N-Acetyl-l -aspartyl-l -glutamate changes functional and structural properties of rat blood–brain barrier

Neuroscience Letters 317 (2002) 85–88 www.elsevier.com/locate/neulet

N-Acetyl-l-aspartyl-l-glutamate changes functional and structural properties of rat blood–brain barrier Lioudmila Pliss a,b, Daniela Jezˇova´ c, Vladislav Maresˇ a, Vladimir J. Balcar d,e,*, Frantisˇek Sˇt’astny´ a,b a

Institute of Physiology, Academy of Sciences, Prague, Czech Republic Department of Brain Pathophysiology, Prague Psychiatric Centre, Prague, Czech Republic c Institute of Experimental Endocrinology, Academy of Sciences, Bratislava, Slovakia d Department of Anatomy and Histology, The University of Sydney, Sydney, NSW 2006, Australia e Department of Molecular Pharmacology, Faculty of Pharmaceutical Sciences, Kanazawa University, Ishikawa, Japan b

Received 30 August 2001; received in revised form 22 October 2001; accepted 25 October 2001

Abstract Intracerebroventricular administration of N-acetyl-l-aspartyl-l-glutamate (NAAG), an agonist at group II metabotropic and NR1/NR2D-containing N-methyl-d-aspartate (NMDA) ionotropic glutamate receptors, increased the permeability of the blood–brain barrier (BBB) to serum albumin in the striatum, but produced no similar effects in the entorhinal cortex or in the hippocampal formation. Electron microscopy showed that NAAG, but not its hydrolytic products l-glutamate and N-acetyl-l-aspartate, increased the number of transport vesicles in the hippocampal endothelial cells. Furthermore, immunocytochemistry detected NR2D subunits on hippocampal capillaries. Consequently, NAAG may have influenced the vesicular transport via NMDA receptors. There was, however, no correlation with the regional pattern of BBB changes (increased permeability in the striatum) that, in turn, could not be directly related to the NAAG-induced neurodegeneration described previously in the hippocampus where no significant changes in BBB permeability were detected. q 2002 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Blood–brain barrier; Brain capillaries; l-Glutamate; N-Acetyl-l-aspartate; N-Acetyl-l-aspartyl-l-glutamate; N-Methyl-d-aspartate receptor; Neurotoxicity; Metabotropic glutamate receptors

It has been suggested that the properties of the endothelial cells, such as the presence of enzymes and transporters, that are essential for the function of the blood–brain barrier (BBB) [15], are determined and regulated by the adjacent astrocytes [6]. This implies the existence of a communication between the two types of cells at a molecular level and, indeed, receptors that could serve as targets for neuroactive substances have been identified on the major components of the BBB [8,9]. The present study tests the hypothesis that the neuropeptide N-acetyl-l-aspartyl-l-glutamate (NAAG), a compound with both neuroprotective and neurotoxic characteristics ([13,20]; for a review see [10]), can influence the BBB in vivo. The levels of NAAG in rat brain tissue are comparable * Corresponding author. Department of Molecular Pharmacology, Faculty of Pharmaceutical Sciences, Kanazawa University, 13-1 Takara-machi, Kanazawa, Ishikawa 920-0934, Japan. Tel.: 181-76-234-4473; fax: 181-76-234-4418. E-mail address: [email protected] (V.J. Balcar).

with those of the most abundant inhibitory neurotransmitter g-aminobutyrate [10]. Its precise function remains unknown, though NAAG activates metabotropic glutamate receptors Group II (mGluR II) [10,16] and could be involved in neuron–glia and glia–endothelial cell communication [1,3]. It has also been suggested that NAAG plays a role in the protection of neurons against neurotoxic assault. Selective activation of mGluR II on astroglial cells by NAAG was shown to stimulate the synthesis of transforming growth factor-b, which could be a mediator of the neuroprotective effect [10]. NAAG can, however, also have an opposite, neurotoxic effect in vitro [20] and in vivo [13]. This finding stands in contrast to the suggested neuroprotectivity of NAAG, but there is a plausible explanation for this apparent discrepancy. NAAG could induce an increase in the BBB permeability that would result in the passage of neurotoxic substances from plasma to brain tissue. Some of those neurotoxins could kill neurons by overactivating N-methyl-daspartate (NMDA)-type receptors (NMDA-R) [18]. The neurotoxicity of NAAG in vivo has been shown to be inhib-

0304-3940/02/$ - see front matter q 2002 Elsevier Science Ireland Ltd. All rights reserved. PII: S03 04 - 394 0( 0 1) 02 42 3- 5

86

L. Pliss et al. / Neuroscience Letters 317 (2002) 85–88

ited by NMDA receptor antagonists [13], but this could reflect the direct interaction of NAAG with NMDA receptors (review [10]). The crucial test of the hypothesis involving the effect of NAAG on the BBB, would be to examine whether the permeability of the BBB increases after the administration of NAAG in vivo and, if it does, whether such changes correlate with the neurotoxic effects of NAAG. This is the principal objective of the present study. Fifty-day-old male SPF Wistar rats (body weight, 200– 300 g) were purchased from Charles River WIGA, Germany. Animals were housed in a 12 h light/dark regime with free access to food and water. The anaesthetic, used in all experiments, was sodium pentobarbital (Nembutal, 45 mg/kg body weight, i.p.). All animal experiments were carried out in accordance with the instructions of the National Committee for the Care and Use of Laboratory Animals and were approved by the Local Animal Care Committee of the Institute of Physiology, Academy of Sciences of the Czech Republic. NAAG, N-acetyl-l-aspartate (NAA), or l-glutamate (LGLU), all purchased from Sigma Aldrich, St Louis, MO, USA, were administered intracerebroventricularly (i.c.v.) at 0.25 mmol each and injected successively into both ventricles, as described in detail elsewhere [13,17]. The needle of a 1.0 ml Hamilton syringe was inserted 1.2 mm caudally from the bregma and 2.0 mm laterally from the sagittal suture, then passed to a depth of 3.6 mm from the skull surface. The volume of 0.25 ml of test solution was injected over a period of 3 min, and the needle was left in place for another 3 min before withdrawal. Control rats received 0.25 ml of the vehicle, i.e. sterile isotonic saline solution. Four days after the operations, the animals were perfused, under deep anaesthesia, either with Ringer’s solution for albumin determination [18], or with Ringer’s solution followed by Karnovsky’s solution for electron microscopy. The brains of animals perfused with Ringer’s solution were rapidly dissected on ice and the entorhinal cortex, striatum and hippocampal formation (divided into the hippocampus proper, dentate gyrus and subicular regions) were isolated and frozen. On the day of the analysis, not more than 10 days after the dissection of the tissue, the samples were thawed, homogenized in barbiturate buffer and centrifuged to remove cell debris. The concentrations of albumin in the supernatants were estimated by the rocket immunoelectrophoresis method [18]. Rat albumin (used as a standard, fraction V. KEBO Lab.OH) and rabbit-raised antibodies against albumin were kindly supplied by Dr O. Fo¨ lders, Institute of Experimental Endocrinology, Bratislava, Slovakia. The data were analyzed by GraphPad Prism version 3.00, (GraphPad, San Diego, CA). In the studies, using electron microscopy, hippocampal formations were dissected on ice and embedded in Epon-812, cut on an ultramicrotome, contrast-labelled with lead citrate and examined using a JEM 100 electron microscope under primary magnification, 5000 £ . In immunohistochemical studies, anaesthetised rats were perfused transcardially

with 0.05% NaNO3 in 0.9% NaCl, followed by a mixture (1:1) of 4% paraformaldehyde and 0.1% glutaraldehyde in 0.12 M phosphate buffer (pH 7.3) for 20 min [7]. The brains were postfixed overnight in the same mixture, then washed in phosphate-buffered saline (PBS) and kept in 30% sucrose–PBS at 148C overnight. Cryostat coronary sections were cut at the level of the striatum/hippocampus, incubated for 10 min in 1% H2O2, rinsed in PBS, incubated with inactivated 3% bovine foetal serum (20 min), and immunostained either with goat polyclonal anti-NMDA-R epsilon-4 antibody (C-20: sc-1471) or with goat polyclonal antiNMDA-R zeta-1 antibody (C-20: sc-1467; both antibodies were from Santa Cruz Biotech., Inc; diluted at 1:50) at 148C overnight. After washing in PBS, the slices were incubated in biotin-SP-conjugated affiniPure F(ab 0 )2 fragment donkey anti-goat *IgG (H 1 L), dilution 1:500, for 60 min, followed by ExtrAvidin-Peroxidase (Sigma), at a dilution of 1:100, for 60 min. Afterwards, the slices were subjected to the diaminobenzidine (DAB) reaction enhanced with 8% NiCl2. In control sections, primary antibodies were omitted or replaced by an ‘ineffective’ polyclonal goat antibody. The stained specimens were analyzed by an AXIOPHOT light microscope (Opton). Administration of NAAG increased the tissue level of albumin in the striatum, but not in the entorhinal cortex or in subregions of the hippocampal formation compared with saline-injected brains (Fig. 1). Neither NAA nor L-GLU influenced the levels of albumin in any of the brain structures studied (Fig. 1). Examination of hippocampal tissue by electron microscopy revealed clear differences in ultrastructure between controls and the tissue from animals injected with the tested compounds (Fig. 2). Tight junctions in controls and in brains injected with NAAG, L-GLU or NAA were not damaged. There was evidence of limited endothelial transcytosis in the hippocampal capillaries of NaCl-treated brains (Fig. 2A), as well as in brains treated with L-GLU or NAA (not shown). The vesicular transport was more pronounced in most endothelial cells from NAAG-treated brains (Fig. 2B). However, the density of

Fig. 1. Regional variations in microvascular permeability to plasma albumin in brains of rats 4 days after i.c.v. injection of sterile saline and 0.25 mmol NAAG or 0.25 mmol L-GLU and/or NAA into each lateral ventricle. EC, entorhinal cortex; HIP, hippocampus proper; DG, dentate gyrus; SUB, subiculum; STR, striatum. Data are expressed as means ^ SEM; *P , 0:05.

L. Pliss et al. / Neuroscience Letters 317 (2002) 85–88

87

reached in the hippocampus proper and dentate gyrus (0.59 and 0.44, respectively; data obtained after all types of i.c.v. injection, i.e. those from the vehicle controls and from the brains of rats given the non-neurotoxic compound NAA were also included in the calculations). Previous experiments showed that neither the administration of L-GLU nor NAA resulted in any important neurotoxic events [12] and, similarly, neither L-GLU nor NAA had any significant effects on the BBB in the present experiments. In the case of L-GLU, the lack of effect can be explained by the presence of an efficient transport system that removes L-GLU rapidly from the extracellular space. The transport is, in turn, linked to brain metabolism [14]

Fig. 2. Electron microscopy of hippocampal capillaries obtained from rats 4 days after the i.c.v. injection into each lateral ventricle of: (A), 0.25 ml of sterile saline; and (B), 0.25 mmol of NAAG. The capillaries from the brains treated with NAAG displayed the type of vesicles (arrows) characteristic for the transendothelial transport [19]. Astroglial endfeet (AE) are surrounding the abluminal side of the capillary basement membrane. Scale bar, 0.5 mm.

endothelial vesicles alone cannot be used to predict brain capillary permeability to albumin because of the absence of an albumin-binding protein in the brain capillaries [19]. In parallel experiments, immunocytochemistry revealed the binding of both anti-NMDA-R zeta-1 and anti-NMDA-R epsilon-4 antibodies in a dot-like staining pattern on large neuronal perikarya in all studied regions. In addition, using the anti-NMDA-R epsilon-4 antibody, we observed distinct staining of capillary walls, especially in the hippocampus (Fig. 3). It may seem that the anti-NMDA-R-IR is located closer to the luminal side of the basement membrane. However, because of the small thickness of endothelial cells (,1 mm) combined with the accumulation of the robust DAB staining product (see Fig. 3), we have not yet been able to draw a general conclusion as to the subcellular location of NMDA-IR, specifically, whether it is localized predominantly on the luminal or abluminal face of the cells. The regional variability in the regulation of BBB could be an intrinsic property of rat brain capillaries reflecting differences in the metabolic rate within the tissue. It might be based on variable expression of specific endothelial transporters and/or enzymes involved in the transport of solutes from plasma into the brain interstitium. It should be stressed, however, that the changes in the permeability of hippocampal capillaries induced by the i.c.v. injections of saline [18] were not accompanied by increased neuronal death [13]. In fact, statistical comparison of the present data with the previously published quantitative analysis of NAAG neurotoxicity [13] indicated poor correlation between the increase in the permeability of the BBB to serum albumin and the extent of neurodegeneration caused by each type of i.c.v. injection. The highest values for Pearson’s coefficient (normal correlation, GraphPad Prism) were

Fig. 3. Hippocampal capillary wall immunolabelled with epsilon4 antibody to reveal NR2D subunit of NMDA receptor (A). Negative control (B) in which primary antibody was omitted and/or epsilon-4 antibody was replaced by an unrelated primary antibody. The apparent vasodilation may have been caused by the presence of 0.05% NaNO3 in the perfusion fluid. Scale bar, 9 mm.

88

L. Pliss et al. / Neuroscience Letters 317 (2002) 85–88

and, if the metabolism is not disturbed at the same time, the neurotoxic effects of glutamate may not eventuate [5]. NAA, though undoubtedly active in the brain metabolism [1], has no documented effects on L-GLU receptors, nor has it ever been shown to be neurotoxic. Thus, neither the neurotoxicity nor the changes in the BBB observed after the i.c.v. administration of NAAG could be accounted for by the presence of L-GLU and NAA generated by the hypothetical hydrolysis of NAAG [10,20]. As NAAG exerted its effects on the BBB only in the striatum, and not in the hippocampus where it is known to be neurotoxic [13], the neurotoxicity of NAAG cannot be explained by the changes in the permeability of the BBB that have been observed in the present experiments. Therefore, the hypothesis that the neurotoxicity of NAAG is directly linked to the actions of NAAG on L-GLU receptors — whether they are of metabotropic (mGluR II) or ionotropic (NMDA-R) nature, located on neurons or glial cells [4,11] — has not been refuted by the present data. If the neurotoxicity of NAAG is also related to the NR2D-containing NMDA receptors (which are known to be sensitive to NAAG) that have been detected, in the present study, on hippocampal capillaries, the hypothetical mechanism would be unlikely to involve an increase in the permeability of the BBB to large molecules such as plasma albumin. In conclusion, there is no correlation between the changes in the permeability of the BBB to albumin caused by the i.c.v. administration of NAAG and the previously reported neurotoxic effects of NAAG in the hippocampal formation. Nevertheless, the present observations, particularly when the immunocytochemical detection of NMDA receptors on brain capillaries is considered, should serve as a poignant reminder of the possibility that the disturbances in the BBB may contribute to actions of certain neurotoxins in vivo [18]. This should be kept in mind when interpreting data obtained in animal models of neurodegenerative diseases or when discussing causes of neuronal loss in pathological conditions in the human brain, especially when NMDA receptors are thought to be involved in the underlying mechanisms [2]. This work was supported by GA/CR grant 309/99/0211 (F.S.) and by Clive and Vera Ramaciotti Foundation, New South Wales, Australia (V.J.B.). [1] Baslow, M.H., Functions of N-acetyl-l-aspartate and Nacetyl-l-aspartylglutamate in the vertebrate brain: role in the glial cell-specific signaling, J. Neurochem., 75 (2000) 453–459. [2] Brace, H., Latimer, M. and Winn, P., Neurotoxicity, blood– brain barrier breakdown, demyelination and remyelination associated with NMDA-induced lesions of the rat lateral hypothalamus, Brain Res. Bull., 43 (1997) 447–455. [3] Braet, K., Paemeleire, K., D’Herde, K., Sanderson, M.J. and Leybaert, L., Astrocyte–endothelial cell calcium signals conveyed by two signalling pathways, Eur. J. Neurosci., 13 (2001) 79–91.

[4] Cull-Candy, S., Brickley, S. and Farrant, M., NMDA receptor subunits: diversity, development and disease, Curr. Opin. Neurobiol., 11 (2001) 327–335. [5] Garcı´a, O. and Massieu, L., Strategies for neuroprotection against l-trans-2,4-pyrrolidine dicarboxylate-induced neuronal damage during energy impairment in vitro, J. Neurosci. Res., 64 (2001) 418–428. [6] Janzer, R.C. and Raff, M.C., Astrocytes induce blood–brain barrier properties in endothelial cells, Nature, 325 (1987) 253–257. [7] Joelson, D. and Schwartz, I.R., Development of N-methyl-daspartate receptor subunit immunoreactivity in the neonatal gerbil cochlear nucleus, Microsc. Res. Tech., 41 (1998) 246–262. [8] Krizbai, I.A., Deli, M.A., Pestena´ cz, A., Siklo´ s, L., Szabo´ , C.A., Andra´ s, I. and Joo´ , F., Expression of glutamate receptors on cultured cerebral endothelial cells, J. Neurosci. Res., 54 (1998) 814–819. [9] Mineff, E. and Valtschanoff, J., Metabotropic glutamate receptors 2 and 3 expressed by astrocytes in rat ventrobasal thalamus, Neurosci. Lett., 270 (1999) 95–98. [10] Neal, J.H., Bzdega, T. and Wroblewska, B., N-Acetylaspartylglutamate: the most abundant peptide neurotransmitter in the mammalian central nervous system, J. Neurochem., 75 (2000) 443–452. [11] Petralia, R.S., Wang, Y.-X., Niedzielski, A.S. and Wenthold, R.J., The metabotropic glutamate receptors show unique postsynaptic, presynaptic and glial localisation, Neuroscience, 71 (1996) 949–976. [12] Pliss, L., Balcar, V.J., Pokorny´ , J. and Sˇ t’astny´ , F., Neuronal damage in hippocampus induced by N-acetyl-aspartylglutamate, an endogenous agonist at group II metabotropic glutamate receptors, In J. Sikora and Z. Fisˇ ar (Eds.), Neurobiology of Mental Disorders, Galen, Prague, 1999, pp. 196– 199. [13] Pliss, L., FitzGibbon, T., Balcar, V.J. and Sˇ t’astny´ , F., Neurotoxicity of NAAG in vivo is sensitive to NMDA antagonists and mGluR II ligands, NeuroReport, 11 (2000) 3651–3654. [14] Rae, C., Lawrence, M.L., Dias, L.S., Provis, T., Bubb, W. and Balcar, V.J., Strategies for studies of neurotoxic mechanisms involving deficient transport of l-glutamate: antisense knockout in rat brain in vivo and changes in the neurotransmitter metabolism following inhibition of glutamate transport in guinea pig brain slices, Brain Res. Bull., 53 (2000) 373–381. [15] Schlosshauer, B., The blood–brain barrier: morphology, molecules, and neurothelin, Bioessays, 15 (1993) 341–345. [16] Shave, E., Pliss, L., Lawrence, L.L., FitzGibbon, T., Sˇ t’astny´ , F. and Balcar, V.J., Regional distribution and pharmacological characteristics of [ 3H]N-acetyl-aspartyl-glutamate (NAAG) binding sites in rat brain, Neurochem. Int., 38 (2001) 53–62. [17] Sˇ t’astny´ , F., Dvorˇ a´ kova´ , L. and Lisy´ , V., Biochemical characteristics of g-glutamyl transpeptidase in capillaries from entorhinohippocampal complex of quinolinate-lesioned rat brain, Mol. Chem. Neuropathol., 32 (1997) 143–161. [18] Sˇ t’astny´ , F., Sˇ kulte´ tyova´ , I., Pliss, L. and Jezˇ ova´ , D., Quinolinic acid enhances permeability of rat brain microvessels to plasma albumin, Brain Res. Bull., 53 (2000) 415–420. [19] Stewart, P.A., Endothelial vesicles in the blood–brain barrier: are they related to permeability? Cell. Mol. Neurobiol., 20 (2000) 149–163. [20] Thomas, A.G., Vornov, J.J., Olkowski, J.L., Merion, A.T. and Slusher, B.S., N-Acetylated a-linked acidic dipeptidase converts N-acetylaspartylglutamate from a neuroprotectant to a neurotoxin, J. Pharmacol. Exp. Ther., 295 (2000) 16–22.