Quinolinic acid promotes albumin deposition in Purkinje cell, astrocytic activation and lipid peroxidation in fetal brain

Neuroscience 134 (2005) 867– 875

QUINOLINIC ACID PROMOTES ALBUMIN DEPOSITION IN PURKINJE CELL, ASTROCYTIC ACTIVATION AND LIPID PEROXIDATION IN FETAL BRAIN E. YAN,a M. CASTILLO-MELÉNDEZ,a G. SMYTHEb AND D. WALKERa*

indole 2,3-dioxygenase (IDO) by pro-inflammatory cytokines (Kudo et al., 2000, 2001). Although QUIN crosses the intact blood– brain barrier (BBB) only slowly (Heyes et al., 1997), 85% of QUIN in extracellular fluid and CSF is thought to derive from blood (Beagles et al., 1998). QUIN is an agonist at glutamate receptors and at high concentrations or after long exposure to moderate concentrations, is neurotoxic (Stone, 1993). QUIN toxicity has been implicated in several neurodegenerative conditions, such as Parkinson and Huntington diseases (reviewed by Stone, 2001), and in the dementia associated with HIV infection (Heyes et al., 1989b) and cerebral malaria (Sanni et al., 1998; Medana et al., 2002). In addition to these well-characterized excitotoxic actions, in vitro and in vivo experiments show that QUIN inhibits glutamate uptake by astrocytes (Tavares et al., 2002), promotes seizures (Losch de Oliveira et al., 2004), induces formation of reactive oxygen species and causes lipid peroxidation (Santamaria et al., 2001a,b). QUIN also increased the penetration of endogenous albumin into the brain after i.c.v. QUIN infusion in the rat (St’astny et al., 2000), raising the possibility that QUIN might act directly on the cerebral microvasculature to alter permeability of the BBB. We have recently shown that plasma QUIN in fetal sheep is increased after maternal tryptophan loading (Nicholls et al., 1999), and in the chronically hypoxic conditions induced by partial embolization of the placenta (Nicholls et al., 2001). Plasma QUIN is also increased by administration of lipopolysaccharide (LPS) to fetal sheep, a treatment that results in compromise of the fetal BBB permeability (Yan et al., 2004). It is possible that hypoxic and inflammatory conditions in pregnancy result in increased levels of QUIN in fetal blood and the brain, with increased likelihood of brain damage arising from the excitotoxic and pro-oxidative stress effects of QUIN. Thus, we hypothesized that infusion of QUIN into the circulation of healthy fetal sheep would lead to cellular effects in the fetal brain that are sometimes seen after hypoxic and inflammatory challenges. QUIN infusion resulted in evidence of astrocytic activation in several brain regions, increased lipid peroxidation and accumulation of albumin in Purkinje cells of the cerebellum.

a Fetal and Neonatal Research Group, Department of Physiology, Monash University, Clayton, Victoria 3800, Australia b Ray Williams Biomedical Mass Spectrometry Facility, University of New South Wales, NSW 2052, Australia

Abstract—In high concentrations or after prolonged exposure, the N-methyl-D-aspartate receptor agonist quinolinic acid (QUIN) induces lipid peroxidation, oxidative stress, and cell death in the adult brain, and after i.c.v. injection induces seizures and increases blood– brain barrier permeability. As QUIN is substantially increased in plasma and brain of fetal sheep after endotoxin treatment or maternal tryptophan loading, we examined the effects of increasing plasma QUIN concentrations on the brain of late gestation fetal sheep. Continuous fetal infusion of QUIN (0.1 mmol/h i.v.; nⴝ4) for 12 h increased plasma QUIN concentrations from 22.3ⴞ6.0 – 210.8ⴞ31.4 ␮M; the infusion of vehicle [normal saline] had no effect on QUIN concentrations (nⴝ4). At 24 h after QUIN infusion glial fibrillary acidic protein immunoreactivity was significantly increased in cerebral gray matter and the granule cell layer of cerebellum, and the lipid peroxide product 4-hydroxynonenal-immunoreactivity and albumin-immunoreactivity were present throughout the cytoplasm of cerebellar Purkinje cells. Extravasation of albumin into the brain was not observed, indicating the cerebral microvasculature with respect to permeability to plasma proteins was normal at the time of analysis. We suggest that increased glial fibrillary acidic protein and 4-hydroxynonenal result from oxidative stress induced by QUIN, and that the increased intracellular albumin in cerebellar Purkinje cells may be an adaptive response. © 2005 IBRO. Published by Elsevier Ltd. All rights reserved. Key words: 4-hydroxynonenal, blood– brain barrier, glial fibrillary acidic protein, fetus, striatum, cerebellum.

Quinolinic acid (QUIN) is an endogenous metabolite of L-tryptophan produced by kynurenine pathway enzymes in brain and peripheral tissues, particularly by the liver and macrophages. In adult animals, basal concentration of QUIN in blood is normally low (⬍0.5 ␮M), but can increase approximately 10-fold in plasma and 18-fold in the brain extracellular fluid in response to infection (Heyes et al., 1988, 1989a; Babcock and Carlin, 2000; Fujigaki et al., 2001) due to induction of the enzyme

EXPERIMENTAL PROCEDURES

*Corresponding author. Tel: ⫹61-3-9905-2534; fax: ⫹61-3-9905-2547. E-mail address: [email protected] (D. Walker). Abbreviations: BBB, blood–brain barrier; DAB, diaminobenzidine; GFAP, glial fibrillary acidic protein; IR, immunoreactive; LPS, lipopolysaccharide; PBS, phosphate-buffered saline; QUIN, quinolinic acid; 4-HNE, 4-hydroxynonenal.

Animals and experimental protocol Eight Merino/Border-Leicester cross time-mated pregnant sheep were housed in individual cages and kept at constant temperature (22 °C), under a 12-h light/dark cycle, and with free access to

0306-4522/05$30.00⫹0.00 © 2005 IBRO. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.neuroscience.2005.04.056

867

868

E. Yan et al. / Neuroscience 134 (2005) 867– 875

water. They were fed daily between 09:00 and 11:00 h. All the experimental procedures had received prior approval from the Standing Committee on Ethics and Animal Experimentation of Monash University. All experiments and animal care were according to the Australian Code of Practice for the Care and Use of Animals for Scientific Purposes. At 124 –126 days of gestation (term is ⬃147 days) each ewe was anesthetized by inhalation of 2% halothane (Fluothane; Merial, Australia) in O2 for surgical implantation of catheters into a maternal and fetal carotid artery and jugular vein using aseptic procedures, as previously described (Nicholls et al., 2001). A bolus i.v. injection of ampicillin (Aspen Pharmacare Australia Pty Ltd) was given to the ewe during the surgery, and analgesia (buprenorphine) was administered intramuscularly to the ewe after the surgery. Thereafter, no further antibiotics were administered. Each catheter was flushed daily with heparinized (50 U/ml; Heparin, Pharmacia, Australia) sterile saline to prevent clotting. At 132–133 days gestation the fetuses received either an i.v. infusion of QUIN (Sigma-Aldrich, USA) or an equivalent volume of saline (n⫽4) at beginning of the experiment. Fetal plasma QUIN concentration was rapidly increased by infusing 15 mM of QUIN in saline at a rate of 6.7 ml/h for 2 h, and then the QUIN infusion rate was changed to 0.7 ml/h to match the excretion rate of QUIN (unpublished observation), this was maintained for the next 10 h before the infusion stopped. Fetal and maternal arterial blood samples (0.5 ml, fluid replaced with equal volume of saline) were taken at ⫺1.5, ⫺1.0, ⫺0.5, 0.5, 1.0, 2, 3, 6, 9, 12, 12.5, 13.5, 15.5, and 24 h with respect to the time of starting the infusion for measurement of blood gases and pH, and recovery of plasma for measurement of QUIN, glucose and lactate. Fetal plasma glucose and lactate were measured using YSI 2300 STAT Glucose and Lactate Analyzer (YSI Life Sciences, USA). Fetal arterial pressure was measured throughout the experiment by a strain gauge transducer connected to an analogdigital converter (PowerLab, ADInstrument, Australia) and computer system. Heart rate was computed from the pulse frequency. All fetuses and ewes were killed by an i.v. overdose of pentobarbitone sodium (Lethabarb; Virbac Pty Ltd, Australia) given to the ewe 24 h after treatment. Fetal brains were perfused transcardially with approximately 1 l heparinized saline at 90 ml/min and 40 mmHg, followed by 1 l 4% paraformaldehyde (Probing & Structure, Australia) in 0.1 M phosphate buffer at pH 7.4. The brains were then removed, and post-fixed in 4% paraformaldehyde overnight. Each brain was then dissected into blocks containing the striatum, parietal cortex, or cerebellum and embedded in paraffin wax. Tissue blocks were cut into 10 ␮m sections for immunohistochemical staining, and Cresyl Violet and acid fuchsin staining to identify pyknotic cells.

QUIN measurement The concentration of QUIN in fetal and maternal plasma was measured by gas chromatography–mass spectrometry, as described in detail previously (Smythe et al., 2002).

Immunohistochemistry Glial fibrillary acidic protein (GFAP). Paraffin-embedded brains were cut at 10 ␮m and mounted on SuperFrost⫹ (MenzelGlaser, Germany) slides. Sections were first incubated in 50% methanol containing 0.3% hydrogen peroxide for 10 min at room temperature to reduce endogenous peroxidase activity, washed in phosphate-buffered saline (PBS; 0.1 M, pH 7.4), and then blocked with PBS containing 5% normal rabbit serum and 0.3% Triton X-100 (Sigma-Aldrich) for 1 h at room temperature. The sections were then incubated with mouse GFAP antibody (1:400 dilution; Sigma-Aldrich) in 2.5% normal rabbit serum and 0.3% Triton in PBS overnight at 4 °C. After washing in PBS, the sections were incubated with biotinylated rabbit anti-mouse secondary antibody (1:500 dilution; DAKO, Denmark) for 1 h at room temperature,

followed by 30 min incubation with streptavidin horseradish peroxidase (1:100 dilution; Amersham Pharmacia Biotech, Sweden). Color was developed using diaminobenzidine (DAB; Pierce Chemical, USA) as the chromogen. The number of GFAP positive cell bodies was counted for each bright field of view using an image analysis program (Image-Pro Plus, MediaCybernetics, USA). Four fields from each of two sections were counted and averaged for each brain region within animals, and the results combined to arrive at mean values for each group. Albumin. Sections (10 ␮m) adjacent to those used for GFAP were stained for endogenous albumin. The staining procedures were performed as described above except: (a), Protein Block (DAKO) was used to block the non-specific binding; (b), 0.5% fish gelatin (Sigma-Aldrich) was added to a polyclonal albumin primary antibody raised in rabbits (species specific for sheep, 1:6000 dilution; Accurate Chemical & Scientific Corporation, USA); and (c), biotinylated goat anti-rabbit IgG (1:200 dilution; Vector Laboratories, USA) was used as the secondary antibody. 4-Hydroxynonenal (4-HNE). Adjacent sections (also 10 ␮m) to those used for albumin staining were dewaxed and rehydrated. The sections were then incubated in 50% methanol containing 0.3% hydrogen peroxide for 10 min, and then in 5% normal goat serum in PBS for 1 h at room temperature. The procedure was then as described above for albumin immunohistochemistry, but using a 4-HNE polyclonal primary antibody raised in rabbit (Alexis, USA) diluted 1:500 in PBS containing 2.5% normal goat serum. Lectin. Corresponding sections (10 ␮m) to the areas examined above were stained for lectin reactivity using biotinylated lectin from Bandeiraea simplicifolia (isolectin B4; Sigma-Aldrich) diluted 1:200 In PBS. Immunoreactivity was visualized by reacting with DAB as chromogen.

Data analysis Data are presented as mean⫾S.E.M. Two-way repeated measures ANOVA was used to determine differences between the saline and QUIN treatment groups for blood gas parameters, pH, blood pressure, heart rate, and plasma QUIN concentrations. Where the ANOVA indicated significant differences between treatment and time, a least square difference test was used post hoc to identify the significance of differences between the treatment groups at each sampling time. The non-parametric Mann-Whitney U-test was used to identify differences between the number of GFAP positive cell bodies in each brain region of the saline and QUIN treatment groups. Differences were considered significant when P⬍0.05.

RESULTS The infusion of QUIN increased fetal arterial plasma QUIN from 22.3⫾6.0 –210.8⫾31.4 ␮M by the end of the 2 h loading dose period, and to 159.4⫾28.1 ␮M by the end of the 12 h infusion period (Fig. 1). Plasma QUIN concentrations remained elevated for 2 h after the end of infusion and then decreased; although the rate of decrease varied between animals, plasma QUIN concentrations at 15 and 24 h were not significantly different from the pre-infusion control values, or from values at the same time points for the saline-infused fetuses (Fig. 1). The resting, or basal concentrations of QUIN in fetal plasma were significantly higher than maternal concentrations (22.3⫾6.0 vs. 1.20⫾ 0.63 ␮M). The infusion of QUIN into the fetal circulation had no effect on maternal plasma concentrations (1.5⫾0.7 ␮M at ⫹12 h, P⬎0.1). QUIN infusion caused a transient but signif-

E. Yan et al. / Neuroscience 134 (2005) 867– 875

869



Fig. 1. QUIN concentrations (␮M) in fetal (〫⽧) and maternal ( ) plasma measured prior to, during and after fetal i.v. infusion of either QUIN (open symbols) or saline (closed symbols). Asterisks indicate significant differences (P⬍0.05) of fetal plasma QUIN between saline- and QUIN-treated groups. Symbol
indicates that fetal plasma QUIN concentrations were significantly higher than maternal QUIN concentrations for the saline-infused group (P⬍0.05). Shaded bar indicates the period of QUIN or vehicle infusion.

icant increase in both glucose and lactate concentrations in fetal plasma blood (Table 1), but there were no effects on maternal plasma values (data not shown). Both parameters were significantly increased after 2 h of QUIN infusion and remained significantly elevated until 6 h before returning to control levels. At the end of the 12 h QUIN infusion both glucose and lactate were not different to plasma values for the saline infused group. The QUIN infusion had no effect on fetal blood gases, pH,

blood pressure and heart rate which were all normal for fetal sheep at this gestational age (Table 1). Low numbers of GFAP-immunoreactive (IR) cells were observed in cortical gray and intragyral white matter, caudate nucleus of striatum, and in the granular layer of the cerebellum in the control (saline-infused) fetuses (Fig. 2). GFAP-IR, was present in cells with a morphology typical of astrocytes; i.e. stellate cells with a small cell body. After QUIN infusion, GFAP-IR was significantly increased in cortical gray matter

Table 1. Fetal blood gases (PO2, partial pressure of oxygen; PCO2, partial pressure of carbon dioxide), pH, mean arterial pressure (MAP) and heart rate (HR) measured prior to (⫺2 h), during (⫹2 h, ⫹6 h and ⫹12 h), and at 12 h after (Post) infusion of QUIN

PO2 (mm Hg) PCO2 (mm Hg) pH MAP (mm Hg) HR (beats/min) Glucose (mM) Lactate (mM)

QUIN Saline QUIN Saline QUIN Saline QUIN Saline QUIN Saline QUIN Saline QUIN Saline

⫺2 h

⫹2 h

⫹6 h

⫹12 h

Post

24.9⫾1.9 25.4⫾2.3 43.2⫾2.9 49.4⫾1.1 7.365⫾0.011 7.374⫾0.008 39.2⫾4.6 39.2⫾5.8 143.9⫾10.5 134.6⫾8.7 0.70⫾0.04 0.53⫾0.02 1.48⫾0.09 1.58⫾0.18

26.5⫾2.4 24.8⫾2.5 42.6⫾2.9 45.9⫾0.8 7.369⫾0.011 7.387⫾0.010 41.2⫾3.7 41.3⫾5.8 142.3⫾10.3 129.3⫾7.5 0.78⫾0.06 0.78⫾0.01 1.60⫾0.24 1.40⫾0.24

23.8⫾2.8 24.0⫾2.6 48.9⫾2.1 48.4⫾1.6 7.374⫾0.010 7.380⫾0.005 39.0⫾3.1 41.3⫾3.7 147.6⫾5.2 136.9⫾10.5 1.00⫾0.07* 0.68⫾0.07 2.15⫾0.35* 1.35⫾0.13

22.2⫾3.2 24.7⫾2.2 48.0⫾1.4 46.7⫾1.3 7.369⫾0.004 7.379⫾0.010 35.8⫾4.8 42.4⫾4.4 164.9⫾18.3 130.9⫾11.3 0.83⫾0.06 0.85⫾0.05 1.90⫾0.27 1.63⫾0.18

23.8⫾2.2 22.9⫾2.0 47.8⫾0.8 48.8⫾2.2 7.375⫾0.009 7.368⫾0.007 40.0⫾3.0 42.2⫾3.0 136.1⫾14.5 135.7⫾6.6 0.70⫾0.08 0.65⫾0.09 1.45⫾0.05 1.43⫾0.13

The data were averaged for the 30 min period prior to each of the time points shown. Asterisks indicate significant differences (P ⬍ 0.05) between QUIN and saline-treated groups.

870

E. Yan et al. / Neuroscience 134 (2005) 867– 875

Fig. 2. GFAP immunohistochemistry of cortical gray matter for a saline-infused fetus (left-hand panels) and a QUIN-infused fetus (right-hand panels). A higher magnification view of the boxed areas in A and B are shown in C and D, respectively. In the saline-infused fetuses the morphology of GFAP-positive cells was of small cell bodies with thin, ramified processes (panels A and C). In the QUIN-treated fetuses many more GFAP-positive cells were present, with a hyperplastic cell body and short, thickened processes (panels B and D). Scale bar⫽50 ␮m in panels A and B; 20 ␮m in panels C and D.

and the granule cell layer of the cerebellum. There was a trend for increased numbers of GFAP-IR cells in cortical white matter and caudate nucleus of striatum after QUIN infusion, but this did not reach significance. After QUIN infusion, the morphology of the GFAP-IR cells was distinctly different, changing from the normal appearance of a small cell body with long processes to an enlarged body with short

or no stellate processes (Figs. 3 and 4). These changes in morphology were observed in all the QUIN-treated fetuses. The number of microglia and macrophages, identified as lectin positive cells, was not changed in both saline- and QUIN-infused fetus. Extravascular or parenchymal albumin-IR was not observed in any forebrain region of the saline or QUIN in-

E. Yan et al. / Neuroscience 134 (2005) 867– 875

871

Fig. 3. Micrograph of GFAP immunohistochemistry in the granule cell layer of cerebellum (indicated by a box in cerebellum illustration) at 12 h after the end of the 12 h infusion of either saline (panel A) or QUIN (panel B). GFAP IR cell bodies were clearly present after the QUIN infusion, whereas no GFAP positive cells were observed in the saline-treated fetus. Scale bar⫽40 ␮m.

fused fetuses. However, diffuse albumin-IR was observed in the molecular layer of the cerebellum in QUIN-treated fetuses (Fig. 5B). In addition, Purkinje cells located in these regions were immunopositive for albumin, with distinct albumin-IR present throughout the cell body but not

extending into the dendrites (Fig. 5D). This pattern of albumin-IR was observed in all of the QUIN-treated fetuses, but not in any of the saline-treated fetuses. 4-HNE-IR was also observed in cerebellar Purkinje cells after QUIN treatment (Fig. 5F), but not in the saline

Fig. 4. Bar graph of the number of GFAP positive cells per mm2 (mean⫾S.E.M.) in cortical gray and intragyral white matter, caudate nucleus and granule layer of cerebellum at 12 h after the end of a 12 h continuous infusion of either QUIN (solid or gray shaded bar) or saline (open or gray bar) (n⫽4). Asterisks indicate significant differences between the saline- and QUIN-treated groups; Mann-Whitney U test, P⬍0.05.

872

E. Yan et al. / Neuroscience 134 (2005) 867– 875

Fig. 5. Immunohistochemistry of albumin (panels A–D) and 4-HNE (panels E–H) in the cerebellum of fetuses at 12 h after the end of a 12 h infusion of either saline (left) or QUIN (right). Purkinje cell staining for albumin (panel B) and 4-HNE (panel F) was observed in the QUIN infused fetuses, but was not present in the saline-infused fetuses (panels A and E, respectively). At higher magnifications the albumin immunoreactivity was seen to be cytoplasmic (panel D) and for 4-HNE it was present in the cell body and dendrites (panel H). No IR cells for albumin or 4-HNE were observed in the cerebellum of saline-treated fetuses (panels A, C, E, G). Magnification in panels A, B, E, F 100⫻, scale bar⫽100 ␮m as shown in panel A. Magnification in panels C, D, G, H is 400⫻, scale bar⫽40 ␮m as shown in panel C.

E. Yan et al. / Neuroscience 134 (2005) 867– 875

treated fetuses, or elsewhere in the brain of either saline or QUIN-treated fetuses (Fig. 5E, G). The 4-HNE-IR product was present throughout the cytoplasm of the cell body and also in dendrites reaching into the molecular layer (Fig. 5H). Pyknotic cells, identifiable as cells with condensed chromatin and intensely pink staining cytoplasm, were not observed in any brain regions that were examined.

DISCUSSION This study has demonstrated that the fetal brain responds to high circulating concentrations of the kynurenine metabolite, QUIN. The function of QUIN in the brain has been debated for many years since its neurotoxic properties were first shown by Stone and Perkins (1981). While QUIN concentrations in circulation are normally very low (⬍1 ␮M), several neuro-inflammatory and degenerative conditions are associated with high QUIN concentrations in the brain, possibly due to both increased de novo synthesis by glia (in particular, microglia), and release from macrophages recruited from the circulation or having passed through a compromised BBB (During et al., 1989). While this may be true also for the fetus (for example, infection causes BBB compromise in fetal sheep and increases the number of microglia/macrophage cells identified by lectin; Yan et al., 2004), a difference is that the resting QUIN concentrations in the fetal circulation are higher than maternal concentrations. The source of QUIN in fetal sheep plasma is not clear, but it is likely to include release from the fetal liver and cells of monocyte origin. The sheep placenta may also release QUIN, as the human placenta has been shown to capable of producing QUIN under normal conditions and in the presence of intrauterine infection (Manuelpillai et al., 2005). One of the aims of this study was to examine the effects of high levels of plasma QUIN on the fetal BBB. Increased penetration of plasma albumin into the fetal sheep brain has been observed following treatment with LPS (Yan et al., 2004), and a study in adult rats indicated that QUIN itself could compromise the BBB (St’astny et al., 2000), although in that study QUIN was administered into a cerebral ventricle. In the present study we found no evidence that QUIN infusion increased the albumin-IR around blood vessels in any brain region, when examined 12 h after the end of the infusion. It is possible that QUIN did affect the cerebral microvasculature and that some plasma albumin entered the brain during or after the QUIN infusion, but by 12 h after the end of the infusion it had diffused within the extracellular space or had been cleared so that it was no longer detectable by immunohistochemical staining. The basal plasma QUIN concentration in the human fetus is relatively low 1–2 ␮M and increases to 50 ␮M in the condition of intrauterine infection (P. Ligam and D. Walker, unpublished observation). In the current study, the QUIN infusion produced a 10-fold increase in fetal plasma QUIN concentration to reach at 210.8 ␮M. The levels achieved in this experiment are the result of high basal plasma QUIN concentration (22.3 ␮M) in normal sheep

873

fetus, but the magnitude of this increase is physiological. Little of the infused QUIN appears to have crossed the placenta into the maternal circulation because there was no change of maternal plasma QUIN concentrations after fetal QUIN infusion. The presence of strong albumin-IR in cerebellar Purkinje cells was observed in the QUIN-infused fetuses, and not in any of the saline-infused animals. Assuming that albumin, or an albumin-like protein, is not synthesized by Purkinje cells, this result suggests that albumin can be taken up from the extracellular space, and had entered the cerebellum from the circulation. This may explain the diffuse albumin-IR that was observed in some regions of the cerebellum, where the Purkinje cells showed strong staining for albumin. Albumin uptake by neurons has been noted as a response to cellular stress that may be protective (Tabernero et al., 2002), and suggests that high QUIN concentrations place these cells under oxidative stress. QUIN infusion produced a significant effect on brain astrocytes but not for other glial cells such as microglia and macrophages. The GFAP-IR increased in many parts of the QUIN-infused fetal brain. GFAP, a 50 kDa protein, is present in the intermediate filaments of the cytoskeleton of differentiated astrocytes (Schmidt-Kastner et al., 1993). Increased GFAP immunoreactivity may be due to previously GFAP-negative astrocyte cell bodies becoming GFAP-positive, or because changes of cell shape concentrate the protein near the cell body and make it more easily identified by immunocytochemistry (Norenberg, 1996). Regardless of the mechanism involved, increased GFAP-IR has become widely used as a marker of reactive astrocytes, and of the response of astrocytes to oxidative and hypoxic stress (Eng et al., 1992; Norenberg, 1994, 1996). Astrocytes provide metabolic and trophic support to the neurons and other cells within the brain, and it is widely believed that astrocytic activation is an adaptive response that regulates the nutrient, ionic and neurotransmitter levels in the extracellular environment of the neuron (reviewed by Takuma et al., 2004). A consistent observation in this study was that the QUIN treatment resulted in a change of astrocyte morphology from a stellate form with numerous processes to a non-stellate form with a hypertrophic ‘ameboid’ cell body. The significance of this response is unclear, but if it involves changes in the intimate contact that astrocytes normally have with neurons and cerebral blood vessels, effects on BBB function and energy supply to neurons might be expected as a consequence. Astrocytic activation has been seen in number of pathological conditions such as hypoxia–ischemia, inflammation and excitotoxin exposure (see review by Acarin et al., 2001). In vivo studies have shown that astrocytes are the important sources of pro-inflammatory cytokines such as tumor necrosis factor-␣, interleukin-1, -6, and NF-␬B (reviewed by Acarin et al., 2001) which play a vital role in the neuronal function and survival following an injury. QUIN infusion also resulted in increased lipid peroxidation in cerebellar Purkinje cells. The peroxidant effects of QUIN have been shown in vivo in the adult rat brain (Rios and Santamaria, 1991), and in vitro in rat brain

874

E. Yan et al. / Neuroscience 134 (2005) 867– 875

synaptosomes (Santamaria et al., 2001b). QUIN also induces hydroxyl radical ( · OH) release in the rat brain (Santamaria et al., 2001a), that could account for the increase of 4-HNE in the Purkinje cell. 4-HNE is one of the toxic aldehyde products of lipid peroxidation (Zhang et al., 1999) and its appearance in many Purkinje cells suggests that this cell population had been subjected to oxidative stress as a result of the QUIN infusion. In addition, QUIN acts on NMDA receptors leading to change in cellular ionic influx (especially Ca2⫹) which results in mitochondrial dysfunction, · OH generation and possible cell death. The Purkinje cell has an exceptionally high metabolic demand as it receives extensive excitatory input and has a high level of calcium uptake (Welsh et al., 2002). A high cerebellum metabolic rate may increase the vulnerability of Purkinje cells to oxidative and metabolic stress. Purkinje cell damage has been linked to many adult diseases such as epilepsy, Huntington’s disease, Alzheimer’s disease and mitochondrial disorders (reviewed by Sarna and Hawkes, 2003). Furthermore, a consistent neurological abnormality present in autistic individuals is Purkinje cell loss, which is thought to arise not from cell death, but as a developmental abnormality during gestation (reviewed by Kern, 2003). It is unclear from these experiments whether QUIN entered the cerebellum more readily than for other brain regions, or whether Purkinje cells are more sensitive to oxidative stress than neurons elsewhere in the brain. Alternatively, the cerebellar BBB may be more vulnerable to damage, as shown in experimental models of autoimmune encephalitis (Tonra et al., 2001). Postmortem studies of still-born human fetal brains revealed strong 4-HNE immunoreactivity in cerebellar Purkinje cells (Itakura et al., 2002), although no other sign of neurological damage was observed.

CONCLUSIONS While the present study found no evidence that high concentrations of QUIN in blood impaired the BBB in the fetal brain, it has shown that QUIN caused significant increases of GFAP-IR in astrocytes, and lipid peroxidation and albumin accumulation in cerebellar Purkinje cells. This suggests that a possible contribution to perinatal fetal brain damage could be the increased production of QUIN by fetal tissues and the placenta, particularly when intra-uterine or placental infection is present. Such damage is likely to be more pronounced if combined with other insults to the fetus, such as intrauterine hypoxia or asphyxia at birth. This study only examined the effects of a relatively short (24 h) exposure to elevated levels of QUIN, whereas clinically it is possible that the fetus is exposed to increased QUIN, possibly released from the placenta, for days or weeks before birth (Manuelpillai et al., 2005). Acknowledgments—We thank Alex Satragno for help with the surgery and Sonia Bustamante for conducting the QUIN measurements. This project was supported by a grant from the National Health and Medical Research Council of Australia and March of Dimes Birth Defects Foundation (USA) to D.W.

REFERENCES Acarin L, Gonzalez B, Castellano B (2001) Glial activation in the immature rat brain: implication of inflammatory transcription factors and cytokine expression. In: Progress in brain research, Vol. 132, Glial cell function (Castellano-Lopez B, Nieto-Sampedro M, eds), pp 376 –389. Amsterdam: Elsevier. Babcock TA, Carlin JM (2000) Transcriptional activation of indoleamine dioxygenase by interleukin 1 and tumor necrosis factor alpha in interferon-treated epithelial cells. Cytokine 12:588 –594. Beagles KE, Morrison PF, Heyes MP (1998) Quinolinic acid in vivo synthesis rates, extracellular concentrations, and intercompartmental distributions in normal and immune-activated brain determined by multiple-isotope microdialysis. J Neurochem 70:269 – 291. During MJ, Heyes MP, Freese A, Markey SP, Martin JB, Roth RH (1989) Quinolinic acid concentrations in striatal extracellular fluid reach potentially neurotoxic levels following systemic L-tryptophan loading. Brain Res 476:384 –387. Eng LF, Yu AC, Lee YL (1992) Astrocytic response to injury. Prog Brain Res 94:353–365. Fujigaki S, Saito K, Sekikawa K, Tone S, Takikawa O, Fujii H, Wada H, Noma A, Seishima M (2001) Lipopolysaccharide induction of indoleamine 2,3-dioxygenase is mediated dominantly by an IFNgamma-independent mechanism. Eur J Immunol 31:2313–2318. Heyes MP, Kim P, Markey SP (1988) Systemic lipopolysaccharide and pokeweed mitogen increase quinolinic acid content of mouse cerebral cortex. J Neurochem 51:1946 –1948. Heyes MP, Quearry BJ, Markey S (1989a) Systemic endotoxin increases L-tryptophan, 5-hydroxyindoleacetic acid, 3-hydroxykynurenine and quinolinic acid content of mouse cerebral cortex. Brain Res 491:173–179. Heyes MP, Rubinow D, Lane C, Markey SP (1989b) Cerebrospinal fluid quinolinic acid concentrations are increased in acquired immune deficiency syndrome. Ann Neurol 26:275–277. Heyes MP, Saito K, Chen CY, Proescholdt MG, Nowak TS, Li L, Beagles KE, Proescholdt MA, Zito MA, Kawai K, Markey SP (1997) Species heterogeneity between gerbils and rates: quinolinate production by microglia and astrocytes and accumulations in response to ischemic brain injury and systemic immune activation. J Neurochem 69:1519 –1529. Itakura A, Kurauchi O, Takashima S, Uchida K, Ito M, Mizutani S (2002) Immunological detection of 4-hydroxynonenal protein adducts in developing pontine and Purkinje neurons and in karyorrhexis in pontosubicular neuronal necrosis. Early Hum Dev 67:19 –28. Kern JK (2003) Purkinje cell vulnerability and autism: a possible etiological connection. Brain Dev 25:377–382. Kudo Y, Boyd CA, Sargent IL, Redman CW (2000) Modulation of indoleamine 2,3-dioxygenase by interferon-gamma in human placental chorionic villi. Mol Hum Reprod 6:369 –374. Kudo Y, Boyd CA, Sargent IL, Redman CW (2001) Tryptophan degradation by human placental indoleamine 2,3-dioxygenase regulates lymphocyte proliferation. J Physiol 535:207–215. Losch de Oliveira D, Horn JF, Rodrigues JM, Frizzo MES, Moriguchi E, Souza DO, Wofchuk S (2004) Quinolinic acid promotes seizures and decreases glutamate uptake in young rats: reversal by orally administered guanosine. Brain Res 1018:48 –54. Manuelpillai U, Ligam P, Smythe G, Bustamante S, Wallace EM, Hirst JJ, Walker DW (2005) Identification of kynurenine pathway enzyme mRNAs and metabolites in human placenta: upregulation by inflammatory stimuli and with clinical infection. Am J Obstet Gynecol 192:280 –288. Medana IM, Hien TT, Day NP, Phu NH, Mai NT, Chu’ong LV, Chau TT, Taylor A, Salahifar H, Stocker R, Smythe G, Turner GD, Farrar J, White NJ, Hunt NH (2002) The clinical significance of cerebrospinal fluid levels of kynurenine pathway metabolites and lactate in severe malaria. J Infect Dis 185:650 – 656.

E. Yan et al. / Neuroscience 134 (2005) 867– 875 Nicholls T, Nitsos I, Walker DW (1999) Tryptophan metabolism in pregnant sheep: increase fetal kynurenine production in response to maternal tryptophan loading. Am J Obstet Gynecol 181: 1452–1460. Nicholls T, Lacey B, Nitsos I, Smythe G, Walker DW (2001) Regional changes in kynurenic acid, quinolinic acid, and glial fibrillary acidic protein concentrations in the fetal sheep brain after experimentally induced placental insufficiency. Am J Obstet Gynecol 184: 203–208. Norenberg MD (1994) Astrocyte responses to CNS injury. J Neuropathol Exp Neurol 53:213–220. Norenberg MD (1996) Reactive astrocytosis. In: The role of glia in nuerotoxicity (Aschner M, Kimelberg HK, eds), pp 93–115. Lancaster: CRC Press. Rios C, Santamaria A (1991) Quinolinic acid is a potent lipid peroxidant in rat brain homogenates. Neurochem Res 16:1139 –1143. Sanni LA, Thomas SR, Tattam BN, Moore DE, Chaudhri G, Stocker R, Hunt NH (1998) Dramatic changes in oxidative tryptophan metabolism along the kynurenine pathway in experimental cerebral and noncerebral malaria. Am J Pathol 152:611– 619. Santamaria A, Jimenez-Capdeville ME, Camacho A, Rodriguez-Martinez E, Flores A, Galvan-Arzate S (2001a) In vivo hydroxyl radical formation after quinolinic acid infusion into rat corpus striatum. Neuroreport 12:2693–2696. Santamaria A, Galvan-Arzate S, Lisy V, Ali SF, Duhart HM, OsorioRico L, Rios C, St’astny F (2001b) Quinolinic acid induces oxidative stress in rat brain synaptosomes. Neuroreport 12:871– 874. Sarna JR, Hawkes R (2003) Patterned Purkinje cell death in the cerebellum. Prog Neurobiol 70:473–507. Schmidt-Kastner R, Wietasch K, Weigel H, Eysel UT (1993) Immunohistochemical staining for glial fibrillary acidic protein (GFAP) after deafferentation or ischemic infarction in rat visual system: features of reactive and damaged astrocytes. Int J Dev Neurosci 11: 157–174. Smythe GA, Braga O, Brew BJ, Grant RS, Guillemin GJ, Kerr SJ, Walker DW (2002) Concurrent quantification of quinolinic, picolinic, and nicotinic acids using electron-capture negative-ion gas chromatography-mass spectrometry. Anal Biochem 301:21–26.

875

St’astny F, Skultetyova I, Pliss L, Jezova D (2000) Quinolinic acid enhances permeability of rat brain microvessels to plasma albumin. Brain Res Bull 53:415– 420. Stone TW (1993) Neuropharmacology of quinolinic and kynurenic acids. Pharmacol Rev 45:309 –379. Stone TW (2001) Kynurenines in the CNS: from endogenous obscurity to therapeutic importance. Prog Neurobiol 64:185–218. Stone TW, Perkins MN (1981) Quinolinic acid: a potent endogenous excitant at amino acid receptors in CNS. Eur J Pharmacol 72:411– 412. Tabernero A, Granda B, Medina A, Sanchez-Abarca LI, Lavado E, Medina JM (2002) Albumin promotes neuronal survival by increasing the synthesis and release of glutamate. J Neurochem 81: 881– 891. Takuma K, Baba A, Matsuda T (2004) Astrocyte apoptosis: implications for neuroprotection. Prog Neurobiol 72:111–127. Tavares RG, Tasca CI, Santos CE, Alves LB, Porciuncula LO, Emanuelli T, Souza DO (2002) Quinolinic acid stimulates synaptosomal glutamate release and inhibits glutamate uptake into astrocytes. Neurochem Int 40:621– 627. Tonra JR, Reiseter BS, Kolbeck R, Nagashima K, Robertson R, Keyt B, Lindsay RM (2001) Comparison of the timing of acute bloodbrain barrier breakdown to rabbit immunoglobulin G in the cerebellum and spinal cord of mice with experimental autoimmune encephalitis. J Comp Neurol 430:131–144. Welsh JP, Yuen G, Placantonakis DG, Vu TQ, Haiss F, O’Hearn E, Molliver ME, Aicher SA (2002) Why do Purkinje cells die so easily after global brain ischemia? Aldolase C, EAAT4, and the cerebellar contribution to posthypoxic myoclonus. Adv Neurol 89:331–359. Yan E, Castillo-Melendez M, Nicholls T, Hirst J, Walker DW (2004) Cerebrovascular responses in the fetal brain to low dose endotoxin. Pediatr Res 55:855– 863. Zhang D, Dhillon HS, Mattson MP, Yurek DM, Prasad RM (1999) Immunohistochemical detection of the lipid peroxidation product 4-hydroxynonenal after experimental brain injury in the rat. Neurosci Lett 272:57– 61.

(Accepted 18 April 2005) (Available online 18 July 2005)