Cellular localization of adenosine A1 receptors in rat forebrain: Immunohistochemical analysis using adenosine A1 receptor-specific monoclonal antibody

Cellular Localization of Adenosine A1Receptors in Rat Forebrain:

Immunohistochemical Analysis UsingAdenosine A1 Receptor-Specific

Monoclonal Antibody

TOMOYO OCHIISHI,1* LING CHEN,1 AYA YUKAWA,1 YOSHIKO SAITOH,2 YUKO

SEKINO,3 TAKAO ARAI,4 HIROYASU NAKATA,2 AND HIROSHI MIYAMOTO1

1Biosignalling Department, National Institute of Bioscience and Human-Technology,Tsukuba, Ibaraki 305–8566, Japan

2Department of Molecular and Cellular Neurobiology, Tokyo Metropolitan Institute forNeuroscience, Fuchu, Tokyo 183–8526, Japan

3Gunma University School of Medicine, Maebashi, Gunma 371–8511, Japan4Department of Applied Biological Science, Faculty of Science and Technology, Science

University of Tokyo, Noda, Chiba 278–8510, Japan

ABSTRACTMonoclonal antibodies were generated against the adenosine A1 receptor (A1R) purified

from rat brain. In immunoblot analyses of purified or partially purified A1R preparations fromrat brain, these antibodies recognized a solitary band, the size of which corresponded to thatexpected for A1R. These antibodies recognized not only the native form of A1R but also thedeglycosylated form of A1R. Immunocytochemical analysis of Chinese hamster ovarian cellsthat were transfected stably with rat A1R cDNA showed that their cell bodies were stainedintensely by these antibodies, whereas nontransfected Chinese hamster ovarian cells werenot. These antibodies detected the A1R naturally present in the DDT1 MF-2 smooth musclecells. One of these antibodies (the 511CA antibody) was then used to examine the immunohis-tochemical distribution of A1Rs in rat forebrain. On light microscopy, A1R immunoreactivitywas observed in the cerebral cortex, septum, basal ganglia, hippocampal formation, andthalamus. However, in some regions of the forebrain, regional differences in staining intensitywere found as follows: In the cerebral cortex, the strongest immunoreactivity was found in thelarge pyramidal neurons of layer V. This immunoreactivity was detected in the pyramidal cellbodies, dendrites, and axon initial segments. In the hippocampus, A1R immunoreactivity wasdetected mainly in the stratum pyramidale. The pyramidal cells in fields CA2–CA3 of thehippocampus were stained more intensely or more clearly than those in field CA1 or thedentate gyrus. More intense A1R immunoreactivity of the apical dendrites was detected infield CA2 compared with other hippocampal fields and the dentate gyrus. Many interneuronsof the hippocampus were stained by the 511CA antibody. The subcellular distribution of A1Rsin the forebrain was examined by using a digital deconvolution system and electronmicroscopy. In the cerebral cortex, the view obtained by removing the background haze bydeconvolution revealed that the immunofluoresence-labeled A1Rs were distributed on thesurfaces of the cell bodies and dendrites and in the cytoplasm of layer V neurons as smallspots. In field CA1, immunoreactivity was detected in the areas surrounding pyramidal cells.

Grant sponsors: COE Research Program of the Science and TechnologyAgency of Japan, and Ministry of Education, Science, Sports and Culture ofJapan (to Yoshiko Saitoh and Hiroyasu Nakata).

*Correspondence to: Tomoyo Ochiishi, Biosignalling Department, Na-tional Institute of Bioscience and Human-Technology, Agency of IndustrialScience and Technology, 1–1, Higashi, Tsukuba, Ibaraki 305–8566, Japan.E-mail: [email protected]

Received 13 April 1998; Revised 3 November 1998; Accepted 22 March1999

THE JOURNAL OF COMPARATIVE NEUROLOGY 411:301–316 (1999)

r 1999 WILEY-LISS, INC.

Electron microscopy revealed the presence of A1R-immunoreactive products in both thepresynaptic terminals and the postsynaptic structures. The specific cellular distribution ofA1Rs is consistent with the physiological premise that endogeneously released adenosineexerts control over the excitability of forebrain neurons at both presynaptic and postsynapticsites through A1Rs. J. Comp. Neurol. 411:301–316, 1999. r 1999 Wiley-Liss, Inc.

Indexing terms: cerebral cortex; hippocampus; basal ganglia; thalamus; electron microscopy

In the central nervous system, adenosine acts as anextracellular signal influencing neuronal activity. Thefunction of adenosine has been determined by the specificadenosine receptors coupled to guanine nucleotide bindingproteins (G-proteins; Londos et al., 1980; Stiles, 1992).Activation of the adenosine A1 receptor (A1R), whichcouples with the inhibitory G-protein, leads to potentinhibition in the central nervous system by reducingadenylate cyclase activity (Dunwiddie and Fredholm, 1989;Mahan et al., 1991; Townsend-Nicholson and Shine, 1992;Libert et al., 1993; Rivkees et al., 1995). Physiologicalstudies reveal that activation of A1Rs stimulates theconductance of K1 current (Trussell and Jackson, 1987;Gerber et al., 1989; Thompson et al., 1992) and reducesCa21 influx (Proctor and Dunwiddie, 1983; Scholz andMiller, 1991; Wu and Saggau, 1994) in pre- and postsynap-tic regions of the forebrain. Through these functions, A1Rsmediate reduction in the presynaptic release of neurotrans-mitters and modulate synaptic transmission (Siggins andSchubert, 1981; Segal, 1982; Fredholm and Dunwiddie,1988; Gerber et al., 1989; Lupica et al., 1992; Prince andStevens, 1992; Yamamoto et al., 1993; Klishin et al., 1995).Recent studies have reported that A1Rs are involved in thepathogenesis of epilepsy (During and Spencer, 1992) andin the regulation of synaptic plasticity, such as long-termpotentiation, in the hippocampus (de Mendonca andRibeiro, 1993, 1994). Autoradiographic, light microscopicstudies of agonist or antagonist binding (Goodman et al.,1983; Lee and Reddington, 1986; Lee et al., 1986; Snowhilland Williams, 1986; Deckert and Jorgensen, 1988; Dragu-now et al., 1988; Fastbom and Fredholm, 1990; Weber etal., 1990; Olah and Stiles, 1995; Swanson et al., 1995) andin situ hybridization studies (Reppert et al., 1991; Swan-son et al., 1995) have revealed that A1Rs are expressed inalmost all regions of the forebrain. The cellular distribu-

tion of A1Rs in the forebrain has been revealed by usingpolyclonal antiserum generated against synthetic peptidesof rat and human A1R (Rivkees et al., 1995; Swanson et al.,1995). The results of these studies have shown that A1Rsare distributed in the cell bodies and axons of neurons inthe cerebral cortex and hippocampus, but not in presynap-tic terminals. Therefore, physiologically characterizedA1R,which is now believed to be located both in the presynapticterminals and/or postsynaptic structures and to regulatesynaptic transmission, as described above, had not beenidentified immunohistochemically. Therefore, to define thecellular and subcellular localization of A1R in the ratforebrain in detail, we developed specific A1R monoclonalantibodies by using highly purified A1R from rat brainmembrane as the antigen. With one of these monoclonalanti-A1R antibodies (the 511CA antibody), we determinedthe distribution of A1Rs in presynaptic terminals andpostsynaptic structures by electron microscopy and re-vealed the regional differences in the expression of A1Rs inthe rat forebrain.

MATERIALS AND METHODS

Production of monoclonal antibodies

Anti-A1R monoclonal antibodies were raised againstA1R purified from rat brain (Nakata, 1989b). The foot-padimmunization procedure (Rolink et al., 1989; Ohuchi et al.,1994; Yoshida et al., 1995) was used to immunize 6-week-old female BALB/c mice with purified A1R protein, whereinthree injections (20 µg of protein in total) were adminis-tered in the same hind footpad of each mouse. Three daysafter the final injection, the regional lymph nodes wereremoved, and lymphocytes were fused with the mousemyeloma cell line PAI (Stocker et al., 1982) by using a cellelectrofusion method (Hui and Stenger, 1993) with a cellelectrofusion pulser [3-mm electrode gap; 50 Vp-p; 1 MHzalternating current (AC) field for five seconds; 700 V directcurrent (DC) field for 30 µsec; BM-850; Kurabo, Kurashiki,Japan]. Fused hybrids were selected in hypoxanthine/aminopterin/thymidine medium (Arai and Matsumoto,1988). Hybridoma cells producing anti-A1R antibodieswere screened by using the microplate enzyme-linkedimmunosorbent assay procedure (Yoshida et al., 1995)with partially purified A1R (Nakata, 1989a). A 50-µlaliquot of partially purified A1R protein was adsorbed ontoeach well of a 96-well plate (Sumilon, Tokyo, Japan)overnight at 4°C. After removing the A1R-containing solu-tion from each well, phosphate-buffered saline, pH 7.4(PBS; Dulbecco’s PBS; Dainippon Pharmaceutical Co.,Ltd., Osaka, Japan) containing 10% fetal bovine serum(FBS; Gibco, Grand Island, NY) was added for blockingnonspecific binding and then incubated for one hour. Afterremoving the blocking solution, each well was incubatedwith 50 µl of fluid from a hybridoma culture for one hour at

Abbreviations

alv alveus hippocampuscc corpus callosumCP caudate-putamend dendriteDG dentate gyrusdhc dorsal hippocampal commissureGP globus-pallidusGrDG granule cell layer of dentate gyrusmf mossy fiberMit mitochondriaMol molecular layer of dentate gyrusNuc nucleusOc2L occipital cortex, area 2, lateralPoDG polymorphic layer of dentate gyrusSLM stratum lacunosum-moleculareSO stratum oriensSP stratum pyramidaleSR stratum radiatumT presynaptic terminal

302 T. OCHIISHI ET AL.

room temperature. Each well was washed three times withPBS and then incubated with horseradish peroxidase(HRP)-conjugated goat anti-mouse immunoglobulin (Amer-sham, Buckinghamshire, England; 1:200 dilution in PBScontaining 10% FBS) for one hour at room temperature.After the wells were washed again with PBS, enzymeactivity was detected by incubation with 0.4 mg/ml o-phenylenediamine (Sigma, St. Louis, MO) and 0.01% H2O2in 0.1M citrate-phosphate buffer, pH 4.8, for 10 minutes.The hybridomas giving rise to positive reactions werecloned and cultured and, subsequently, were injectedintraperitoneally into eight-week-old male BALB/c mice toinduce the production of ascitic fluid that contained themonoclonal antibodies.

Confirmation of the specificity of themonoclonal anti-A1R antibodies

The specificity of the monoclonal antibodies thus devel-oped in this study was confirmed by immunoblot analysesusing purified or partially purified A1Rs from rat brain,immunocytochemical analyses of Chinese hamster ovar-ian (CHO) cells expressing A1Rs or DDT1 MF-2 smoothmuscle cells, and a preadsorption test.

Immunoblot analysis. Purified or partially purifiedA1Rs from rat brain membrane (Nakata, 1989a) wereseparated by sodium dodecyl sulphate-polyacrylamide gelelectrophoresis (SDS-PAGE; 12% polyacrylamide; Laemmli,1970) followed by silver staining (2D silver stain kit:Daiichi Pure Chemicals, Tokyo, Japan) or immunoblotting.For immunoblotting, the separated proteins were trans-ferred electrophoretically onto an Immobilon membranefilter (Millipore, Bedford, MA). After precoating the filterwith 5% skim milk in Tris-buffered saline, pH 8.0, contain-ing 0.1% Tween 20 (TBST), immunoblotting was per-formed with monoclonal anti-A1R antibodies in skimmilk/TBST (1:200 dilution) overnight at 4°C. Polyclonalanti-A1R antibody, which was raised against the highlypurifiedA1R protein from rat brain and has been character-ized previously (Nakata, 1993), was used as a positivestandard (1:500 dilution). The Immobilon membrane filterwas rinsed five times with TBST and incubated withHRP-conjugated goat anti-mouse immunoglobulin G (IgG)or anti-rabbit IgG (Amersham; 1:500 dilution in skimmilk/TBST, respectively) for one hour. After the membranefilter was rinsed five times with TBST, the immunoreac-tive bands were detected by chemiluminescence using theenhanced chemiluminesence (ECL) system (Amersham).

Immunoblot analysis of the deglycosylated A1R.

Aliquots of the purified A1R from rat brain membrane(Nakata, 1989a) were treated with 0 U, 0.1 U, or 0.25 U ofendoglycosidase F/N-glycosidase F (Boehringer Mann-heim Biochemica, Mannheim, Germany) in 10 mM Tris-acetate buffer, pH 6.8, containing 20 mM NaCl, 0.04%SDS, and 0.4 mM ethylenediamine tetraacetic acid for fivehours at room temperature (Nakata, 1993). The reactionmixtures were then subjected to SDS-PAGE followed byimmunoblotting with monoclonal anti-A1R antibodies, asdescribed above.

Immunofluorescence labeling of the A1R proteins in

CHO cells expressing the A1R protein and in DDT1

MF-2 smooth muscle cells. The cDNA coding for the ratA1R (Mahan et al., 1991) was subcloned into the HindIIIsite of the pRC/CMV expression vector containing theneomycin-resistant gene. Stable transfection of cDNA intoCHO cells was achieved by calcium-phosphate precipita-

tion. G-418-resistant CHO cell clones were then screenedfor intact cells by [3H]1,3-dipropyl-8-cyclopentylxanthinebinding. Positive clones covering a range of A1R expres-sion levels were expanded further to establish a cell line.The stable cell lines that expressed approximately 2pmol/mg of A1R protein were maintained in Ham’s F12medium (Nissui, Tokyo, Japan) supplemented with 10%FBS and 0.01% kanamycin at 37°C in a 5% CO2 atmo-sphere.

DDT1 MF-2 smooth muscle cells originally were isolatedfrom a steroid-induced leiomyosarcoma of Syrian hamstervas deferens (Norris et al., 1974) and express A1R natu-rally (Ramkumar et al., 1991; Gerwins and Fredholm,1995). Cells were grown in Dulbecco’s modified Eagle’smedium (Nissui) supplemented with 1.5 g NaHCO3/liter,10% FBS, 100 units/ml penicillin, and 100 µg/liter strepto-mycin at 37°C in a 5% CO2 atmosphere.

For immunocytochemical analysis, the stably trans-fected CHO cells or DDT1 MF-2 smooth muscle cells werecultured in Lab-Tek tissue culture chambers (Nunc, Naper-ville, IL). When the cells were 30–50% confluent, they werewashed with PBS, fixed with cold ethanol (220°C) for fiveminutes, and then with cold ethanol (80%) and acetone(20%; 220°C) for five minutes. The cells on the slides weredried for 10 minutes and incubated in the ascitic fluidcontaining the monoclonal anti-A1R antibodies [1:50 dilu-tion in PBS containing 1% normal goat serum (NGS);Funakoshi, Tokyo, Japan; and 0.02% Triton-X 100 (Sigma)]for three hours at 37°C. After the antibodies were washedout with PBS, the cells on chamber slides were incubatedwith biotinylated goat anti-mouse IgG (Vector Laborato-ries; 1:200 dilution in PBS containing 1% NGS) for twohours at room temperature. After being rinsed again withPBS, the cells were incubated with fluorescein isothiocya-nate (FITC)-conjugated streptavidin (Amersham; 1:100dilution in PBS) for one hour at room temperature. Onceagain, after being rinsed with PBS, the cells were cover-slipped with PermaFluor (Lipshow, Pittsburgh, PA), andphotomicrographs were then obtained by using an Olym-pus microscope (BX50WI; Olympus, Tokyo, Japan). Non-transfected cells were used as negative controls. To exam-ine the cross reactivity of these antibodies against theother types of adenosine receptors, we cultured PC12 cells,which express A2a adenosine receptor, and CHO cells,which overexpress A2b or A3 adenosine receptors, andimmunostained as described above.

Preadsorption test. TheA1R preparation purified fromthe rat forebrain (Nakata, 1989a) was used for the pread-sorption test. The monoclonal anti-A1R antibodies wereincubated with the purified A1R for one hour at roomtemperature, and the mixture was then diluted in PBScontaining 1% NGS and 0.02% Triton X-100 (1:100). Thesections prepared for immunohistochemistry (as describedbelow) were then incubated with the preadsorbed antibodyovernight at 4°C. After washing with PBS, the sectionswere incubated with FITC-conjugated goat anti-mouseIgG (KPL, Gaithersburg, MD; 1:200 dilution in PBS con-taining 1% NGS) for two hours at room temperature,rinsed with PBS, coverslipped in PermaFluor, and viewedunder an Olympus microscope (BX50WI).

Immunohistochemical staining

Eighteen three-month-old female Wistar rats (180–230g) were anesthetized by intraperitoneal injection of chloralhydrate (3.5 mg/10 g body weight) and perfused transcardi-

ADENOSINE A1 RECEPTOR EXPRESSION IN FOREBRAIN 303

ally with PBS at room temperature for 5–7 minutes,followed by 4% paraformaldehyde and 7% saturated picricacid in 0.1 M phosphate buffer (PB), pH 7.4, at 4°C for 15minutes. After the perfusion, the brains were removed andplaced in the same fixative at 4°C for two hours andimmersed overnight in 0.1 M PB containing 20% sucrose at4°C for cryoprotection. A freezing microtome (Micron,Germany) was used to prepare 20-µm-thick coronal sec-tions. Immunostaining was carried out by using a modifiedavidin-biotin-peroxidase method (Hsu et al., 1981; Ochii-shi et al., 1994). In brief, the sections were washed withPBS and then blocked for nonspecific labeling with 1%NGS for one hour. After an overnight incubation at 4°Cwith monoclonal anti-A1R antibody (511CAantibody; 1:100dilution in PBS containing 0.02% Triton X-100 and 1%NGS), the sections were washed with PBS and thenincubated with biotinylated goat anti-mouse IgG (1:200dilution in PBS containing 1% NGS) for one hour at roomtemperature. After washing again with PBS, the sectionswere incubated with avidin-biotin-peroxidase complex (Vec-tor Laboratories; 1:100 dilution in PBS) for 30 minutes atroom temperature and then rinsed three times with 0.1 MPB for 10 minutes. Immunoreactive products were visual-ized after a 15-minute incubation with 0.05% 3, 38-diaminobenzidine (DAB; Dojin, Kumamoto, Japan) and0.003% H2O2 dissolved in 0.1 M PB. After another washwith 0.1 M PB, the sections were mounted on gelatin-coated slides, dried, dehydrated in a graded series ofethanol, cleaned in xylene, coverslipped, and photo-graphed under an Olympus microscope (BX50WI).

The immunofluoresence labeling procedure used wassimilar to that described above for DAB staining. In brief,brain sections were incubated sequentially with 1) 511CAantibody diluted in PBS containing 1% NGS and 0.02%Triton X-100 for one hour at room temperature followed byan additional overnight incubation at 4°C, and 2) FITC-conjugated goat anti-mouse IgG (KPL; 1:200 dilution) fortwo hours at room temperature. Between these steps, thesections were rinsed three times for 10 minutes each inPBS. The sections were then coverslipped with Perma-Fluor. The immunoreactivity was observed under a fluores-cence microscope (Axiophot; Zeiss, Oberkochen, Germany)and analyzed by using the CELLscan image-acquisitionsystem (Scanalytics, Billerica, MA). Three-dimensionaldigital deconvolved views were obtained for a series ofoptical sections covering a distance of 0.22 µm on theZ-axis, with a total thickness of about 10 µm.

Electron microscopic examination

The brains were cut coronally on a microslicer to obtain100-µm-thick sections (Dosaka E.M., Kyoto, Japan). Thesections were incubated with 511CA antibody in PBScontaining 1% NGS (1:300 dilution) for two days at 4°C,and visualized using a Vectastain anti-mouse peroxidasekit (Vector Laboratories) with DAB as substrate. Thesections were then postfixed with 1% OsO4 for 30 minutesat 4°C, dehydrated in a graded series of ethanol andpropylene oxide, and embedded in an Epon 812 mixture.Ultrathin sections (60–80 nm) were counterstained withuranyl acetate and lead citrate and were observed underan electron microscope (model H7000; Hitachi, Tokyo,Japan) at an acceleration voltage of 75 kV. The protocolsused in this study were approved by our institutionalAnimal Care and Use Committee.

RESULTS

Production and verification of the specificityof the monoclonal antibodies against purified

A1R protein

Three hybridoma cell lines producing monoclonal anti-A1R antibodies were obtained. In the present study, weused one of these, the 511CA antibody, for immunohisto-chemical analysis.

The specificity of the 511CA antibody was demonstratedfirst by immunoblotting (Figs. 1, 2). This antibody revealeda major immunoreactive band at approximately 35 kDa(Fig.1, lane 2), which corresponded to the silver-stained,highly purified A1R band (Fig. 1, lane 1). Figure 2 showsthat, in the partially purified A1R preparation, this anti-body also recognized a solitary protein band with a molecu-lar mass of approximately 35 kDa (Fig. 2, lane 1). Silverstaining demonstrated that the preparation included manybands representing unpurified proteins (Fig. 2, lane 3).The immunoblotting pattern was identical to that ob-tained for polyclonal anti-A1R antiserum that specificallyrecognized A1R, as described previously by Nakata (1993;Fig. 2, lane 2). The broad-band pattern is a commoncharacteristic of glycoproteins (Stiles, 1986; Nakata, 1990).However, a difference was found between the stainingpatterns obtained by using the two types of antibodies. Theextent of immunoreactivity of the 511CA antibody (Fig. 2,

Fig. 1. Immunoblot analysis of highly purified rat brain adenosineA1 receptor (A1R) with a monoclonal anti-A1R antibody (511CAantibody). Approximately 5 ng of purified A1R were subjected tosodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and immunoblotted. Lane 1 is silver-stained, highly purifiedA1R protein. A band with an approximate molecular weight of 35 kDawas detected. Lane 2 is the immunoblot of purified A1R using the511CA antibody. A major band corresponding to the silver-stained,purified A1R was detected. Std: molecular weight markers labeled bysilver staining.


lane 1) was far less than that of the polyclonal antiserum(Fig. 2, lane 2), suggesting that, because glycosylation ofthe protein is heterogeneous, the monoclonal antibodyrecognizes only restricted glycosylated species of the A1Rprotein, which is in contrast to the polyclonal antiserum.

Next, we examined the immunoreactivity of the 511CAantibody against deglycosylated A1R protein by immuno-blotting. Purified A1R from rat brain membranes wastreated with Flavobacterium meningosepticum-endoglyco-sidase F/N-glycosidase F to remove the sugar residuesbound to the A1R protein. An apparent decrease in themolecular mass of A1R from approximately 35 kDa to 30kDa was observed after deglycosylation (Nakata, 1993).Figure 3 shows that the 511CA antibody recognized boththe native form of A1R (approximately 35 kDa band; Fig. 3,lane 1) and the deglycosylated form (approximately 30kDaband; Fig. 3, lanes 2 and 3).

To confirm the specificity of the 511CA antibody for A1R,we immunostained A1R-transfected CHO cells and DDT1MF-2 smooth muscle cells, which express A1R naturally(Ramkumar et al., 1991; Gerwins and Fredholm, 1995).Intense A1R immunoreactivity with the 511CA antibodywas found in A1R-transfected cells (Fig. 4A); in contrast,no immunoreactivity was observed in nontransfected cells

(Fig. 4C). Figure 4D shows the phase-contrast view of thesame visual field shown in Figure 4C. Dense immunoreac-tivity also was observed in DDT1 MF-2 smooth muscle cells(Fig. 4B). A1R immunoreactivity was not observed inA1R-transfected cells or in DDT1 MF-2 smooth musclecells when the primary antibody was omitted (data notshown). To determine whether or not the 511CA antibodyreacts with other types of adenosine receptors, immuno-staining of PC12 cells that express A2a adenosine recep-tors naturally (Hide et al., 1992) and CHO cells transfectedwith A2b (Pierce et al., 1992) or A3 (Nakata, unpublished)adenosine receptor cDNAs were performed. In these cells,no immunoreactivity was seen (data not shown). Theseresults, combined with those from immunoblotting, indi-cate that the 511CA antibody is highly specific for A1Rsand can detect the native A1R that is expressed naturallyin cells.

Immunohistochemistry

To determine the characteristic distribution pattern ofimmunoreactivity obtained with the monoclonal anti-A1Rantibody (511CA antibody) in the rat forebrain, Table 1lists the relative intensities of A1R labeling within indi-vidual areas or nuclei. A1Rs were expressed in almost allregions of the forebrain; however, regional differences inthe degree of staining were evident in the cerebral neocor-

Fig. 2. Immunoblot analysis for the specificity of the monoclonalanti-A1R antibody (511CA antibody) and the detection of A1R proteinin rat brain. Partially purified A1R from the rat brain membrane wereseparated by SDS-PAGE (lanes 1–3) and immunoblotted. Lane 1 wasprobed by using the 511CA antibody. Lane 2 was probed by usingpreviously characterized polyclonal anti-A1R antiserum raised againstpurified A1R protein (Nakata, 1993). Lane 3 shows the silver-stained,partially purified A1R protein. The arrow indicates the A1R receptor.

Fig. 3. Effect of endoglycosidase F/N-glycosidase F treatment onimmunoreactivity of 511CA antibody for purified A1R protein. Using0.0 U (lanes 1 and 4), 0.1 U (lanes 2 and 5), and 0.25 U (lanes 3 and6) of endoglycosidase F/N-glycosidase F, deglycosylation was per-formed in the reaction mixture with (lanes 1–3) or without (lanes 4–6)A1R protein. The reaction mixtures were then subjected to SDS-PAGEfollowed by immunoblotting with the 511CA antibody. An apparentdecrease in molecular mass of A1R from approximately 35 kDa to 30kDa was observed after the treatment, and the 511CA antibodyrecognized both native and deglycosylated forms of A1R. No immuno-reactivity was observed in the samples that did not include A1R.


Fig. 4. Immunocytochemical analysis of the specificity of the monoclonal anti-A1R antibody(511CA antibody) using Chinese hamster ovarian (CHO) cells that were stably transfected withA1R cDNAs, and DDT1 MF-2 smooth muscle cells. A: Immunostaining of A1R-transfected CHOcells using the 511CA antibody. Intense immunoreactive signals were detected in the cell bodiesof transfected CHO cells. B: Immunostaining of DDT1 MF-2 smooth muscle cells using the

511CA antibody. Immunoreactive signals were detected in the cell bodies of these cells,indicating that this antibody recognized the native A1R present in these cells. C: Immunostain-ing of nontransfected CHO cells using the 511CA antibody. No immunoreactivity was observedin these cells. D: Phase-contrast view of the same visual field shown in C indicating that manynontransfected CHO cells exist in the visual field of C. Scale bars 5 100 µm.

306

T.

OC

HIIS

HI

ET

AL

.

tex and the hippocampal formation. Furthermore, weexamined the subcellular distribution of A1Rs by using thedigital deconvolution system and electron microscopy.When the 511CA antibody was preadsorbed with purifiedA1R and sections were incubated with it, there was noresidual immunoreactivity was observed.

Cerebral neocortex. The typical immunostaining pro-file in the cerebral neocortex is shown in Figure 5. Al-though the findings described here were derived from area2 of the lateral occipital cortex (Oc2L; Fig. 5A–C) and theretrosplenial agranular cortex (RSA; Paxinos and Watson,1986; Fig. 5D), they also are applicable to other areas ofthe cerebral neocortex. A1R immunoreactivity was de-tected in almost all of the pyramidal neurons in corticalcell layers II–VI. However, the intensity of the immunore-activity was distributed heterogeneously among the corti-cal cell layers (Fig. 5A,B). Compared with other layers,layer V neurons were the most intensely stained with thisantibody. The cell bodies, apical dendrites, and basaldendrites of large pyramidal cells of layer V were allstrongly immunostained (Fig. 5B,C). Immunopositive api-cal dendrites could be seen clearly extending from the cellbodies of layer V neurons to layer II (Fig. 5B, arrows). Inaddition to the apical and basal dendrites, the axon initialsegments that extended from the immunopositive largepyramidal cell bodies also were stained intensely (Fig. 5C,arrows). To demonstrate the immunoreactivity of A1Rs inthe large pyramidal neurons in detail, we performedimmunofluorescence labeling of A1Rs and analyzed theresults by using the digital deconvolution system. The

deconvolved three-dimensional images of FITC-labeledA1Rs in layer V neurons (Fig. 8A) showed that the A1Rimmunoreactivity here was distributed as many smallspots on the surfaces of dendrites, cell bodies (Fig. 8A,large arrows), and axon initial segments (Fig. 8A, smallarrows) and in the cytoplasm. The nuclei were immu-nonegative.

The other layers (I–IV and VI) were stained moderatelyor faintly (Fig. 5B). Layer I, which contains many fibers,horizontal cells, and Golgi cells, showed no A1R-immu-nopositive structures. Layers II/III showed moderate im-munoreactivity in the small pyramidal cell bodies. LayersIV and VI showed faint immunoreactivity and smallgranule-type cells that were immunopositive.

In the RSA, the immunoreactivity for A1Rs exhibited aheterogeneous distribution pattern among the cell layers(Fig. 5D) similar to the pattern seen in Oc2L. In layer V,the pyramidal cell bodies and their apical dendrites, basaldendrites, and axon initial segments (Fig. 5D, arrow) werestained strongly by the 511CA antibody. Layers II/III ofRSA also were stained relatively intensely (data notshown).

Basal ganglia. The two main components of basalganglia are the caudate-putamen (CP) and the globuspallidus (GP). The perikarya of neurons in both the CP andthe GP were immunostained by the 511CA antibody (Fig.6). In the CP, these immunopositive neurons were polygo-nal or triangular in shape, with diameters ranging from 10µm to 20 µm (Fig. 6B, arrows), and were embedded in themoderately stained neuropil. Conversely, intensely labeledneurons were present in the GP (Fig. 6C). These immu-nopositive neurons were large fusiform or triangular cellsand had two or three primary dendrites that were long,thick, smooth, and sparsely branched (Fig. 6C, arrows).A1R immunoreactivity was present in the cell bodies anddendrites, but the neuropil around the cell bodies in theGP was stained only faintly. Massive myelinated fiberbundles penetrating the CP and GP also were stained veryweakly by the 511CA antibody.

The subthalamic nucleus and substantia nigra generallyare considered to be a part of the basal ganglia, becausethey are related closely to the striatopallidal neuronalcircuitry. The subthalamic nucleus was immunostainedmoderately by the 511CA antibody (Table 1). The substan-tia nigra was not examined in this study.

Hippocampal formation. Immunohistochemical dis-tribution of A1Rs in the Ammon’s horn and dentate gyrusof the hippocampal formation is shown in Figure 7. Low-power microscopic observation revealed that the stratumpyramidale (SP) of fields CA1–CA4 and the granule celllayer of the dentate gyrus (GrDG) were immunopositivefor A1R (Fig. 7A). Individual cell bodies were stained moreclearly in fields CA2–CA3 than in other fields or in thedentate gyrus. Many A1R-positive interneurons were scat-tered in the stratum oriens (SO), SP, and stratum lacuno-sum-moleculare (SLM) of fields CA1–CA3 and in thedentate gyrus. The mossy fibers were immunonegative.Under higher magnification, regional differences in A1R-immunostaining patterns of pyramidal cells in fields CA1,CA2, and CA3 and of granule cells in the dentate gyrus(Fig. 7B–E, respectively) were observed. In field CA1 (Fig.7B), the pyramidal cells were stained faintly by the 511CAantibody, and the individual cell bodies could be distin-guished only with difficulty. A1R immunoreactivity wasevident in the areas surrounding the pyramidal cell bodies

TABLE 1. Distribution of Adenosine A1 Receptor in Rat Forebrain1

Brain region Staining intensity

Cerebral cortexLayer I 2/1Layer II/III 111 (RSG, RSA: 1111)Layer IV 1Layer V 1111Layer VI 1

Corpus callosum 1Piriform cortex 1111Septum

Lateral septal nucleus, dorsal 1111Lateral septal nucleus, intermediate 1111Fornix 2/1

Basal gangliaCaudate putamen 111Globus pallidus 1111

Hippocampal formationCA1 11CA2 1111CA3 111Dentate gyrus 11Entorhinal cortex 111Subiculum 111

ThalamusCentral group 1Lateral group 111Medial group 1Paracentral nucleus 1Paraventricular nucleus 1111Parafascicular nucleus 111Posterior nucleus group 11Reticular nucleus 111Ventrolateral nucleus 1111Ventromedial nucleus 1111Ventral posterolateral nucleus 1111Ventral posteromedial nucleus 1111Subthalamic nucleus 111

HabenulaMedial nucleus 1Lateral nucleus 1

1The number of (1) indicates the degree of the A1R labeling intensity. Five grades havebeen assigned to the labeling intensity, with 2/1 indicating very weak labeling and1111 indicating the maximum. Only regions with labeling are noted. RSA, retrosple-nical agranular cortex; RSG, retrosplenial granular cortex.


Figure 5


rather than in the cytoplasm. These findings were con-firmed by the deconvolved analysis of FITC-labeled A1Rsby using the three-dimensional digital deconvolution sys-tem (Fig. 8B). This analysis showed that A1R-immunoposi-tive spots were concentrated around the pyramidal cellbodies, producing an outline of these cell bodies, and that afew immunopositive spots were present in the cytoplasm ofpyramidal neurons in field CA1. Figure 7B shows that theapical dendrites extending from the pyramidal cells werestained rather faintly. A1Rs were immunostained in-tensely in many interneurons (Fig. 7B, arrows 1–3) andwere localized at the border of the alveus and SO, withinthe SO, within the SP, or within the SLM. These immu-nopositive interneurons were larger, were either fusiform,triangular, or polygonal in shape, and had many dendrites.Astrocytes also were stained by the 511CA antibody (Fig.7B, arrow). Individual large pyramidal cells were stainedmore clearly by the 511CA antibody in field CA2 (Fig. 7C)than in other fields or the dentate gyrus, and some verydensely stained cell bodies were dispersed throughout thisfield (Fig. 7C, numbered arrows). A peculiar feature in thisfield was the high immunoreactivity of apical dendritesextending from immunopositive pyramidal cells. Denselystained branches of these dendrites could be followed intothe SR and the SLM (Fig. 7C, arrows). In field CA3 (Fig.7D), individual pyramidal cells were stained moderatelyby the 511CA antibody. A1R immunoreactivity was presentin the perikarya and dendritic shafts of the apical den-drites. In the dentate gyrus (Fig. 7E), each granule cellshowed only faint immunoreactivity, therefore, individualcells could not be distinguished. However, the interneu-rons were stained very densely (Fig. 7E, arrows). The cellbodies of the interneurons located at the border betweenthe GrDG and the polymorphic layer of the dentate gyrus(PoDG) and their densely stained dendrites penetratedinto the GrDG. Many polymorphic cells in PoDG also werestained faintly by the 511CA antibody.

Electron microscopic distribution

The subcellular distribution of A1R in the cortical andhippocampal neurons was examined by using electronmicroscopy (Fig. 9). Immunoreactive products associatedwith A1Rs were present diffusely in the cytoplasm of thecell bodies and dendritic shafts of cortical layer V neurons(Fig. 9A,B). These immunoreactive products also were

Fig. 5. A1R immunoreactivity in the cerebral neocortex. A: Low-power photomicrograph of the occipital cortex, area 2, lateral (Oc2L).A1R immunoreactivity was distributed heterogeneously among thecellular layers (layers II–VI). Very intense immunoreactivity wasdetected in layer V, layer II/III was stained moderately, and layers IVand VI were stained faintly by the 511CA antibody. B: High-powerphotomicrograph magnified from the area enclosed by the dashedrectangle in A. Intensely stained, large pyramidal cell bodies, apicaldendrites, and basal dendrites are observed in layer V. Immunoposi-tive apical dendrites extending from the large pyramidal cells can betraced throughout layer II–IV (arrows). C: Photomicrograph showingthe immunopositive large pyramidal neurons of layer V magnifiedfrom B. Neurons indicated by numbers 1–3 correspond to those in B. Inaddition to the apical and basal dendrites, the axon initial segmentsalso were clearly stained by the 511CA antibody (arrows). D: Immuno-reactivity of A1Rs in the retrosplenial agranular cortex (RSA). Theimmunostaining pattern is the same as that in the Oc2L. Intenseimmunoreactivity was detected in layer V neurons. Cell bodies, apicaland basal dendrites, and axon initial segments (arrow) were stainedby the 511CA antibody. Scale bars 5 250 µm in A, 200 µm in B, 50 µmin C, 100 µm in D.

Fig. 6. A–C: A1R immunoreactivity in the basal ganglia. A: Low-power photomicrograph of a coronal section through the caudateputamen (CP) and globus pallidus (GP). The areas enclosed by thedashed rectangles (labeled B and C) are shown at a higher magnifica-tion in B and C. In the CP, moderate A1R immunoreactivity wasdetected in the perikarya of small neurons (B, arrows) and in thesurrounding neuropil. In the GP, very intense A1R immunoreactivitywas detected in the large neuronal cell bodies (C, arrows) and in theirlong, thick, smooth dendrites. Immunoreactivity of the neuropilaround the cell bodies was faint. The fibers passing through thisregions were stained very weakly by the 511CA antibody. Scale bars 5200 µm in A, 100 µm in B,C.


Figure 7

likely to be associated with the membranes of organelles,such as mitochondria and endoplasmic reticulum. Thenucleoplasm was immunonegative. Figure 9B–D showsthe pre- or postsynaptic immunoreactivity of A1Rs in theneurons of the cortex (Fig. 9B,C) and the dentate gyrus ofthe hippocampus (Fig. 9D). A postsynaptic cell body (Fig.9B, large arrow) and a postsynaptic dendrite (Fig. 9C,arrow 1) were immunopositive for the 511CA antibody,whereas another postsynaptic dendrite was immunonega-tive (Fig. 9C, arrow 2). In the positive postsynaptic dendrite,immunoreactive products were likely to be associated withthe round-sliced microtubles. A1R-immunoreactive prod-ucts also were associated in the presynaptic axon terminal(Fig. 9D, T1). Fine A1R-immunoreactive products wereobserved in the cytoplasm of the terminal, whereas an-other terminal (Fig. 9D, T2) was immunonegative, despitethe fact that these two axon terminals terminated at thesame dendrite. These results suggest thatA1Rs are distrib-uted at both pre- and postsynaptic sites in synaptic regionsand that not all the pre- and postsynaptic elements couldbe immunostained with the 511CA antibody.

The present electron microscopic analysis again con-firmed that not only neuronal elements but also glialelements are immunoreactive for the 511CA antibody.Figure 9B shows immunopositive deposits in an astrocyte(Fig. 9B, small arrow).

DISCUSSION

In this study, we developed monoclonal antibodies spe-cific for A1R and confirmed that one of these antibodies(the 511CA antibody) is useful as a probe for brain A1Rs incrude preparations and as a tool for detecting the A1Rreceptor proteins in brain tissue. By light microscopicexamination, using an immunohistochemical technique,we succeeded in demonstrating that A1Rs were expressedmainly on the surfaces of cell bodies, apical dendrites, andbasal dendrites and in the cytoplasm and axon initialsegments (axon hillock) and that there were regional

differences in the distribution of A1Rs in the rat forebrain.Furthermore, electron microscopic analysis revealed thepre- and postsynaptic localization of A1Rs.

Specificity of the monoclonal anti-A1R(511CA) antibody

The 511CA antibody was generated by using highlypurified A1R protein from rat brain as the antigen. The

Fig. 7. A1R immunoreactivity in the hippocampal formation.A: Low-power photomicrograph showing that Ammon’s horn (CA1–CA4 fields) and the dentate gyrus (DG) were immunostained with511CA antibody. Immunoreactivity was detected primarily in thestratum pyramidale (SP) of Ammon’s horn and in the granule celllayer of the dentate gyrus (GrDG). Individual cell bodies and theirapical dendrites in field CA2 were labeled more clearly than those inadjacent field CA1. B–E: High-power photomicrographs of fields CA1(B), CA2 (C), and CA3 (D) and the dentate gyrus (E) magnified fromthe areas enclosed by the dashed rectangles in A. In field CA1 (B), theSP neurons are stained moderately by the 511CA antibody, but theshapes of the individual cell bodies is distinguished only with diffi-culty. The apical dendrites were stained only very faintly by thisantibody. Many interneurons were A1R immunoreactive and werescattered throughout the stratum oriens (SO), SP, and stratumlacunosum-moleculare (SLM; B, arrows 1–3, respectively). In fieldCA2 (C), apical dendrites and their branches extending from theA1R-positive pyramidal cells were stained intensely in the stratumradiatum (SR) and the SLM (C, arrows). Some intensely A1R-immunopositive, large pyramidal cells also were observed (C, neuronsindicated arrows 1–3). In field CA3 (D), moderately A1R-labeled, largepyramidal cells were detected. Each positive cell was distinguishable,and their thick dendritic shafts also were stained faintly. In thedentate gyrus (E), faint A1R immunoreactivity was detected in theGrDG, and the individual cell bodies in the granule cell layer were notdistinguishable. However, intensely A1R-immunolabeled interneu-rons and their dendrites that penetrated into the GrDG were detected(E, arrows). Polymorphic cells also were stained faintly. Astrocytesalso were A1R immunopositive (B and D, arrows). For other abbrevia-tions, see list. Scale bars 5 500 µm in A, 100 µm in B–E.

Fig. 8. Deconvolved, three-dimensional views of fluorescein isothio-cyanate (FITC)-labeled A1Rs in the cerebral neocortex (A) and thehippocampus (B). A: Large pyramidal neurons of layer V in Oc2L.Immunoreactivity was distributed as small spots on the surfaces ofdendrites (large arrows), cell bodies, and axon initial segments (smallarrows) and in the cytoplasm. Nuclei were immunonegative. B: CA1pyramidal neurons of the hippocampus. Immunopositive spots wereconcentrated in the areas surrounding the pyramidal cell bodies(arrows), with a few detected in the cytoplasm. Scale bars 5 20 µm.


following lines of evidence suggest that this antibodyrecognized A1Rs specifically: 1) In immunoblot analyses,this antibody recognized a solitary band the size of whichcorresponded to that expected for A1Rs from rat brain

membrane (Figs. 1, 2); 2) the cell staining revealed thatthis antibody recognizes A1Rs expressed in CHO cells (Fig.4) and also the native A1Rs present in DDT1 MF-2 smoothmuscle cells, which naturally express A1R (Fig. 4), but

Fig. 9. Electron photomicrographs of A1R-immunoreactivity in theneurons in the sensory-motor cortex (A–C) and the hippocampus (D).A: Dense A1R-immunoreactive products were distributed in theperikarya and dendrites of layer V large pyramidal neurons. Noimmunoreactivity was detected in their nuclei (Nuc). B: High-powerphotomicrograph magnified from the area enclosed by the dashedrectangle in A. Dense accumulations of immunoreactive products wereobserved in a postsynaptic cell body (large arrow). The products werelikely to be associated with the membranes of cellular organelles, suchas mitochondria and endoplasmic reticulum. An astrocyte also was

immunopositive (small arrow). C: This electron photomicrographshows the A1R-immunopositive (arrow 1) and A1R-immunonegative(arrow 2) postsynaptic elements, indicating that not all of the postsyn-aptic sites are immunopositive for A1R. d, Dendrite. D: This electronphotomicrograph shows the A1R- immunopositive (presynaptic termi-nal 1; T1) and A1R-immunonegative (T2) presynaptic axon terminals,indicating that not all of the presynaptic sites are immunopositive forA1R. For abbreviations, see list. Scale bars 5 2 µm in A, 0.5 µm in B,D,2.5 µm in C.


does not exhibit cross reactivity for antigens related toother types of adenosine receptors, such as A2a, A2b, andA3 adenosine receptors (data not shown); and 3) thepreadsorption test revealed that specific labeling of A1Rsin brain sections was not observed when the antibody waspreadsorbed with highly purified A1R (data not shown).

In immunoblot analyses, the band pattern detected bythe 511CA antibody had the characteristically broad ap-pearance of that of a glycoprotein (Stiles, 1986; Nakata,1990). Deglycosylation treatment revealed that this anti-body recognized both nondeglycosylated and enzymati-cally deglycosylated forms of A1R, suggesting that thisantibody recognizes the native form of A1R, a glycoprotein,and that the recognition sites of this antibody are notlocated on the bound sugar residue. These results demon-strate that the 511CA antibody developed in this study wasspecific for A1Rs.

Comparison with previous ligand-binding,in situ hybridization, and

immunohistochemical studies

Although immunoreactivity of A1Rs was present inalmost all regions of the forebrain, there also were regionaldifferences in the staining intensities. The pattern ofimmunostaining with the 511CA antibody was consistentwith the pattern of A1R mRNA distribution (Reppert et al.,1991; Swanson et al., 1995). For example, in the hippocam-pus, both in situ hybridization and immunohistochemicalstudies revealed heavy labeling of A1R in the SP ofAmmon’s horn and the GrDG in the dentate gyrus andlittle specific labeling over the SO, SR, SLM, and dentatehilus was noted. However, there is a discrepancy in A1Rdistribution when comparing ligand-binding studies (Good-man et al., 1983; Deckert and Jorgensen, 1988; Dragunowet al., 1988; Swanson et al., 1995) with our immunohisto-chemical study. Our immunohistochemical images werethe inverse of the ligand-binding regions, i.e., ligand-binding regions were observed in the SO, SR, and SLM ofAmmon’s horn and not in the SP. However, the mismatchesbetween immunohistochemical localization sites or mRNAexpression sites and ligand-binding sites also have beenreported for other receptors (Ross et al., 1994). The specific-ity of monoclonal antibodies usually is higher than that ofligands. Furthermore, ligand affinities are affected byseveral factors, such as the concentration of guaninenucleotides and the presence of divalent cations, sodiumions, and magnesium ions in the incubation media duringthe reaction between ligands and A1Rs (Goodman et al.,1982; Frotscher and Leranth, 1986; Fastbom and Fred-holm, 1987, 1990; Olah and Stiles, 1990; Parkinson andFredholm, 1992). Therefore, some discrepancies betweenthe results of ligand-binding studies and our immunohisto-chemical study are expected, because ligand affinities wereinfluenced by the factors described above.

Previously, Rivkees et al. (1995) developed a polyclonalantiserum raised against synthetic polypeptides of A1R(hereinafter referred to as the A1R peptide antibody) anddemonstrated immunohistochemically the distribution ofA1Rs in rat forebrain by using this antibody. Comparingtheir results using the A1R peptide antibody and ourresults using the 511CA antibody, similarities were ob-served in the cerebral neocortex, but differences wereobserved in the hippocampus results. In our study, the A1Rimmunoreactivity in the hippocampus was present mainlyin the SP, with regional differences in labeling, and many

A1R-immunopositive interneurons were detected in fieldsCA1–CA3 and in the dentate gyrus. The axons were notdetected by using our monoclonal antibody. In contrast,Swanson et al. (1995), by using the A1R peptide antibody,found few A1Rs in cellular sites in the entire Ammon’shorn and dentate gyrus and found that axons were labeledvery densely, whereas interneurons were immunonega-tive. Because the fixation and labeling techniques used inthese studies were almost identical to those used here, theinconsistencies in the immunostaining pattern observedbetween the studies using these two antibodies can beexplained by the differences in specificity of each of theseantibodies for the A1Rs in brain tissue. Our study wasbased on the use of a monoclonal antibody that wasprepared by using highly purified A1R from rat brain,specifically reacting with A1R protein, whereas Swansonet al. used A1R peptide antibody. Immunoblot analysis forthe A1R peptide antibody (Rivkees et al., 1995) showedthat the immunoreactive band was very narrow and waslocated in a slightly different position (38 kDa) than theband detected by our monoclonal antibody. Another expla-nation for the inconsistency is the possibility of the exis-tence of A1R subtypes, i.e., the monoclonal antibody usedin our study and the A1R peptide antibody possiblyrecognize different subtypes of A1R. Previous ligand-labeling analyses revealed a similar but slightly differentpattern of distribution between agonist (Lee and Redding-ton, 1986) and the antagonist labeling (Fastbom andFredholm, 1990; Swanson et al., 1995). These results madeus anticipate the existence of such discrepancies possiblyreflecting the presence of A1R subtypes, which may havespecific expression patterns in the axons or somata, al-though there is no evidence for the existence of suchreceptor subtypes for A1R, despite the many attempts thathave been made to subclassify A1Rs (Linden, 1991; Coateset al., 1994; Dalziel and Westfall, 1994; Olah and Stiles,1995).

In the monkey hippocampus, a higher level of A1Rexpression was found in CA2 pyramidal neurons and theirdendrites than in rat hippocampus using the 511CA anti-body (data not shown). These results suggest that theneurons in this field may play a subtle but important rolein regulating hippocampal activity through A1Rs in verte-brates.

Subcellular distribution

A1R usually is thought to be localized on the membranesurface to function as a receptor (Olah and Stiles, 1995).Detailed electron microscopic examination revealed thatA1R is present at high density throughout the somatoden-dritic cytoplasm of the cortical neurons (Fig. 9A). The samedistribution patterns also were detected for other recep-tors, such as the receptor subunits of g-aminobutyric acidA (GABAA; Richards et al., 1987; Somogyi et al., 1989;Baude et al., 1992), nicotinic receptors (Hill et al., 1993),and subunits of ionotropic glutamate receptors (Jaarsmaet al., 1995). Although the significance of the cytoplasmicstaining is not understood fully, the immunoreactivity mayreflect receptors in transit to or from the synaptic mem-brane and/or an internal storage pool of receptor precur-sors (Stollberg and Berg, 1987). Therefore, the cytoplasmicimmunoreactivity of A1Rs in our study may indicates thatthe 511CA antibody recognizes the A1R protein that is justbeing transported from the cytoplasm to the membrane


surface after the biosynthesis of the receptor protein in thecytoplasm, which might not have ligand-binding activity.

By using electron microscopic analysis, cytoplasmic A1R-immunoreactive products were likely to be associated withstructures, such as the microtubles and outer mitochon-drial membranes. These structures are not likely to bearreceptor proteins. An explanation for this inappropriatestaining inside the cells may be that the strong immunore-activity of the 511CA antibody-binding region leads to theformation of large amounts of reaction products thatdiffuse for some distance from the binding region.

Presynaptic and postsynapticdistribution of A1Rs

One of the principal findings of the present study is theresult from electron microscopic analysis of the denseaccumulation of immunoreactive products reflecting thedistribution of A1Rs in the intracellular sites of bothpresynaptic terminals and postsynaptic structures in thecortical and hippocampal neurons (Fig. 9B–D). In electronmicroscopic autoradiographic studies using an A1R ago-nist, Tetzlaff et al. (1987) showed that significant numbersof silver grains representing the binding sites of theagonist accumulated at synaptic clefts and dendritic mem-branes, but not at preterminal axon membranes. However,their report was unclear regarding the existence of agonist-binding sites at the presynaptic membrane. The differ-ences between the results of our study and those ofprevious autoradiographic studies (both employing elec-tron microscopy) can be explained by the differences inaffinities between the antibodies and ligands for the A1Rs,as described above. Another explanation for the differencesis the difficulty of radiohistochemical studies at an ultra-structural level, because the technique still suffers fromlimitations in resolution, namely, radiation scatter, diffu-sion, and nonspecific binding of the radioligand. Receptorlocalization by immunohistochemistry overcomes the limi-tations in resolution of the radiohistochemical technique.However, diffusion of the enzyme (DAB) reaction productsfrom the strong immunoreactive region also conceivablymay occur in immunohistochemical analysis as describedabove. Although, in our immunohistochemical study, PSDand the pre-and postsynaptic membranes in the cerebralcortex were likely to be A1R immunopositive, the immuno-reactivity of these sites will have to be confirmed by the useof gold-labeled secondary antibodies as a marker.

By using the A1R peptide antibody, Swanson et al.(1995) showed that presynaptic terminals were A1R-immunonegative, even though the axons were immunoposi-tive. Our results, demonstrating the presynaptic expres-sion of A1Rs by electron microscopy, are the firstimmunohistochemical evidence supporting the physiologi-cal premise that A1Rs that exert an inhibitory influence onneurotransmitter release may exist in presynaptic termi-nals. Our results also support the hypothesis of Fredholmand Dunwiddie (1988) that the same A1R type exertsdifferent functions in pre- and postsynaptic sites by cou-pling with different types of G-proteins.

A1R immunoreactivity in hippocampalinterneurons

Many interneurons in Ammon’s horn and the dentategyrus of the hippocampus with diverse morphologies anddistribution sites expressed A1Rs as recognized by the511CA antibody. This is the first immunohistochemical

evidence of A1R expression in interneurons. Hippocampalinterneurons are divided into many groups according tothe location of their cell bodies, their dendritic and axonalarborizations, their afferent and efferent connections inthe hippocampus, and the neurotransmitters they contain(Freund and Buzaki, 1996). A1R immunoreactivity inhippocampal interneurons was observed mainly in the cellbodies and proximal dendritic arbors, but not in axons.This selective staining makes it difficult to classify theseA1R-immunopositive neurons according to their morpholo-gies. These immunopositive interneurons could be classi-fied to a certain extent, however, according to the positionof their cell bodies. The interneurons in the SO or SP maybe divided into basket cells, chandelier cells, or O-LM cellswhich are interneurons with soma and dendrites in SOand axons in SO and SLM (see Freund and Buzaki, 1996).

Previous physiological studies suggest that the interneu-rons in the CA1 region effect presynaptic inhibition throughthe release of adenosine from their presynaptic terminalswhen activated by glutamate through N-methyl-D-aspar-tate receptors (Manzoni et al., 1994). The A1Rs expressedin interneurons may subtly control the CA1 presynapticinhibition by modulating the presynaptic release of adeno-sine or other neurotransmitters. To clarify the classifica-tion of A1R-expressing interneurons and to further investi-gate the function of A1Rs in these neurons, we mustexamine the colocalization of A1Rs with other neurotrans-mitters or with calcium-binding proteins, such as GABA,acetylcholine, and parvalbumin, that are expressed inhippocampal interneurons in a specific distribution pat-tern (Frotscher and Leranth, 1986; Kosaka et al., 1987;Freund and Somogyi, 1989; Somogyi and Freund, 1989).

CONCLUSIONS

In this study, we succeeded in developing a monoclonalanti-A1R antibody for highly purified A1R from rat brainand, by using immunohistochemical techniques, we demon-strated the regional differences in the distribution of A1Rsin the rat forebrain. Our observations confirm the expres-sion of A1R in interneurons of the hippocampus and inboth pre- and postsynaptic structures. The finding ofsynaptic distribution is the first immunohistochemicalevidence supporting the physiological surmise that A1R islocalized at presynaptic terminals to inhibit the release ofneurotransmitters or at postsynaptic sites to inhibit theexcitation of postsynaptic cells. At present, the significanceof the regional differences in A1R immunoreactivity re-mains unclear. The precise structural localization of A1Rsrevealed in this study is important for further clarifica-tions of the modulatory effects of adenosine in variousbrain regions.

ACKNOWLEDGMENTS

We thank Dr. M. Kaizu, Science University of Tokyo, forhis excellent advice on the production of monoclonalantibody. We also thank, Professor T. Terashima (KobeUniversity School of Medicine) and Dr. M. Ichikawa (TokyoMetropolitan Institute for Neuroscience) for valuable dis-cussions.

LITERATURE CITED

Arai T, Matsumoto G. 1988. Axolinin localization in the nervous tissue ofsquid revealed by monoclonal antibodies specific for axolinin: character-


ization of monoclonal antibodies against axolinin. Hybridoma 7:583–593.

Baude A, Sequier J-M, McKernan RM, Olivier KR, Somogyi P. 1992.Differential subcellular distribution of the a6 subunit vs. the a1 andb2/3 subunits of the GABAA/benzodiazepine receptor complex in gran-ule cells of the cerebellar cortex. Neuroscience 51:739–748.

Coates J, Gurden MF, Harris C, Kennedy I, Sheehan MJ, Strong P. 1994.Adenosine receptor classification: quo vadimus? Nucleosides Nucleo-tides 13:1953–1976.

Dalziel HH, Westfall DP. 1994. Receptors for adenine nucleotides andnucleosides: subclassification, distribution, and molecular characteriza-tion. Pharmacol Rev 46:449–466.

Deckert J, Jorgensen MB. 1988. Evidence for pre- and postsynapticlocalization of adenosine A1 receptors in the CA1 region of rat hippocam-pus: a quantitative autoradiographic study. Brain Res 446:161–164.

de Mendonca A, Ribeiro JA. 1993. Adenosine inhibits the NMDA receptor-mediated excitatory postsynaptic potential in the hippocampus. BrainRes 606:351–356.

de Mendonca A, Ribeiro JA. 1994. Endogenous adenosine modulateslong-term potentiation in the hippocampus. Neuroscience 62:385–390.

Dragunow M, Murphy K, Leslie RA, Robertson HA. 1988. Localization ofadenosine A1-receptors to the terminals of the perforant path. BrainRes 462:252–257.

Dunwiddie TV, Fredholm BB. 1989. Adenosine A1 receptors inhibit adenyl-ate cyclase activity and neurotransmitter release and hyperpolarizepyramidal neurons in rat hippocampus. J Pharmacol Exp Ther 249:31–37.

During MJ, Spencer DD. 1992. Adenosine: a potential mediator of seizurearrest and postictal refractoriness. Ann Neurol 32:618–624.

Fastbom J, Fredholm BB. 1987. Regional differences in the GTP-dependence of adenosine receptor binding in rat brain? Acta PhysiolScand 131:467–469.

Fastbom J, Fredholm BB. 1990. Regional differences in the effect of guaninenucleotides on agonist and antagonist binding to adenosine A1-receptors in rat brain, as revealed by autoradiography. Neuroscience34:759–769.

Fredholm BB, Dunwiddie TV. 1988. How does adenosine inhibit transmit-ter release? Trends Pharmacol Sci 9:130–134.

Freund T F, Buzaki G. 1996. Interneurons of the hippocampus. Hippocam-pus 6:347–470.

Freund TF, Somogyi P. 1989. Synaptic relationships of Golgi impregnatedneurons as identified by electrophysiological or immunocytochemicaltechniques. In: Heimer L, Zaborszky L, editors. Neuroanatomicaltract-tracing methods, vol. 2: recent progress. New York: Plenum Press,p 201–238.

Frotscher M, Leranth C. 1986. The cholinergic innervation of the rat fasciadentata: identification of target structures on granule cells by combin-ing choline acetyltransferase immunocytochemistry and Golgi impreg-nation. J Comp Neurol 243:58–70.

Gerber U, Greene RW, Haas HL, Stevens DR. 1989. Characterization ofinhibition mediated by adenosine in the hippocampus of the rat in vitro.J Physiol 417:567–578.

Gerwins P, Fredholm BB. 1995. Activation of adenosine A1 and bradykininreceptors increase protein kinase C and phospholipase D activity insmooth muscle cells. Naunyn-Schmiedeberg’s Arch Phamacol 351:186–193.

Goodman RR, Cooper MJ, Gavish M, Snyder SH. 1982. Guanine nucleotideand cation regulation of the binding of [3H] cyclohexyladenosine and[3H] diethylphenylxanthine to adenosine A1 receptors in brain mem-branes. Mol Pharmacol 21:329–335.

Goodman RR, Kuhar MJ, Hester L, Snyder SH. 1983. Adenosine receptors:autoradiographic evidence for their location on axon terminals ofexcitatory neurons. Science 220:967–969.

Hide I, Padgett WL, Jacobson KA, Daly JW. 1992. A2A adenosine receptorsfrom rat striatum and rat pheochromocytoma PC12 cells: characteriza-tion with radioligand binding and by activation of adenylate cyclase.Mol Pharmacol 41:352–359.

Hill JA, Zoli M, Bourgeois J-P, Changeux J-P. 1993. Immunocytochemicallocalization of a neuronal nicotinic receptor: the b2-subunit. J Neurosci13:1551–1568.

Hsu S-M, Raine L, Fanger H. 1981. Use of avidin-biotin-peroxidase complex(ABC) in immunoperoxidase techniques: a comparison between ABCand unlabeled antibody (PAP) procedures. J Histochem Cytochem29:577–580.

Hui SW, Stenger DA. 1993. Electrofusion of cells: hybridoma production byelectrofusion and polyethylene glycol. Methods Enzymol 220:212–227.

Jaarsma D, Wenthold RJ, Mugnaini E. 1995. Glutamate receptor subunitsat mossy fiber-unipolar brush cell synapses: light and electron micro-scopic immunocytochemical study in cerebellar cortex of rat and cat. JComp Neurol 357:145–160.

Klishin A, Lozovaya N, Krishtal O. 1995. A1 adenosine receptors differen-tially regulate the N-methyl-D-aspartate and non-N-methyl-D-aspar-tate receptor-mediated components of hippocampal excitatory postsyn-aptic current in a Ca21/Mg21-dependent manner. Neuroscience 65:947–953.

Kosaka T, Katsumaru H, Hama K, Wu JY, Heizmann CW. 1987. GABAergicneurons containing the Ca21-binding protein parvalbumin in the rathippocampus and dentate gyrus. Brain Res 419:119–130.

Laemmli U K. 1970. Cleavage of structural proteins during the assembly ofthe head of bacteriophage T4. Nature 227:680–685.

Lee KS, Reddington M. 1986. Autoradiographic evidence for multiple CNSbinding sites for adenosine derivatives. Neuroscience 19:535–549.

Lee KS, Schubert P, Reddington M, Kreutzberg GW. 1986. The distributionof adenosine A1 receptors and 58-nucleotidase in the hippocampalformation of several mammalian species. J Comp Neurol 246:427–434.

Libert F, Lefort A, Okimoto R, Womack J, Georges M. 1993. Construction ofa bovine genomic library of large yeast artificial chromosome clones.Genomics 18:270–276.

Linden J. 1991. Structure and function of A1 adenosine receptors. FASEB J5:2668–2676.

Londos C, Cooper DME, Wolff J. 1980. Subclasses of external adenosinereceptor. Proc Natl Acad Sci USA 77:2551–2554.

Lupica CR, Proctor WR, Dunwiddie RV. 1992. Presynaptic inhibition ofexcitatory synaptic transmission by adenosine in rat hippocampus:analysis of unitary EPSP variance measured by whole-cell recording. JNeurosci 12:3753–3764.

Mahan LC, McVittie LD, Smyk-Randall EM, Nakata H, Monsma FJ Jr.,Gerfen CR, Sibley DR. 1991. Cloning and expression of an A1 adenosinereceptor from rat brain. Mol Pharmacol 40:1–7.

Manzoni OJ, Manabe T, Nicoll RA. 1994. Release of adenosine by activationof NMDA receptors in the hippocampus. Science 265:2098–2101.

Nakata H. 1989a. Affinity chromatography of A1 adenosine receptors of ratbrain membranes. Mol Pharmacol 35:780–786.

Nakata H. 1989b. Purification of A1 adenosine receptor from rat brainmembranes. J Biol Chem 264:16545–16551.

Nakata H. 1990. A1 adenosine receptor of rat testis membranes. Purifica-tion and partial characterization. J Biol Chem 265:671–677.

Nakata H. 1993. Development of an antiserum to rat-brain A1 adenosinereceptor: application for immunological and structural comparison of A1adenosine receptors from various tissues and species. Biochim BiophysActa 1177:93–98.

Norris JS, Gorski J, Kohler PO. 1974. Androgen receptors in a Syrianhamster ductus deferens tumor cell line. Nature 248:422–424.

Ochiishi T, Terashima T, Yamauchi T. 1994. Specific distribution of Ca21/calmodulin-dependent protein kinase II a and b isoforms in somestructures of the rat forebrain. Brain Res 659:179–193.

Ohuchi T, Katoh N, Ueno T, Arai T. 1994. Application of image processing toscreening and characterizing of monoclonal antibodies recognizingPC12 cell surface antigens. Bioimages 2:101–109.

Olah M, Stiles GL. 1990. Agonists and antagonists recognize different butoverlapping populations of A1 adenosine receptors: modulation ofreceptor number by MgC12, solubilization, and guanine nucleotides. JNeurochem 55:1432–1438.

Olah ME, Stiles GL. 1995. Adenosine receptor subtypes: characterizationand therapeutic regulation. Annu Rev Pharmacol Toxicol 35:581–606.

Parkinson FE, Fredholm BB. 1992. Magnesium-dependent enhancement ofendogenous agonist binding to A1 adenosine receptors: a complicatingfactor in quantitative autoradiography. J Neurochem 58:941–950.

Paxinos G, Watson C. 1986. The rat brain in stereotaxic coordinates, 2nded. San Diego: Academic Press.

Pierce KD, Furlong TJ, Selbie LA, Shine J. 1992. Molecular cloning andexpression of an adenosine A2b receptor from human brain. BiochemBiophys Res Commun 187:86–93.

Prince DA, Stevens CF. 1992. Adenosine decreases neurotransmitterrelease at central synapses. Proc Natl Acad Sci USA 89:8586–8590.

Proctor WR, Dunwiddie TV. 1983. Adenosine inhibits calcium spikes inhippocampal pyramidal neurons in vitro. Neurosci Lett 35:197–201.

Ramkumar V, Olah ME, Jacobson KA, Stiles GL. 1991. Distinct pathwaysof desensitization of A1- and A2-adenosine receptors in DDT1 MF-2 cells.Mol Pharmacol 40:639–647.

Reppert SM, Weaver DR, Stehle JH, Rivkees SA. 1991. Molecular cloning


and characterization of a rat A1-adenosine receptor that is widelyexpressed in brain and spinal cord. Mol Endocrinol 5:1037–1048.

Richards JG, Schoch P, Haring P, Takacs B, Mohler H. 1987. ResolvingGABA-A/benzodiazepine receptors: cellular and subcellular localizationin the CNS with monoclonal antibodies. J Neurosci 7:1866–1886.

Rivkees SA, Price SL, Zhou FC. 1995. Immunohistochemical detection of A1adenosine receptors in rat brain with emphasis on localization in thehippocampal formation, cerebral cortex, cerebellum, and basal ganglia.Brain Res 677:193–203.

Rolink AG, Melchers F, Palacios R. 1989. Monoclonal antibodies reactivewith the mouse interleukin 5 receptor. J Exp Med 169:1693–1701.

Ross AH, Lachyankar MB, Poluha DK, Loy R. 1994. Axonal transport of thetrkA high-affinity NGF receptor. Progr Brain Res 103:15–21.

Scholz KP, Miller RJ. 1991. Analysis of adenosine actions on CA21 currentsand synaptic transmission in cultured rat hippocampal pyramidalneurons. J Physiol 435:373–393.

Segal M. 1982. Intracellular analysis of a postsynaptic action of adenosinein the rat hippocampus. Eur J Pharmacol 79:193–199.

Siggins GR, Schubert P. 1981. Adenosine depression of hippocampalneurons in vitro: an intracellular study of dose-dependent actions onsynaptic and membrane potentials. Neurosci Lett 23:55–60.

Snowhill EW, Williams M. 1986. [3H]cyclohexyladenosine binding in ratbrain: a pharmacological analysis using quantitative autoradiography.Neurosci Lett 68:41–46.

Somogyi P, Freund TF. 1989. Immunocytochemistry and synaptic relation-ships of physiologically characterized, HRP-filled neurons. In: HeimerL, Zaborszky L, editors. Neuroanatomical tract-tracing methods, vol. 2:recent progress. New York: Plenum Press, p 239–264.

Somogyi P, Takagi H, Richards JG, Mohler H. 1989. Subcellular localizationof benzodiazepine/GABAA receptors in the cerebellum of rat, cat, andmonkey using monoclonal antibodies. J Neurosci 9:2197–2209.

Stiles GL. 1986. Photoaffinity cross-linked A1 adenosine receptor-bindingsubunits: homologous glycoprotein expression by different tissues. JBiol Chem 261:10839–10843.

Stiles GL. 1992. Adenosine receptors. J Biol Chem 267:6451–6454.Stocker JW, Forster HK, Miggiano V, Stahli C, Staiger G, Takacs B,

Staehelin T. 1982. Generation of 2 new mouse myeloma cell lines ‘‘PAI’’and ‘‘PAI-O’’ for hybridoma production. Res Disclosure 217:155–157.

Stollberg J, Berg DK. 1987. Neuronal acetylcholine receptors: fate ofsurface and internal pools in cell culture. J Neurosci 7:1809–1815.

Swanson TH, Drazba JA, Rivkees SA. 1995. Adenosine A1 receptors arelocated predominantly on axons in the rat hippocampal formation. JComp Neurol 363:517–531.

Tetzlaff W, Schubert P, Kreutzberg GW. 1987. Synaptic and extrasynapticlocalization of adenosine binding sites in the rat hippocampus. Neurosci-ence. 21:869–875.

Thompson SM, Haas HL, Gahwiler BH. 1992. Comparison of the actions ofadenosine at pre- and postsynaptic receptors in the rat hippocampus invitro. J Physiol 451:347–363.

Townsend-Nicholson A, Shine J. 1992. Molecular cloning and characterisa-tion of a human brain A1 adenosine receptor cDNA. Mol Brain Res16:365–370.

Trussell LO, Jackson MB. 1987. Dependence of an adenosine-activatedpotassium current on a GTP-binding protein in mammalian centralneurons. J Neurosci 7:3306–3316.

Weber RG, Jones CR, Lohse MJ, Palacios JM. 1990. Autoradiographicvisualization of A1 adenosine receptors in rat brain with [3H] 8-cyclo-pently-1, 3 dipropylxanthine. J Neurochem 54:1344–1353.

Wu L, Saggau P. 1994. Adenosine inhibits evoked synaptic transmissionprimarily by reducing presynaptic calcium influx in area CA1 ofhippocampus. Neuron 12:1139–1148.

Yamamoto C, Sawada S, Ohno-Shosaku T. 1993. Quantal analysis ofmodulating action of adenosine on the mossy fiber synapse in hippocam-pal slices. Hippocampus 3:87–92.

Yoshida K, Arai T, Yokota T, Komatsu N, Miura Y, Yanagisawa K, Tetsuka T,Tominaga S. 1995. Studies on natural ST2 gene products in the humanleukemic cell line UT-7 using monoclonal antihuman ST2 antibodies.Hybridoma 14:419–427.


Documents

Cellular localization of adenosine A1 receptors in rat forebrain: Immunohistochemical analysis using adenosine A1 receptor-specific monoclonal antibody