Immunohistochemical andEnzyme-Histochemical Study on theAccessory Olfactory Bulb of the Dog

TAKAYUKI NAKAJIMA, MOTOHARU SAKAUE, MIHOKO KATO,SHOUICHIRO SAITO, KAZUSHIGE OGAWA, AND KAZUYUKI TANIGUCHI*

Department of Veterinary Anatomy, Faculty of Agriculture, Iwate University,Iwate 020-8550, Japan

ABSTRACTThe accessory olfactory bulb (AOB) is a primary center of the vomerona-

sal system. In the dog, the position and morphology of the AOB remainedvague for a long time. Recently, the morphological characteristics of the dogAOB were demonstrated by means of lectin-histochemical, histological, andimmunohistochemical staining, although the distribution of each kind ofneuron, especially granule cells, remains controversial in the dog AOB. Inthe present study, we examined the distribution of neuronal elements in thedog AOB by means of immunohistochemical and enzyme-histochemicalstaining. Horizontal paraffin or frozen sections of the dog AOB wereimmunostained with antisera against protein gene product 9.5 (PGP 9.5),brain nitric oxide synthase (NOS), glutamic acid decarboxylase (GAD),tyrosine hydroxylase (TH), substance P (SP), and vasoactive intestinalpolypeptide (VIP) by avidin-biotin peroxidase complex method. In addition,frozen sections were stained enzyme-histochemically for NADPH-diapho-rase. In the dog AOB, vomeronasal nerve fibers, glomeruli, and mitral/tuftedcells were PGP 9.5-immunopositive. Mitral/tufted cells were observed in theglomerular layer (GL) and the neuronal cell layer (NCL). In the NCL, asmall number of NOS-, GAD-, and SP-immunopositive and NADPH-diaphorase positive granule cells were observed. In the GL, GAD-, TH-, andVIP-immunopositive periglomerular cells were observed. In the GL and theNCL, TH-, and VIP-immunopositive short axon cells were also observed. Inaddition to these neurons, TH- and SP-immunopositive afferent fibers wereobserved in the GL and the NCL. We could distinctly demonstrate thedistribution of neuronal elements in the dog AOB. Since only a small numberof granule cells were present in the dog AOB, the dog AOB did not displaysuch a well-developed GCL as observed in the other mammals. Anat. Rec.252:393–402, 1998. r 1998 Wiley-Liss, Inc.

Key words: accessory olfactory bulb; enzyme-histochemistry; immuno-histochemistry; dog

The vomeronasal system appears from amphibian genea-logically (Crosby and Hamphrey, 1939; Hoffman, 1963)and is known to be associated with reproductive and socialbehaviors (Halpern, 1987; Wysocki, 1979). The primarycenter of this system is the accessory olfactory bulb (AOB),which receives projections from the vomeronasal organ(McCotter, 1912; Scalia and Winans, 1975).

In the rodents, the AOB is well-developed and dividedinto five laminae: vomeronasal nerve layer (VNL); glomeru-lar layer (GL); mitral/tufted cell layer (MTL); lateralolfactory tract (LOT); and granule cell layer (GCL; Takami

et al., 1992). The mitral/tufted cells—the output neuronsof the AOB—make synaptic contacts with axons of thevomeronasal receptor neurons within the glomeruli(Takami and Graziadei, 1990, 1991). The periglomerularand granule cells—the interneurons of the AOB—control

*Correspondence to: Dr. Kazuyuki Taniguchi, Department ofVeterinary Anatomy, Faculty of Agriculture, Iwate University,3-18-8 Ueda, Morioka, Iwate 020-8550, Japan.

Received 17 March 1998; Accepted 28 May 1998

THE ANATOMICAL RECORD 252:393–402 (1998)

r 1998 WILEY-LISS, INC.

excitation of the mitral/tufted cells (Macrides and Davis,1983). A small number of short axon cells—a kind ofinterneuron—are also present in the AOB (Alonso et al.,1995; Nakajima et al., 1996; Porteros et al., 1994; Sanides-Kohlrausch and Wahle, 1990).

In the dog, however, the position and morphology of theAOB remained vague for a long time (Cajal, 1902; McCot-ter, 1912, 1913; Miodonski, 1968). Recently, Salazar et al.(1992) first demonstrated the dog AOB clearly by lectin-histochemistry, a histochemical method for detection ofspecific sugar residues on the cell surface. Furthermore,Salazar et al. (1994) showed that the dog AOB wascomprised of a thick layer containing glomeruli and thin-ner layer containing a number of different cell types andthat its structure was simpler than that of the rodent AOB,although the distribution of each kind of neuron, especiallygranule cells, remains controversial in the dog AOB.

In the present study, we examined the distribution ofneuronal elements in the dog AOB by means of immunohis-tochemical and enzyme-histochemical staining to betterunderstand the general organization of processing of phero-monal signals in mammalian AOB.

MATERIALS AND METHODSTissue Preparation

Eleven adult mongrel dogs of either sex weighing 9.5–15.5 kg were used in the present study. The animals weredeeply anesthetized by intravenous injection of sodiumpentobarbital (60 mg/kg body weight; Abbott, North Chi-cago, IL) and perfused through both right and left commoncarotid arteries with physiological saline, followed byBouin’s solution without acetic acid for paraffin sections orby 4% paraformaldehyde-0.1 M phosphate buffer (PB; pH7.4) for frozen sections. For paraffin sections, the olfactorybulbs were removed from four dogs, immersed in the samefixative as in the perfusion for 24–48 hr, and embedded inparaffin. Paraffin sections were cut horizontally at 5 µm.For frozen sections, the olfactory bulbs from seven dogswere immersed in the same fixative as in the perfusion for5–6 hr at 4°C, and transferred to 30% sucrose-0.1 M PB(pH 7.4). Frozen sections were cut horizontally at 6–8 µmon a cryostat, thaw-mounted on slide glasses coated with3-aminopropyltriethoxysilane (Shin-etsu Chemicals, Tokyo,Japan), and stored at 220°C until use.

ImmunohistochemistrySix kinds of antisera were used in the present study.

These antisera contained rabbit polyclonal antibodiesagainst protein gene product 9.5 (PGP 9.5; 1:2,000; Ultra-Clone, Wellow, UK), brain nitric oxide synthase (NOS;1:1,000; Euro-Diagnostica, Malmo, Sweden), glutamic aciddecarboxylase (GAD; 1:4,000; Chemicon, Temecula, CA),tyrosine hydroxylase (TH; 1:1,000; Chemicon), substanceP (SP; 1:2,000; Incstar, Stillwater, MN), and vasoactiveintestinal polypeptide (VIP; 1:2,000; Incstar). We selectedthese antisera in the present study, since PGP 9.5, NOS,GAD, TH, SP, and VIP were previously demonstrated to belocalized in each kind of neuron in mammalian olfactorybulb (Baker, 1986; Davis and Macrides, 1983; Gall et al.,1986; Kishimoto et al., 1993; Kosaka et al., 1985; Lopez-Mascaraque et al., 1989; Matsutani et al., 1988, 1989;Mugnaini et al., 1984; Nakajima et al., 1996, 1998; Sanides-Kohlrausch and Wahle, 1990; Taniguchi et al., 1993; Wahleet al., 1990).

Immunohistochemical staining was performed by theavidin-biotin peroxidase complex (ABC) method using a

commercial ABC kit (Vector, Burlingame, CA) and in-volved the following steps: 1) incubation with 10% normalgoat serum at 32°C for 30 min; 2) rinse in 0.02 M phosphatebuffered saline (PBS; pH 7.25) for 15 min; 3) incubationwith primary antiserum at 4°C for 48 hr; 4) rinse in PBSfor 15 min; 5) incubation with 1:600 biotinylated goatanti-rabbit IgG (Vector) at 32°C for 40 min; 6) rinse in PBSfor 15 min; 7) incubation with ABC at 32°C for 40 min; 8)rinse in PBS for 15 min; 9) incubation with 0.05 M tris-HCl(pH 7.6) containing 0.01% 3-38diaminobenzidine tetrahy-drochloride (DAB) and 0.003% hydrogen peroxide for20–30 min; 10) rinse in distilled water. The specificity ofthe antisera against PGP 9.5, GAD, and TH was previ-ously described by Chen et al. (1994), Behar et al. (1993),and Ren and Sagar (1992), respectively, and checked byimmunoblotting in the present study. The NOS, VIP, andSP immunoreactivities were abolished by adding 100 µg ofthe respective antigens to 1 ml of each antiserum. Whenthe primary antibody was replaced by nonimmune rabbitor goat serum, no specific reaction products were observed.

Enzyme-HistochemistryIn the present study, we also performed enzyme-

histochemical staining for NADPH-diaphorase known tobe equal to NOS activity (Hope et al., 1991) in order todetect NOS activity in the dog AOB.

Frozen sections were incubated for 50–60 min at 37°Cwith 1 mM b-NADPH (Oriental Yeast, Osaka, Japan), 0.1mM nitroblue tetrazolium (NBT; Wako, Osaka, Japan),and 0.3% triton X-100 in 0.1 M PB (pH 7.4), rinsed indistilled water, and mounted with glycerin jelly. Negativecontrol staining was performed in the absence of b-NADPH.No specific reaction products were observed in the nega-tive control staining.

RESULTSAccessory Olfactory Bulb (AOB)

The AOB was observed on the medial surface of theolfactory peduncle at the caudal margin of the MOB in alldogs examined. This finding was almost consistent withthe descriptions of McCotter (1912, 1913) and Salazar etal. (1992, 1994). In the present report, the dog AOB wasdivided into three layers: vomeronasal nerve layer (VNL);glomerular layer (GL); and neuronal cell layer (NCL) forconvenience (Fig. 1A). The VNL in the superficial layer ofthe AOB was composed of vomeronasal nerve fiber bundles.Under the VNL, a large number of glomeruli were ar-ranged to form the GL. In the dog, the VNL and the GLoccupied most of the AOB. The NCL was the deepest layerof the AOB, and contained several kinds of neurons. Theimmunoreactivities of antisera and NADPH-diaphorasereactivity in the AOB are described below.

PGP 9.5 ImmunoreactivityPGP 9.5-immunopositive neurons were frequently ob-

served in the GL and the NCL (Fig. 1A,B). These neuronshad round or oval somata, varied in size from 10 to 17 µmin diameter, and seemed to correspond to mitral/tuftedcells according to their morphology (Fig. 1B,C). In thepresent study, only short primary dendrites of the mitral/tufted cells were observed (Fig. 1C). The mitral/tufted cellswere smaller in size and scantier in cytoplasm, comparedto the mitral cells of the MOB, and seemed to resemble thetufted cells rather than the mitral cells of the MOB inmorphology. In addition, the vomeronasal nerve fibers and

394 NAKAJIMA ET AL.

the glomeruli were PGP 9.5-immunopositive in the AOB(Fig. 1A).

NOS ImmunoreactivityIn the NCL, NOS-immunopositive fibers were observed

to form dense plexuses without a clear orientation (Fig.2A). A small number of neurons were also NOS-immu-nopositive. They had small round or oval somata and were7 to 10 µm in diameter. Since they were observed mainly inthe inner portion of the NCL adjoining the olfactorypeduncle, this portion seemed to be intensely stained (Fig.2A). Although their distal dendritic ramifications could notbe identified, these neurons seemed to correspond togranule cells according to their morphology (Fig. 2C).NOS-immunopositive plexuses in the NCL seemed to becomposed of the dendritic processes of these granule cells.The staining intensity for NOS in the granule cells of the

AOB was almost equal to that of the MOB. No immunore-activity was observed in the VNL and the GL.

GAD ImmunoreactivitySince GAD immunoreactivity was more predominant

than NOS immunoreactivity in nerve fibers, GAD immuno-reactivity seemed to be more intense than NOS immunore-activity in the NCL (Fig. 3A). A small number of GAD-immunopositive neurons were observed among plexusescomposed of GAD-immunopositive fibers, mainly in theinner portion of the NCL. These neurons seemed tocorrespond to granule cells according to their morphologyand distribution (Fig. 3B). GAD-immunopositive neuronswere also observed around the glomeruli. These neuronshad small round or oval somata, 7 to 10 µm in diameter,and their morphology and distribution seemed to corre-spond to periglomerular cells (Fig. 3C).

Fig. 1. A: PGP 9.5 immunoreactivity in the accessory olfactory bulb(AOB). The AOB was divided into three layers: vomeronasal nerve layer(VNL); glomerular layer (GL); and neuronal cell layer (NCL). PGP 9.5immunoreactivity was detected in vomeronasal nerve fibers, glomeruli,

and mitral/tufted cells. B: PGP 9.5-immunopositive mitral/tufted cells inthe GL (arrowheads) and the NCL (arrows) of the AOB. C: A dendriteoriginating from mitral/tufted cell (arrowheads). M, medial; R, rostral;G, glomerulus. Scale bars 5 200 µm in A and 25 µm in B,C.

395ORGANIZATION OF CANINE ACCESSORY OLFACTORY BULB

Fig. 2. A: Nitric oxide synthase (NOS) immunoreactivity in theaccessory olfactory bulb (AOB). In the neuronal cell layer (NCL), NOS-immunopositive fibers formed dense plexuses without clear orientation.Since a small number of NOS-immunopositive granule cells were mainlyobserved in the inner portion of the NCL, this portion seemed to beintensely stained (thick arrows). B: NADPH-diaphorase reactivity in the

AOB. A small number of moderately stained granule cells were observedmainly in the inner portion of the NCL to form dark blue streak (thickarrows). C: NOS-immunopositive granule cells (arrows) in the NCL of theAOB. D: NADPH-diaphorase positive granule cells (arrows) in the NCL ofthe AOB. VNL, vomeronasal nerve layer; GL, glomerular layer; M, medial;R, rostral. Scale bars 5 200 µm in A,B and 25 µm in C,D.

396 NAKAJIMA ET AL.

TH ImmunoreactivityA small number of TH-immunopositive neurons were

observed in the GL and the NCL. TH-immunopositiveneurons in the GL had small round or oval somata (6 to 9µm in diameter) and their morphology and distributionseemed to correspond to periglomerular cells (Fig. 5A),although their frequency was much fewer than that ofGAD-immunopositive periglomerular cells. TH-immu-nopositive neurons in the NCL had round (6 to 10 µm indiameter) or spindle-shaped (10 to 16 µm in major diam-eter, and 5 to 6 µm in minor diameter) somata, and sentone or two long primary dendrites running parallel to thelayer. They seemed to correspond to short axon cellsaccording to their morphology (Fig. 5C). In addition, theTH immunoreactivity was detected in a few varicose fibersin the GL and the NCL.

SP ImmunoreactivityIn the GL and the NCL, numerous fibers with character-

istic varicosity showed intense immunoreactivity for SP

and formed dense plexuses (Fig. 4A–C). In the NCL, asmall number of SP-immunopositive neurons were alsoobserved. Since they were similar to NOS- and GAD-immunopositive granule cells in morphology and distribu-tion, they also seemed to correspond to granule cells (Fig.4B).

VIP ImmunoreactivityVIP immunoreactivity was detected in a small number

of neurons. These VIP-immunopositive neurons were di-vided into two types. One type was distributed in the GLand had small round or oval somata (6 to 9 µm indiameter). These neurons seemed to correspond to periglo-merular cells according to their morphology and distribu-tion (Fig. 5B). The frequency of the VIP-immunopositiveperiglomerular cells was much fewer than that of GAD-immunopositive periglomerular cells, but a little morethan that of TH-immunopositive periglomerular cells. Theother type was distributed in the GL and the NCL. Thistype of neuron had round or oval somata (6 to 10 µm in

Fig. 3. A: Glutamic acid decarboxylase (GAD) immunoreactivity in theaccessory olfactory bulb (AOB). B: GAD-immunopositive granule cells(arrows) in the neuronal cell layer (NCL) of the AOB. C: GAD-

immunopositive periglomerular cells (arrows) in the glomerular layer (GL)of the AOB. VNL, vomeronasal nerve layer; M, medial; R, rostral. Scalebars 5 200 µm in A and 25 µm in B,C.

397ORGANIZATION OF CANINE ACCESSORY OLFACTORY BULB

diameter) and sent one or two long primary dendritesrunning parallel to the layer. These neurons seemed tocorrespond to short axon cells and resembled the TH-immunopositive short axon cells in morphology (Fig. 5D).

NADPH-Diaphorase StainingA small number of granule cells in the NCL showed

moderate reactivity for NADPH-diaphorase (Fig. 2D).These granule cells were mainly observed in the innerportion of the NCL as NOS-, GAD-, and SP-immunoposi-tive granule cells to form dark blue streak (Fig. 2B). TheNADPH-diaphorase reactivity in the granule cells of theAOB was almost equal to that in the granule cells of theMOB. In the GL, glomeruli were stained light blue (Fig. 2B).

Main Olfactory Bulb (MOB)The laminar organization, distribution of each kind of

neuron and pattern of immunoreactivities in the MOB,were almost the same as those of the rodents (Baker, 1986;

Davis and Macrides, 1983; Gall et al., 1986; Kishimoto etal., 1993; Kosaka et al., 1985; Macrides and Davis, 1983;Matsutani et al., 1988, 1989; Mugnaini et al., 1984;Taniguchi et al., 1993). NADPH-diaphorase activity wasintensely positive in the periglomerular, granule, andshort axon cells as in the rodents (Davis, 1991; Kishimotoet al., 1993). Although some glomeruli were darkly stainedat the caudal portion of the MOB in the NADPH-diaphorasestaining, the other glomeruli were lightly stained (Fig. 6).

DISCUSSIONIn the present study, we demonstrated the distribution

of neuronal elements in the dog AOB by means of immuno-histochemical and enzyme-histochemical staining.

Salazar et al. (1994) performed immunohistochemicalstaining for neuron specific enolase (NSE) and neurofila-ments in order to identify the mitral/tufted cells in the dogAOB, but did not detect the NSE and neurofilamentsimmunoreactivity in any kind of neurons in the AOB.

Fig. 4. A: Substance P (SP) immunoreactivity in the accessoryolfactory bulb (AOB). B: SP-immunopositive granule cells (arrow) andfibers (arrowheads) in the neuronal cell layer (NCL) of the AOB.

C: SP-immunopositive fibers (arrowheads) in the glomerular layer (GL) ofthe AOB. VNL, vomeronasal nerve layer; M, medial; R, rostral; G,glomerulus. Scale bars 5 200 µm in A and 25 µm in B,C.

398 NAKAJIMA ET AL.

Instead, they demonstrated that the mitral/tufted cellswere distributed at the boundary between the GL and theNCL according to the nuclear size by means of severalkinds of histological methods: haematoxylin/eosin; Tolivia;Nissl; or silver staining. In our previous studies, PGP 9.5,one of the neuron-specific protein, was demonstrated to bepredominantly expressed in mitral and tufted cells of theMOB and mitral/tufted cells of the AOB in the rat andhamster (Nakajima et al., 1998; Taniguchi et al., 1993).PGP 9.5 is a soluble protein isolated from human brain(Doran et al., 1983) and known to correspond to ubiquitincarboxyl hydrolase (Wilkinson et al., 1989). In the presentstudy, we distinctly demonstrated the distribution of themitral/tufted cells of the dog AOB by means of immunohis-tochemical staining for PGP 9.5. In the dog AOB, themitral/tufted cells frequently occurred in the GL and theNCL, but the number of these cells seemed to be less in thedog AOB than that of the rodent AOB (Nakajima et al.,1998; Taniguchi et al., 1993). The distribution pattern ofthe mitral/tufted cells was different in the dog AOB fromthat in the rat AOB (Takami and Graziadei, 1991). Sincethe tufted cells of the MOB are distributed not only in theexternal plexiform layer, but also in the GL (Baker, 1986;

Davis and Macrides, 1983; Davis, 1991; Macrides andDavis, 1983), the distribution pattern of the dog mitral/tufted cells seems to be similar to that of the tufted cells.Takami et al. (1991) indicated that the large mitral/tuftedcells in the rat AOB were homologous to the mitral andinternal tufted cells of the MOB on the basis of the size oftheir somata in their Golgi study. In the dog AOB, however,most of the mitral/tufted cells seemed to resemble thetufted cells rather than the mitral cells in morphology.These findings suggest that the mitral/tufted cells of thedog AOB may be similar in nature to the tufted cells of theMOB despite the previous report in the rat. Although weobserved the primary dendrites originating from the PGP9.5-immunopositive mitral/tufted cells, it is unclearwhether these neurons extend plural dendrites to someglomeruli or only one dendrite to a specific glomelurus(Takami and Graziadei, 1990, 1991).

Since the presence or absence of the GCL in the dog AOBhas remained undetermined so far, the goal of this study isto determine the pattern of distribution of granule cells inthe dog AOB. Although Salazar et al. (1994) also tried toidentify granule cells according to nuclear size throughhistological methods, they could not distinguish granule

Fig. 5. A: Tyrosine hydroxylase (TH)-immunopositive periglomer-ular cell of the accessory olfactory bulb (AOB). B: Vasoactive intestinalpolypeptide (VIP)-immunopositive periglomerular cell of the AOB.

C: TH-immunopositive short axon cell in the neuronal cell layer of theAOB. D: VIP-immunopositive short axon cell in the glomerular layer of theAOB. Scale bar 5 25 µm.

399ORGANIZATION OF CANINE ACCESSORY OLFACTORY BULB

cells from grial cells and could not tell whether the GCL ispresent in the dog AOB. In the present study, however, wedistinctly demonstrated the distribution of granule cells inthe dog AOB by means of immunohistochemical stainingfor NOS, GAD, and SP, and enzyme-histochemical stainingfor NADPH-diaphorase. Since NOS, GAD, SP, or NADPH-diaphorase was detected in granule cells of the AOB inmany mammalian species (Alonso et al., 1995; Baker,1986; Davis, 1991; Kishimoto et al., 1993; Matsutani et al.,1989; Porteros et al., 1994; Takami et al., 1992; Wahle etal., 1990), we believe these substances could be used asmarkers to identify granule cells of the dog AOB. Sinceonly a small number of granule cells were present in thedog AOB, the dog AOB displayed less-developed GCLcompared with other mammals (Alonso et al., 1995; Baker,1986; Davis, 1991; Kishimoto et al., Matsutani et al., 1989;McCotter, 1912; Takami et al., 1992; Wahle et al., 1990).Therefore, the dog AOB seemed to be divided into onlythree layers: VNL; GL; and NCL, unlike other mammals,which is fairly consistent with the reports of Salazar et al.(1992, 1994).

Previous results in the rat, hamster, mouse, and hedge-hog showed that NOS immunoreactivity or NADPH-diaphorase reactivity in the granule cells was more intensein the AOB than in the MOB (Alonso et al., 1995; Davis,1991; Kishimoto et al., 1993; Porteros et al., 1994), suggest-ing that NOS activity in the granule cells was higher in theAOB than in the MOB in these animals. In the dog,however, NOS activity in the granule cells of the AOBseemed to be almost the same as that of the MOB judgingfrom the NOS immunoreactivity and NADPH-diaphorasereactivity. Since the granule cells play a role as inhibitoryinterneurons in the olfactory transduction, these findingsmay suggest that there is some difference in inhibitoryfunction of the AOB between the dog and other animals.

Miodonski (1968) described that the AOB of the dogformed a kind of finish or rim of the MOB and showed theposition of the AOB in his figures. In the NADPH-

diaphorase staining, we constantly observed some glo-meruli stained dark blue as shown in Figure 6 at the mostcaudal portion of the MOB, which Miodonski (1968) identi-fied as the dog AOB, but we confirmed that these glomerulidid not belong to the GL of the AOB, but to the GL of theMOB. Neurochemically homogeneous olfactory receptorcells were revealed to project restrictively to specificglomeruli by means of immunohistochemistry using mono-clonal antibody (Allen and Akeson, 1985; Mori et al., 1985).The difference of NADPH-diaphorase reactivity betweenthe glomeruli at the most caudal portion and the otherglomeruli may reflect projections from subsets of neuro-chemically different olfactory receptor cells.

In the present study, although NADPH-diaphorase reac-tivity was detected in glomeruli of the MOB and AOB, noNOS immunoreactivity was detected in any glomerulus ofthe MOB and AOB. This finding seems to suggest that theNADPH-diaphorase reactivity in the glomeruli is attrib-uted to some enzyme other than NOS as described in therodent MOB and AOB (Kishimoto et al., 1993; Spessert etal., 1994).

In the GL of the AOB, GAD-, TH-, and VIP-immunoposi-tive periglomerular cells were observed in the presentstudy. In the rat and hedgehog MOB, a large number of theperiglomerular cells were demonstrated to be VIPergic,although the VIPergic periglomerular cells of the AOBwere not reported (Gall et al., 1986; Lopez-Mascaraque etal., 1989). In the cat AOB, VIP-immunopositive neuronswere observed in the region corresponding to the MTL andidentified as short axon cells (Sanides-Kohlrausch andWahle, 1990). To the best of our knowledge, this is the firstreport suggesting the existence of the VIPergic periglo-merular cells in the mammalian AOB. Previous resultsindicated that the catecholaminergic periglomerular cellswere much fewer than the GABAergic periglomerular cellsin the rat AOB (Davis and Macrides, 1983; Takami et al.,1992). This is consistent with the current results. In thepresent study, the frequency of the GAD-, TH-, and VIP-

Fig. 6. NADPH-diaphorase reactivity in glomeruli at the caudal portion of the MOB. Some glomeruli werestained dark blue (asterisks). GL, glomerular layer; EPL, external plexiform layer; CR, cerebral cortex; M,medial; R, rostral. Scale bar 5 200 µm.

400 NAKAJIMA ET AL.

immunopositive periglomerular cells varied, suggestingthat the periglomerular cell population of the dog AOBmay be divided neurochemically into several subpopula-tions as in the MOB (Davis, 1991; Kosaka et al., 1985).

In the dog AOB, one type of neuron morphologicallydifferent from periglomerular cells were also TH- andVIP-immunopositive. These neurons resembled each otherin their morphology and seemed to correspond to shortaxon cells. Sanides-Kohlrausch and Wahle (1990) reportedthat the VIP-immunopositive neurons in the cat AOBresembled Van Gehuchten cells, a kind of short axon cell.The TH- and VIP-immunopositive short axon cells in thedog AOB were different from these neurons in morphologyand seemed to correspond to horizontal cells (Macrides andDavis, 1983).

The SP immunoreactivity was detected in numerousvaricose fibers in addition to the granule cells. These fibersseemed to be centrifugal afferents according to theircharacteristic profiles. The TH immunoreactivity was alsodetected in the afferent fibers, whereas the amount of theTH-immunopositive fibers was much fewer than that ofthe SP-immunopositive fibers. SP- and TH-immunoposi-tive afferents were demonstrated in the cat AOB by Wahleet al. (1990) and in the rat AOB by Davis and Macrides(1983), respectively, and were distributed in the GCL inboth species. In the dog AOB, however, the SP- andTH-immunopositive afferents invaded not only the NCL,but also the GL, suggesting more extensive innervation ofboth the SPergic and catecholaminergic fibers in the dogAOB than in the cat and rat AOB.

LITERATURE CITEDAllen WK, Akeson R. Identification of a cell surface glycoprotein family

of olfactory receptor neurons with a monoclonal antibody. J. Neuro-sci. 1985;5:284–296.

Alonso JR, Arevalo R, Garcıa-Ojeda E, Porteros A, Brinon JG, Aijon J.NADPH-diaphorase active and calbindin D-28k-immunoreactiveneurons and fibers in the olfactory bulb of the hedgehog (Erinaceuseuropaeus). J. Comp. Neurol. 1995;351:307–327.

Baker H. Species differences in the distribution of substance P andtyrosine hydroxylase immunoreactivity in the olfactory bulb. J.Comp. Neurol. 1986;252:206–226.

Behar TA, Schaffner A, Laing P, Hudson L, Komoly S, Barker J. Manyspinal cord cells transiently express low molecular weight forms ofglutamic acid decarboxylase during embryonic development. Dev.Brain Res. 1993;72:203–218.

Cajal SR. Textura del lobulo olfativo accessorio. Rev. Micros. 1902;1:141–150.

Chen ST, von Bussman KA, Gray LJ, Jen LS. Protein gene product9.5-immunoreactive retinal neurons in normal developing rats andrats with optic nerve or tract lesion. Dev. Brain Res. 1994;78:265–272.

Crosby EC, Humphrey T. Studies of the vertebrate telencephalon. I.The nuclear configuration of the olfactory and accessory olfactoryformation and the nucleus olfactorius anterior of certain reptiles,birds and mammals. J. Comp. Neurol. 1939;71:121–213.

Davis BJ, Macrides F. Tyrosine hydroxylase immunoreactive neuronsand fibers in the olfactory system of the hamster. J. Comp. Neurol.1983;214:427–440.

Davis BJ. NADPH-diaphorase activity in the olfactory system of thehamster and rat. J. Comp. Neurol. 1991;314:493–511.

Doran JF, Jackson P, Kynoch PAM, Thompson RJ. Isolation of PGP9.5, a new human neurone-specific protein detected by high-resolution two-dimensional electrophoresis. J. Neurochem. 1983;40:1542–1547.

Gall C, Seroogy KB, Brecha N. Distribution of VIP- and NPY-likeimmunoreactivities in rat main olfactory bulb. Brain Res. 1986;374:389–394.

Halpern M. The organization and function of the vomeronasal system.Annu. Rev. Neurosci. 1987;10:325–362.

Hoffman HH. The olfactory bulb, accessory olfactory bulb and hemi-sphere of some animals. J. Comp. Neurol. 1963;120:317–368.

Hope BT, Michael GJ, Knigge KM, Vincent SR. Neuronal NADPHdiaphorase is a nitric oxide synthase. Proc. Natl. Acad. Sci. 1991;88:2811–2814.

Kishimoto J, Keverne EB, Hardwick J, Emson PC. Localization ofnitric oxide synthase in the mouse olfactory and vomeronasalsystem: a histochemical, immunological and in situ hybridizationstudy. Eur. J. Neurosci. 1993;5:1684–1694.

Kosaka T, Hataguchi Y, Hama K, Nagatsu I, Wu J-Y. Coexistence ofimmunoreactivities for glutamate decarboxylase and tyrosine hy-droxylase in some neurons in the periglomerular region of the ratmain olfactory bulb: possible coexistence of gamma-aminobutyricacid (GABA) and dopamine. Brain Res. 1985;343:166–171.

Lopez-Mascaraque L, Villalba RM, de Carlos JA. Vasoactive intestinalpolypeptide-immunoreactive neurons in the main olfactory bulb ofthe hedgehog (Erinaceus europaeus). Neurosci. Lett. 1989;98:19–24.

Macrides F, Davis BJ. The olfactory bulb. In: Emson PC, ed. ChemicalNeuroanatomy. New York: Raven Press, 1983:391–426.

Matsutani S, Senba E, Tohyama M. Neuropeptide- and neurotransmit-ter-related immunoreactivities in the developing rat olfactory bulb.J. Comp. Neurol. 1988;272:331–342.

Matsutani S, Senba E, Tohyama M. Distribution of neuropeptidelikeimmunoreactivities in the guinea pig olfactory bulb. J. Comp.Neurol. 1989;280:577–586.

McCotter RE. The connection of the vomeronasal nerves with theaccessory olfactory bulb in the opposum and other mammals. Anat.Rec. 1912;6:299–318.

McCotter RE. The nervus terminalis in the adult dog and cat. J. Comp.Neurol. 1913;23:145–152.

Miodonski RE. Bulbus olfactorius of the dog (Canis familiaris). ActaBiol. Cracov. 1968;11:65–78.

Mori K, Fujita SC, Imamura K, Obata K. Immunohistochemical studyof subclasses of olfactory nerve fibers and their projections to theolfactory bulb in the rabbit. J. Comp. Neurol. 1985;242:214–229.

Mugnaini E, Oertel WH, Wauterlood F. Immunocytochemical localiza-tion of GABA neurons and dopamine neurons in the rat main andaccessory olfactory bulbs. Neurosci. Lett. 1984;47:221–226.

Nakajima T, Okamura M, Ogawa K, Taniguchi K. Immunohistochemi-cal and enzyme histochemical characteristics of short axon cells inthe olfactory bulb of the golden hamster. J. Vet. Med. Sci. 1996;58:903–908.

Nakajima T, Murabayashi C, Ogawa K, Taniguchi K. Immunoreactiv-ity of PGP9.5 in the developing hamster olfactory bulb. Anat. Rec.1998;250:238–244.

Porteros A, Alonso JR, Arevalo R, Garcıa-Ojeda E, Crespo C, Aijon J.Histochemical localization of NADPH-diaphorase in the rat acces-sory olfactory bulb. Chem. Senses 1994;19:413–424.

Ren T, Sagar SM. Induction of c-fos immunostaining in the rat brainafter the systemic administration of nicotine. Brain Res. Bull.1992;29:589–597.

Salazar I, Barber PC, Cifuentes JM. Anatomical and immunohistologi-cal demonstration of the primary neural connections of the vomero-nasal organ in the dog. Anat. Rec. 1992;233:309–313.

Salazar I, Cifuentes JM, Quinteiro PS, Caballero TG. Structural,morphometric, and immunohistological study of the accessory olfac-tory bulb in the dog. Anat. Rec. 1994;240:277–285.

Sanides-Kohlrausch C, Wahle P. VIP- and PHI-immunoreactivity inolfactory centers of the adult cat. J. Comp. Neurol. 1990;294:325–339.

Scalia F, Winans SH. The differential projections of the olfactory bulband accessory olfactory bulb in mammals. J. Comp. Neurol. 1975;161:31–56.

Spessert R, Wohlgemuth C, Reuss S, Layes E. NADPH-diaphoraseactivity of nitric oxide synthase in the olfactory bulb: co-factorspecificity and characterization regarding the interrelation to NOformation. J. Histochem. Cytochem. 1994;42:569–575.

Takami S, Graziadei PPC. Morphological complexity of the glomerulus

401ORGANIZATION OF CANINE ACCESSORY OLFACTORY BULB

in the rat accessory olfactory bulb—a Golgi study. Brain Res.1990;510:339–342.

Takami S, Graziadei PPC. Light microscopic Golgi study of mitral/tufted cells in the accessory olfactory bulb of the adult rat. J. Comp.Neurol. 1991;311:65–83.

Takami S, Fernandez GD, Graziadei PPC. The morphology of GABA-immunoreactive neurons in the accessory olfactory bulb of rats.Brain Res. 1992;588:317–323.

Taniguchi K, Saito H, Okamura M, Ogawa K. Immunohistochemicaldemonstration of protein gene product 9.5 (PGP 9.5) in the primaryolfactory system of the rat. Neurosci. Lett. 1993;156:24–26.

Wahle P, Sanides-Kohlrausch C, Meyer G, Lubke J. Substance P- andopioid-immunoreactive structures in olfactory centers of the cat:adult pattern and postnatal development. J. Comp. Neurol. 1990;302:349–369.

Wilkinson KD, Lee K, Deshpande S, Duerksen-Hughes P, Boss JM,Pohl J. The neuron-specific protein PGP 9.5 is a ubiquitin carboxyl-terminal hydrolase. Science 1989;246:670–673.

Wysocki CJ. Neurobehavioural evidence for the involvement of thevomeronasal system in mammalian reproduction. Neurosci. Biobe-hav. Rev. 1979;3:301–341.

402 NAKAJIMA ET AL.