The neurons of the transient subplate zone, considered important for the prenatal development of the cerebral cortex, were shown here to express kynurenine aminotransferase (KAT)-I from embryonic day (E) 16 until postnatal day (P) 7 in the rat. No other cells of brain tissue exerted KAT-I immunoreactivity during this period. From P3 on, the neurons of the subplate gave rise to KAT-I immunoreactive, varicose axons, which entered the thalamus and terminated around thalamic nerve cells that are devoid of KAT-I immunoreactivity. Other subplate markers displayed a different expression pattern during development. Thus, subplate neurons displayed parvalbumin (PV) immuno-reactivity from E16 to P10 and an intense NPY immunoreaction from P7 to P1. They also exhibited nitric oxide synthase immunoreactivity from E16 to P10, whereas on the surface of the subplate neurons, the α7 subunit of the nicotinic acetylcholine receptor (nAChR) was present from P1 to P10. The cells of Cajal–Retzius were nAChR-immunoreactive during this period. Between P1 and P7, the perikarya of subplate neurons also showed an intense immuno-reaction with the N-methyl-d-aspartate (NMDA) receptor subtype R2A. After the first postnatal week, many of the KAT-I positive subplate neurons display a gradual decrease of immunoreactivity and undergo programmed cell death. Since KAT-I persists in the subplate through the period E16–P7, we conclude that KAT-I is a useful and reliable subplate marker in the rat. Since it is assumed that migration of nerve cells is regulated by NMDA receptors, and since kynurenic acid — the only naturally occurring NMDA receptor antagonist — is synthesized by KAT, we suggest that a temporary breakdown of the delicate equilibrium between NMDA and KAT might induce abnormal neuronal migration, giving rise to developmental abnormalities.
The subplate layer is a transient embryonic zone that plays an important role in the ontogeny and phylogeny of the cerebral cortex (Kostovic and Rakic, 1990). Postmitotic neurons, migrating from the ventricular zone up to the superficial layer of the cerebral wall, divide the early preplate into a superficial superplate and a deep subplate (Sidman and Rakic, 1973; Kostovic and Rakic, 1980). Transient neurons of the subplate are considered to be instrumental in the development of the cortex and in the establisment of corticothalamic and thalamocortical connections (see Discussion for details). While, in most mammalian species, including primates, the majority of subplate neurons are lost in the development of the cortex (Kostovic and Rakic, 1980), the subplate persists into adulthood in rodents (Woo et al., 1991). Subplate dysfunction in the developing brain may give rise to diseases such as epilepsy and schizophrenia by disturbing neuronal migration and generating abnormal function of circuits (Rakic, 1988; Akbarian et al., 1993, 1996; Bloom, 1993).
As the only known endogenous N-methyl-d-aspartate (NMDA) receptor antagonist, kynurenic acid displays a neuroprotective effect in neurodegenerative diseases (Du et al., 1992; Nozaki and Beal, 1992; Knyihár-Csillik et al., 1999). Kynurenic acid is known to be produced by the activity of the enzyme kynurenin aminotransferase (KAT). According to the studies of Guidetti et al. (Guidetti et al., 1997), there are two subtypes of this enzyme — KAT-I and KAT-II — with distinctive properties. KAT-I, identical with glutamine transaminase K, has an optimal pH of 9.5, prefers pyruvate as a cosubstrate and is inhibited by glutamine. KAT-II, identical with l-α-aminoadipate transaminase, has an optimal pH of 7.0, shows no preference for pyruvate and is essentially insensitive to inhibition by glutamine. While the presence of KAT-II mRNA in the brain has been identified with Northern blot analysis (Okuno et al., 1991), to date, antibodies for immunohistochemical localization are only available against KAT-I (Knyihár-Csillik et al., 1999). Therefore, in the present study we examined KAT-I immunoreactivity in the cells of the subplate and its remnants in rat embryos, newborn rats and rat pups, at the light and electron microscopic level. Since migration of nerve cells has been proposed to be regulated by NMDA receptors (Komuro and Rakic, 1993), we also studied the expression of the NMDA receptor subtype R2A, and its relation to KAT-I expression in the cells of the subplate. At the same time, we compared the existence of KAT-I immunoreactivity to those of several known markers of the subplate, such as parvalbumin (PV), neuropeptide Y (NPY), nitric oxide synthase (NOS) and the α7-subunit of the nicotinic acetylcholine receptor (nAChR).
Materials and Methods
Investigations were performed on Wistar and Sprague–Dawley rats: 20 rat embryos (embryonic days E14–E21), 9 newborns and 22 rat pups of ages 1–14 days of both sexes. Experiments were carried out in accordance with the European Communities Council Directive (24 November 1986; 86/609/EEC). Rats were anesthetized with ether and subjected to transcardial perfusion, first with 10 ml of 0.1 M PBS (pH 7.4) followed by 200 ml of fixative containing paraformaldehyde, lysine and periodate (PLP) (McLean and Nakane, 1974). Perfusion was carried out at a rate of 10 ml/min. Brains were removed and postfixed in the same fixative for 3 h at 4°C. Vibratome sections, 50 mm thick, were obtained from the parietal cortical area immediately after postfixation. Alternatively, after incubation in PBS containing 15% sucrose for 48 h, the samples were processed in a graded series of sucrose (10%, 20%, 30%) dissolved in PBS, placed in Tissue-Tek embedding medium (Miles, Diagnostic Devision Elkhart, IN) and frozen in a chamber cryostat at –20°C. In this case, freezing was immediately followed by sectioning. Serial, 40 mm thick cryostat sections, were obtained from the paramedian sagittal plane of the whole brain. Following pretreatment with 2% H2O2, incubation with primary antiserum was performed either at 4°C for 3 days or overnight at room temperature, both on free-floating sections and on sections mounted on gelatine-pretreated slides. Primary antiserum used was an anti-KAT polyclonal antibody raised in rabbit against rat kidney KAT-I (Okuno et al., 1990), diluted to 1:1500 in 0.01 M PBS (pH 7.4) containing 1% bovine serum albumin and 5% normal goat serum (NGS). Following a 90 min incubation in biotinylated anti-rabbit secondary antibody, the sections were processed according to the avidin–biotin system (Vecta-stain ABC Elite, Vector Laboratories, Burlingame, CA).
Essentially the same technique was used for the visualization of PV, NOS and NPY immunoreactivity. The PV antibody raised in mouse (Sigma-Aldrich, St Louis, MO) was diluted to 1:20 000; the bNOS antibody raised in mouse (Sigma-Aldrich) was diluted to 1:3000; the NPY antibody raised in rabbit (Sigma-Aldrich) was diluted to 1:2000.
For the detection of the a7-subunit of nAChR, biotinylated α-bungaro-toxin (Molecular Probes, Eugene, OR) was used at a dilution of 1:2000.
KAT-I, PV, NOS, NPY and nAChR immunoreactivity was visualized with diaminobenzidine to which hydrogen peroxide had been added (3 μl of 30% H2O2 to 10 ml of 0.05% diaminobenzidine). Sections were dehydrated in a graded series of ethanol and processed in carbol–xylene. Slides were coverslipped with Permount (Fisher, Fair Lawn, NJ).
The specificity of the immunohistochemical reaction was assessed by means of one of the following treatments: (i) Omission of the first (specific) antiserum. (ii) Use of normal rabbit serum instead of the specific antiserum. (iii) Treatment according to the avidin–biotin complex method, from which one of the steps had been omitted. (iv) Preabsorbtion of the specific antibody with pure rat kidney KAT-I (Okuno et al., 1990) at 4°C for 24 h. None of these specimens showed any immunoreactivity.
Immunocytochemical Localization of the R2A Subunit of the NMDA Receptor
Fixation, preparation and sectioning of the brain was carried out exactly as previously described, except that postfixation in PLP was performed only for 1 h at 4°C. Microscopic analysis of NMDA-R2A immunolabeling showed that the best NR2A labeling with minimal background could be achieved by fixation with periodate–lysine–paraformaldehyde for 1 h. After sectioning either with a vibrotome or with a cryostat, primary antibody incubation with anti-NR2A serum (RBI Research Biochemicals International, Natick, MA) diluted to 1:200 in Tris-buffered saline (TBS) containing 2% NGS, was carried out overnight at 4°C, followed by 1 h incubation at room temperature. The samples were washed three times for 5 min in TBS (pH 8.2) containing 1% NGS. For immunogold labeling, sections were treated with ∼1 nm gold-labeled goat anti-rabbit IgG (H+L) 1 ml gelatine (IGSS) quality (Amersham Life Sciences, UK) diluted to 1:20 in TBS (pH 8.2) containing 0.4% lysine, 0.1% gelatine and 1% NGS, for 90 min at room temperature. After incubation, sections were rinsed 3 × 15 min in TBS (pH 8.2) containing 1% NGS, and 3 × 5 min in distilled water. This was followed by 10–20 min silver intensification. The silver enhancement solution was prepared by mixing equal parts of Enhancer and Initiator of the IntenSE M kit (Amersham Life Sciences). After the silver enhancement process, which was monitored under a microscope with bright-field illumination, specimens were rinsed 3 × 5 min in excess distilled water. Sections were dehydrated in a graded series of ethanol and processed in carbol–xylene. Slides were coverslipped with Permount (Fisher).
Combined KAT-I Immunohistochemistry and Immunogold–Silver Staining Method on the Same Section
Sections of the subplate were double labeled for KAT-I and NMDA-R2A. Sections were labeled for NMDA-R2A with the immunogold–silver staining method as described above, then they were immunolabeled for KAT-I with the ABC method. Sections were photographed on Leitz Diaplan and Nikon Nomarski microscopes.
Double Staining of KAT-I and Parvalbumin Immunoreactivity
Cryostat sections (30 mm thick) of the subplate were incubated in the cocktail medium containing KAT-I antiserum (dilution 1:1500) and parvalbumin antiserum (dilution 1:20 000). After 48 h incubation at 4°C KAT-I was visualized with diaminobenzidine (DAB) using the ABC technique and PV was visualized with Nickel–DAB, using the PAP method.
For electron microscopic immunohistochemistry, KAT-I and NMDA R2A were visualized in Vibratome sections (50 mm), which were treated in the same manner as the light microscopic specimens. After 30 min of osmic acid fixation followed by dehydration, the sections were flat-embedded on liquid release-pretreated slides in Durcupan ACM. Relevant areas were excised with a razor blade under a microscope, remounted to pre-polymerized blocks and sectioned with a diamond knife on a Reichert Ultrotome. Serial sections of silver interference colour were collected on copper slot grids and stained with lead citrate and uranyl acetate. Sections were examined and photographed using a Zeiss Opton 902 electron microscope.
At the light microscopic level it is evident that neurons and glial cells of the embryonic and early postnatal brain did not stain for KAT-I except for the subplate neurons (Fig. 1a,d–f) which displayed an intense KAT-I immunoreaction. The KAT-I-immunopositive cells of the subplate were initially apolar, and later acquired a unipolar morphology. Based on electron microscopy, KAT-I in subplate neurons was bound to ribosomes of the rough endoplasmatic reticulum (Fig. 1b). Immunopositive dendrites of embryonic subplate cells were in synaptic contact with immunonegative axons (Fig. 1b). Distribution of the KAT-I-immunopositive neurons in the subplate layer was not homogeneous, especially at the beginning of extrauterine life. In some places they were arranged in columns or tended to form cell groups consisting of two or three cells attached to each other (Fig. 2a). Later in development, subplate neurons gave rise to KAT-I-immunopositive varicose axons (Fig. 2b). Synapses of KAT-I-immunonegative axons were formed both on the bodies and on the dendrites of subplate cells (Fig. 2c). In both the axo-somatic and the axo-dendritic synapses, the presynaptic elements were KAT-I-immunonegative axons that revealed an immature structure. It should be emphasized that the KAT-I immunoreactivity was present only in postsynaptic localizations. Axons of KAT-I-immunopositive subplate cells were never found to establish synapses with any of the KAT-I-immunopositive cells.
From the second postnatal week on, an increasing number of KAT-I-immunopositive subplate cells exhibited signs of cell death. The apoptotic cells were characterized by a shrinkage of the nucleus and an increased density of the cytoplasm (Fig. 4b,c). However, the subplate itself, i.e. the layer between the sub-ventricular zone and the cortical plate, still persisted.
In the dorsal thalamus, KAT-I-immunopositive axons approaching thalamic cells were varicose; the varicosities of the axons surrounding the cell surfaces of the thalamic cells were loaded with the end-product of the KAT-I immunoreaction. KAT-I-immunopositive axons established pericellular nests around thalamic cells (Fig. 2d). Under the electron microscope it could clearly be seen that KAT-I-immunopositive axons establish synapses both with somata and dendrites of thalamic cells (Fig. 2e).
The cells of the subplate expressed the R2A subunit of the NMDA receptor (Fig. 3a). Using the gold–silver labeling method, the grains representing the receptor were found to be present in the cytoplasm as well as on the surface of subplate cells. It could readily be demonstrated, especially if using the Nomarski optics, that neither the nucleus nor the nucleolus contained any silver grains; they were restricted partly to the cytoplasm and partly to the surface of the cells (Figs. 3b,c). In between the cells presenting NMDA receptors on their surfaces, non-reactive cells could also be seen.
Double staining with the NMDA-R2A receptor antibody and KAT-I antiserum revealed that NMDA receptors were present both in the cytoplasm (Fig. 4a) and in dendritic processes of KAT-I-immunopositive cells.
The presence of KAT-I immunoreactivity characterized the subplate during the embryonic period of the developing rat cortex from E16 to E21 and during the postnatal (P) period P1–P7. PV immunoreactivity and NOS were present in sub-plate cells from E16 to P10. However, the localization and morphological appearance of PV- and NOS-immunoreactive cells differed considerably from that of KAT-I-immunoreactive cell populations. The fusiform PV- and NOS-immunopositive cells were located above KAT-I cells and exhibited a rich arborization of neuronal processes directed against the cortical plate (Fig. 5a–d). Double staining clearly delineated the KAT-I- and PV-immunoreactive sublayers of the subplate (Fig. 5g), which proves that different cell populations express KAT-I and PV without any trace of coexistence. On the other hand, NPY immunoreactivity was characteristic of the subplate only during the postnatal period from P7 to P10 (Fig. 5e). We found expression of the α7-subunit of nAChR in the spherical (mostly unipolar) subplate cells from E16 until postnatal day P10 (Fig. 5f). The α7-subunit of nAChR was also present in embryonic Cajal– Retzius cells.
By the time neurons originating in the ventricular and sub-ventricular zone begin to migrate outwards to establish the initial cortical plate below the pia (Hicks and D’Amato, 1964; Rakic, 1974), there already exists a sizable network of mono-aminergic fibers, the primordial plexiform layer (Marin-Padilla, 1971, 1972, 1978). The function of the earliest generated neurons that form the subplate zone is not fully understood, but several hypotheses have been suggested that underscore their further development of the cerebral cortex. Thus, it has been shown that the developing thalamocortical axons establish temporary synapses with neurons of the subplate, in a so-called ‘waiting (holding) compartment’, until their target cells in the cortical plate acquire the necessary degree of maturation (Rakic, 1976, 1977, 1982, 1988; Kostovic and Rakic, 1980, 1984, 1990; Ghosh et al., 1990; Molnar and Blakemore, 1995; Molnar et al., 1998), similar to callosal axons (Goldman-Rakic, 1982). Neurons of the subplate zone also emit so-called ‘pioneer axons’ which serve as itineraries for developing corticothalamic fibers (McConnell et al., 1989; Allendoerfer and Shatz, 1994). Thus, subplate neurons pioneer the first axonal pathways from the cerebral cortex. According to Molnar (Molnar, 2000), the interactions of the thalamocortical projections with the early generated, largely transient cells of the subplate play a crucial role in their development. Our observation that axons of subplate cells, containing the enzyme KAT-I and expressing NMDA-R2A, establish close contact with developing neurons of the thalamus, is in accord with this notion. Cells of the subplate regulate neuronal migration in the developing cortex through a process controlled by NMDA receptors (Komuro and Rakic, 1993; Feldmeyer and Cull-Candy, 1996). According to Hanganu et al. (Hanganu et al., 2001) subplate neurons receive functional synaptic inputs mediated, inter alia, by NMDA receptors. The density of NMDA receptors is low at birth and increases in the postnatal period (Smith and Thompson, 1999). In this context, it may be significant that migration of nerve cells is regulated by NMDA receptors (Komuro and Rakic, 1993). Double staining proves that the enzyme KAT-I and the R2A subunits of the NMDA receptor are, for the most part, co-localized in nerve cells of the subplate. Other studies suggested the importance of GABA and GABA receptors in cell proliferation and development of the subplate (Meinecke and Rakic, 1992; Haydar et al., 0,0,0,02000). Notwithstanding the above roles, blockade of activity in the subplate produces misguided thalamic connections (Catalano and Shatz, 1998). In this respect, it should be emphasized that not only the presence, but also the neuronal activity of subplate neurons may be factors which play an important role in cortical development (Yuste and Sur, 1999). A spontaneous correlated activity among spatially separated neurons found in the marginal zone of intact hemisphere preparations (Schwartz et al., 1998) might, with due extrapolation, reflect the importance of spontaneous functional activity of the subplate neurons (Hanganu et al., 2001). Therefore, the notion that neurotrophins and activity-dependent development of the cerebral cortex are inseparable entities seems to be a logical consequence of the above premises (Katz and Shatz, 1996), despite recent studies (Crowley and Katz, 1999; Verhage et al., 2000) suggesting that neural activity may not be essential for the establishment of cortical networks. Further evidence against activity-dependent cortical development is the observation that neocortical regionalization can develop in the absence of thalamic innervation (Miyashita-Lin et al., 1999), perhaps due to the effects of local concentrations of regulatory genes (Bishop et al., 2000). In spite of such reservations, it still appears that the subplate may play important roles in the establishment of corticothalamic and thalamocortical pathways as well (Molnar et al., 1998; Molnar, 2000).
Due to the importance of the subplate for normal development of the cerebral cortex, several attempts have been made to find specific markers and morphoregulatory molecules that characterize its function. A fibronectin-like molecule was found both in the developing cerebral cortex and in the subplate zone (Chun and Shatz, 1988), and the antigen SP1 could be successfully used to analyze the fate of the cells constituting the subplate (Wahle et al., 1994). Also, cells of the subplate zone were shown to exhibit strong, transient immunoreactivity of the nerve growth factor-inducible gene VGF (Lombardo et al., 1995), which reflects the presence of the antigen in axon terminals originating from thalamic neurons. Since the transmitter nitric oxide also characterizes the subplate (Derer and Derer, 1993; Luth et al., 1995; Yan and Ribak, 1997; Fatemi et al., 1998; Judas et al., 1999; Downen et al., 1999), attempts were made to designate NOS, an enzyme readily accessible for immuno-histochemical studies, as the specific marker substance of the subplate. Finally, transient immunoglobulin-like molecules (Dunn et al., 1995) and the sulfoglucuronyl carbohydrate-binding protein-1 (Nair et al., 1998) was observed in immature neurons of the subplate. Also, calcium-binding proteins like PV (Hogan and Berman, 1994; Honig et al., 1996) and calbindin-D (Berger et al., 1993; Hogan and Berman, 1994) were shown to be present in the subplate. The PV-immunoreactive interneurons follow an inside-out pattern of maturation of cortical laminae (Hogan and Berman, 1994). According to Honig et al. (Honig et al., 1996) molecular markers are expressed in spatial and temporal patterns that characterize humans, non-human primates, carnivores and rodents. According to Finney et al. (Finney et al., 1998), subplate neurons provide a major glutamatergic synaptic input to the cortical plate. Remnants of the subplate neuronal population comprise the interstitial cells of adult cortical white matter (Dunn et al., 1995).
In our present studies, we found that KAT-I is not only a novel marker for the localization of the subplate, but also a marker that seems to have important functional aspects, since kynurenic acid, the product of kynurenine aminotransferase (KAT-I) is, as far as is currently known, the only endogenous antagonist of NMDA receptors. In this capacity, KAT-I contributes to neuronal protection against the cytotoxic effect of endogenous excito-toxins (Schwarcz et al., 1984, 1992).
According to our present studies, the morphological entity of the subplate does not represent a homogeneous cell population. This is especially striking in double-stained specimens where the PV- and the KAT-I-expressing cell populations appear in different colours (Fig. 5g). Cells expressing PV, NOS and NPY appear to migrate to the cortical plate. At the same time, cells displaying KAT-I and the nicotinic AChR seem rather to stay in loco and support the guiding of corticothalamic and thalamo-cortical pathways.
The role of KAT-I is transient, since later, during the course of development, as the subplate cells enter the process of programmed cell death (Kostovic and Rakic, 1980; Al-Ghoul and Miller, 1989; Ferrer et al., 1990), this enzyme disappears. This again is in complete harmony with the role of KAT-I as proposed earlier. According to recent investigations, the neurodevelopmental origin of schizophrenia once again seems to hold its ground (Bloom, 1993), as follows from the studies of Akbarian et al. (Akbarian et al., 1993, 1996). Dysfunctions of the NMDA receptors seem to play a crucial role in the development of numerous neurological and psychiatric diseases, including epilepsy, Parkinson’s disease, Alzheimer’s disease, depression, post-traumatic stress disorder, alcoholism and schizophrenia (Heresco-Levy and Javitt, 1998), while those of calcium-binding proteins are supposed to be involved in the development of hydrocephalus (Ulfig et al., 2001). Further investigations may shed light on the role of dysfunctioning and surviving KAT-I-containing subplate neurons in structural and functional irregularities of the cerebral cortex. In this respect, it seems to be of major importance that kynurenic acid, the product of the enzyme KAT, inhibits α7-nicotinic receptor activity (Hilmas et al., 2001), since the a7-nicotinic receptor, a ligand-gated ion channel that admits calcium ions into cells, has various developmental roles (Freedman et al., 2000). It has to be assumed that in the course of normal development, a7-nicotinic receptor and kynurenic acid are in a delicate equilibrium. Our observation that the a7-nicotinic receptor is present in the subplate and in Cajal–Retzius cells (Meyer et al., 1999) seems to be closely related to the developmental role of α7-nicotinic receptor, the malfunction of which has recently been implicated in the developmental pathogenesis of schizophrenia (Freedman et al., 2000). Recent studies in this laboratory, performed in collaboration with gynecologists and psychiatrists on human embryos, may prove or disprove the validity of this assumption.
Our thanks are due to the Electron Microscopic Laboratory of the Biological Research Centre (SzBK), and to Prof. József Hámori and Dr József Takáis (Laboratory for Neurobiology, Hungarian Academy of Sciences, Budapest) for help with the Nomarski technique. We are indebted to Mrs Valeria Vékony for expert histotechnical assistance, to Mr János Szeles for electron microscopic technical assistance, to Mr Mihály Dezsô for photographic work and to Mr John Balog (University of Waterloo, Canada) for editing the text. These studies were supported by the Hungarian Scientific Research Fund (OTKA) grants no. T-022724, no. T-32039, by fund no. FKFP 1077/1997 and no. 1/27 from the NKFP and the Hungarian Medical Research Fund (ETT), funds no. 04/2000 and no. 10/2000.
- congenital abnormality
- cerebral cortex
- neuropeptide y
- nitric oxide synthase
- nicotinic receptors
- n-methyl-d-aspartate receptors
- thalamic structure
- kynurenic acid
- prenatal development
- brain tissue
- nmda receptor antagonist
- neuron migration