THE JOURNAL OF COMPARATIVE NEUROLOGY 291~468-489(1990)

Morphology of Neuropeptide YIxnmunoreactive Neurons in the Cat Olfactory Bulb and Olfactory Peduncle: Postnatal Development and Species Comparison CLAUDIA SANIDES-KOHLRAUSCHAND PETRA WAHLE Max-Planck Institut fur Biophysikalische Chemie, Abt. Neurobiologie, D-3400 Gottingen, Federal Republic of Germany

ABSTRACT The distribution and morphology of Neuropeptide Y-immunoreactive (NPY-ir)neurons in the olfactory bulb and the olfactory peduncle was studied in the adult cat and rat, and the common marmoset Callithrix jacchus. Significant species differences were not observed. In all three species, the population of NPY-ir neurons is localized in the white matter extending from the main olfactory bulb to the border of the striatum. The neurons are characterized by a conspicuously looping axonal ramification pattern with some major collaterals running toward the olfactory bulb and others running toward the internal olfactory tract. The former, ipsilateral projection terminates in the granule cell layer of the main and accessory olfactory bulb and in layer II/III of the anterior olfactory nucleus. Reconstruction of the latter projection has revealed that the fibers are continuous with the olfactory limb of the anterior commissure and the anterior commissure proper suggesting a commissural contralateral projection. The analysis of the postnatal development of the cat NPY neuron system supports this assumption in a very clear-cut way. In young animals growing fibers are observed to cross the brachium of the commissure. The NPY neuron system develops postnatally. The maximum cell number is reached during the third postnatal week. The appearance of more and more NPY-ir neurons slightly precedes the formation of the terminal fields and of the fiber projection in the internal olfactory tract. The density of this early fiber projection by far exceeds the fiber density observed in the adult. Later in development the fiber density in the olfactory limb and the anterior commissure becomes considerably reduced. In contrast, the plexus density in the anterior olfactory nucleus and the granule cell layer of the main and accessory olfactory bulb undergoes only a slight reduction, and the NPY-ir cell number remains roughly constant. These observations suggest that the ipsilateral NPY-ir projection remains largely unchanged, in contrast to the contralateral projection, which exists to a large extent only for the first four postnatal months. The observation that the NPY neuron system gives rise to a contralateral projection does not support a classification of NPY neurons as short axon cells.

Accepted August 10,1989. Address reprint requests to Dr. Petra Wahle, Abt. Neurobiologie, MaxPlanck-Institut fiir Biophysikalische Chemie, Am Fessberg, Postfach 2841, D-3400 Gottingen, FRG.

0 1990 WILEY-LISS, INC.

469

DEVELOPMENT OF NPY IN CAT OLFACTORY BULB

Key words: ontogenesis, NPY-ir neuron system, transient axonal projection, anterior commissure, rat, marmoset

Earlier studies on the distribution of Neuropeptide Y in the mammalian brain mentioned NPY-ir neurons and fibers in the olfactory bulb and the olfactory peduncle (Chronwall et al., '85; Nakagawa et al., '85; De Quid and Emson; '86). However, more detailed structural analyses focused only on NPY-ir neurons of the main olfactory bulb of the rat and guinea pig (Gall et al., '86; Scott et al., '87; Matsutani et al., '89). NPY-ir neurons of the olfactory bulb and peduncle are, to date, described only for the human olfactory bulb (Ohm et al., '88). According to this study, human bulbar NPY-ir neurons differ from those of the rat and guinea pig. The difference could reflect anatomical or neurochemical species differences. Anatomically, the primate olfactory bulb and peduncle differ in several respects from the rodent. The lamination of the primate olfactory bulb is not so distinct compared to other mammalian families (Stephan, '75). The anterior olfactory nucleus is reduced in primates (Crosby and Humphrey, '39). At least the Old World primates lack an accessory olfactory bulb (Crosby and Humphrey, '39). Neurochemical species differences are reported for substance P, Neurotensin, and Cholecystokinin-8 immunoreactive neurons in the main olfactory bulb (Baker, '86; Matsutani et al., '89). Knowledge of NPY-ir structures of the olfactory bulb and peduncle is still incomplete. For example, the morphology of the NPY-ir neurons in the anterior olfactory nucleus has not yet been analyzed in species other than human. Moreover, the origin of the NPY-ir fibers in the olfactory peduncle (Gall et al., '86) and the olfactory bulb is still unknown. Our present study is focused on the following topics. Since nothing is known about the NPY neuron system in olfactory regions of the cat, we analyzed the distribution and morphological features of the NPY-ir neurons in the olfactory bulb and the anterior olfactory nucleus of this species paying special attention to the axonal pattern. In order to obtain direct information about possible species differences, we decided to reinvestigate the rat, and also analyzed a primate species, the common marmoset Callithrix jacchus. In addition, we studied the NPY neuron system in the cat olfactory regions during postnatal development. It has been demonstrated for the rat that the majority of the NPY-ir neurons appear postnatally (Woodhams et al., '85; Matsutani et al., '88). Our developmental study, presented here, puts special emphasis on the postnatal development of the NPY-ir fiber system, which has not yet been documented in detail for any species. We demonstrate that the results of our postnatal developmental approach help to answer questions about the origin of the NPY-ir fibers. A preliminary report has appeared elsewhere (Sanides-Buchholtz, '88).

METHODS Our material consisted of 18 cats of either sex and of the following postnatal ages: postnatal day (P) 3, P10, P11, P15, P17, P18, P23, P24, P25, P48, P60; and 3.5, 5.5, and 8 months, and 1and 3 years. In addition, 2 adult male Wistar rats and 2 adult male marmosets were investigated. The animals were from an in-house animal colony. They were

deeply anesthetized with sodium pentobarbitone (NembutaP; 60 mg/kg bodyweight) and perfused through the heart with 0.9 5% saline and 1% ' sucrose in 0.05 M phosphate buffer pH 7.4 followed by 4 % paraformaldehyde, 1O/c sucrose, and 0.2% picric acid in 0.1 M phosphate buffer, pH 7.4. The brains were removed from the skull. Both hemispheres were blocked into pieces containing the MOB, the AON, and the AC. The blocks were soaked overnight in 20% sucrose for cryoprotection and were then cut on a freezing microtome. Cat and rat tissue were cut into 80-pm-thick sections. Marmoset tissue was cut into 40-pm-thick sections in order to obtain more sections from the very tiny bulbs of this species. Of every cat, one hemisphere was cut in the frontal, the other in the sagittal or horizontal plane. Of rats and monkeys, one hemisphere was cut in the parasagittal, the other in the horizontal plane. The sections were collected in icecold 0.1 M phosphate buffer and subsequently processed for immunocytochemistry on freefloating sections. Antibody penetration was enhanced by 0.2% Triton X100, and nonspecific binding sites were blocked with 3% normal swine serum for 1 hour, prior to incubation in the primary antisera. The following primary antisera were used: antiserum against NPY raised in rabbits, diluted 1:400, a gift from Prof. Dr. J. Polak, University of London (for specification see Allen et al., '83a,b). Antiserum against NPY, raised in goat, diluted 1:700, was a gift from Drs. B.M. Chronwall and T. O'Donohue, Bethesda (for specifications, see Chronwall et al., '84, '85; DiMaggio et al., '85). The sections were incubated in the primary antisera for at least 12 hours at room temperature. For NPY (rabbit) a PAPmethod was carried out with affinity purified swine-antirabbit IgG as linking antibody, diluted l:lO, and PAP-complex (rabbit), diluted 1:50. For NPY (goat), a peroxidase conjugated rabbit-antigoat antibody, diluted l:lO, was used. Incubation times for the secondary antibodies (purchased from DAKO PATTS, Hamburg) were 2 hours. The peroxidase activity was developed with 0.03 Yo Diaminobenzidinetetrahydrochloride and 0.001qi H,O,. The sections were mounted and dried. In order to intensify the reaction product, the sections were treated with 1!?& OsO, in 0.1 M phosphate buffer pH 7.4. Thereafter, they were dehydrated, cleared, and coverslipped. Controls: Staining was totally abolished by adding 100 pg of Neuropeptide Y (Sigma) to 1,000 p1 diluted antiserum 30 min prior to incubation, or by omitting the primary antiserum. Series of alternate sections were stained with 0.2 !?& thionine. Anatomical structures were identified with the atlas of Berman and Jones (%a),the reports of Lohman ('63), and of Crosby and Humphrey ('39). Sections were analyzed with a Leitz Dialux 20 photo microscope. NPY-ir neurons were drawn with a camera lucida at a final magnification x800. The distribution of the NPY-ir fibers was charted with a camera lucida at a final magnification of x 200. Cell counts were carried out in the young and adult cat material in plots, prepared with an XY-plotter attached to the stage of the microscope. For every animal, the cell numbers were counted in 4 parasagittal sections close to the parasagittal midline of the MOB because in this plane the white matter, where the majority of the NPY-ir neurons is localized, displays its widest extension.

C. SANIDES-KOHLRAUSCH AND P. WAHLE

470

The AON was identified in all the 3 species. In the cat, it was wedge-shaped with its rostrally oriented apex penetrating into the white matter of the MOB (Fig. la). In the rat, it was located caudally adjacent to the MOB (Fig. lb). In the rat and cat AON we distinguished two laminae, layer I, a cell-poor molecular layer and layer IIDII, a cellular layer containing larger and smaller cells, which may represent a mixed pyramidal/polymorph layer (shown for the cat, see Fig. 5a). Superficially to layer I of the lateral division of the AON, the LOT was located (see Fig. 5a). In the marmoset, the AON was integrated into the peduncle intermingling with fiber bundles. It was wedge-shaped with its apex oriented rostrally (as shown by Reyher, ’88). The white matter deep to the cellular layers of MOB, AOB, and AON extending to the border of the striatum could he identified in the cat and rat (Fig. 1).It contained the TOI, which was continuous with the olfactory limb of the AC, and thus with the AC proper (Stephan, ’75). In the marmoset and other primates, the LOT and TO1 fibers form one common tract representing the long peduncle (Stephan, ’75). We use the term “peduncle” synonymously for the AON. In addition to the above mentioned structures, in the young kittens we regularly identified an olfactory ventricle, surrounded by a layer of very densely packed and very darkly Nissl-stained small cells, which we identified as the subependymal layer (according to Kishi, ’87). The olfactory ventricle was closed in kittens older than P48, hut residues of the subependymal layer still persisted as a narrow stripe in the 3-year-old cat.

b \

2mm Fig. 1. Outline drawings of parasagittal sections show the anatomy and size differences of the olfactory bulb and peduncle of cat (a),rat (b), and marmoset ( c ) drawn at a midline level through the olfactory bulb and peduncle. The position of the AOB and the AON can be identified. The distribution of NPY-ir somata is plotted, and in (a,b), every dot represents two cells, in (c) every dot represents one cell (the lower cell density is due to the thinner sections; see Methods).

RESULTS Anatomical observations and terminology According to our Nissl-stained material, the MOB is 6layered in all 3 species (Fig. 1: cat: 2a; marmoset: 6a; rat is extensively documented in the literature, for example Bayer, ’83). In all 3 species, an AOB could he identified. Its position is caudally adjacent to the MOB, and it is dorsally integrated into the olfactory peduncle (Fig. 1).According to Stephan (’75), the AOB is organized into the same 6 layers. We found these 6 layers clearly distinguishable only in the rat (see Fig. 6g). In the cat, EPL and IPL were scarcely distinguishable in the AOB. Thus we combined EPL, ML, and IPL in the cat AOB to one layer, termed EPL/ML/IPL (Fig. 2d). In the marmoset, the AOB was even less well laminated (see Fig. 6f). We observed no plexiform layers. Furthermore, the ML and the GRL occurred as one layer containing a mixed population of large cells (probably mitral cells) and small cells (probably granule cells). In all 3 species, the cell bodies in the AOB granular layer were not arranged in clusters.

Abbreviations AC ACC ACN AOR C AON AONd AONe AONl AONp AONv CA DB EP EPL FL GL GRL

HLA

IC IPL LOT LV MFB ML

MOB

OT OLV OP SI STM STR SUB TO1 VP wm I I1/111

anterior commissure nucleus accumbens nucleus of the anterior commissure accessory olfactory bulb cortex anterior olfactory nucleus AON dorsal division AON external division AON lateral division AON posterior division AON ventral division caudate nucleus diagonal band of Rroca ependymal layer external plexiform layer layer of the olfactory nerve fibers glomerular layer granule cell layer lateral hypothalamic area internal capsule internal plexiform layer lateral olfactory tract lateral ventricle medial forebrain bundle mitral cell layer main olfactory bulb olfactory tubercle olfactory ventricle olfactory peduncle substantia innominata stria medullaris thalami stria terminalis subependymal layer internal olfactory tract ventral extension of the globus pallidus white matter layer I (molecular layer) layer II/III (pyramidal polymorph layer)

DEVELOPMENT OF NPY IN CAT OLFACTORY BULB

Distribution and morphology of NPY-ir neurons in the adult cat The majority of NPY-ir neurons were localized in the bulbar and peduncular white matter, extending from the MOB to the border of the striatum (Fig. la). Some NPY-ir neurons were also localized in the deep GRL of the MOB and in the deep layer I of the AON (see Fig. 5b). Very occasionally, NPY-ir neurons were found in the EPL of the MOB. There is no morphological difference in NPY-ir neurons located caudally or rostrally. They had large somata of 25-30 pm in diameter. The majority had a multipolar dendritic configuration (Fig. 3d,e,f) in parasagittal sections. In frontal and horizontal sections they appeared more bitufted, with the longest axis oriented parallel to the olfactory ependyma. Occasionally, unitufted NPY-ir neurons were observed with the tuft oriented toward the pial surface. Dendrites were smooth, stout, and infrequently branched. They followed straight or slightly sinuous courses. The axon arose from the soma (Fig. Ya,e,f) or from a primary dendrite (Fig. 3b, 4e). It was smooth or slightly beaded (Fig. 4e). Despite the dendritic variation, NPY-ir neurons displayed the same conspicuous axonal ramification pattern: after a short initial segment the axon gave off a first recurrent collateral, which often formed a hairpin loop of 180'. The collateral and the main axon immediately gave off collaterals of a higher order, which again formed narrow loops. This ramification pattern produced configurations difficult to survey, especially in cases in which the distances between the bifurcations were very short (about 10 pm; Fig. 3e,f,g). Often, main axon and first collateral were of the same caliber, and it could not be determined which one actually represented the main axon. After several ramifications within or near the dendritic domain, the neurons produced a number of collaterals, some of which descended and entered the TOI, whereas others ascended into the cellular layers of the MOB and the AOB, and layer II/III of the AON (Fig. 3f).

NPY-ir fiber system in the adult cat In the MOB, the collaterals ascended into the GRL. Their initially smooth appearance gradually changed into a beaded character and they branched occasionally. In the GRL, the varicose fibers finally bent off. They wound through or wrapped up whole granule cell clusters and in this way produce a dense plexus (Fig. 2b,c). The terminal plexus extended into the IPL, and single fibers pass through the ML into the EPL (Fig. 2c). The GL was void of immunoreactive processes (Fig. 2b). In the AOB, a similar terminal pattern was observed: Smooth fibers ascended and became varicose in the GRL. In contrast t o the MOB, the plexus was largely concentrated in the deep aspect of the GRL, and the pattern was more diffuse. This was probably due to the fact that AOB granule cells were not organized into dense clusters (Fig. 2d). In the AON (Fig. 5a,b), the plexus covered layer II/III and extended into layer I. We observed no distinct accumulations of NPY-ir fibers related to subdivisions of the AON. The morphology of the fibers was similar to those of AOB and MOB. Their target cells in the AON, however, could not be identified. The dominant terminal form in the AON as well as in the MOB and in the AOB seemed to be en passent boutons. As described above, the NPY-ir neurons also gave rise to collaterals that did not ascend into the cellular layers of either MOB and AOB, or layer II/III of the AON. Rather,

471

they remained within the white matter, where they ran into an anterio-posterior direction following the course of the TOI. Single isolated fibers could be followed to the border of the striatum. These fibers occasionally branched into recurrent collaterals, coursing back toward the MOB, but they did not form varicose terminal elements in the TOI. Since the TO1 continued with the olfactory limb of the AC, we examined the AC for the presence of NPY-ir fibers, where few NPY-ir fibers can be observed in adult cats.

NPY-ir structures in rat and marmoset The distribution of the NPY-ir neurons in the rat was similar to the cat (Fig. Ib). The soma diameter of the NPYir neurons was about 20 pm (Fig. 6h) and, thus, smaller compared to the cat. However, the dendritic configuration was multipolar or unitufted (Fig. 7a,b). The dendrites were smooth, stout, and not frequently branched. The axons were smooth or slightly beaded and branched, forming hairpin loop collaterals (Fig. 7a). The neurons gave rise to collaterals, some of which entered the TOI, whereas others ran towards the cellular layers of MOB and AOB, and towards layer II/III of the AON. NPY-ir collaterals ascending into the cellular layers of the MOB, produced a plexus mainly localized in the GRL (Fig. 8a). Single NPY-ir fibers ascended into the IPL, ML and EPL. In contrast to the cat, single fibers also reached the GL. The plexuses, produced by NPY-ir collaterals ascending into the AOB and the AON, were similar in extent to those found in the cat. In general, terminal fibers displayed a finer caliber and a lower density compared to the cat. NPY-ir collaterals following the course of the TO1 were present in the white matter. They were oriented parallel to the olfactory ependyma and display recurrent collaterals. Some NPY-ir fibers were found in the AC. In the marmoset, NPY-ir neurons were localized within the MOB white matter mainly in the part where the olfactory peduncle arises, but also along the entire olfactory peduncle proper (Fig. lc). The soma diameter of the NPY-ir neurons was about 20 pm. The NPY-ir neurons in the white matter of the MOB displayed a multipolar dendritic pattern (Fig. 3b), whereas those localized in the olfactory peduncle were bitufted with spindlelike somata, oriented with their longest axis parallel to the course of the peduncle (Figs. 6c, 7c,d). The dendrites were smooth and stout. The axons were smooth or slightly beaded. Dendritic and axonal ramifications were less frequently observed compared to cat and rat. This i s most probably due to the thinner sections (see Methods) as many processes were cut off. Nevertheless, a hairpin loop axonal branching pattern could be observed in this species (Fig. 7c,d,e). A NPY-ir terminal plexus covered the region where the peduncle arises. It extended into the GRL of the MOB. In the MOB proper, the distribution and density of the NPY-ir plexus were similar to the rat (Fig. 6e; 8b). A few NPY-ir fibers and terminals were found in the AOB in the GRL. In the TOI, NPY-ir fibers were present, some of which had straight courses (Fig. 6d). Single fibers could be traced over distances of more than 1mm giving off recurrent collaterals (Fig. 8e). Some NPY-ir fibers were present in the AC. To summarize our findings, the NPY neuron system in the adult rat and marmoset strikingly resembled the cat NPY neuron system in distribution, morphology, innervation, and projection pattern.

Figure 2

DEVELOPMENT OF NPY IN CAT OLFACTORY BULB

473

100pm

Fig. 3. Camera lucida drawings of NPY-ir neurons in the cat MOR (a-d,f,g) and AON ( e ) .(a+) Immature cells without recognizable axon. Note growth cones (asterisks) in (c). (b,d,e,f) Mature neurons; axons are indicated by arrows. Note the looping axonal pattern. (e) A neuron with growing collaterals (asterisk). ( 8 ) The axonal pattern of cell (0 in higher magnification.

Postnatal development of the NPY-ir structures in the cat olfactory bulb and peduncle In order to learn more about the postnatal development

of the NPY-ir fiber system, we analyzed kittens aged 3 days

to 3 years. We found that the majority of the NPY-ir neurons appeared postnatally. Only a low number of NPY-ir neurons

Fig. 2. Cat olfactory bulb. (a) Nissl-stained M O B photomontage; the layers are indicated. (b)Low power photomicrograph shows the distribution of NPY-ir somata and fibers in the MOB layers; photomontage. (c) NPY-ir fibers in IPL and GRL at higher magnification. Two granule cell clusters (indicated by asterisks) are surrounded by immunoreactive processes. Note single axon ascending into EPL. (d) Nisslstained AOB; the layers are indicated. EPL and ML do not form two distinct layers. Granules are not organized in clusters. (e) Long projection axon8 in the TOI.Fibers are straight and slightly beaded. Bars: 250 p m in a,b; 50 fim in c,e; 200 pm in d.

were already present at P3, the earliest age examined, but their number had steeply increased until P25, followed by a slight reduction. Thereafter, the number of NPY-ir neurons remained more or less constant (Fig. 9). The increase between P 3 and P25 was obviously due to newly differentiating neurons. These were stained by immunohistochemistry during this period and were characterized by the following features: Immature NPY-ir neurons had small somata (about 12-18 fim in diameter), which consisted of the unstained nucleus surrounded by a very narrow rim of lightly immunoreactive cytoplasm. These somata gave rise to very short processes (Fig. 3a; 4a). In more mature stages, they give rise to longer processes, which occasionally bore growth cones (Fig. 4c). However, we classified these neurons as immature, since the processes could not yet be distinguished as dendrites or axons. More advanced and mature neurons had soma diameters of 25-30 pm. They were intensely labeled and immunoreactive material was equally distributed within the neuron. They displayed an identifi-

474

C. SANIDES-KOHLRAUSCH AND P. WAHLE

Fig. 4. NPY-ir structures in the cat. (a)Small immature neuron. (b,c)Growth cones of lamellipodial and bulletlike shapes. Note in (c) axonal branches in a granule cell cluster (outlined; the cell bodies can be identified in this Nomarski micrograph). (d) Axons in the olfactory limb of the AC. The arrows marks two growth cones at tips of axons that grow towards the AC and towards the olfactory bulb, respectively. ( e ) Fully differentiated cell. Note looping axon (arrow). Bars: 50 pm in a,d,e; 20 pm in b,c.

able axon. Often it bore growth cones a t the most distal tips of collaterals (Fig. 3e). The axons displayed the recurrent ramification pattern. During the first 2 postnatal weeks, neurons at different stages of maturation were present concurrently in the same section (Fig. 3a,b). Thereafter, immature neurons became rare, and during the third week, eventually all neurons appeared mature by soma size, dendritic length, and looping

axonal pattern. However, we still observed axonal growth cones during the seventh postnatal week, which indicated that the axons were still elongating (Fig. 4b). Thereafter, growth cones were not observed. A series of camera lucida drawings illustrated the development of the NPY-ir innervation of the MOB (Fig. 10). The NPY-ir plexus density in the MOB increased postnatally. At P3, only single NPY-ir fibers were present within

DEVELOPMENT OF NPY IN CAT OLFACTORY BULB

475

The changes in NPY-ir plexus density in the AOB are shown in Figure 11. NPY-ir fibers were not identified in the AOB layers before P17. At P25, some NPY-ir fibers were present in the deep GRL. At P48, the plexus had expanded enormously. It now covered GRL and EPL/ML/IPL. Single fibers even ascended into the GL, winding through the glomerular intersticies. This high density persisted until 3.5 months. At 5.5 months i t was reduced to the adult extent and density. It could be further seen that the projection fibers, which were dense in the white matter at P25 and P48, had become reduced. The fiber plexus in the AON is shown in Figure 12. Many horizontal fibers were present in the white matter below the AON. In anterior parts, fibers were present in layer II/III and in the deep half of layer I already a t P10. In dorso-caudal parts, fibers invaded layer II/III until P17 (Fig. 13). The fiber density in layer I and layer II/III of all AON divisions increased dramatically until P48. Fibers now occupied the entire layer I and a few invaded the LOT. The plexus persisted until 3.5 months (Fig. 5b), thereafter decreasing to its adult level (5.5 months, Fig. 12). The horizontal fibers in the white matter had largely disappeared.

Evidence for a contralateralprojection of the NPY-ir neurons

Fig. 5. (a)Nissl-stained parasagittal section of the 3.5 months old cat AON; photomontage; the layers are indicated. (b)Distribution of NPY-ir fibers in the AON. Note cells deep in layer II/III and in white matter. Note that the TO1 projection fibers have already disappeared from the white matter at this age. Fibers are also present in deep and superficial layer I, but not in the LOT; photomontage. Bar: 250 pm in a,b.

the GRL. From P3 to P48, the fiber density increased, and many fibers with growth cones were identified at P17 and P25 (Fig. 4c). During this period, the plexus also expanded into the subependymal layer, the superficial GRL, the IPL, and the ML. From P48 onward, single collaterals even extended into the EPL. Now the plexus had reached its maximal density. A meshwork of long smooth axons was present in the white matt'er and subependymal layer ascending into the GRL: where terminal plexuses entwined granule cell clusters (Fig. 2c). After 3.5 months the plexus slightly decreased until its adult density was reached at around 5.5 months. No further decrease in older animals was observed.

We have described above that in the adult cat NPY-ir fibers were found in the TOI, which have been traced to the anterior border of the striatum. These fibers are the caudally oriented collaterals of the NPY-ir neurons in the white matter below MOB, AOB, and AON. When we focused our attention on the caudal part of the peduncle and the connection of the white matter/TOI to the AC, we almost instantly noted differences in the immature material compared to the adult stage. In kittens, the density of NPY-ir fibers in the TO1 (Fig. Ze), the olfactory limb of the AC, and in the AC proper by far exceeded the density in adults. Since the high density makes it so much easier to trace, we reconstructed this projection system in kittens at age PI 7. Against the surrounding striatum characterized by finer, varicose NPY-ir terminal plexuses, the projection tract stands out by long parallel oriented, mostly unbranched axons. This is best visualized in parasagittal sections, as shown in Figure 13. In Figure 13a NPY-ir fiber bundles are present superfical and deep to the subependymal layer (arrows). They are oriented in rostrocaudal direction. More caudally, these bundles join the TOI. The fiber bundle can be followed to the border of the striatum, which is easy to identify. The processes of the striatal NPY-ir neurons are oriented perpendicular to the NPY-ir fibers of the TO1 and thus demarcate the striatal border. In front of the border the NPY-ir fiber bundle changes its orientation by bending slightly dorsalward. Now, it has the same orientation as the NPY-ir fibers in the AC, visible as a semilunate structure in the Figure 13a. In the AC, the NPY-ir fibers follow the semilunate course. The piece missing between the AC and the TO1 in Figure 13a is shown in Figure 13b and is reproduced from a more lateral level. Here, it can be seen that the NPY-ir fibers of the TO1 retain their orientation, until they have reached the AC. Thus the NPY-ir fibers of the TO1 and in the AC are continuous in their course. We conclude that NPY-ir fibers project into the AC. Do they indeed cross the AC? To answer this question, we ana-

C. SANIDES-KOHLRAUSCH AND P. WAHLE

476

Figure 6

477

DEVELOPMENT OF NPY IN CAT OLFACTORY BULB

Fig. 7. Camera lucida drawings of NPY-ir neurons in the MOB of the rat (a,b)and marmoset (c,d,e). Note looping axonal pattern, and long fiber that was traced through the peduncle in (e). Arrows mark axons, the open arrows on (e) indicate that the fiber is continuous.

lyzed horizontal sections containing the brachium of the AC (Fig. 14). The material of the P10 kitten was used, since at this age, the density of the NPY-ir fibers is still relatively low, making it easier to trace single fibers over longer distances. Growth cones indicate the direction the fibers grow. Figure 14 shows that fibers grow into both directions, thus crossing the commissure. This was also observed in the TOI, where growing fibers are oriented into both, the rostra1 and the caudal direction (Fig. 4d). Fig. 6. Marmoset and rat MOB and AOB. (a)Nissl-stained MOB of marmoset; the layers are indicated. (b)Multipolar NPY-ir neurons in the marmoset MOB. (c) NPY-ir neuron in the marmoset olfactory peduncle. (d) Long, straight, beaded axon in the marmoset peduncle; arrows indicate swellings. (e) NPY-ir fibers in the GRL of marmoset MOB. (f) Nissl-stained marmoset AOB, layers are indicated. Note that the AOB is small and deeply embedded into the peduncle. (g) Nisslstained rat AOB layers are indicated. (h) NPY-ir neuron in the white matter of rat MOB. Note the looping axon (large arrow). Small arrows mark branch points. Bars: 200 pm in a,f,g: 100 q n in b,c: 50 p m in d,h,e.

The developmental analysis revealed the changes in the number of NPY-ir processes. This is shown in Figure 15 on sections of the AC at different ages. At P3, only a few ir fibers are visible in the AC, but their density increases steeply at P17 and P48. Growth cones can be observed during this period and indicate the long period of collateral outgrowth. Between the third and fifth month, the fiber density in the AC decreases drastically, and no growth cones are present a t 5.5 months or at 3 years. The changes in fiber density coincides with the changes in NPY-ir terminal plexus density in MOB, AOB, and AON. Material from older animals showing lower plexus densities always has a low fiber density in the TO1 and in the AC.

DISCUSSION The present study describes a population of NPY-ir neurons in the white matter of MOB, AOB, and AON. The neu-

478

C. SANIDES-KOHLRAUSCHAND P. WAHLE

Fig. 8. Distribution of NPY-ir fibers in rat (a)and marmoset MOB layers. (b)Note that some fibers pass through ML into the EPL and GL. Hatched cell bodies are mitral cells. The small sketches indicate position of the panels.

DEVELOPMENT OF NPY IN CAT OLFACTORY BULB I

1801

150-

[r

W

m

100-

3

z

_I

1

W

O

50-

10F-3

PIC

pi5

P20

PZ5

PLS

P35rnO

P3a

AGE Fig. 9. The development of NPY-ir neuron number in the cat MOB and peduncle. Cells were counted in plots of parasagittal sections near the parasagittal midline of the bulb (indicated by presence of olfactory ventricle or ependyma). Every point marks the mean cell number counted in four sections. Vertical bars indicate standard deviation. Note a dramatic increase in cell number during the first two postnatal weeks. Maximal numbers are reached by the end of the first month and slightly decrease to adulthood.

ions display the same axonal characteristics: looping axons produce a system of collaterals, some of which remain ipsilaterally and innervate layer II/III of the AON and the granule layers of the MOB and AOB, whereas others project across the AC to the very same structures on the contralatera1 side. The developmental study performed in the cat has shown that the ipsilateral innervation persists into adult life, whereas a large proportion of the contralateral projection disappears during postnatal development. A comparison of the adult cat, rat, and marmoset reveals striking similarities in NPY neuronal morphology, distribution, innervation, and projection pattern, suggesting that the NPY neuron system is phylogenetically conserved in mammals. The first part of the discussion focuses on the dynamic changes of the cat NPY-ir neuron system. The second part summarizes our results on the organization of the NPY-ir neuron system with respect to species differences and cell classification as discussed in the recent literature.

Postnatal development of the NPY-ir neuron system in the cat olfactory bulb and peduncle We have shown that with immunohistochemistry the process of morphological differentiation can be demonstrated in bulbar/peduncular NPY-ir neurons. The somata increase in size, dendrites, and axons grow, as indicated by growth cones at their tips, and dendritic and axonal fields enlarge, Such features are considered as indices for neuronal maturation (Maslim et al., '86; Kishi, '87). We did not observe migrating NPY-ir neurons with leading or trailing processes. The morphological and neurochemical maturation pattern in the bulb are very similar to peptidergic neurons of the kitten visual cortex (Wahle and Meyer, '87). Bulbar NPY-ir neurons differ, for example, from tyrosine hydroxy-

479

lase (TH)-ir periglomerular cells of the rat MOB. In these neurons, the process of maturation could not be observed with TH-immunohistochemistry. They apparently do not start enzyme synthesis before they are fully mature and have accomplished synapse formation (Mc1,ean and Shipley, '88). The majority of NPY-ir neurons start differentiation early postnatally, thus confirming the results that Woodhams et al. ('86) and Matsutani et al. ('88) obtained in the rat. The presence of few fully differentiated neurons in the P3 material suggests that the first NPY-ir neurons mature late prenatally. Maximal NPY-ir neuron numbers are reached at P25. This means that the NPY-ir neuronal population differentiates heterochronously over a period of about 4 weeks. NPY-ir somata and dendrites display mature shapes by the end of the first postnatal month. Axonal maturation continues far into the second postnatal month because axon elongation, indicated by growth cones, is still observed at P48. A heterochronous differentiation could therefore be expected for the development of the NPY-ir fiber density in the AC. Indeed, we did not observe a homogeneous front of axons growing through the AC within a short period of time. Rather, the fiber density continuously increases, reaching maximal density a t the seventh postnatal week. The cell counts have shown that the majority of NPY-ir neurons are present at around P10/15, and many cells appear well differentiated and have sent out an axon. Assuming an axonal growth rate of 1mm per day (Katz, '85) and a distance of 11 mm from the AON/MOB border to the brachium of the AC (measured in the Pl'i material), we calculated that many axons should have reached the AC during the third postnatal week. Indeed, the AC in the P17 material already displays a high NPY-ir fiber density. Taking another 11 mm from the brachium of the AC to the contralateral bulb, NPY -ir axons could have reached the other bulb by the end of the first postnatal month. Indeed, the GRL of the MOB at P25 contain many long axons tipped by growth cones. Now the formation of the terminal plexus in the GRL starts. NPY--ir fibers develop a beaded appearance and entwine the granule clusters. Other NPY-ir axons from later differentiating cells arrive and pass the AC during the fourth and fifth week, and they reach the contralateral bulb a t around P48. They cause the further increase in fiber density in the AC and in the bulb observed a t this age. The calculation outlined above supports our view that during development the NPY-ir neurons send collaterals to the contralateral bulb and AON, whereas local collaterals ascend into the ipsilateral bulb and AON. This suggests that fibers of ipsiand contralateral origin contribute to the innervation of the MOB, AOB, and AON. The contralateral projection becomes drastically reduced after the third postnatal month. A large number of fibers disappear from the AC and the TOI. This fiher loss seems not to be due to myelinaytion. We still observe the proximal Fig. 10. Development of NPY-ir innervation pattern in the cat MOB at different ages, drawn from parasagittal sections. Note that GRL is nearly void of fibers at P?. Differentiated cells and increasing fiber density are evident at P17 and P25. Maximal density is reached at P48. Note that fibers and terminal entwine granule clusters in GRL. Density slightly decreases until adulthood. The size difference between the P3 to 48 and older stages is due to the late postnatal shrinkage of the subependymal layer. Hatched cell bodies are mitral cells. The sketch indicates position of the panels in the MOB.

P3

P 17

Figure 10

P25

P5.5mo

Pb0

P3a FL

GL EPL

VL IPL 3RL

vm

\

il p

1OLV

rs

Figure 10 continued

C. SANIDES-KOHLRAUSCH AND P. WAHLE

482

P25

P4 8

P5.5mo

---I

200pm Fig. 11. Development of NPY-ir innervation of the cat AOB, drawn from parasagittal sections. Fibers have reached the deep GRL at P25, and an exuberant innervation has appeared until P48. Fiber density in the 5.5 months and adult (not shown) material is rather low, and fibers are scattered evenly in GRL and EPL/ML/IPL. Sketch indicates position of the panels.

portion of the axon and the ipsilaterally ascending collaterals and can follow single axons over long distances (more than 500 pm) within the TO1 in adult material. They have the regularly beaded morphology of unmyelinated immunolabeled fibers. Thus the NPY-ir fiber system itself appears to be unmyelinated. Another possibility is that myelination of the AC renders antiserum penetration difficult, so that the NPY-ir axons are not labeled. The high concentration of detergent used on our material makes this possibility less likely. Some AC and TO1 fibers are present and remain labeled with high staining intensities in young adult cats and in adult animals of all 3 species. Also, the somata and dendrites remain well labeled, arguing against a reduction of NPY expression. The disappearance of an axonal projection pattern could be caused by the death of the cells of origin. Such a process occurs, for example, in the developing cat visual cortex, where NPY-ir subplate cells with long cortical projections

degenerate and die (Wahle and Meyer, '87). In the olfactory bulb, however, there is no evidence for a comparable process. Our cell counts revealed only a slight decrease in cell number. The majority of the population persists. Thus cell death does not account for the severe reduction of NPY-ir fibers in the AC. Instead, it appears that the NPY-ir cells eliminate the contralaterally projecting branches. Such a process has been described by Innocenti et al. ('86) in the cat neocortex, where layer 111 pyramids selectively eliminate their projection to the contralateral hemisphere. Other neurons of the occipital cortex have been reported to project transiently into the pyramidal tract (O'Leary and Stanfield, '85). Other examples for transient projections and axonal reorganization are summarized in O'Leary ('87). The reduction of the NPY-ir fibers in the AC occurs concurrently with the reduction in terminal plexus density of the AON, MOB, and AOB. One could assume that the latter reduction is

483

DEVELOPMENT OF NPY IN CAT OLFACTORY BULB

P 5.5mo

P 10

200pm

Lrost

Fig. 12. Development of NPY-ir innervation of the cat AON from horizontal sections. Note the dramatic overshoot in the P48 material and the rather loose, persisting innervation. An array of long projecting axons can be identified at P10,also at P48 in the white matter. These are TO1 fibers which are continuous with the anterior limb of the AC. Note

that they have largely disappeared at 5.5 months. A slightly denser fiber and terminal tier is present in the deep molecular layer. This plexus is continuous in layer I and is not correlated with the presence or absence of the AONe (see also Fig. 13).

caused by the elimination of the contralateral component. We cannot, however, exclude shrinkage of the ipsilateral part of the axon plexuses. A similar phenomenon has been described for developing axon plexuses of intrinsic cell types in the cat neocortex (Meyer and Ferres-Torres, '84). In summary, we favor the hypothesis that a large proportion of the NPY-ir neuron population eliminates its contralaterally projecting branches after the second postnatal month, but retains its ipsilateral connection. The NPY-ir projection system should not be confounded with other ipsiand contralateral connections of the AON and MOB (Ha-

berly and Price, '78; Alheid et al., '84; Reyher et al., '88; De Carlos et al., '89). These are projections persisting into adulthood, since they were analyzed in adult animals. Furthermore, they originate mainly from spiny, pyramidal-type cells clustered in layer II/III of the AON (Haberly and Price, '78). We have presented evidence that NPY-ir axons terminate on granule cells, probably the somata. However, the fibers do not begin to form terminal plexuses immediately after arriving in the GRL of MOB, but start during the second postnatal month. In the rat, the majority of MOB

DEVELOPMENT OF NPY IN CAT OLFACTORY BULB

granule cells are generated and differentiate postnatally (Altman, '69; Bayer, '83; Kishi, '87). The proportion of morphologically immature granule cells decreases drastically after the fourth week (Kishi, '87). We assume a similar time course for the cat, since for instance substance P-ir granule cells reach maximal density a t around this time (own unpublished data). So it is possible that by the end of the first postnatal month the granule cell population has reached the necessary degree of structural and functional maturity to support a dense NPY-ir plexus. The reduction of the ipsilateral NPY-ir terminal plexus density and the elimination of the NPY-ir contralateral projection may coincide with the reduction of the spine density of granule cells in the MOB. I t is known from the ferret olfactory bulb that granule cells lose spines after the third postnatal month, a time that represents the end of prey imprinting in this species (Apfelbach, '86). Ferrets, like cats, are carnivores. In cats, spine reduction has not been investigated, but behavioral data indicate that the prey imprinting phase in cats also ends between the third and fifth postnatal month (Leyhausen, '75).So, the end of this sensitive period is characterized by structural modifications of not only spines of granule cells, but also the NPY-ir projection system. However, the functional role of the transient contralateral projection and significance of a strong reciprocal connection of the olfactory bulbs during development is unknown.

The NPY-ir system in comparison to previous studies The axonal ramification pattern is the most distinctive feature of neuronal types (Jones, '75). We have shown that all NPY-ir neurons in the peduncular and bulbar white matter have the same axonal morphology. Therefore, we consider them as one neuronal type that is characterized by looping axons. The axon collaterals innervate the GRL of the MOB and the AOB, and layer II/III of the AON. We have discussed the suggestion that the NPY-ir fibers in the

Fig. 13. Camera lucida drawn overview of parasagittal sections a t two medio-lateral levels of the P17 cat. (a) NPY-ir fibers in the white matter (arrows) of the hlOB (appears to the right side) fuse to form the TO1 below the AON (center) and course towards the striatal (CA and ACC) border. They have the same orientation as the axons present in the AC (the lunate structure appears to the left). The NPY-ir neurons with looping axons are distributed along with the TO1 (large dots). The innervation of layer IIDII of MOB, AOB, and AON has not yet formed completely. Fibers traveling in the deep half of layer I can be identified surrounding ail AON divisions without preference for the AONe (adjacent to the LOT). Dorsally, the layer I plexus overlaps with fibers scattered in layer I of the adjacent cortical gyrus (C). Ventro-caudally, it extends into layer I of the OT. NPY-ir neurons of cortical and striatal types are indicated by small dots: some are represented in deeper cortical layers, many are present in the CA and ACC. These neurons differ in their morphology from the AON/MOB population. They have smaller somata. In the striatum, they have locally ramifying axons arranging in a discontinuous pattern of densely innervated and terminal-poor patches (the latter are outlined, and it is evident that the neurons sit outside these patches). Ventrally, the ACC is continuous with layer I1 of the OLT, and neurons and terminal patches are present in the cortical parts of this structure (Wahle and Meyer, '86).We did not represent the entire striatal plexus, and NPY-ir structures present in the cortex and in more caudal parts (e.g. in HLA, STM). (b) At this alternate level, the olfactory limb of the AC is represented containing many long NPY-ir axons. TO1 fibers ascend in the AC. Striatal neurons (small dots) are indicated, as well as the most posterior AON/MOB-type NPY-ir cells (large dots). Terminal plexuses, drawn in (a) are not given.

486

AC, which were already shown in the diagrams of Woodhams et al. ('86), arise from the NPY-ir neuronal type of the bulbar and peduncular white matter. This is supported by our anatomical reconstructions, by the developmental time course of the projection, and by the innervation pattern. It is very unlikely that the NPY-ir fibers originate from sources other than the bulbar and peduncular NPY-ir population. Afferents to the olfactory bulb and olfactory peduncle have, so far, not been shown to contain NPY. Also, other neurons in neighboring structures, for example, striatal NPY-is cells, do not contribute to the NPY-ir fiber tracts in TO1 and AC or to the bulbar and peduncular plexuses. The striatal neurons are local axon cells (DiFiglia and Aronin, '82; Vincent e t al., '831, and their axon plexuses are confined to the anterior striatal border (see our overview drawings). The NPY-is fibers in TO1 and olfactory limb of the AC are oriented almost perpendicular to this border. In the cat, rat, and marmoset the NPY-ir neurons are localized in the white matter extending from the MOB to the border of the striatum. Furthermore, the dendritic variation and the looping axonal ramification pattern of the NPY-ir neurons are identical, as are NPY-ir fibers in TO1 and AC, and the NPY-ir plexuses innervate the MOB, the AOB, and the AON. The only difference occurred in soma diameters of the NPY-ir neurons, which are larger in the cat compared to the rat and marmoset. It seems to us that the NPY-ir neuron system is highly conserved in mammals and that species differences, as discussed by Ohm et al. ('88), do not exist. NPY-ir neurons in the MOB have been classified as short axon cells (Gall et al., '86; Scott et al., '87; Mntsutani et al., '88, '89) mainly according to their location in the deepest layer of the MOB. This result was seemingly confirmed by the analysis of the proximal axonal patterns of the NPY-ir neurons (Scott et al., '87), which, however, entirely missed the long, posteriorly oriented collaterals. The studies focused on NPY-ir neurons in the MOB proper and did not recognize the entire NPY-system in the olfactory bulb and peduncle. We have demonstrated in the present study that NYY-ir neurons localized in the olfactory bulb are only part of a much larger NPY-ir neuronal population, which is present in the bulbar and peduncular white matter. As a consequence, we propose here that the NPY-ir type should not be regarded as "deep short axon cells," because these cells represent true interneurons of the MOB (Price and Powell, "70; Schneider and Macrides, '78) and should not be expected in structures outside the MOB. For the few NPY-ir neurons found within the E P L of the MOB in our material, also described by Scott et al. ('87), a classification of "superficial short axon cells" (Scott et al., '87) may be discussed. We did not succeed in tracing the axons far enough to demonstrate that they enter the white matter. On the other hand, they differ neither in their proximal axonal pattern, nor in their soma size nor dendritic morphology from NPY-ir neurons of the white matter. Moreover, superficial short axon cells are reported to have spiny dendrites (Pinching and Powell, '71; Schneider and Macrides, '78). Tn our NPY-ir material, however, we did not observe spiny dendrites. Scott et al. ('87) pointed out that the spines arc probably not labeled, because their content of peptide is too low to be detected by immunohistochemistry. This is extremely unlikely. For instance, our analysis of enkephalinergic dwarf cells of the cat olfactory tubercle clearly revealed the spiny dendrites of this cap region cell type (Meyer and Wahle, '86). It is much more likely that the

500~1 m

P10

Fig. 14. NPY-ir axons growing through the AC in the P10 cat; horizontal plane. The sketch indicates the anatomical position of the brachium of the AC. The gray shading across the midline is the part of the AC represented in camera lucida drawing. Many fibers can be identified, and arrows mark growth cones. Note that fibers grow into both directions, crossing the midline.

3

i

C. SANIDES-KOHLRAUSCH AND P. WAHLE

488

intensely labeled dendrites of the NPY-ir neurons are indeed lacking spines and that the NPY-ir neurons of the EPL in fact do not represent “superficial short axon cells,” but probably are displaced individuals of the white matter cell population. The quality of the NPY-immunohistochemistry allows a correlation with cell types classified in Golgi-impregnated material. Intrinsic neurons with distinctive axonal patterns were recently described by L6pez-Mascaraque et al. (’86) in the hedgehog MOB. One type (Fig. 8 of their paper) shows a striking resemblance to the NPY-ir type with respect to the smooth dendrites and the ipsilateral axonal ramification pattern. The authors already described that axon plexuses of this cell type may not be entirely local. This suggests that a NPY-ir type neuron system may also be present in the hedgehog. Support comes from another recent Golgi study analyzing the hedgehog AON. It describes neurons with smooth dendrites and axon plexuses very similar to the NPY-ir neurons in the cat, rat, and marmoset AON (Valverde et al. ’89; their Figs. 8a,b, lle). In conclusion, we have demonstrated that the NPY neuron system in the cat does not form part of the short axon cell class as originally concluded from data obtained in rodents. In contrast, our reinvestigation of the rat NPY neuron system reveals striking similarities that also account for a primate species, the marmoset. In our opinion, the NPY-ir neurons are representatives of a phylogenetically conserved neuron system present in the white matter of the mammalian olfactory bulb and peduncle. The analysis of the postnatal development gives insight to the organization of the projection pattern and its changes in ontogenesis. The similarity of the NPY system in adult mammals suggests a similarity of the developmental processes leading to the adult pattern. The present study will contribute to an understanding of the functional role of the bulbar/peduncular NPY neuron system.

ACKNOWLEDGMENTS We thank Drs. J.M. Polak, B. Chronwall, and T. O’Donohue for generous supply of antiserum. We thank Drs. K. Albus and H. Redies for discussion and comments on the manuscript, and Dr. John Yin for improving the English. This work was supported by Deutsche Forschungsgemeinschaft Schwerpunkt “Verhaltensontogenie.”

LITERATURE CITED Alheid, G.F., J . Carlsen, J. De Olmos, and L. Heimer (1984) Quantitative determination of collateral anterior olfactory nucleus projections using a fluorescent tracer with an algebraic solution to the problem of double retrograde labeling. Brain Res. 29217-22. Allen, J.M., T.J. Crow, and J.M. Polak (1983b) Costorage of a novel peptide, neuropeptide Y, with catecholamines in the rat locus coeruleus. J. R. Microscop. Sci. 18r77. Allen, J.M., T.E. Adrian, K. Tatemoto, T.J. Crow, S.R. Bloom, and J.M. Polak (1983a) Neuropeptide Y distribution in the rat brain. Science 221r877-879. Altman, J. (1969) Autoradiographic and histological studies of postnatal neurogenesis. IV. Cell proliferation and migration in the anterior forebrain with special reference to persisting neurogenesis in the olfactory bulb. J. Comp. Neurol. 137~433458. Apfelbach, R. (1986) Imprinting on prey odours in ferrets (Mustela putorius j.fur0 L.)and its neural correlates. Behav. Proc. I2:363-381. Baker, H. (1986) Species differences in the distribution of substance P and tyrosine hydroxylase immunoreactivity in the olfactory bulb. J. Comp. Neurol. 252:206-226. Bayer, S.A. (1983) 3H-Thymidine-radiographic studies of neurogenesis in the rat olfactory bulb. Exp. Brain Res. 50~329-340.

Berman, A.L., and E.G. Jones (1982) The Thalamus and Basal Telencephalon of the Cat. A Cytoarchitectonic Atlas with Stereotruic Coordinates, 1st ed. Madison: University of Wisconsin. Chronwall, B.M., T.N. Chase, and T. O’Donohue (1984) Coexistence of neuropeptide Y and somatostatin in rat and human cortical and rat hypothalamic neurons. Neurosci. Lett. 52:213-218. Chronwall, B.M., D.A. DiMaggio, V.M. Massary, V.M. Pickel, D. Ruggiero and T. O’Donohue (1985) The anatomy of neuropeptide Y containing neurons in the rat brain. Neuroscience 15r1159-1182. Crosby, E.C., and T. Humphrey (1939) Studies of the vertebrate telencephaIon. I. The nuclear configuration of the olfactory and accessory olfactory formations and of the nucleus olfactorius anterior of certain reptiles, birds, and mammals. J. Comp. Neurol. 71r121-213. De Carlos, J.A., L. L6pez-Mascaraque, and F. Valverde (1989) Connections of the olfactory bulb and nucleus olfactorius anterior in the hedgehog (Erinaceus europaeus): Fluorescent tracers and HRP study. J. Comp. Neurol. 279601418. De Quidt, M.E., and P.C. Emson (1986) Distribution of neuropeptide Y-like immunoreactivity in the rat central nervous system-11. Immunohistochemical analysis. Neuroscience 18:545-618. DiFiglia, M., and N. Aronin (1982) Ultrastructural features of immunoreactive somatostatin neurons in the rat caudate nucleus. J. Neurosci. 2:12671274. DiMaggio, D.A., B.M. Chronwall, K. Buchanan, and T. O’Donohue (1985) Pancreatic polypeptide immunoreactivity in rat brain is actually neuropeptide Y. Neuroscience 15,1149-1158. Gall, C.M., K. Seroogy, and N. Brecha (1986) Distribution of VIP- and NPYlike immunoreactivity in rat main olfactory bulb. Brain Res. 374:389394. Haberly. L.B., and J.L. Price (1978) Association and commissural fiber systems of the olfactory cortex of the rat. 11. System originating in the olfactory peduncle. J. Comp. Neurol. 181:781-808. Innocenti, G.M., S. Clarke, and R. Kraftsik (1986) Interchange of callosal and association projections in the developing visual cortex. J. Neurosci. 6:1384-1409. Jones, E.G. (1975) Varieties and distribution of non-pyramidal cells in the somatic sensory cortex of the squirrel monkey. J. Comp. Neurol. 243r205268. Katz, M.J. (1985) How straight do axons grow? J. Neurosci. 5:589-595. Kishi, K. (1987) Golgi studies on the development of granule cells of the rat olfactory bulb with reference to migration in the subependymal layer. J. Comp. Neurol. 258:112-124. Leyhausen, P. (1975) Verhaltensstudien an Katzen. Berlin: Parey. Lohman, A.H.M. (1963) The anterior olfactory lobe of the guinea pig. A descriptive and anatomical study. Acta Anat. (Basel) 53 (Suppl. 49):l109. L6pez-Mascaraque, L., J.A. De Carlos, and F. Valverde (1986) Structure of the olfactory bulb of the hedgehog (Erinanaceus europaeus): Description of cell types in the granular layer. J. Comp. Neurol. 253:135-152. McLean, J.H., and M.T. Shipley (1988) Postmitotic, postmigrational expression of tyrosine hydroxylaae in olfactory bulb dopaminergic neurons. J. Neurosci. 8:365%3669. Maslim, J., M. Webster, and J. Stone (1986) Stages in the structural differentiation of retinal ganglion cells. J. Comp. Neurol. 254,382-402. Matsutani, S., E. Senba, and M. Tohyama (1988) Neuropeptide- and neurotransmitter-related immunoreactivities in the developing rat olfactory bulb. J. Comp. Neurol. 272r331-342. Matsutani, S., E. Senba, and M. Tohyama (1989) Distribution of neuropeptide-like immunoreactivities in the guinea pig olfactory bulb. J. Comp. Neurol. 280.577-586. Meyer, G., and R. Ferres-Torres (1984) Postnatal maturation of non-pyramidal neurons in the visual cortex of the cat. J. Comp. Neurol. 228~226-244. Meyer, G., and P. Wahle (1986) The olfactory tubercle of the cat. I. Morphological components. Exp. Brain Res. 62515-527. Nakagawa, Y., S. Shiosaka, P.C. Emson, and M. Tohyama (1985) Distribution of neuropeptide Y in the forebrain and diencephalon: an immunohistochemical analysis. Brain Res. 361:524. Ohm, T.G., E. Braak, A. Prohst, and A. Weindl (1988) Neuropeptide Y-like immunoreactive neurons in the human olfactory bulb. Brain Res. 451:29&300. O’Leary, D.D.M., and B.B. Stanfield (1985) Occipital cortical neurons with transient pyramidal tract axons extend and maintain collaterah to subcortical but not intracortical targets. Brain Res. 336:32&333. OLeary, D.D.M. (1987) Remodelling of early axonal projections through the selective elimination of neurons and long axon collaterals. In: Selective

DEVELOPMENT OF NPY IN CAT OLFACTORY BULB Neuronal Death. Chichester: Wiley. (Ciha Foundation Symposion 126), pp. 113-142. Pinching, A.J., and T.P.S. Powell (1971) The neuron types of the glomerular layer of the olfactory bulb. J. Cell Sci. 9:305-345. Price, J.L., and T.P.S. Powell (1970) The mitral and short axon cells of the olfactory bulb. J. Cell Sci. 7:631-651. Reyher, C.K.H. (1988) Persistence of the pars externa system of the anterior olfactory nucleus in a microsmatic primate, Callithrix jacehus. Brain Res. 457:169-175. W.K. Schwerdtfeger, and H.G. Baumgarten (1988) InterbulReyher, C.K.H., bar axonal collateralization and morphology of anterior olfactory nucleus neurons in the rat. Brain Res. Bull. 20:549-566. Sanides-Buchholtz, C. (1988) Postnatal development of NPY-ir structures in the cat olfactory bulb. In N. Elsner and F.G. Barth (eds): Sense Organs: Interfaces between environment and behavior; Proceedings of the 16th Gottinger Neurobiology Conference. Stuttgart: Thieme, p. 74. Schneider, S.P., and F. Macrides (1978) Laminar distribution of interneurons in the main olfactory bulb of the adult hamster. Brain Res. Bull. 3:7382. Scott, J.W., J.K. McDonald, and J.L. Pemberton (1987) Short axon cells of the rat olfactory bulb display NADPH-Diaphorase activity, neuropeptide

489 Y-like immunoreactivity, and somatostatin-like immunoreactivity. J. Comp. Neurol. 260:378-391. Stephan, H. (1975) Handbuch der mikroskopischen Anatomie des Menschen IV/9 Allocortex. 2nd ed. Berlin: Springer. Valverde, F., L. L6pez-Mascaraque, and J.A. De Carlos (1989) Structure of the nucleus olfactorius anterior of the hedgehog Erinaceus europaeus. J. Comp. Neurol. 279:581400. Vincent, S.R., 0.Johansson, T. Hokfelt, L. Skirholl, R.P. Elde, L. Terenius, J. Kimmel, and M. Goldstein (1983) NADPH-Diaphorase: A selective histochemical marker for striataJ neurons containing both somatostatin- and avian pancreatic polypeptide (APP)-like immunoreactivities. J. Comp. Neurol. 21 7r252-263. Wahle, P., and G. Meyer (1986) The olfactory tubercle of the cat. 11: Immunohistochemical compartmentation. Exp. Brain Res. 62:52%540. Wahle, P., and G. Meyer (1987) Morphology and quantitative changes of transient NPY-ir neuronal populations during early postnatal development of the cat visual cortex. J. Comp. Neurol. 261:165-192. Woodhams, P.L., Y.S.Allen, J. McGovern, J.M. Allen, S.R. Bloom, R. Balazs, and J.M. Pol& (1985) Immunohistochemical analysis of the early ontogeny of the neuropeptide Y system in rat brain. Neuroscience 15t173202.