THE JOURNAL OF COMPARATIVE NEUROLOGY 321:477487 (1992)

Localization of Kynurenine Aminotransferase Immunoreactivity in the Rat Hippocampus FU DU, WERNER SCHMIDT, ETSUO OKUNO, RYO KIDO, CHRISTER KOHLER, AND ROBERT SCHWARCZ Maryland Psychiatric Research Center, University of Maryland School of Medicine, Baltimore, Maryland 21228 (F.D., W.S., R.S.); Department of Biochemistry, Wakayama Medical College, Wakayama 640, Japan (E.O., R.K.); ASTRA Research Center, Department of Pharmacology, 151 85 Sodertalje, Sweden (C.K.)

ABSTRACT The localization and distribution of kynurenine aminotransferase (KAT), the biosynthetic enzyme of the excitatory amino acid receptor antagonist, kynurenic acid, was studied in the rat hippocampal formation with immunohistochemical methods. The enzyme was found mainly in glial cells that could be distinguished as 3 types on the basis of their shapes and locations. Typically, these cells shared the morphological features of astrocytes and exhibited glial fibrillary acidic protein immunoreactivity as demonstrated by a double-labeling technique. The distribution of KAT-containing glial cells was heterogeneous throughout the hippocampal formation. In the hippocampus, the stratum lacunosum-moleculare of Ammon's horn and the hilus contained a higher density of KAT-positive glial cells than other regions, whereas the lowest density of KAT glial cells was observed in the granule cell layer of the dentate gyrus and in the stratum radiatum of CA subfields. In the subicular complex, the density of KATcontaining glial cells was generally higher in the superficial than in the deep layer. Hippocampal neurons exhibiting KAT immunoreactivity, distinguished as nonpyramidal cells, were very few in number and mainly distributed in strata oriens and pyramidale of Ammon's horn. Substantially more KAT-positive neurons were observed in layers I1 and I11 of the subicular complex. The organization of cellular elements containing KAT may be of relevance for the function and possible dysfunction of kynurenic acid in the rat hippocampal formation. o 1992 Wiley-Liss, Inc. Key words: astrocytes, excitotoxicity, limbic, kynurenic acid

Kynurenine aminotransferase (KAT) is the biosynthetic enzyme of kynurenic acid (KYNA), a broad spectrum antagonist of cerebral excitatory amino acid ( E M ) receptors. In the rat brain, only a single KAT is responsible for the production of KYNA at physiological concentrations of its bioprecursor, kynurenine (Okuno et al., '91). This enzyme has been purified to homogeneity, and its catalytic properties have been characterized in some detail (Okuno et al., '90, '91). In addition, polyclonal anti-rat KAT antibodies were raised in rabbits and were used for preliminary immunohistochemical studies in the rat brain (Okuno et al., '90). In excellent agreement with the results from biochemical work in the rat striatum (Turski et al., '891, the data indicated a meferential localization of KAT in astrocytelike cells. As a normal constituent of the mammalian brain (Carla et al., '88; Turski et al., '88) and a Possible endogenous modulator of EAA receptor function (Schwarcz et d.,'92),

o 1992 WILEY-LISS, INC.

KYNA may play a role in pathological brain processes, particularly in diseases that may be caused by endogenous excitotoxins such as glutamate, aspartate, homocysteate, or quinolinate (Meldrum and Garthwaite, '90; Do et al., '91; Schwarcz and Du, '91). Frequently, these diseases are associated with selective neuronal necrosis in the hippocampal formation, a brain region containing discretely distributed EAA receptors (Monaghan et al., '83) and a complex but well defined EAA network (Ottersen and StormMathisen, '89). KYNA dysfunction in the hippocampus may therefore be involved, e.g., in the pathogenesis of cerebral ischemia, temporal lobe epilepsy, and Alzheimer's disease (Schwarcz et al.,'84, '92; Stone and Burton, '88). Accepted March 9,1992. Address reprint requests to Robert Schwarcz, Maryland Psychiatric Research Center, P.O. Box 21247, Baltimore, MD 21228.

478

F. DU ET AL.

In preparation of future work which will examine a possible participation of KYNA in the pathogenesis of neuropsychiatric diseases afflicting the limbic system, we have now performed a detailed immunohistochemical study of KAT-containing cells in the normal rat hippocampus. A preliminary account of this work has been published in abstract form (Schwarcz et al., '90).

MATERIALS AND METHODS Ten adult male Sprague-Dawley rats (250-300 g) were anesthetized (chloral hydrate, 400 mg/kg, i.p.) and perfused via the ascending aorta with 50 ml saline followed by 500 ml of a fixative containing paraformaldehyde, lysine, and periodate (McLean and Nakane, '74). Brains were removed and postfixed in the same fixative for 3 hours at 4°C. Following washes in 0.1 M phosphate-buffed saline (PBS; pH 7.4) containing 15%sucrose for 48 hours (4"C), the brains were frozen and stored at -80°C. Twenty or thirty micron cryostat sections were cut either coronally or horizontally from the whole hippocampi of 7 rats, and every second and third sections of each series of 6 were separately collected in cold PBS. The first set of sections was used either for Nissl-staining or for control experiments (cf. below). Sections from the second set were incubated free-floating at 4°C for 3 days with anti-rat KAT antiserum diluted 1:1500 in 0.01 M PBS (pH 7.4) containing 0.3% Triton X-100 (Sigma, St. Louis, MO) and 1% normal goat serum. Subsequently, KAT immunoreactivity (KAT-i) was visualized according to the avidin-biotincomplex (ABC) method of Hsu et al. ('811, as described in a previous study of human hippocampal quinolinic acid phosphoribosyltransferase (Du et al., '90). In addition, some immunostained sections were counterstained with thionin. For the study of colocalization of KAT and glial fibrillary acidic protein (GFAP), 10-20 pm cryostat sections were cut coronally at the dorsal level and horizontally at the ventral level of hippocampi from 3 rat brains, and every second section of each series of 10 was collected in cold PBS. All sections were then incubated in a mixture containing rabbit anti-KAT antiserum and a mouse monoclonal antibody to GFAP (Boehringer Mannheim Biochemicals, Indianapolis, IN) for 3 days at 4°C. Following the primary incubation, the sections were processed first for GFAP immunoreactivity (GFAPd) according to the indirect irnmunofluorescence

Abbreviations

al ABC

EAA CA1,2,3a-c GFAP-i GFAP gl

3HAO

KAT-i JSYNA

KAT mi NMDA Pl

QPRT

slm so SP sr S

Alveus avidin-biotin-complex excitatory amino acid fields of the hippocampus GFAP immunoreactivity glial fibrillary acidic protein granule cell layer of the dentate gyrus 3-hydroxyanthranilic acid oxygenase KAT immunoreactivity kynurenic acid kynurenine aminotransferase molecular layer of the dentate g y m s N-methyl-D-aspartate polymorphic layer of the dentate gyrus quinolinic acid phosphoribosyltransferase stratum lacunosum-moleculare stratum oriens stratum pyramidale stratum radiatum subiculum proper

methods of Coons ('58). Briefly, sections were incubated with fluorescein isothiocyanate-conjugated goat anti-mouse immunoglobulin G (Boehringer Mannheim Biochemicals) for 1 hour at room temperature. The sections were thoroughly washed and mounted in glycerin-PBS (1:3) and photographed under a Nikon fluorescence microscope. Subsequently, the sections were removed from the slides, and processed for KAT-i with the ABC method as described above. After being dehydrated and coverslipped in Permount (Fisher, Fair Lawn, NJ), the sections were photographed a second time under a brightfield microscope. To evaluate the relative densities of glial cells containing KAT-i in the hippocampal formation, cell counts were performed in all hippocampal subfields by using a brightmagnification with an ocular grid. field microscope at 4 0 0 ~ A total of 8 sections cut from both the dorsal and ventral hippocampi of 2 rats were selected for cell counting. KAT-i glial cells were generally counted twice, by moving the grid randomly in the same lamina of each subfield, and the mean numbers were used for computation. In addition, 4 coronal sections cut from the dorsal hippocampi of 2 rats were chosen for the size measurement of the cell bodies of both KAT-i glial cells and neurons. The size of all KAT-positive neurons observed in the entire hippocampal formation of these sections was measured. However, size measurements of KAT-i glial cells were performed only in the stratum radiatum of CA1 and in the molecular and polymorphic layers of the dentate gyrus, since most typical KAT-positive glial cells were observed in these areas. The anti-KAT antiserum used in this study was raised in rabbit by immunization with homogeneous rat kidney KAT, and has been used previously for immunohistochemistry (Okuno et al., '90). Again, in the course of the present study, controls were made in adjacent sections as follows: (1) omission of the first specific antiserum, (2) use of normal rabbit serum instead of anti-KAT antiserum, and (3) preabsorption of the specific antibody with pure rat kidney KAT at 4°C for 24 hours. In addition, heterologous absorption controls were also made in sections taken from different levels of the hippocampal formation. In these tests, the anti-KAT serum used was preincubated at 4°C for 24 hours with purified rat 3-hydroxyanthranilic acid oxygenase (3HAO), quinolinic acid phosphoribosyltransferase (QPRT), two enzymes which have been shown previously to be present in cellular elements similar to those containing KAT-i (cf. Kohler et al., '88; Okuno et al., '90) and pure aspartate aminotransferase. The terminology used in this paper is mainly based on the studies of Lorente de NO ('34). To facilitate the description of KAT immunostaining, a summary of the hippocampal parcellations is given. Thus, the hippocampal formation refers here to the brain areas including the dentate gyrus, Ammon's horn (CA subfields 1-3), and the subicular complex (presubiculum, subiculum proper, parasubiculum). The CA subfields can be further laminated into the alveus, stratum oriens, stratum pyramidale, stratum lucidum (or layer of mossy fibers), stratum radiatum, and stratum lacunosum-moleculare. The dentate gyrus is composed of only 3 layers: the molecular, the granule cell, and the polymorphic layers (cf. Fig. 1A).

RESULTS In this study, KAT-i was detected throughout the entire rat hippocampal formation (Fig. lA,B). No KAT-i was

Fig. 1. Photomicrographs of 20 pm sections cut through the dorsal (A) and ventral (B,C) portions of the rat hippocampal formation. A An immunostained coronal section, showing the general distribution of KAT-i. B and C: Two adjacent horizontal sections were incubated with

anti-KAT antibody (B) and anti-KAT antibody preabsorbed with pure KAT ( C ) . Note the total absence of KAT-i in C. b and c in A and B indicate CA3b and CA3c, respectively. Scale bar = 500 pm and applies to A-C.

480

F. DU ET AL.

observed either in the absence of the primary antibody or after substitution of normal rabbit serum for anti-KAT antiserum. Preabsorption with purified rat kidney KAT totally abolished KAT-i observed in adjacent sections (Fig. 1C). In addition, heterologous preabsorption with 3HA0, QPRT or aspartate aminotransferase did not noticeably affect either the distribution pattern or the staining intensity of KAT-i observed in adjacent sections (micrographs not shown). Observations of coronal sections through the dorsal hippocampus and horizontal sections through the ventral hippocampus revealed an identical pattern in the organization of KAT-i (Fig. 1A,B). The same pattern of KAT-i was seen in the hippocampal formation of all rats examined. Notably, KAT-i was preferentially associated with glial cells and sporadically with neurons (see below).

KAT-i glial cells in the rat hippocampal formation Morphological features. Glial cells exhibiting KAT-i were generally intensely immunostained. Notably, the cell bodies and processes of these cells had various sizes and shapes and were obviously well organized. Based on their morphological features, they were distinguished into 3 types at the light microscopic level. The first type of KAT-i glial cells represented the majority of immunoreactive glial cells observed in the rat hippocampal formation (Fig. 2A). Most of these cells had stellate cell bodies, though elongated cells were also noticed. The cells of this type were relatively large in size with 90% measuring 6-9 pm between the most distant points of the cell body. In general, the cells showed numerous processes radiating from the cell body. These processes were mostly long, smooth, and straight, and some were clearly seen coursing towards and reaching blood vessels (Fig. 2A, inset). The second type also represented a large population of KAT-i glial cells (Fig. 2B). Compared with the first type described above, the cells of this type were substantially smaller. As measured between the most distant points of the cell body, 90% of these cells had a cell body diameter between 4.5 and 6 pm. These cells generally had a round or oval cell body. In KAT-immunostained sections, most of them showed only a few processes, which were usually thin and short. It was interesting, however, that many processes of these cells, as demonstrated by a GFAP double-labeling (see below), did not display KAT-i. There were fewer KAT-i glial cells of the third type in the hippocampal formation (Fig. 2C). The cell bodies of this type were very similar, both in size and shape, to those of the first type, but their processes showed a characteristic branching pattern. Typically, 1-3 major processes emanated from one end of the cell body. These major processes further ramified repeatedly and formed an arbor with highly branched processes, most of which were oriented towards the pial surface. Distribution pattern. Glial cells containing KAT-i were found to be present in all subfields of the hippocampal formation, but their distribution was not homogeneous throughout the regions examined (Figs. 1A,B; Fig. 3). The relative densities of KAT-i glial cells were therefore determined by counting KAT-i glial cells in individual laminae of each subfield (Table 1).In the dentate gyrus, the highest density of cells was found in the polymorphic layer. In contrast, the granule cell layer contained substantially

fewer KAT-i glial cells, constituting the lowest density observed in the entire hippocampal formation. A moderate number of KAT-i glial cells was also present in the outer two-thirds of the molecular layer of the dentate gyrus, while the portion of this layer immediately adjacent to the granule cell layer showed only very few KAT-i glial cells (Fig. lA,B). In Ammon’s horn, the cell densities differed not only between layers, but also among different subfields. Thus, CA3c contained a higher density of KAT-i glial cells in the stratum pyramidale than CAl-3a,b. Glial cells exhibiting KAT-i were particularly numerous in the stratum lacunosum-moleculare of all CA subfields, though the cell density gradually decreased from CA1 to CA3. Numerous KAT-i glial cells were also observed in the alveus and in strata oriens and pyramidale, while the stratum radiatum showed a significantly lower density of KAT-i glial cells. In the subicular complex, the density of KAT-i glial cells was comparable to that of Ammon’s horn and the dentate gyrus (Table 1).The presubiculum contained the highest density of KAT-i glial cells, whereas the lowest cell density was found in the subiculum proper. Remarkable differences were also noticed between layers; thus, the molecular layer showed a higher density of KAT-i glial cells than layers I1 and 111, except that layer I1 of the presubiculum contained the highest number of KAT-i glial cells in the subicular complex. In addition to the relative densities described above, other features were observed with regard to the distribution pattern of KAT-i glial cells. Most strikingly, the three types of KAT-i glial cells were apparently distributed in a well organized fashion in the hippocampal formation (Fig. 2D,E). Thus, glial cells of the first type were located in most layers of CAl-3a,b of Ammon’s horn, including the alveus, stratum oriens, stratum pyramidale and stratum radiatum. The majority of KAT-i glial cells observed in the subicular complex also belonged to this type (cf. Fig. 6C). Cells of the second type were predominantly present in the hilar area including both the polymorphic layer of the dentate gyrus and the CA3c region of Ammon’s horn (Fig. 2D,E). In the polymorphic layer, many KAT-i glial cells were located close to the granule cell layer, with processes often dispersed among granule cells (Figs. 2B,D,E; 5C,D). On the other hand, KAT-i cells in the deep polymorphic layer were frequently found to be closely associated with neuronal perikarya (Fig. 4A). In addition, glial cells containing KAT-i in the stratum lacunosum-moleculare of all CA fields also had the morphological features of the second type. In CA1, most of these cells were located along the border of strata lacunosum-moleculare and radiatum, where their processes were often oriented parallel to the long axis of the layer (Figs. 1A; 2D,E). The third type of glial cells was only observed in the molecular layer of the dentate gyrus where their highly branched processes were organized perpendicular to the pial surface (Fig. 2D,E). Apart from immunoreactive glial cells, specific diffuse KAT-i was consistently found in several regions which generally showed a high density of KAT-i glial cells. In Ammon’s horn, evenly distributed diffuse KAT-i was particularly obvious in the stratum lacunosum-moleculare of CA2 and CA3a,b, though it also occurred as a thin layer in CA1 along the border between the stratum lacunosummoleculare and stratum radiatum (Fig. 1A). Diffuse KAT-i was also present in the outer two-thirds of the molecular layer of the dentate gyrus, a distribution identical to that of most KAT-i glial cells in this layer (Fig. 1A). In addition,

Fig. 2. Photomicrographs of coronal sections (20 pm) taken from the dorsal portion of the rat hippocampal formation. A-C: High magnification of glial cells exhibiting KAT-i. A Section from the stratum radiatum of CA1, showing the first type of KAT-i glial cells observed in the rat hippocampal formation. These cells generally have relatively long processes radiating from their large cell bodies. Inset:A KAT-i glial cell (arrow) with processes coursing and terminating on a blood vessel (indicated by an arrowhead). B: Section taken from the granule cell (gl) and polymorphic (PI) layers of the dentate gyms, showing the second type of hippocampal KAT-i glial cells. Note their small cell bodies with relatively few and short processes. C: Section

from the molecular layer of the dentate gyms, showing the third type of KAT-i glial cells. Typically, their major processes emanate from one end of the cell bodies and course towards the pial surface (marked by stars). D: An immunostained section through all layers of CA1 and dentate gyrus, showing the distribution of KAT-i glial cells (some indicated by solid arrows) and KAT-i neurons (arrowheads). Note KAT-positive neurons in strata oriens (so) and pyramidale (sp). The open arrow (also seen in E) indicates the pial surface in the hippocampal fissure. E: The same section as in D, counterstained with thionin. Solid arrows and arrowheads in D and E indicate the same cells. Scale bars: A-C = 50 pm; D,E = 100 pm.

482

F. DU ET AL.

Fig. 3. Camera lucida drawing from one immunostained coronal section (20 bm), showing the locations and relative densities of KAT-i glial cells (represented by dots) and neurons (indicated by triangles) in

TABLE 1. Relative Densities of KAT-i Glial Cells in the Rat Hippocampal Formation’

Ammon’s horn

CA1 CA2 CABa,b CA3c

a1

so

su

sr

slm

1 9 3 c 23 1882 24 170 i- 19

153 2 15 1 3 6 ? 10 152c 7

103 i- 7 112 + 7 129 i 14

313 2 14 222 f 17 206 t 15

-

-

157 t 5 157i- 21 144 i 17 217 c 11

-

-

Dentate gyrus

d

mi 153 c 10

84

i

PI 2 1 3 c 12

8

Subicular complex

Presubiculum Subiculum proper Parasubiculum

Layer I (ml)

Layer I1

Layer 111 (deep layer)

241 ? 23 201 c 9 259 c 11

272 c 22 160 c 9 192 i 7

185 e 8 186 i 14

-

‘Mean number of cells per 0.25 mm2

diffuse KAT-i with less staining intensity was noticed in the extracellular space of stratum pyramidale of CAl-3a,b (cf. Fig. 6A), but was rarely detected in the same layer of CA3c. Instead, punctate KAT-i structures were conspicuous around the nonimmunoreactive pyramidal cells of CA3c (Fig. 4B) and, in some cases, CA3a,b.

Double-labeling study with a monoclonal anti-GFAP antibody With the indirect immunofluorescence method, GFAP-i displayed intense fluorescence that prominently labeled glial processes and, in some cases, cell bodies. In the double-immunostained sections, GFAP-i was clearly exhibited in the processes of glial cells demonstrating KAT-i (Fig. 5A-F). However, a substantially larger number of processes of individual cells was visualized with GFAP immunofluorescence than with the ABC method for KAT.

the rat hippocampal formation. Note that KAT-i glial cells are distributed heterogeneously throughout the region, while KAT-positive neurons are mainly located in strata oriens and pyramidale.

Since three types of KAT-i glial cells are present in the hippocampal formation, attempts were made to examine the possible colocalization of KAT-i and GFAP-i in glial cells of each type. Examination of various subfields which contained different types of KATi glial cells clearly showed an identical pattern of double-staining throughout the hippocampal formation (Fig. 5A-F). Thus, GFAP-i was found to be present in virtually every KAT-i glial cell of all types. In contrast, GFAP-i glial cells lacking KAT-i were occasionally noticed in almost all hippocampal subfields examined.

KAT-i neurons in the rat hippocampal formation Only a small number of neurons exhibited intense KATi. The reaction product was generally granular in appearance and was present throughout the cytoplasm and frequently in the primary processes. The cell nucleus was usually unstained (Fig. 6A-C). In coronal sections, most neurons containing KAT-i had an oval or elongated soma, though KATi perikarya of other shapes were also observed (Fig. 6A-C). The sizes of KAT-i neurons were determined by measuring the longest dimension of the cell body. Thus, 237 KAT-i neurons sampled throughout the hippocampal formation had cell body diameters ranging from 10.5 to 22.5 pm, with more than half of them (58%)measuring 13.5-16.5 km. The morphology of KAT-i neurons was found to be similar in all hippocampal subfields, though the neurons observed in CA3 were generally larger than those in CA1 and more often triangular in shape. The distribution of KAT-i neurons was heterogeneous as revealed in both coronal (cf. Fig. 3) and horizontal sections. Notably, the subicular complex contained the highest density of KAT-i neurons in the hippocampal region (Fig. 6C). These neurons were located in the pyramidal and particularly in the deep layer. In the Ammon’s horn, KAT-i neurons were mainly observed in CA1, and only a few were present in CA2-3 (Fig. 3).Notably, most KAT-i neurons in

483

KAT-CONTAINING CELLS IN RAT HIPPOCAMPUS

and aspartate aminotransferase, an enzyme which is capable of producing KYNA from KYN at very high substrate concentrations [Okuno et al., '911) did not affect KAT-i, suggesting a lack of significant cross-reaction with these enzymes.

KAT-i in the rat hippocampus

Fig. 4. Photomicrographs of 10 p,m horizontal sections taken from the ventral portion of the rat hippocampus. A: Section through the polymorphic layer of the dentate gyrus. Note that many KAT-i glial cells (indicated by arrows) are closely associated with slightly immunoreactive neurons (n). B: Section through the stratum pyramidale of CA3c, showing dense KAT-i granules around nonimmunoreactive pyramidal cells (n). The arrow indicates a KAT-i glial cell. Scale bar = 25 p,m for both A and B.

the CA fields were located in strata oriens and pyramidale. In addition, KAT-i neurons, often stained lightly, were occasionally seen in the polymorphic layer of the dentate gyrus.

DISCUSSION Methodological considerations The specificity of the primary antibody used in the present study has been examined previously by Western blotting and immunotitration (cf. Okuno et al., '90). Here, antibody specificity was further tested using different controls, including preabsorption with pure rat kidney KAT (homologous absorption). The fact that no immunoreactivity was observed in any control experiment strongly supports the authenticity of the antibody and its usefulness as a marker of KAT in brain tissue sections. In addition, three brain enzymes (the astrocytic enzymes 3HAO and QPRT,

The major finding of the present study was that KAT is preferentially contained in glial cells, probably astrocytes, which are distributed heterogeneously throughout the rat hippocampal formation. In addition, intense KAT-i was detected in some hippocampal neurons. Glial cells containing KAT. Glial cells, principally classified as neuroglia and microglia, have been studied extensively using metallic impregnations, as well as hematoxylin and eosin methods (for reviews, see Penfield, '32; Vaughan, '84). Neuroglia can be further categorized into astrocytes and oligodendroglia, and their morphology has been well documented (Carpenter and Sutin, '83; Vaughan, '84).The identification of astrocytes was facilitated by the establishment of GFAP as a specific structural cell marker (Eng et al., '71; Bignami et al., '72). In the astrocyte family, at least 2 types (protoplasmic and fibrous) have been recognized (Privat and Rataboul, '86; Peters et al., '91). In the present study, all glial cells containing KAT were found to be GFAP-positive, and their morphological features closely resembled those of astrocytes (Duffy, '83). Moreover, on the basis of morphological criteria, KAT-i glial cells in the rat hippocampal formation could be ascribed to both types of astrocytes. For example, most KAT-i glial cells observed in the hilus and in the stratum lacunosum-moleculare of CA sectors are likely to be protoplasmic astrocytes, whereas KAT-i glial cells with relatively large cell bodies and long processes, including those seen in the molecular layer of the dentate gyrus, probably are fibrous astrocytes. Neurons Containing KAT. In the present study, only a small number of immunoreactive neurons was detected in the rat hippocampal formation. Although the methodological limitations of immunohistochemistry should be considered, it is likely that the rat hippocampus does not harbor as many KAT-i neurons as other brain regions, such as the brainstem (Du et al., '91a). The morphology and distribution of these hippocampal neurons suggest that they belong to a subpopulation of local circuit neurons, since they are nonpyramidal cells located mainly in strata oriens and pyramidale of the Ammon's horn (cf. Sloviter and Nilaver, '87 and references therein). This characterization of KAT-i neurons is further supported by the fact that no KATpositive fibers were detected either in the alveus or in the fimbria, the main efferent fiber systems of the hippocampal formation. Diffuse KAT-i. Because electron microscopy was not used, it was not possible to determine unequivocally the nature of the diffuse immunoreactivity which was reliably observed in several hippocampal regions (cf. Fig. 1A). Diffuse KAT-i was mainly observed in areas with dense KAT-i glial cells, This type of staining could represent very fine processes of KAT-containing cells or may indicate the extracellular presence of the enzyme, possibly subsequent to cellular release. Notably, the release of brain proteins into the extracellular milieu has been described, and they may serve specific roles in synaptic neurotransmission (Hesse et al., '84; Appleyard et al., '88; CharriautMarlangue et al., '88).

484

Fig. 5. Photomicrographs of 10 pm sections cut through theventral (A,B) and dorsal (C-F) portions of the rat hippocampus, showing the colocalization of KAT-i (A,C,E) and GFAP-i (B,D,F) in the same glial cells. A and B: The same horizontal section through the stratum radiatum of CA2. Arrowheads in both micrographs indicate the same glial cells. C and D The same coronal section through the granule cell

F. DU ET AL.

(91) and polymorphic (pl) layers of the dentate gyrus. Arrowheads in both micrographs indicate the same glial cells. E and F: The same coronal section through the molecular layer of the dentate gyrus. Arrowheads in both micrographs indicate the same glial cells. Scale bar = 50 K r n for A-F.

485

KAT-CONTAINING CELLS IN RAT HIPPOCAMPUS

Functional considerations KAT is the only enzyme capable of synthesizing KYNA in the rat brain at physiological concentrations of its bioprecursor kynurenine (Okuno et al., '91). The preferential localization of KAT in astrocytes therefore suggests that KYNA in the hippocampal formation derives mainly from glial cells. This interpretation is in excellent agreement with biochemical studies conducted both in vitro and in vivo. Thus, KAT activity and tissue levels of KYNA are doubled in hippocampal homogenates obtained seven days after a massive intrahippocampal quinolinate injection, i.e., at a time of pronounced astroglial proliferation (Wu et al., '91). In addition, assessment by microdialysis has revealed that such lesion-induced KYNA increases are also detectable in the extracellular compartment in vivo, either in the presence or in the absence of kynurenine (Wu et al., '91). Taken together, these data strongly indicate that the transamination of kynurenine to KYNA which takes place in astrocytes is the major determinant of the intra- and extracellular concentration of KYNA in the rat hippocampus. The present results are also in accordance with data obtained in the rat striatum, where most KAT-i resides in astrocytes as well. In the striatum, local excitotoxic lesions also cause astrocytic proliferation and a concomitant increase in KAT activity (Bjorklund et al., '86; Schwarcz et al., '92). Compared to the hippocampus, however, the striatum displays a more complex distribution of KAT-containing cells. In particular, the striatum contains a larger proportion of KAT-positive neurons than the hippocampus. Notably, KAT activity and KYNA production in striatal slices are significantly decreased during the early stages after an intrastriatal excitotoxin injection, i.e., after nerve cells have degenerated but before the onset of the astrocytic reaction. Thus, all biochemical data obtained so far are in full support of the contention that all cells labeled with anti-KAT antibodies (i.e., both glial cells and neurons) are indeed capable of biosynthesizing functionally relevant KYNA. Preliminary analyses in several other rat brain regions have revealed that brainstem nuclei contain particularly large numbers of KAT-i neurons. For example, neurons in the zona compacta of the substantia nigra and in the nucleus ruber are prominently endowed with KAT (unpublished data). As in the few KAT-positive hippocampal neurons, KAT-i in all those cells is also punctate in appearance. On the electron microscopic level, these puncta have recently been identified as densely KAT-positive multivesicular elements (Schwarcz et al., '92). The potential role of neuronal KAT clearly remains to be elaborated, In particular, it needs to be determined if intraneuronal KYNA production is critically dependent on the slow sodiumdependent accumulation of kynurenine into KAT-containing nerve cells (Speciale and Schwarcz, '90). Since KYNA is a broad spectrum antagonist of all three classical ionotropic EAA receptors (Perkins and Stone, '821, Fig. 6. Photomicrographs of coronal sections (20 pm) taken from the topographical distribution of KAT in the hippocampus the dorsal portion of the rat hippocampal formation. A Section through indicates how KYNA may affect the function of these stratum oriens (so) and stratum pyramidale (sp) of CA1, showing receptors under both physiological and pathological condiKAT-containing neurons (indicated by arrows). Note dense granular tions. N-methyl-D-aspartate (NMDA), a-amino-3-hydroxyKAT-i reaction product present in the cytoplasm but not in the nucleus. and kainate receptors are B: Section through the stratum radiatum of CA1, showing a KAT-i 5-methyl-4-isoxazolepropionate, neuron (arrow) which was occasionally seen in this layer. C: Section discretely distributed throughout the structure (Monaghan through the deep layer of the subiculurn proper, showing subicular et al., '83; Monaghan and Cotman, '85; Greenamyre et al., KAT-i neurons (arrows). All arrowheads in A-C indicate KAT-i glial '85), where they are likely to serve specific roles, for cells. Scale bar = 50 pm for A-C. example, in synaptic development during ontogeny and in

486

F. DU ET AL.

learning and memory processes (Collingridge and Singer, '90). While it is currently unclear if the steady-state extracellular KYNA concentrations are sufficiently high to exert a tonic inhibitory influence on the any of the receptors, the strychnine-insensitive glycine site associated with the NMDA receptor complex is likely to be a preferential target for KYNA (Kessler et al., '89). Since NMDA receptors are most prominent in the CA1 sector (Monaghan et al., '83), the relatively high density of KAT-positive astrocytes in the CA1 subfield (cf. Table 1)suggests a possible role of KYNA in the regulation of NMDA receptor activity. In this respect, it is noteworthy that intrahippocampal injections of aminooxyacetic acid, a powerful (though nonspecific) blocker of KYNA production in vitro (Turski et al., '89) and in vivo (Speciale et al., '90; Swartz et al., '901, cause seizures and result in preferential excitotoxic lesions in CA1. Both the convulsive and neurodegenerative effects of the drug can be prevented by the co-administration of 2-amino-7-phosphonoheptanoic acid, a selective NMDA receptor antagonist (McMaster et al., '91). More generally, a reduction in extracellular KYNA levels may be causally involved in one or more diseases which afflict the hippocampal formation and which are widely accepted to be "excitotoxic" in nature. NMDA receptormediated hippocampal damage has been observed in animal models of temporal lobe epilepsy, cerebral ischemia and hypoglycemia (Auer and Siesjo, '88), and can also be induced by measles virus (Andersson et al., '91). In all these conditions, as well as in the aged hippocampus (Bjorklund et al., '851, astrocytes undergo morphological changes, and their role in pathophysiological processes is being increasingly appreciated. Interestingly, the pattern of hippocampal KAT-i is significantly altered in an animal model of epilepsy (Du et al., '91b). Moreover, KAT activity is elevated in situations of experimentally induced nerve cell loss (see above), and both KYNA (Moroni et al., '88) and KAT activity (Gramsbergen et al., '92) are greatly increased in the aged rat brain. In addition to KAT, astrocytes harbor several enzymes with putative links to excitotoxicity, such as glutamine synthetase (Martinez-Hernandez et al., '77) and the quinolinate-metabolizing enzymes 3HAO and QPRT (Kohler et al., '88). The interplay between astrocyte-derived glutamine (Goldberg et al., '88), quinolinate, and KYNA may therefore hold the key to EAAs receptor function and dysfunction. Studies designed to pursue these interactions, including the role of EAA in the regulation of extracellular KYNA concentrations (Gramsbergen et al., '89), are currently being performed in our laboratories.

ACKNOWLEDGMENTS We thank Mrs. Joyce Burgess for excellent secretarial assistance. This work was supported by USPHS grants NS 16102 and NS 28236.

LITERATURE CITED Andersson, T., M. Schultzherg, R. Schwarcz, A. Love, C. Wickman, and K. Kristensson (1991) NMDA-receptor antagonist prevents measles virus induced neurodegeneration. Eur. J. Neurosci. 3:66-71. Appleyard, M.E., J.-L. Vercher, and S.A. Greenfield (1988) Release of acetylcholinesterase from the guinea-pig cerebellum in vivo. Neuroscience 25: 133-138. Auer, R.M., and B.K. Siesjo (1988) Biological differences between ischemia, hypoglycemia, and epilepsy. Ann Neurol. 243399-707.

Bignami, A., L.F. Eng, D. Dahl, and C.T. Uyeda (1972) Localization of the glial fihrillary acidic protein in astrocytes by immunofluorescence. Brain Res. 43:423435. Bjorklund, H., M. Eriksdotter-Nilsson, D. Dahl, G. Rose, B. Hoffer, and L. Olson (1985) Image analysis of GFA-positive astrocytes from adolescence to senescence. Exp. Brain Res. 58t163-170. Bjorklund, H., L. Olson, D. Dahl, and R. Schwarcz (1986) Short- and long-term consequences of intracranial injections of the excitotoxin quinolinic acid, as evidenced hy GFA immunohistochemistry of astrocytes. Brain Res. 371:267-277. Carla, V., G. Lombardi, M. Bensi, P. Russi, G. Moneti, and F. Moroni (1988) Identification and measurement of kynurenic acid in the rat brain and other organs. Anal. Biochem. 169:89-94. Carpenter, M.B., and J. Sutin (1983) Human Neuroanatomy. Baltimore: Williams and Wilkins, pp. 134-154. Charriaut-Marlangue, C., L. Aniksztejn, M.P. Roisin, and Y. Ben-Ari (1988) Release of proteins during long-term potentiation in the hippocampus of the anesthetized rat. Neurosci. Lett. 91:308-314. Collingridge, G.L., and W. Singer (1990) Excitatory amino acid receptors and synaptic plasticity. Trends Pharmacol. Sci. 11:290-296. Coons, A.H. 11958) Fluorescent antibody methods. In J.F. Danielli (ed): General Cytochemical Methods. New York: Academic Press, pp. 399422. Do, K.Q., J.M. Klancnik, B. Gahwiler, H. Perschak, H.-G. Wieser, and M. Cuenod (1991) Release of excitatory amino acids. Animal studies and epileptic foci studies in humans. In B.S. Meldrum, F. Moroni, R.P. Simon, and J.H. Woods (eds): Excitatory Amino Acids. New York: Raven Press, pp. 677-685. Du, F., E. Okuno, and R. Schwarcz (1991a) Kynurenic acid hiosynthesis in the rat brain studied hy kynurenine aminotransferase immunohistochemistry. Third IBRO World Congress of Neuroscience Abstract, p56.11. Du, F., E.W. Lothman, E.H. Bertram, J.M. Williamson, and R. Schwarcz (1991b) Altered kynurenine metabolism in a model of chronic epilepsy: Immunohistochemicd analysis. SOC. Neurosci. Ahstr. 17:369.11. Du, F., E. Okuno, W.O. Whetsell, Jr., C. Kohler, and R. Schwarcz (1990) Distribution of quinolinic acid phosphorihosyltransferase in the human hippocampal formation and parahippacampal gyms. J. Comp. Neural. 295:71-82. Duffy, P.E. (1983) Astrocytes: Normal, Reactive, and Neoplastic. New York: Raven Press, pp. 1-23. Eng, L.F., J.J. Vanderhaeghen, A. Bignami, and B. Gerstl (1971) An acidic protein isolated from fibrous astrocytes. Brain Res. 28:351-354. Goldberg, M.P., Monyer, H., and D.W. Choi (1988) Hypo 'c neuronal injury in vitro depends on extracellular glutamine. Neuroscilett. 94t52-57. Gramshergen, J.B.P., W. Schmidt, W.A. Turski, and R. Schwarcz (1992) Age-related changes in kynurenic acid production in rat brain. Brain Rea. (in press). Gramsbergen, J.B.P., W.A. Turski, and R. Schwarcz (1989) Sodiumdependent inhibition o f kynurenate synthesis by quisqualate in rat brain. Soc. Neurosci. Abstr. 15:328.10. Greenamyre, J.T., J.M.M. Olson, J.B. Penney, and A.B. Young (1985) Autoradiographic characterization of N-methyl-D-aspartate-, quisqualate- and kainate-sensitive glutamate binding sites. J. Pharmacol. Exp. Ther. 233:254-263. Hesse, G.W., R. Hofstein, and V.E. Shashoua (1984) Protein release from hippocampus in vitro. Brain Res. 305%-66. Hsu, S.M., L. Raine, and H. Fanger (1981) Use of avidin-biotin-peroxidase complex (ABC) in immunoperoxidase techniques: A comparison between ABC and unlabeled antibody (PAP) procedures. J. Histochem. Cytochem. 29:577-580. Kessler, M., T. Terramani, G. Lynch, and M. Baudry (1989) A glycine site associated with N-methyl-D-aspartic acid receptors: Characterization and identification of a new class of antagonists. J. Neurochem. 5213191328. Kohler, C., L.G. Eriksson, P.R. Flood, J.A. Hardie, E. Okuno, and R. Schwarcz (1988) Quinolinic acid metabolism in the rat brain. Immunohistochemical identification of 3-hydroxyanthranilic acid oxygenase and quinolinic acid phosphoribosyltransferase in the hippocampal region. J. Neurosci. 8:975-987. Lorente de NO, R. (1934) Studies on the structure of the eerebral cortex. 11. Continuation of the study of the ammonic system. J. Psychol. Neurol. 46: 113-177. Martinez-Hernandez, A,, K.P. Bell, and M.D. Norenberg (1977) Glutamine synthetase: Glial localization in brain. Science 1951356-1358. McLean, I.W., and P.K. Nakane (1974) Periodate-lysine-paraformaldehyde

KAT-CONTAINING CELLS IN RAT HIPPOCAMPUS fixative: A new fixative for immunoelectron microscopy. J. Histochem. Cytochem. 22: 1077-1083. McMaster, O.G., F. Du, E.D. French, and R. Schwarcz (1991) Focal injection of aminooxyacetic acid produces seizures and lesions in rat hippocampus: Evidence for mediation by NMDA receptors. Exp. Neurol. 113:378-385. Meldrum, B., and J. Garthwaite (1990) Excitatory amino acid neurotoxicity and neurodegenerative disease. Trends Pbarmacol. Sci. 11:379-387. Monaghan, D.T., and C.W. Cotman (1985) Distribution of N-methyl-Daspartate-sensitive L-[3Hlglutamate-binding sites in rat brain. J. Neurosci. 5:2909-2919. Monaghan, D.T., V.R. Holets, D.W. Toy, and C.W. Cotman (1983) Anatomical distributions of four pharmacologically distinct WL-glutamate binding sites. Nature 306:176-179. Moroni, F., P . Russi, G. Lombardi, M. Beni, and V. Carla (1988) Kynurenic acid in the mammalian brain increases during the aging process o r after the administration of its precursors. In E.A. Cavalheiro, J. Lehmann, and L. Turski (eds): Frontiers in Excitatory Amino Acid Research. New York: Alan R. Liss, pp. 629-636. Okuno, E., F. Du, T. Ishikawa. M. Tsujimoto, M. Nakamura, R. Schwarcz, and R. Kid0 (1990) Purification and characterization of kynureninepyruvate amino-transferase from rat kidney and brain. Brain Res. 534:3744. Okuno, E., W. Schmidt, D.A. Parks, M. Nakamura, and R. Schwarcz (1991) Measurement of rat brain kynurenine aminotransferase a t physiological kynurenine concentrations. J. Neurochem. 57:533-540. Ottersen, O.P., and J. Storm-Mathisen (1989) Excitatory and inhibitory amino acids in the hippocampus. In V. Chan-Palay and C. Kohler (eds): The Hippocampus: New Vistas. New York Alan R. Liss, pp. 97-117. Penfield, W. (1932) Neuroglia and microglia. The interstitial tissue of the central nervous system. In E.V. Cowdry (ed): Special Cytology Vol. 3. New York: Hoeber, pp. 1445-1482. Perkins, M.N., and T.W. Stone (1982) An iontophoretic investigation of the actions of convulsant kynurenines and their interaction with the endogenous excitant quinolinic acid. Brain Res. 247:184-187. Peters, A., S.L. Palay, and H. de F. Webster (1991) The Fine Structure of the Nervous System: Neurons and Supporting Cells. New York: Oxford Press. Privat, A., and P. Rataboul(1986) Fibrous and protoplasmicastrocytes. In S. Fedoroff and A. Vernadakis (eds): Astrocytes. Vol. 1. Development, Morphology, and Regional Specialization of Astrocytes. New York: Academic Press, pp. 105-129. Schwarcz, R., and F. Du (1991) Quinolinic acid and kynurenic acid in the

487 mammalian brain. In R. Schwarcz, S.N. Young, and R.R. Brown !eds): Kynurenine and Serotonin Pathways: Progress in Tryptophan Research. New York: Plenum Press, pp. 185-199. Schwarcz, R., F. Du, W. Schmidt, and E. Okuno (1990) Preferential astroglial localization of kynurenine aminotransferase in the rat hippocampus. SOC.Neurosci. Abstr. 16:491.17. Schwarcz, R., A.C. Foster, E.D. French, W.O. Whetsell, Jr., and C. Kohler (1984) Excitotoxic models for neurodegenerative disorders. Life Sci. 35:19-32. Schwarcz, R., F. Du, W. Schmidt, W.A. Turski, J.B.P. Gramsbergen, E. Okuno, and R.C. Roberts (1992) Kynurenic acid: A potential pathogen in brain disorders. In R.E. Heikkila, J.W. Langston, and A.B. Young !edsJ: Neurotoxins and Their Potential Role in Neurodegeneration. Ann. N.Y. Acad. Sci., in press. Sloviter, R.S., and G. Nilaver (1987) Immunocytochemical localization of GABA-. cholecystokinin-,vasoactive intestinal polypeptide-, and somatostatin-like immunoreactivity in the area dentata and hippocampus of the rat. J. Comp. Neurol. 256:42-60. Speciale, C., and R. Schwarcz (1990) Uptake of kynurenine into rat brain slices. J. Neurochem. 54:156-163. Speciale, C., H.-Q. Wu, J.B.P. Gramsbergen, W.A. Turski, U. Ungerstedt, and R. Schwarcz (1990) Determination of extracellular kynurenic acid in the striatum of unanesthetized rats: Effect of aminooxyacetic acid. Neurosci. Lett. 116:198-203. Stone, T.W., and N.R. Burton (1988) NMDA receptors and ligands in the vertebrate CNS. Prog. Neurobiol. 10:333-368. Swartz, K.J., M.J. During, A. Freese, and M.F. Beal (1990) Cerebral synthesis and release of kynurenic acid: An endogenous antagonist of excitatory amino acid receptors. J. Neurosci. 10r2965-2973. Turski, W.A., J.B.P. Gramsbergen, H. Traitler, and R. Schwarcz (1989) Rat brain slices produce and liberate kynurenic acid upon exposure to L-kynurenine. J. Neurochem. 52: 1629-1636. Turski, W.A., M. Nakamura, W.P. Todd, B.K. Carpenter, W.O. Whetsell JF., and R. Schwarcz (1988) Identification and quantification of kynurenic acid in human brain tissue. Brain Res. 454r164-169. Vaughan, D.W. (1984) The structure of neuroglial cells. In E.G. Jones and A. Peters (edsJ: Cerebral Cortex. Vol. 2. Functional Properties of Cortical Cells. New York: Plenum Press, pp. 285-329. WU, H.-Q., H. Baran, F. Du, and R, Schwarcz (1991) Kynurenic acid production in the rat hippocampus: effects of quinolinate lesion. Soc. Neurosci. Abstr. 17:356.8.

Localization of kynurenine aminotransferase immunoreactivity in the rat hippocampus.

The localization and distribution of kynurenine aminotransferase (KAT), the biosynthetic enzyme of the excitatory amino acid receptor antagonist, kynu...
2MB Sizes 0 Downloads 0 Views