were confirmed by immunoblot as shown below. The specificity of a rabbit anti-GAD65 (Sigma, St. Louis, MO, USA) was confirmed by comparing the stain pattern following its application to that yielded by application of the mouse anti-GAD65. The specificity of a rabbit anti-nNOS (Sigma, St. Louis, MO, USA) has been confirmed by immunoblot in a previous study (Liang et al., 2013). Antibodies directed against CR, GABA, PV, and SOM have been characterized previously (Ka¨gi et al., 1987; Sun et al., 2007; Mascagni et al., 2009; Gentet et al., 2010). Controls in which the primary antibodies were omitted resulted in no stains. Testing the specificities of mouse anti-GAD67 and mouse anti-GAD65 with immunoblots We performed immunoblot on the rat hippocampal protein extract to verify the specificities of the mouse anti-GAD67 and the mouse anti-GAD65. The procedure for extracting hippocampal proteins has been described elsewhere (Liang et al., 2013). Cytosolic protein samples were loaded onto 8% SDS-polyacrylamide gels and separated by electrophoresis at 200 V for 2 h. Proteins on the gels were then transferred onto nitrocellulose membranes (Bio-Rad, Berkeley, CA, USA) at 120 V for 2 h. Blots were incubated with the mouse anti-GAD67 (1:2500) or mouse antiGAD65 (1:1000) overnight at 4 8C, respectively. After the primary antibody incubations, membranes were washed and then incubated with a horseradish peroxidase-conjugated goat anti-mouse IgG (Comwin Biotech, Beijing, China) for 60 min at room temperature. Immunoreactions were visualized by using an enhanced chemiluminescent horseradish peroxidase substrate (SuperSignal West Pico, Thermo Scientific, Rockford, IL, USA). The experiment was repeated in protein samples extracted from two animals. Intraventricular injection of colchicine Because intraventricular injection of colchicine has been shown to improve immunoreactivity for both GAD67 and GABA (Ko¨hler and Chan-Palay, 1983; Kumoi et al., 1987), 5 rats were given intraventricular injection of colchicine with a stereotaxic technique to determine if inhibition of axonal transport could also enhance GAD65 immunoreactivity and the co-localization of GAD65 and GAD67. Two symmetric injecting sites were located 3.8 mm from the bregma, 5 mm apart from the midline with a depth of 4 mm. The dose injected to each lateral ventricle was 100 mg (dissolved in 5 ml saline). Transcardiac perfusion was done 24 h after colchicine injection. Brain sections were used for GAD65/GABA, GAD67/GABA or GAD65/GAD67 double immunofluorescence. Transcardiac perfusion and tissue preparation Each rat was deeply anesthetized with an intraperitoneal injection of sodium pentobarbital (60 mg/kg) and then transcardially perfused with heparinized saline (20–30 ml) followed by 300–400 ml of 4% paraformaldehyde in phosphate-buffered saline (PBS, 0.1 M, pH 6.8) over a period of 40 min. For mice, approximately 10 ml of heparinized saline and 50 ml of 4% paraformaldehyde in 0.1 M PBS were sequentially perfused. After the completion of transcardiac perfusion, the brain was removed and a tissue block containing hippocampi was made with the aid of a brain matrix in the appropriate size (Alto, World Precision Instruments, Sarasota, FL, USA). The tissue block was post-fixed with the same fixative overnight at 4 8C. Thereafter, the tissue block was sequentially immersed in 10%, 15%, and 20% sucrose (in 0.1 M PBS) for 4 h, 8 h and overnight at 4 8C, respectively. After each tissue block was firmly embedded with a medium that consisted of 30% (w/v) chicken egg albumin (Sigma, St. Louis, MO, USA), 0.5% (w/v) gelatin, and 0.9% (v/v) glutaraldehyde in 0.1 M PBS as described previously (Zhan and Nadler, 2009), it was serially cut into 60 mm-sections in the horizontal plane throughout the dorsoventral extent of the hippocampal formation with a Vibratome 1000 (The Vibratome CO., St. Louis, MO, USA). Because interneurons are more densely distributed in the ventral dentate gyrus, sections that contained the C-shaped dentate gyri were sequentially collected in 0.1 M PBS and used for immunofluorescence. Double immunofluorescence Every five adjacent sections from one hemisphere were double-stained with anti-GAD67/anti-CR, anti-GAD67/anti-GABA, anti-GAD67/anti-nNOS, anti-GAD67/ anti-PV, and anti-GAD67/anti-SOM, whereas every five adjacent sections from

X. Wang et al. / Journal of Chemical Neuroanatomy 61–62 (2014) 51–63

53

Table 1 Information of primary antibodies. Antibody

Clonality/species

Antigen

Manufacturer

Catalog no.

Lot no.

Dilution

Anti-calretinin (CR)

Polyclonal/rabbit

EMD Millipore

AB5054

NG1780667

1:4000

Anti-g-Aminobutyric acid (GABA) Anti-GAD65 (mouse)

Polyclonal/rabbit

Recombinant rat calretinin GABA-BSA

Sigma–Aldrich

A2052

112M4768

1:1000

Purified rat GAD65

BD Biosciences

BDB559931

28139

1:1000

Synthetic peptide corresponding to the C-terminal region of human GAD 65 (amino acids 514–530) Recombinant GAD67

Sigma–Aldrich

G4913

061M4898

1:8000

EMD Millipore

MAB5406

2019961

1:2500

Sigma–Aldrich

N7280

060M4779

1:4000

Swant Penisula Lab

PV25 T-4103

1637 A09211

1:6000 1:4000

Anti-GAD65 (rabbit)

Anti-GAD67 Anti-neuronal nitric oxide synthase (nNOS)

Anti-parvalbumin (PV) Anti-somatostatin (SOM)

Monoclonal (IgG2a)/mouse Polyclonal/rabbit

Monoclonal (IgG2a)/mouse Polyclonal/rabbit

Polyclonal/rabbit Polyclonal/rabbit

Synthetic peptide corresponding to nNOS of rat brain (1409–1429) conjugated to KLH Rat muscle parvalbumin Synthetic somatostatin-14

another hemisphere were double-stained with anti-GAD65/anti-CR, anti-GAD65/ anti-GABA, anti-GAD65/anti-nNOS, anti-GAD65/anti-PV, and anti-GAD65/anti-SOM, respectively. For each antibody pair, three sections were stained from one animal. The procedure for double immunofluorescence in free-floating sections has been described previously (Liang et al., 2013). Briefly, the selected sections were first washed with 0.1 M PBS and then incubated in a blocking buffer that consisted of 5% normal donkey serum, 2.5% bovine serum albumin (BSA), and 0.2% Triton X-100 in 0.1 M PBS for 2.5 h at 4 8C. The sections were then subjected to different pairs of primary antibody incubation overnight at 4 8C. After washing with 0.1 M PBS for 3 times (15 min each), sections were incubated with a mixed Alexa Flour 488conjugated donkey anti-mouse IgG and Alexa Flour 568-conjugated donkey antirabbit IgG (1:600 each, Invitrogen, San Diego, CA, USA) diluted with the blocking buffer for 2.5 h at 4 8C. After rinsing in 0.1 M PBS for 3 times (15 min each), sections were mounted on glass slides and cover-slipped with 75% (v/v) glycerol in 0.1 M PBS. Confocal imaging All immunofluorescently stained sections were scanned with a Carl Zeiss Laser Scanning Microscopy (LSM 780, Jena, Germany). After the dentate gyrus was scanned under a 10 objective, three areas in the hilus adjacent to the granule cell layer/hilus border where at least one CR-, GABA-, nNOS-, PV-, or SOM-positive cell presented were randomly chosen to undergo scanning under a 63 oil objective. The scan zooms under 10 and 63 objectives were set at 0.6 and 1.0, respectively. In sections prepared from colchicine-treated rats, scanning was also done with a 20 objective (zoom = 1). Quantitative analysis of GAD immunofluorescence To quantify relative fluorescence intensity in the molecular layer, granule cell layer, hilus, and the stratum lucidum of the CA3 region, individual images scanned under the 10 objective were converted to 8-bit grayscale images (0–256 range) and mean fluorescence intensities in the molecular layer, granule cell layer, hilus, and stratum lucidum (SL) of the CA3 region of each section were measured by using ImageJ (http://rsb.info.nih.gov/ij/). Fluorescence intensities of the molecular layer, the hilus, and the stratum lucidum of the CA3 region of each section were normalized by the fluorescence intensity in the granule cell layer. Measures of each rat were performed in 8 sections, and values in different sections were averaged and used to represent the animal. In total, 5 animals were used for this analysis.

(ANOVA) or unpaired t-test as appropriate. Percentages of GABAergic somata expressing GAD67 or GAD65 obtained were treated as a single data point. Normal distributions of all data sets were confirmed by Kolmogorov–Smirnov test. p-Values less than or equal to 0.05 are considered to be statistically significant.

Results Specificities of mouse anti-GAD67 and mouse anti-GAD65 Three approaches, including immunoblotting analysis, comparing immunoreactivity in gene-knockout tissue to wild-type tissue, and looking for identical localization of immunoreactivity following two or more antibody applications that are raised against different epitopes of an interest protein, have been recommended for determining the specificity of antibodies to be used for immunohistochemistry (Fritschy, 2008; Lorincz and Nusser, 2008). Because the mouse anti-GAD65 used in the present study has not been widely applied, we performed immunoblot to test its specificity and compared it to that of mouse anti-GAD67. As shown in Fig. 1, similar to the application of mouse anti-GAD67 (left two lanes), applications of mouse anti-GAD65 (right two lanes) to rat hippocampal proteins resulted in single bands at 65-kDa, confirming that the specificity of mouse anti-GAD65 is as good as that of mouse anti-GAD67.

Determining the expression of GAD67 or GAD65 in CR-, GABA-, nNOS-, PV-, and SOMpositive cells To quantitatively analyze the percentage of CR-, GABA-, nNOS-, PV-, or SOMpositive somata that expressed GAD65 or GAD67, the number of cell bodies positive for individual interneuron markers with or without GAD expression in all three scanned hilar areas of each section using the 63 objective were counted. The number of CR-, GABA-, nNOS-, PV-, or SOM-positive somata and the number of corresponding cell bodies that expressed GAD67 or GAD65 in each rat were counted from at least three sections (9 areas). Statistical analysis Quantitative data expressed as mean  standard error of the mean (SEM) were statistically compared by Dunnett test following one-way analysis of variance

Fig. 1. The specificities of mouse anti-GAD67 and mouse anti-GAD65 tested by immunoblot. Rat hippocampus proteins separated by electrophoresis were transferred onto nitrocellulose membranes. Note that applications of both mouse anti-GAD67 (the left two lanes) and mouse anti-GAD65 (the right two lanes) yielded single bands which were at near 67 and 65 kDa, respectively. In lanes 1 and 2, the amounts of proteins loaded were 5 and 10 mg, respectively.

54

X. Wang et al. / Journal of Chemical Neuroanatomy 61–62 (2014) 51–63

Characteristics of GAD67 and GAD65 immunoreactivity in the rat dentate gyrus Immunoreactivity for both GAD67 and GAD65 was visible throughout the dentate gyrus (Fig. 2). At a lower magnification, scattered somata, extensive neural processes, and bouton-like structures appeared to be stained for both GAD67 and GAD65 (Fig. 2A and B). Most GAD67-positive somata appeared very bright; they can be readily identified in the molecular layer and granule cell layer, but defining those in the hilus remains challenging. In the granule cell layer, dense perisomatic GAD67 immunoreactivity was seen in all granule cells (Fig. 2A and C). Across the molecular layer, the intensity of GAD67 immunoreactivity appeared to be moderate and even. GAD67 immunoreactivity in the hilus looked denser than those in the molecular layer and granule cell layer. The dense GAD67 immunoreactivity in the hilus further extended into the stratum lucidum of the CA3 region (Fig. 2A), indicating that the dense hilar GAD67 immunoreactivity has a mossy fiber origin. At higher magnification, the intensities of perikaryal GAD67 immunoreactivity (arrows) appeared variable, ranging from very strong to hardly distinguishable from surrounding stains (Fig. 2C). GAD65 immunoreactivity in the somata was generally weaker than that of GAD67 immunoreactivity; however, perikaryal GAD65 immunoreactivity in many cells was ready to be identified even at a lower magnification. Different from the distribution of GAD67 immunoreactivity, GAD65 immunoreactivity in the hilus appeared to be less dense than that of molecular layer (Fig. 2B). Within the molecular layer, the densest stain of GAD65 was seen in the outer one-third of the molecular layer (Fig. 2B). At a higher magnification, perikaryal

GAD65 immunoreactivity appeared to be characteristically diffuse, tiny, and smooth granules and looked highly identical. Perisomatic GAD65 immunoreactivity in the granule cell layer formed puncta ring-like structures which were generally more smooth than perisomatic GAD67 staining (Fig. 2D). Averaged fluorescence intensities (FI) in the molecular layer, the hilus, and the stratum lucidum of the CA3 region normalized by the value in the granule cell layer for GAD67 and GAD65 are shown in Fig. 2E and F, respectively. Clearly, according to dentate layers, the densest staining for GAD67 was at the hilus whereas the densest staining for GAD65 was seen in the granule cell layer. Perikaryal expression of GAD67 in hilar CR-, nNOS-, PV- or SOMpositive cells Neurochemically, the most prevalent subgroups of interneurons in the rodent dentate gyrus are cells that express SOM, nNOS, PV, neuropeptide Y, or CR (Jinno and Kosaka, 2006; Houser, 2007; Liang et al., 2013). We thus attempted to determine to what percentages, hilar CR-, nNOS-, PV-, and SOM-positive somata express GAD67 and if the distribution pattern and the intensity of GAD67 immunoreactivity are similar within individual subgroups. The representative images are shown in Fig. 3, whereas the quantitative data are presented in Table 2. Two important features are noticeable. First, the distribution pattern and the intensity of perikaryal GAD67 immunoreactivity were not only variable among subgroups but different within the same subgroup also. Among the four subgroups, nNOS- and PV-positive cells expressed GAD67 more densely and consistently, whereas most CR-positive

Fig. 2. Characteristics of GAD67 and GAD65 immunoreactivity in the rat dentate gyrus. (A and B) Images in lower magnification (10) show the distribution of GAD67 (A) and GAD65 (B) immunoreactivity in different layers of dentate gyrus. Note that although GAD65-immunoreactive somata are stained weaker than those of GAD67-immunoreactive somata, most of them are recognizable. ML = molecular layer; GCL = granule cell layer; SL = stratum lucidum of the CA3 region. (C and D) Areas boxed in ‘‘A’’ and ‘‘B’’ were scanned under a higher magnification to show perikaryal GAD67 (C) and GAD65 (D) immunoreactivity, respectively. The perikaryal GAD67 immunoreactivity indicated by arrows in the hilus varies among cells in terms of distribution pattern and intensity. The perikaryal GAD65 immunoreactivity is nearly constant and appears easier to be isolated from the surrounding stains. (E and F) Bar graphs show relative fluorescence intensity (FI) of GAD67 (E) and GAD65 (F) among three layers of the dentate gyrus and the stratum lucidum (SL) of the CA3 region. Fluorescence intensities in the molecular layer (ML), the hilus and the SL of the CA3 were normalized by the value in the granule cell layer (GCL) each section. Bars representing 5 rats are mean  SEM. *p < 0.01, as compared to that of granule cell layer.

X. Wang et al. / Journal of Chemical Neuroanatomy 61–62 (2014) 51–63

cell bodies had only lighter perikaryal GAD67 immunoreactivity. SOM-positive cell bodies showed the most variable perikaryal GAD67 immunoreactivity to the extent that there were SOMpositive cells that showed no detectable perikaryal GAD67 immunoreactivity at all. With 66 and 87 cells randomly sampled,

55

respectively, the percentages of GAD67 expression in nNOS- and PV-positive somata are both 100%. CR-positive cell bodies expressing GAD67 were at 95.4% (83/87) whereas only 86.8% (118/136) of SOM-positive somata had visible perikaryal GAD67 immunoreactivity.

Fig. 3. Representative confocal images show perikaryal expressions of GAD67 in hilar calretinin (CR)-, neuronal nitric oxide synthase (nNOS)-, parvalbumin (PV)-, or somatostatin (SOM)-positive cells. Sections were double-stained with one anti-interneuron marker (left column) and anti-GAD67 (middle column) and scanned by a confocal microscopy using a 63 objective with scan zoom 1. Merged images are shown in the right column. The portion of granule cell layer is in the right side each panel. Arrows indicate cells that are positive for an interneuron marker and GAD67. Arrowheads indicate GAD67 immunoreactive cells that lack expression of targeted interneuron markers. The purple arrowhead indicates a SOM-positive cell lacking GAD67 immunoreactivity. Bar (20 mm) applies to all image panels. Note that the expression of GAD67 varies considerably among different subgroups of cells; nNOS- and PV-positive cells expressed GAD67 consistently and densely (middle rows); CR-positive cells expressed GAD67 at lower levels (upper row); SOM-positive cells expressed GAD67 inconsistently (bottom row).

X. Wang et al. / Journal of Chemical Neuroanatomy 61–62 (2014) 51–63

56

Table 2 Percentages of interneuron marker-positive cells with GAD67 or GAD65 immunoreactivity (IR). Marker

CR nNOS PV SOM

GAD67 IR

GAD65 IR

Number of cells

Number of cells with GAD67 IR

% of cells with GAD67 IR

Number of cells

Number of cells with GAD65 IR

% of cells with GAD65 IR

87 66 87 136

83 66 87 118

95.4 100 100 86.8

82 102 80 90

82 102 80 90

100 100 100 100

A large portion of hilar SOM-positive cells in GAD67-GFP knock-in mice express GAD67-GFP weakly

A fraction of GABA-immunoreactive cell bodies lacks GAD67 immunoreactivity

To verify the GABAergicity of CR-, nNOS-, PV-, and SOM-positive cells observed in the rat dentate gyrus, CR, nNOS, PV, and SOM were individually stained in sections obtained from GAD67-GFP knock-in mice. Representative images and semiquantitative data are given in Fig. 4 and Table 3, respectively. Different from the results observed in the rat dentate gyrus, almost no mouse CRpositive cells had GFP expression (Fig. 4, upper row; Table 3), indicating that hilar CR-positive cells are not GABAergic in mice. Based on their large somata and inner molecular layer projections, hilar CR-positive cells in mice are likely to be mossy cells. More than 95% of nNOS-positive cells and all PV-positive cells had moderate to dense GFP expression (Fig. 4, middle two rows; Table 3). For SOM-positive cells, it appeared that every SOMpositive cell expressed GFP. However, the features of GAD67-GFP in SOM-positive cells closely resembled GAD67 immunofluorescence in variability and intensity observed in the rat; 43% of SOMpositive cells were found to have weak GFP expression (Fig. 4, Table 3).

To determine if all GABAergic somata express GAD67 and GAD65, double stain for GABA/GAD67 and GABA/GAD65 was conducted. Representative images are shown in Fig. 6 and statistical results are shown in Fig. 7. Notably, nearly all (more than 96.2% with all obtained from 3 rats polling together) of the GABA-immunoreactive somata had visible perikaryal GAD65 expression, whereas only 88.1% (92/104) of GABA-immunoreactive cell bodies were positive for GAD67. Data collected from 3 rats clearly showed that the percentage of GABAergic somata with perikaryal GAD65 expression exceeded the percentage of GABAergic somata with perikaryal GAD67 expression, indicating that a certain number of GABAergic somata lacked GAD67 protein in the cytoplasm.

Perikaryal GAD65 expression in hilar CR-, nNOS-, PV-, or SOM-positive cells GAD65 immunoreactivity in representative CR-, nNOS-, PV-, or SOM-positive cell bodies is shown in Fig. 5. Three general features of perikaryal GAD65 expression in the four subgroups of cells have been noticed. First, for any randomly sampled area where one or more cells were positive for one of the aforementioned markers, the number of GAD65 immunoreactive somata is greater than or equal to the number of cell bodies positive for one of those interneuron markers. Another striking feature of perikaryal GAD65 expression is that, without exception, all CR- (n = 82), nNOS(n = 102), PV- (n = 80), and SOM-positive cell bodies (n = 90) counted were found to have visible perikaryal GAD65 immunoreactivity (Table 2). In addition, the distribution pattern and intensity of perikaryal GAD65 immunoreactivity in CR-, nNOS-, PV-, or SOMpositive somata appeared consistent and clearly isolatable from the surrounding stains.

Inhibition of axonal transport dramatically improves GAD65 labeling As shown in Fig. 8, intraventricular injection of colchicine improved perikaryal stains not only for GABA and GAD67 but also for GAD65. Enhancements in GAD67 immunoreactivity were seen in the soma and neurites of both GABAergic cells and granule cells (Fig. 8, upper row). In contrast, the increased GAD65 immunoreactivity was barely seen in granule cells but was robust in the soma and neurites of GABAergic cells (bottom row), leaving GABAergic cells in the granule cell layer more isolatable. Co-labeling of GAD65 and GAD67 shows a lack of GAD67 immunoreactivity in a portion of GAD65 immunoreactive somata Before co-labeling GAD65 and GAD67, we compared the staining pattern yielded by the rabbit anti-GAD65 to that of mouse antiGAD65 in sections obtained from colchicine-treated animals (Fig. 9). Under the visualization of a lower magnification, applications of the two GAD65 antibodies resulted in nearly identical immunoreactive patterns in the dentate gyrus; overlapping positive immunoreactions could be seen in both the soma and the processes (Fig. 9, upper row). Under a higher magnification (63), all somata stained by the mouse anti-GAD65 were found to be labeled by the rabbit anti-GAD65.

Table 3 Intensity of GAD67-GFP expression in hilar cells positive for different interneuron markers in GAD67-GFP knock-in mice. Marker

Number of marker-positive cells

Number of cells with moderate or intense GFP expression

Cells with weak GFP expression

% cells with weak GFP expression

% cells with no visible GFP expression

CR nNOS PV SOM

130 66 47 74

0 63 47 44

0 3 0 30

0 4.8 0 42.8

100 0 0 0

The intensity of GFP expression each cell was graded into very intense (5+), intense (4+), moderate (3+), weak (2+), barely visible (1+) and invisible (0+) categories. Moderate to intense GFP expression: cells with GFP fluorescence equal 3+ or above; Weak GFP expression: 1+ or 2+; No GFP expression: 0+.

X. Wang et al. / Journal of Chemical Neuroanatomy 61–62 (2014) 51–63

57

Fig. 4. Representative confocal images show GAD67-GFP expressions in hilar calretinin (CR)-, neuronal nitric oxide synthase (nNOS)-, parvalbumin (PV)-, or somatostatin (SOM)-positive cells. Sections obtained from GAD67-GFP mice were stained with one anti-interneuron marker (left column) and scanned by a confocal microscopy using a 63 objective with a scan zoom 1. Columns in the middle and right show GAD67-GFP fluorescence and the merged images, respectively. Note that none of hilar CR-positive cells has GFP expression (upper row). Arrows indicate that interneuron marker-positive cells express with moderate or intense GFP expression. The arrowhead indicates a SOM-positive cell with weak but visible GFP expression. Bar (20 mm) applies to all image panels.

In sections prepared from colchicine-untreated animals, perikaryal immunoreactivity in the rabbit anti-GAD65-labeled sections seemed less intense than that in the mouse anti-GAD65-stained ones; however, a portion of GAD65-immunoreactive somata lacking GAD67 immunoreactivity was noticed (Fig. 10, the upper row).

Because of the lighter GAD65 immunoreactivity, a statistical analysis on the overlapping of GAD65 and GAD67 was not performed in colchicine-untreated animals. Double stains done in sections obtained from colchicine-treated animals clearly showed that some GAD65-positive somata lacked visible GAD67 immunoreactivity

58

X. Wang et al. / Journal of Chemical Neuroanatomy 61–62 (2014) 51–63

Fig. 5. Representative confocal images show perikaryal expressions of GAD65 in hilar calretinin (CR)-, neuronal nitric oxide synthase (nNOS)-, parvalbumin (PV)-, or somatostatin (SOM)-positive cells. Sections were double-stained with one anti-interneuron marker (the left column) and the mouse anti-GAD65 (the middle column) and scanned by a confocal microscopy using a 63 objective and scan zoom 1. Merged images are shown in the right column. The portion of granule cell layer is in the right side of each panel. Arrows indicate cells that are positive for one of the interneuron markers and GAD65. Arrowheads indicate GAD65 immunoreactive cells that lack expression of targeted interneuron markers. Note that all CR-, nNOS-, PV-, and SOM-expressing cells are GAD65-immunoreactive. Bar (20 mm) applies to all panels.

(Fig. 10, the lower row) whereas the reverse was not seen. In a randomly collected 80 somata that had perikaryal GAD65 immunoreactivity, 7 cells (9.1%) did not have visible perikaryal GAD67 expression, further confirming that a certain number of GABAergic somata have little GAD67 protein in the cytoplasm.

Discussion Features of GAD65 immunoreactivity in the rat dentate gyrus are different from those of GAD67 immunoreactivity in several aspects. First, unlike GAD67, perikaryal labeling of GAD65 presents a

X. Wang et al. / Journal of Chemical Neuroanatomy 61–62 (2014) 51–63

59

Fig. 6. Representative images show the co-existences of GABA and GAD65 or GAD67 in the hilus of the dentate gyrus. Upper row: Despite the GABA immunoreactivity in a cell indicated by the arrowhead appears faint, all GABA-immunoreactive cells have perikaryal GAD65 expression. Bottom row: All cells arrowed contain GAD 67 except the cell indicated by arrowhead. Bar (20 mm) applies to all image panels.

consistent distribution pattern and similar intensity, making isolation of GAD65 immunoreactive somata easier. Second, the mossy fiber expression of GAD65 is much less in comparison with GAD67. Third, GAD65 but not GAD67 is exclusively expressed in hilar cells positive for CR, nNOS, PV, or SOM. Fourth, approximately onetenth of GABAergic somata have little GAD67 protein. In addition, unlike inhibition of axonal transport-mediated enhancement in GAD67 immunoreactivity, which is seen in the soma and neurites of both GABAergic cells and granule cells, augmentation of GAD65 immunoreactivity appeared to be limited to GABAergic cells. Up-to-date, three approaches, including immunohistochemistry, in situ hybridization, and transgenic animals, have been used for identification of GABAergic cells. There are advantages and disadvantages for either approach. The main problem with the application of anti-GAD antibodies is the weak stain in the somata

Fig. 7. A comparison of percentages of GABAergic cells that express GAD65 or GAD67. The percentage GABAergic cells expressing GAD65 and GAD67 were obtained from 3 rats; 9 areas from 3 sections of each rat were scanned under a 63 objective with a zoom of 1. Bars are mean  SEM. Statistical analysis was done with unpaired t-test. The asterisk indicates p < 0.05.

of a certain portion of cells. Furthermore, anti-GABA antibodies may label non-neuronal cells in addition to the faint stain. To overcome faint stains, efforts have been made to develop better antibodies (Oertel et al., 1981), use more sophisticated fixatives (Wolff et al., 1984; Ha¨rtig et al., 2001), inhibit axonal transport with colchicine (Ko¨hler and Chan-Palay, 1983; Kumoi et al., 1987), and apply confocal imaging for visualization. However, the uses of colchicine are limited because colchicine is neurotoxic (Sutula et al., 1983) and it has to be given hours or days before sacrificing the animals. Over the past decade, transgenic animals knocking in GAD67-GFP or GAD65-GFP have been favored for the identification of GABAergic cells (Tamamaki et al., 2003; Wierenga et al., 2010). Using transgenic animals has numerous advantages over immunohistochemistry or in situ hybridization. However, knowledge obtained from transgenic mice cannot be guaranteed applicable to other species. As shown in Fig. 4, the GABAergicity of hilar CRpositive cells in the rat dentate gyrus is clearly different from that in mice. The dentate gyrus has long been known to contain the higher density of GABAergic neurons within the hippocampal formation (Ho¨rtnagl et al., 1991). Even so, immunohistochemically identifying and counting all GABAergic cell bodies in the dentate gyrus, especially the hilar region, remains challenging (Houser, 2007; Cze´h et al., 2013). Although immunohistochemically targeting GAD67 remains popular (Mu¨ller et al., 2001; Shetty and Turner, 2001; Dı´az-Cintra et al., 2007; Shetty et al., 2009), pitfalls exist in using perikaryal GAD67 immunoreactivity to determine the GABAergicity of a given cell and account for all GABAergic cells in the dentate gyrus. In general, it is unclear if cytoplasmic synthesis of the GAD67 protein takes place in all GABAergic cells. An earlier study has shown that in cells positive for somatostatin, the levels of somatostatin contrasted the lowest GABA and GAD content in the rat

60

X. Wang et al. / Journal of Chemical Neuroanatomy 61–62 (2014) 51–63

Fig. 8. Inhibition of axonal transport with colchicine enhances GABA, GAD67 and GAD65 immunoreactivity. Colchicine was injected into the lateral ventricles 24 h before transcardiac perfusion with 4% paraformaldehyde. Sections were scanned using a 20 objective with a scan zoom 1. Upper row: GABA/GAD67 double immunofluorescence; Bottom row: GABA/GAD65 double immunofluorescence. Bar (100 mm) applies to all image panels.

hippocampal CA1 region (Ho¨rtnagl et al., 1991). Even at the mRNA level, not all somatostatin-positive cells were found to have GAD67 expression (Esclapez and Houser, 1995). Co-localization of GAD67 and individual interneuron markers in the present study revealed

that both the intensity and distribution pattern of GAD67 immunoreactivity are not only variable among subgroups of interneurons but also different within the same subgroup. A particularly notable subgroup is the somatostain-positive cells. In

Fig. 9. Staining patterns yielded by applications of mouse anti-GAD65 and rabbit anti-GAD65 in sections obtained from a colchicine-treated rat. Colchicine was injected into the lateral ventricles 24 h before transcardiac perfusion with 4% paraformaldehyde. The section was initially scanned using a 10 objective with a scan zoom 0.6. The area indicated by a white box was then scanned using a 63 objective with a scan zoom 1.0. Upper row: Stains resulted from the applications of a mouse-anti-GAD65 (in green) and a rabbit anti-GAD65 (in red) are nearly perfectly overlapped under the 10 objective. Bar (200 mm) applies to all image panels. Lower row: Somata stained by the mouse antiGAD65 (in green) were also labeled by the rabbit anti-GAD65 (in red). Arrowheads indicate somata stained with the mouse anti-GAD65 whereas arrows show somata labeled with the rabbit anti-GAD65. Bar (20 mm) applies to all image panels. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of the article.)

X. Wang et al. / Journal of Chemical Neuroanatomy 61–62 (2014) 51–63

61

Fig. 10. Co-labeling of GAD65 and GAD67 shows a portion of GAD65-immunoreactive somata have weak or no GAD67 immunoreactivity. Sections prepared from either colchicine-untreated or treated rats were double-stained with a rabbit anti-GAD65 and a mouse anti-GAD67. Three areas of each section in the hilus near the granule cell layer and the hilus border were randomly scanned with a 63 objective in a scan zoom 1.0. Bar (20 mm) applies to all image panels. Upper row: Representative images show colocalization of GAD65 (in red) and GAD67 (in green) in a section prepared from a colchicine-untreated rat. Note that a GAD65 immunoreactive soma indicated by the arrowhead appears to be lack of GAD67 immunoreactivity. Lower row: Representative images show co-localization of GAD65 (in red) and GAD67 (in green) in a section prepared from a colchicine-treated rat. Despite most GAD65-immunoreactive somata indicated by arrows had GAD67 immunoreactivity, the arrowhead-pointed GAD65-immunoreactive soma lacked visible GAD67 immunoreactivity. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of the article.)

addition to a diverse expression in terms of distribution pattern and immunoreactive intensity, more than one-tenth of somatostatin-positive somata were found to have no detectable perikaryal GAD67 immunoreactivity. Co-labeling GABA and GAD67 and direct co-labeling GAD65 and GAD67 further confirmed that a fraction of GABAergic cells lacked GAD67 immunoreactivity in the cytoplasm. When using perikaryal GAD67 immunoreactivity to count GABAergic neurons, a number of GABAergic cells would certainly be missed. Additionally, using perikaryal GAD67 immunoreactivity to present GABAergic cells in the hilus also suffers from dense mossy fiber expression. The dense mossy fiber expression of GAD67 makes isolation of GABAergic somata difficult if not impossible. In addition, the cytoplasmic level of GAD67 in neurons is closely regulated by the GABA level (Rimvall et al., 1993; Rimvall and Martin, 1994) and it becomes more complicated because the expression of GAD67 is altered by pathological conditions such as seizure activity (Schwarzer and Sperk, 1995; Maqueda et al., 2003; Lau and Murthy, 2012; Sperk et al., 2012). Inhibition of axonal transport with colchicine would be helpful in improving GAD67 labeling; however, its neurotoxicity (Sutula et al., 1983) and the needs of pretreatment may limit its use. As colchicine-induced inhibition of axonal transport enhanced GAD67 immunoreactivity not only in GABAergic cells but also in granule cells, the beneficial effect of colchicine pretreatment on the isolation of GABAergic somata through GAD67 labeling may be marginal. GAD65 has rarely been immunohistochemically targeted; it is contradicted with the findings that the GAD65 mRNA signal is stronger than that of GAD67 mRNA in the dentate gyrus (Houser and Esclapez, 1994.) as well as in the striatum (Mercugliano et al.,

1992). Given the evidence that inhibition of axonal transport dramatically improved GAD65 labeling, the faint immunoreactivity may result from a fast axonal transport of GAD65 from the cytoplasm to the axonal terminals. Due to the following reasons, immunohistochemically targeting GAD65 should be considered more suitable than targeting GAD67 for identifying GABAergic somata in confocal imaging. First, perikaryal GAD65 labeling looked very characteristic with almost identical distribution pattern and similar immunoreactive intensity, clearly distinguishable from surrounding stains. Second, different from a robust hilar GAD67 immunoreactivity that originates from mossy fiber expression, GAD65 immunoreactivity in the hilus looked moderate, making isolation of GAD65 immunoreactive somata easier. Third, GAD65 is expressed in every hilar CR-, nNOS-, PV-, or SOM-positive somata, which have been recognized as neurochemically predominant types of interneurons in the rodent dentate gyrus (Jinno and Kosada, 2006; Houser, 2007; Liang et al., 2013). Inhibition of axonal transport further boosts GAD65 immunoreactivity to a level close to the intensity of GAD67’s; different to the improvement seen in GAD67, the augmentation of GAD65 immunoreactivity by colchicine was barely seen in granule cells. The recent development of sophisticated techniques such as optogenetics used to activate or suppress neurochemically homogeneous interneuron populations leaves one issue that needs to be answered urgently: to what extent is a subgroup of interneurons delineated by the expression of a particular molecular marker GABAergic? PV-positive interneurons have been the most extensively studied subgroup of interneurons in various species as well as

62

X. Wang et al. / Journal of Chemical Neuroanatomy 61–62 (2014) 51–63

in different brain regions. In the dentate gyrus, PV-positive cells are considered to be purely GABAergic (Nitsch et al., 1990). In the present study, PV-positive somata were found to express both GAD67 and GAD65 exclusively and densely. The intense perikaryal expression of GAD65 in PV-positive cells is different from an observation made in the hippocampal CA1 region in the mouse (Fukuda et al., 1997). For cells expressing somatostatin, it has been a controversy whether somatostatin-positive cells in the dentate gyrus are all GABAergic. In an early study, co-localization of somatostatin and GAD revealed that 90% of somatostatin-positive cells expressed GAD (Kosaka et al., 1988). An in situ hybridization study found that approximately 95% of cells positive for presomatostatin expressed GAD65 mRNA (Esclapez and Houser, 1995). Co-localization of GAD65 and somatostatin in the present study revealed that all somatostatin-positive somata had perikaryal GAD65 immunoreactivity, confirming that SOM-positive cells are purely GABAergic cells in the rat dentate gyrus. Labeling SOM in sections obtained from GAD67-GFP knock-in mice also supports that all hilar SOM-positive cells express GAD67. In contrast to the consistent perikaryal expression of GAD65 in somatostatin-expressing cells, perikaryal GAD67 immunoreactivity varied considerably in terms of distribution patterns and intensity. More than one-tenth of somatostatin-positive somata did not have detectable perikaryal GAD67 immunoreactivity. Indeed, the intensity of GFP expression in hilar SOM-positive cells is reminiscent of GAD67 immunofluorescence in the rat. Although all hilar SOM-positive cells seemed to have GFP expression in GAD67-GFP knock-in mice, weak GFP expression was seen in more than 40% of them. If the result obtained from mice is applicable to the rat, the lack of GAD67 immunoreactivity in hilar GABAergic cells of rats might have not been due to a complete absence of GAD67 transcription, but due to fewer copies of GAD67 mRNA or inactive transcription in the cytoplasm and/or a fast anterograde axonal transport of GAD67. As somatostatin-positive cells in the dentate gyrus are vulnerable to a variety of neurological diseases and insults, including epilepsy (Tallent, 2007), it remains to be known if the difference in perikaryal GAD67 expression may make SOM-positive cells functionally diverse and variable in response to pathological alterations. CRpositive cells present a subgroup of cells with small soma sizes in the rat dentate gyrus. It has been argued whether all CR-positive cells in the dentate gyrus are GABAergic (Houser, 2007). Earlier studies reported that approximately three in four of CR-positive cells are positive for GABA in the rat dentate gyrus (Gulya´s et al., 1992; Miettinen et al., 1992). Randomly examining nearly 90 CRpositive cells revealed that CR-positive cells without GAD67 expression were very rare and, in addition, all CR-positive somata expressed GAD65, clearly implicating that CR-positive cells in the rat dentate gyrus are GABAergic. However, this may be not the case in the mouse. In the mouse dentate gyrus, the majority of CR-positive cells have been proven to be mossy cells which target the inner ˜ez and Freund, 1997; Fujise molecular layer densely (Blasco-Iba´n et al., 1998). The nNOS-positive interneurons are newly recognized as a prevalent subgroup of interneurons in the neocortex and the hippocampus (Tricoire et al., 2010; Perrenoud et al., 2012) and they are particularly crowded in the rodent dentate gyrus (Jinno and Kosaka, 2006; Liang et al., 2013). Similar to the perikaryal expression of GADs in PV-positive cells, nNOS-positive somata were found to express both GAD67 and GAD65 exclusively. We conclude that in the application of confocal imaging, immunofluorescently targeting GAD65 for identifying GABAergic somata in the hilar region of dentate gyrus has advantages over targeting GAD67 in terms of easier recognition of perikaryal labeling and more consistent expression in neurochemically prevalent subgroups of interneurons. Inhibition of axonal transport further improves perikaryal GAD65 labeling of GABAergic cells, making GABAergic cells more isolatable.

Acknowledgments This work was supported by grants from National Natural Science Foundation of China (81171231) and from the Ph.D. Programs Foundation of Ministry of Education of China (20100131110038). The authors would like to thank Mr. Stuart Hedgpeth of Brody School of Medicine at East Carolina University and Ms. Senmiao Zhan at University of North Carolina School of Medicine for English editing. References Amaral, D.G., Scharfman, H.E., Lavenex, P., 2007. The dentate gyrus: fundamental neuroanatomical organization (dentate gyrus for dummies). Prog. Brain Res. 163, 3–22. Andrews-Zwilling, Y., Gillespie, A.K., Kravitz, A.V., Nelson, A.B., Devidze, N., Lo, I., Yoon, S.Y., Bien-Ly, N., Ring, K., Zwilling, D., Potter, G.B., Rubenstein, J.L., Kreitzer, A.C., Huang, Y., 2012. Hilar GABAergic interneuron activity controls spatial learning and memory retrieval. PLoS ONE 7, e40555. ˜ ez, J.M., Freund, T.F., 1997. Distribution, ultrastructure, and connectivity Blasco-Iba´n of calretinin-immunoreactive mossy cells of the mouse dentate gyrus. Hippocampus 7, 307–320. Cze´h, B., A´braha´m, H., Tahtakran, S., Houser, C.R., Seress, L., 2013. Number and regional distribution of GAD65 mRNA-expressing interneurons in the rat hippocampal formation. Acta Biol. Hung. 64, 395–413. Committee for the update of the guide for the care and use of laboratory animals, 2011. Guide for the Care and Use of Laboratory Animals, 8th ed. The National Academic Press, San Diego. Dı´az-Cintra, S., Gonza´lez-Maciel, A., Morales, M.A., Aguilar, A., Cintra, L., PradoAlcala´, R.A., 2007. Protein malnutrition differentially alters the number of glutamic acid decarboxylase-67 interneurons in dentate gyrus and CA1-3 subfields of the dorsal hippocampus. Exp. Neurol. 208, 47–53. Erlander, M.G., Tillakaratne, N.J., Feldblum, S., Patel, N., Tobin, A.J., 1991. Two genes encode distinct glutamate decarboxylases. Neuron 7, 91–100. Esclapez, M., Houser, C.R., 1995. Somatostatin neurons are a subpopulation of GABA neurons in the rat dentate gyrus: evidence from colocalization of pre-prosomatostatin and glutamate decarboxylase messenger RNAs. Neuroscience 64, 339–355. Esclapez, M., Tillakaratne, N.J., Kaufman, D.L., Tobin, A.J., Houser, C.R., 1994. Comparative localization of two forms of glutamic acid decarboxylase and their mRNAs in rat brain supports the concept of functional differences between the forms. J. Neurosci. 14, 1834–1855. Ewell, L.A., Jones, M.V., 2010. Frequency-tuned distribution of inhibition in the dentate gyrus. J. Neurosci. 30, 12597–12607. Fritschy, J.-M., 2008. Is my antibody-staining specific? How to deal with pitfalls of immunohistochemistry. Eur. J. Neurosci. 28, 2365–2370. Freund, T.F., Buzsa´ki, G., 1996. Interneurons of the hippocampus. Hippocampus 6, 347–470. Fujise, N., Liu, Y., Hori, N., Kosaka, T., 1998. Distribution of calretinin immunoreactivity in the mouse dentate gyrus: II. Mossy cells, with special reference to their dorsoventral difference in calretinin immunoreactivity. Neuroscience 82, 181–200. Fukuda, T., Aika, Y., Heizmann, C.W., Kosaka, T., 1998. GABAergic axon terminals at perisomatic and dendritic inhibitory sites show different immunoreactivities against two GAD isoforms, GAD67 and GAD65, in the mouse hippocampus: a digitized quantitative analysis. J. Comp. Neurol. 395, 177–194. Fukuda, T., Heizmann, C.W., Kosaka, T., 1997. Quantitative analysis of GAD65 and GAD67 immunoreactivities in somata of GABAergic neurons in the mouse hippocampus proper (CA1 and CA3 regions), with special reference to parvalbumin-containing neurons. Brain Res. 764, 237–243. Gentet, L.J., Avermann, M., Matyas, F., Staiger, J.F., Petersen, C.C., 2010. Membrane potential dynamics of GABAergic neurons in the barrel cortex of behaving mice. Neuron 65, 422–435. Gulya´s, A.I., Miettinen, R., Jacobowitz, D.M., Freund, T.F., 1992. Calretinin is present in non-pyramidal cells of the rat hippocampus – I. A new type of neuron specifically associated with the mossy fibre system. Neuroscience 48, 1–27. Ha¨rtig, W., Singer, A., Grosche, J., Brauer, K., Ottersen, O.P., Bru¨ckner, G., 2001. Perineuronal nets in the rat medial nucleus of the trapezoid body surround neurons immunoreactive for various amino acids, calcium-binding proteins and the potassium channel subunit Kv3.1b. Brain Res. 899, 123–133. Ho¨rtnagl, H., Berger, M.L., Sperk, G., Pifl, C., 1991. Regional heterogeneity in the distribution of neurotransmitter markers in the rat hippocampus. Neuroscience 45, 261–272. Houser, C.R., 2007. Interneurons of the dentate gyrus: an overview of cell types, terminal fields and neurochemical identify. Prog. Brain Res. 163, 217–232. Houser, C.R., Esclapez, M., 1994. Localization of mRNAs encoding two forms of glutamic acid decarboxylase in the rat hippocampal formation. Hippocampus 4, 530–545. Houser, C.R., Esclapez, M., 1996. Vulnerability and plasticity of the GABA system in the pilocarpine model of spontaneous recurrent seizures. Epilepsy Res. 26, 207–218.

X. Wang et al. / Journal of Chemical Neuroanatomy 61–62 (2014) 51–63 Jinno, S., Kosaka, T., 2006. Cellular architecture of the mouse hippocampus: a quantitative aspect of chemically defined GABAergic neurons with stereology. Neurosci. Res. 56, 229–245. Ka¨gi, U., Berchtold, M.W., Heizmann, C.W., 1987. Ca2+-binding parvalbumin in rat testis. J. Biol. Chem. 262, 7314–7320. Ko¨hler, C., Chan-Palay, V., 1983. Distribution of gamma aminobutyric acid containing neurons and terminals in the septal area. An immunohistochemical study using antibodies to glutamic acid decarboxylase in the rat brain. Anat. Embryol. (Berl.) 167, 53–65. Kosaka, T., Wu, J.Y., Benoit, R., 1988. GABAergic neurons containing somatostatinlike immunoreactivity in the rat hippocampus and dentate gyrus. Exp. Brain Res. 71, 388–398. Kumoi, K., Saito, N., Tanaka, C., 1987. Immunohistochemical localization of gammaaminobutyric acid- and aspartate-containing neurons in the guinea pig vestibular nuclei. Brain Res. 416, 22–33. Lau, C.G., Murthy, V.N., 2012. Activity-dependent regulation of inhibition via GAD67. J. Neurosci. 32, 8521–8531. Li, Y., Aimone, J.B., Xu, X., Callaway, E.M., Gage, F.H., 2012. Development of GABAergic inputs controls the contribution of maturing neurons to the adult hippocampal network. Proc. Natl. Acad. Sci. U. S. A. 109, 4290–4295. Liang, Z., Zhang, L., Wang, X., Gao, F., Wang, X., Zhou, S., Yu, S., Zhan, R.-Z., 2013. Distribution and neurochemical features of neuronal nitric oxide synthaseexpressing interneurons in the rat dentate gyrus. Brain Res. 1505, 11–21. Lorincz, A., Nusser, Z., 2008. Specificity of immunoreactions: the importance of testing specificity in each method. J. Neurosci. 28, 9083–9086. Maqueda, J., Ramı´rez, M., Lamas, M., Gutie´rrez, R., 2003. Glutamic acid decarboxylase (GAD)67, but not GAD65, is constitutively expressed during development and transiently overexpressed by activity in the granule cells of the rat. Neurosci. Lett. 353, 69–71. Marı´n-Burgin, A., Mongiat, L.A., Pardi, M.B., Schinder, A.F., 2012. Unique processing during a period of high excitation/inhibition balance in adult-born neurons. Science 335, 1238–1242. Mascagni, F., Muly, E.C., Rainnie, D.G., McDonald, A.J., 2009. Immunohistochemical characterization of parvalbumin-containing interneurons in the monkey basolateral amygdala. Neuroscience 158, 1541–1550. Mercugliano, M., Soghomonian, J.J., Qin, Y., Nguyen, H.Q., Feldblum, S., Erlander, M.G., Tobin, A.J., Chesselet, M.F., 1992. Comparative distribution of messenger RNAs encoding glutamic acid decarboxylases (Mr 65,000 and Mr 67,000) in the basal ganglia of the rat. J. Comp. Neurol. 318, 245–254. Miettinen, R., Gulya´s, A.I., Baimbridge, K.G., Jacobowitz, D.M., Freund, T.F., 1992. Calretinin is present in non-pyramidal cells of the rat hippocampus—II. Coexistence with other calcium binding proteins and GABA. Neuroscience 48, 29–43. Morellini, F., Sivukhina, E., Stoenica, L., Oulianova, E., Bukalo, O., Jakovcevski, I., Dityatev, A., Irintchev, A., Schachner, M., 2010. Improved reversal learning and working memory and enhanced reactivity to novelty in mice with enhanced GABAergic innervation in the dentate gyrus. Cereb. Cortex 20, 2712–2727. Mu¨ller, G.J., Møller, A., Johansen, F.F., 2001. Stereological cell counts of GABAergic neurons in rat dentate hilus following transient cerebral ischemia. Exp. Brain Res. 141, 380–388. Nitsch, R., Soriano, E., Frotscher, M., 1990. The parvalbumin-containing nonpyramidal neurons in the rat hippocampus. Anat. Embryol. 181, 413–425. Oertel, W.H., Schmechel, D.E., Tappaz, M.L., Kopin, I.J., 1981. Production of a specific antiserum to rat brain glutamic acid decarboxylase by injection of an antigen– antibody complex. Neuroscience 6, 2689–2700. Patel, A.B., deGraaf, R.A., Martin, D.L., Battaglioli, G., Behar, K.L., 2006. Evidence that GAD65 mediates increased GABA synthesis during intense neuronal activity in vivo. J. Neurochem. 97, 385–396. Perrenoud, Q., Geoffroy, H., Gauthier, B., Rancillac, A., Alfonsi, F., Kessaris, N., Rossier, J., Vitalis, T., Gallopin, T., 2012. Characterization of type I and type II nNOS-expressing interneurons in the barrel cortex of mouse. Front. Neural Circuits 6, 36.

63

Piatti, V.C., Ewell, L.A., Leutgeb, J.K., 2013. Neurogenesis in the dentate gyrus: carrying the message or dictating the tone. Front. Neurosci. 7, 50. Rimvall, K., Martin, D.L., 1994. The level of GAD67 protein is highly sensitive to small increases in intraneuronal gamma-aminobutyric acid levels. J. Neurochem. 62, 1375–1381. Rimvall, K., Sheikh, S.N., Martin, D.L., 1993. Effects of increased gamma-aminobutyric acid levels on GAD67 protein and mRNA levels in rat cerebral cortex. J. Neurochem. 60, 714–720. Schwarzer, C., Sperk, G., 1995. Hippocampal granule cells express glutamic acid decarboxylase-67 after limbic seizures in the rat. Neuroscience 69, 705–709. Shetty, A.K., Hattiangady, B., Rao, M.S., 2009. Vulnerability of hippocampal GABAergic interneurons to kainate-induced excitotoxic injury during old age. J. Cell. Mol. Med. 13, 2408–2423. Shetty, A.K., Turner, D.A., 2001. Glutamic acid decarboxylase-67-positive hippocampal interneurons undergo a permanent reduction in number following kainic acid-induced degeneration of CA3 pyramidal neurons. Exp. Neurol. 169, 276–297. Sperk, G., Wieselthaler-Ho¨lzl, A., Pirker, S., Tasan, R., Strasser, S.S., Drexel, M., Pifl, C., Marschalek, J., Ortler, M., Trinka, E., Heitmair-Wietzorrek, K., Ciofi, P., Feucht, M., Baumgartner, C., Czech, T., 2012. Glutamate decarboxylase 67 is expressed in hippocampal mossy fibers of temporal lobe epilepsy patients. Hippocampus 22, 590–603. Spiegel, A.M., Koh, M.T., Vogt, N.M., Rapp, P.R., Gallagher, M., 2013. Hilar interneuron vulnerability distinguishes aged rats with memory impairment. J. Comp. Neurol. 521, 3508–3523. Sun, C., Mtchedlishvili, Z., Bertram, E.H., Erisir, A., Kapur, J., 2007. Selective loss of dentate hilar interneurons contributes to reduced synaptic inhibition of granule cells in an electrical stimulation-based animal model of temporal lobe epilepsy. J. Comp. Neurol. 500, 876–893. Sutula, T., Goldschmidt, R., Steward, O., 1983. Mechanisms of colchicine neurotoxicity in the dentate gyrus: dissociation of seizures and cell death. Exp. Neurol. 81, 683–693. Tallent, M.K., 2007. Somatostatin in the dentate gyrus. Prog. Brain Res. 163, 265–284. Tamamaki, N., Yanagawa, Y., Tomioka, R., Miyazaki, J., Obata, K., Kaneko, T., 2003. Green fluorescent protein expression and colocalization with calretinin, parvalbumin, and somatostatin in the GAD67-GFP knock-in mouse. J. Comp. Neurol. 467, 60–79. Tian, N., Petersen, C., Kash, S., Baekkeskov, S., Copenhagen, D., Nicoll, R., 1999. The role of the synthetic enzyme GAD65 in the control of neuronal gamma-aminobutyric acid release. Proc. Natl. Acad. Sci. U. S. A. 96, 12911–12916. Tricoire, L., Pelkey, K.A., Daw, M.I., Sousa, V.H., Miyoshi, G., Jeffries, B., Cauli, B., Fishell, G., McBain, C.J., 2010. Common origins of hippocampal Ivy and nitric oxide synthase expressing neurogliaform cells. J. Neurosci. 30, 2165–2176. Walls, A.B., Nilsen, L.H., Eyjolfsson, E.M., Vestergaard, H.T., Hansen, S.L., Schousboe, A., Sonnewald, U., Waagepetersen, H.S., 2010. GAD65 is essential for synthesis of GABA destined for tonic inhibition regulating epileptiform activity. J. Neurochem. 115, 1398–1408. Wierenga, C.J., Mu¨llner, F.E., Rinke, I., Keck, T., Stein, V., Bonhoeffer, T., 2010. Molecular and electrophysiological characterization of GFP-expressing CA1 interneurons in GAD65-GFP mice. PLoS ONE 5, e15915. Wolff, J.R., Bo¨ttcher, H., Zetzsche, T., Oertel, W.H., Chronwall, B.M., 1984. Development of GABAergic neurons in rat visual cortex as identified by glutamate decarboxylase-like immunoreactivity. Neurosci. Lett. 47, 207–212. Yu, E.P., Dengler, C.G., Frausto, S.F., Putt, M.E., Yue, C., Takano, H., Coulter, D.A., 2013. Protracted postnatal development of sparse, specific dentate granule cell activation in the mouse hippocampus. J. Neurosci. 33, 2947–2960. Zhan, R.-Z., Nadler, J.V., 2009. Enhanced tonic GABA current in normotopic and hilar ectopic dentate granule cells after pilocarpine-induced status epilepticus. J. Neurophysiol. 102, 670–681.