Characterisation of the Expression of NMDA Receptors in Human Astrocytes

1N-methyl D-aspartate (NMDA) receptors (NMDARs) are ligand-gated ion channels which form one group of ionotropic glutamate receptors in the CNS. These receptors are known to be expressed in neurons and are activated by neurotransmitters including glutamate and NMDA as well as endogenous excitotoxins such as quinolinic acid. NMDARs exist as heterotetrameric complexes at the surface membrane. Currently, seven known subunits have been identified: one NR1, four NR2 (A–D) and two NR3 (A–B) subunits. In neurons, NMDARs play an important role in facilitating learning and memory. However, recent studies have revealed that these membrane proteins may also exist in other cell types,,,.

2Although research into astrocyte NMDARs is still controversial and functional expression of these receptors in humans is yet to be confirmed,, there has been some evidence from animal models suggesting the involvement of astrocytic glutamate receptors in glial cell signalling,,,,. This form of glial communication involves the induction of an intracellular calcium wave, which was first elicited through the stimulation of cultured astrocytes with glutamate. The study showed that astrocytes could in fact respond to extracellular neurotransmitters such as those released by neurons via a signalling pathway that could potentially be used for physiological glial communication although more research is required to elucidate the exact pathways of activity.

3The astrocytic glutamatergic system has also been implicated in several neuropathological conditions including amyotrophic lateral sclerosis (ALS) and Alzheimer's disease (AD). We have previously shown that extracellular levels of the NMDAR agonist and neurotoxin quinolinic acid (QUIN) are significantly increased in both AD and ALS,. QUIN is an endogenous metabolite of L-tryptophan, which is produced via the kynurenine pathway. In the brain, the amino acid L-tryptophan is normally used in protein synthesis and metabolised to compounds such as 5-hydroxytryptamine and other indoleamines. However, during neuroinflammatory conditions, increased activity of the enzyme, indoleamine-2,3-dioxygenase (IDO-1) directs metabolism down the kynurenine pathway and increases the formation of QUIN. Following immune activation in the brain, QUIN is produced by activated microglia and invading macrophages, with high levels of this neurotoxin associated with increased neuronal and glial cell death. These changes are seen in a number of diseases including AD and AIDS dementia complex. As QUIN is an NMDAR agonist, the activation of glutamatergic pathway is likely to be responsible for the cytotoxic effects observed in astrocytes,.

4The aim of this study was to investigate the expression of NMDARs in human primary astrocytes as well as characterise the response of these receptors to physiological and excitotoxic concentrations of known NMDAR agonists. Confirming whether NMDARs exist in astrocytes and examining the function of these receptors in glial cells will further enhance our understanding of the functional role astrocytes play in normal, healthy environments, as well as their potential involvement in neuropathological conditions,.

5Dulbecco's phosphate buffered saline (PBS) 1×, RPMI medium 1640 1×, 0.5% Trypsin-EDTA 10×, Glutamax-1 100×, Antibiotic-Antimycotic (AA) 100×, Trizol were obtained from GIBCO Invitrogen (Victoria, Australia). Glucose intravenous infusion BP 50% was from AstraZeneca (Sydney, Australia). Foetal calf serum (FCS) was obtained from Bovogen Australia. Culture flasks and plates were purchased from Becton Dickinson lab ware (New Jersey, USA). QUIN and glutamate were obtained from Sigma Chemical (Sydney, Australia). PCR reaction buffer 10×, MgCl2, dNTP and Taq DNA polymerase were obtained from Roche (Mannheim, Germany). Permanox chamber-slides were obtained from Lab-Tek (California, USA) and Fluoromount-G was obtained from Southern Biotech (Alabama, USA.). The BD Calcium Assay Kit was obtained from BD Biosciences (California, USA) and the Fluostar Optima Fluorometer was obtained from BMG LABTECH (Victoria, Australia). The Hepes-buffered Krebs solution, Fura-2 and pluronic F-127, glutamate, QUIN, MK-801, and memantine were purchased from Sigma-Aldrich (Missouri, USA). Trans-ACBD was obtained from Tocris Biosciences (Bristol, UK). The Microplate reader 680XR was obtained from Bio-Rad (California, USA).

6Human adult brain tissue and human foetal brain tissue were obtained following informed written consent. This has been respectively approved by the Human Ethics Committees from St Vincent's Hospital (HREC 08284) and from the University of New South Wales (UNSW Ethic approval HREC 03187).

7Human brain tissue was obtained from adult donors post-surgery and 16- to 19-week-old foetuses. Astrocytes were prepared using a previously described protocol. Briefly, cerebral portions were washed thoroughly with PBS and dissociated by repeated pipettings. The suspension was centrifuged at 500 g for 5 minutes and the cell pellet resuspended in RPMI 1640 medium containing 10% heat-inactivated FCS, 1% glutamax-1, 1% antibiotic-antimicrobial liquid, 0.5% glucose, then plated onto 75 cm2 culture flasks and incubated at 37°C. Medium was changed on the 3rd, 5th and 10th day. The cells became confluent after 10–12 days. Microglia were detached from the cultures by mechanically shaking the flasks for 2 hours at 220 rpm at room temperature and aspirated. The astrocytes were trypsinized and replated at least three times to further purify and isolate astrocytes from contaminating microglia and neurons. Astrocytes were left to recover for 3 days after each passage. The astrocytes were rinsed twice with PBS and cultured as above in uncoated flasks with the culture medium and maintained for up to 6 weeks. For the all the following experiments, we have used cell cultures at 3–4 weeks. The medium was changed twice a week. Culture purity was determined by immunofluorescence analysis with antibodies against GFAP and more than 95% of the cells stained positive for GFAP (Figure S1). We did not detect any staining for CD68 (microglia) and for 5B5 (fibroblasts); data not shown.

8Using the same brain samples mentioned above, human primary neurons have been isolated and grown as we have previously described.

9RNA extraction was undertaken using the Invitrogen Trizol protocol. The cDNA was subsequently quantified via spectrophotometry. Original primers for the seven NMDAR subunits were designed using ‘Primer3’ primer design software, then tested and optimized for the PCR. The individual primers were analysed via the BLAST database to ensure specificity of binding. Refer to Table 1 for list of primers. For each NMDAR subunit being tested, 1 µl of cDNA sample (1 µg/µl) was added to a PCR reaction mixture of 49 µl containing: 10 µl of PCR reaction buffer 10×, 5 µl of 2.5 mM MgCl2, 2 µl of 10 mM dNTP, 1 µl of 25 mM forward primer, 1 µl of 25 mM backward primer, 0.5 µl of 5 U/µl Taq DNA polymerase and 29.5 µl of DEPC water. PCR was subsequently run on samples for 40 cycles under the following conditions: denaturation (94°C for 60 s), annealing (60°C for 60 s) and extension (72°C for 90 s). Positive controls for primers of the NMDAR subunits involved using reverse transcribed cDNA from a mixture of neuronal and glial brain cells.

10Human foetal astrocytes prepared from 17 to 20-week-old foetuses (n = 4) were used for real-time PCR. Human foetal neurons were used for primer optimisation and as a positive control for target gene expression. Primer optimisation was performed using a standard curve derived from serial dilutions of human foetal neuron cDNA (0.5 pg–50 ng). Multiple primer pairs for each gene were tested and the most suitable primer candidates were chosen by assessing the specificity of the PCR product and efficiency. Specificity was verified by a single peak in melting curve analysis. The real-time PCR assay for unknown samples was performed simultaneously with positive control samples (neurons) in the same plate.

11Total RNA was extracted using the PureLink RNA Mini Kit according to the manufacturer's protocol. RNA quantity was evaluated spectrophotometrically. RNA integrity was confirmed by the Agilent 2100 electrophoresis bioanalyzer. 1 ug of total RNA in a final volume of 20 ul was used for the synthesis of cDNA using the Superscript Vilo cDNA Synthesis Kit in accordance with the manufacturer's recommendations. cDNA was diluted to a concentration of 2.5 ng/µl with water for real-time PCR.

12The NR2A, NR2B, NR2C, NR2D, NR3A, and GAPDH mRNA transcripts were quantified using the oligonucleotide primers for real-time PCR amplification designed based on sequences published in the Harvard Primer Bank and the NCBI primer-designing tool where the sequences were blasted. GAPDH was used as the housekeeping gene. The forward and reverse primer sequences used in this study are given in Table 1.

13Primers were used at 300 nM final concentration. Briefly, 5 µl of the diluted synthesized cDNA together with the appropriate primers was added to 10 ul Brilliant III UltraFast SYBR green QPCR Master mix (Agilent Technologies) to a total volume of 20 µl. A PCR reaction master mix was prepared for each primer before dispensing into 96× PCR plate cells. No-template control (NTC) reactions were also prepared for each gene. Real-time PCR was carried out using a Stratagene Mx3000P (Agilent Technologies). The cycling parameters for all genes were the following: initial denaturation at 95°C for 3 min, then 40 cycles of 95°C for 20 s, and 60°C for 20 s. All 6 transcripts were measured in each unknown sample in duplicates. Expression values were normalised to GAPDH and are reported in units of 2−ΔCt, where ΔCt is the difference in Ct values between the target and GAPDH transcripts in the same sample.

14Data are presented as fold differences of mean normalised expression values ± standard error of mean (SEM). Differences in the relative expression of each gene in HFA were verified by applying the one-way ANOVA Tukey's multiple comparison test using GraphPad Prism 5 (GraphPad software, San Diego, CA, USA)). Statistical significance was accepted at P<0.05; n = 4.

15The method is previously described,. Briefly, purified foetal astrocytes were grown in Permanox chamber-slides until around 70–75% confluence. Cells were then fixed with methanol-acetone (1∶1) solution. Membranous permeabilization was subsequently undertaken using 0.1% Triton-X in PBS for 10 minutes at room temperature. The cells were then washed twice with PBS and incubated in PBS 5% NGS overnight at 4°C. Dilutions of the primary antibodies were made in PBS 5% NGS. Refer to Tables 2 for list of primary and secondary antibodies. Primary and secondary antibody incubations were subsequently undertaken at 37°C for 1 hr each followed by nuclear staining using DAPI (1 µg/ml) diluted 1/500 in water. Cover slips were mounted with Fluoromount-G and the slides were then viewed under the Olympus microscope BX60. Visualisation of the cells under confocal microscopy was also undertaken using the same protocol as immunocytochemistry. Slides were subsequently read under the Olympus FV1000 Laser Scanning microscope.

16Positive controls for immunocytochemistry involved primary antibody staining of NMDARs in human primary neurons using the same protocol. Negative controls were also employed via incubations with only secondary antibodies, to detect any non-specific secondary antibody binding as well as exclude the presence of auto-fluorescent cells (data not shown).

17Human foetal astrocytes were transferred into 96-well, flat-bottom plates and allowed to proliferate until confluence was reached. Astrocytic dye loading with calcium indicator was subsequently undertaken using the BD Calcium Assay Kit following the manufacturer's technical data sheet. 4 mM Probenecid was also added to the mixture followed by a 1 hr incubation period at 37°C. The loading solution was then removed and replaced with HBSS containing 50 mM glycine. Addition of selective NMDAR antagonists was undertaken 5-minutes prior to the experiment to ensure adequate diffusion time was provided to attain equilibrium. The calcium influx experiments were subsequently performed using the Fluostar Optima Fluorometer. Filter excitation and emission was set at 485 nm and 520 nm wavelength respectively. For each well, fluorescence was measured via orbital scanning of 10 locations at a 3 mm radius every 0.5 seconds. The average of these readings was recorded. Baseline fluorescence was always measured during the first 10 seconds of the experiment, which was followed by treatment with an NMDAR agonist (prepared in HBSS solution) via automated syringe injection at 100 µL/s. Fluorescent readings were subsequently taken for an additional 90 seconds. Negative controls included the injection of only HBSS solution without any agonist.

18Concerning the NMDAR agonists used in the experiment, glutamate and QUIN were chosen as they are endogenous ligands found commonly in the CNS. The synthetic compound, trans-ACBD was also selected as a treatment option due to its chemical properties of being a very potent and highly selective NMDAR-specific agonist (Lanthorn et al; 1990).

19Human foetal astrocytes were trypsinized and transferred into 50 mm glass-bottom micro well dishes. When confluence was reached, the astrocytic growth media was removed and the cells were incubated for 1 hr at 37°C in Hepes-buffered Krebs solution containing 2 µM Fura-2 and 4 µM pluronic F-127. After dye loading, cells were washed thoroughly with Hepes-buffered Krebs solution. HBSS containing 50 mM glycine was used as the incubation medium during the experiment. Fluorescent changes caused by calcium influx were detected via a video-based imaging system in conjunction with an Olympus IX81 inverted microscope and a 20× Nikon Fluor objective lens. The emission at 520 nm was measured following excitation at 340 nm and 380 nm. Fluorescent images were recorded using the ratio of excitation by the two different wavelengths to index true changes in calcium entry. Baseline fluorescence was taken every 15 seconds for 2 minutes prior to addition of the NMDAR agonist using a manual micropipette. Fluorescent measurements were subsequently taken every 5 seconds for an additional 2 minutes following treatment. The negative control included detecting any changes in fluorescence when only HBSS solution was added.

20Human foetal astrocytes were trypsinized and equal volumes of cell suspension were plated into 24-well tissue culture plates. Cells were monitored regularly and allowed to reach full confluence. Selective NMDAR antagonists, MK-801, and memantine were then added in triplicates to each well at varying concentrations, followed 5 minutes later by the administration of the NMDAR agonists, QUIN and glutamate at excitotoxic levels of 500 nM and 500 µM respectively. The cells were then incubated at 37°C for 24 hours after which the media was collected and analysed for levels of LDH release.

21For the LDH analysis, samples were initially diluted 1/100 and LDH activity was assayed using a standard spectrophotometric technique described by Koh and Choi.

22LDH levels were also adjusted to take into account variations in cell number using the Bradford protein assay described by Bradford. Briefly, each well was filled with 200 µl of PBS and sonication of the cells was undertaken at 20 kHz for 15 seconds in each well. These sonicated samples were collected and a set of standard protein was prepared using varying amounts of BSA. Next, 60 µl of Bradford reagent (0.01% Coomassie brilliant blue G-250, 5% ethanol, 8.5% phosphoric acid) was added to 240 µl of each sample and standard. After incubating for 30 minutes at room temperature, the samples and standards were measured for their absorbance at 595 nm using the Microplate reader 680XR (Bio-Rad; California, USA).

23Results obtained are presented as the means ± the standard error of measurement (SEM). Significant differences between results were verified using the two-tailed t-test with equal variance. Differences between treatment groups were considered significant if p was less than 0.05 (p<0.05).

24End-point PCR results showed that both human foetal and adult astrocytes express the mRNA of all known NMDAR subunits including NR1, 2a, 2b, 2c, 2d, 3a, 3b (Figure 1A). Mixed brain cell cultures (including neurons, astrocytes, microglia, oligodendrocytes) were used as positive controls (first column). Semi-quantitative studies using ANOVA analysis did not reveal a significant overall difference between adult and foetal expression of the seven sub-types of NMDARs (p = 0.13) (figure 1B).

25Real-time PCR has been performed human foetal astrocytes validating the above End-point PCR results and providing more quantitative data. Target genes and the housekeeping gene show high specificity as verified by a single peak in melting curve analysis (Figures 1C). Target genes and the housekeeping gene show high specificity as verified by a single peak in melting curve analysis (Figure 1C). mRNA expression of NR2A, NR2B, NR2C, NR2D, and NR3A was detected in HFA (Figure 1D). No more than 10% difference in primer efficiency between the reference gene and the target genes was obtained. NR3A mRNA expression was significantly lower compared to NR2D mRNA expression (p<0.05). GAPDH expression did not differ between all samples (data not shown).

26Immunocytochemistry confirmed our RT-PCR data showing that all NMDAR subunits are also expressed in human foetal astrocytes at the protein level (Figures 2B and 2C). Staining of human foetal neurons were used as positive control and revealed the presence of NMDAR expression as expected (Figure 2A). Negative controls were performed using incubations with just the secondary antibodies (Figures 2A, 2B and 2C). The presence of only blue nuclei fluorescence excludes any non-specific secondary antibody binding and also rules out the presence of auto-fluorescent cells. Incubations with an IgG irrelevant antibody as well as incubations with normal rabbit serum were also undertaken and revealed no non-specific staining. Three images were taken for each NMDAR subunit and the best image has been displayed. In addition, single staining for MAP-2 (Figure 1A) and GFAP (Figures 2A and 2C) was also undertaken to show the cells were of neuronal and astrocytic origin respectively (see also Figure S1). We then performed a double staining for GFAP and NR1 and NR2A/B. We observed a characteristic fibrous staining of a structural protein for GFAP and fine-pigmented immunolabelling for the NMDAR subunits (Figure 2C). This double staining confirms the cells expressing NMDARs are human primary astrocytes.

27Confocal microscopy revealed that the majority of NR1 and NR2A subunits are located on the surface membrane of human foetal astrocytes (Figure 2C). In addition, a rim of cytoplasm mostly devoid of fluorescence can be seen separating the plasma membrane from the central lying blue-stained nucleus. From the confocal microscopy results, the detection of NMDARs on the surface membrane suggests a much higher likelihood that these ion channels are functional in astrocytes.

28Stimulation of astrocytes using glutamate (Figure 3A) and QUIN (Figure 3B) results in calcium influx into the cells, with the intensity of the response increasing as the concentration of the NMDAR-agonist increases. The intracellular movement of calcium was significantly inhibited by pre-incubating the astrocytes with NMDAR-specific antagonists, memantine 20 µM or MK-801 20 µM for 5 min prior to stimulation. The inhibitory activity of these NMDAR-specific antagonists on glutamate and QUIN are shown in Figures 3C and 3D respectively. The glutamate concentration of 500 µM used in the experiment is well within the normal levels reached at the synaptic junction following neuronal depolarisation. QUIN 1 µM is a pathophysiological concentration found in the CNS of patients with neuroinflammatory conditions. ANOVA analysis on the quantified amplitudes of astrocytic response showed that treatment with both glutamate and QUIN in the presence of specific NMDA antagonists resulted in a calcium influx response that was statistically lower than the control (p<0.05) (Figure 3).

29Real-time fluorescent imaging of astrocytes showed similar results, with calcium influx and subsequent cell fluorescence being detected after stimulation with trans-ACBD, which is a highly selective NMDAR agonist (Figure 4 and avi movie Video S1). One interesting observation when the experiments were being performed was that not all astrocytes responded with calcium entry following treatment with trans-ACBD. In the images taken using real-time microscopy, the cell on the bottom left did not fluoresce even after stimulation with an NMDAR agonist. This finding may reflect the heterogeneity in astrocytic NMDAR expression observed by immunocytochemistry.

30Significantly higher levels of LDH were detected in the culture supernatant following astrocyte stimulation with QUIN and glutamate (Figure 5A), reflecting increased cell death. Linear regression analysis showed that increasing concentrations of both NMDAR agonists resulted in higher levels of LDH release; glutamate (p<0.05), QUIN (p<0.05). A two-tailed t-test also revealed that the different levels of LDH release were statistically significant when treatment with glutamate (50 µM) and QUIN (200 nM) was compared to control. The data were also analysed to determine the agonist concentrations that would be most suitable for subsequent experiments involving the addition of selective NMDAR antagonists. The criteria for an appropriate excitotoxic agonist concentration included ensuring sufficient amount of cell death would occur whilst trying to maintain at least some viable astrocytes after treatment. Using the above graphs as well as observations of cell morphology after the excitotoxic incubation period, the QUIN concentration of 500 nM and glutamate concentration of 500 µM were chosen.

31Pre-incubation of the astrocytes with MK-801 and memantine prior to incubation with NMDAR agonists was able to reduce LDH release, suggesting that the excitotoxic effects of both QUIN and glutamate act at least partially via the NMDAR pathway (Figure 5B). Linear regression analysis showed that both antagonists were individually effective at reducing astrocyte LDH release when added in the presence of excitotoxic concentrations of glutamate (MK-801 p<0.05; memantine p<0.01) and QUIN (MK-801 p<0.01; memantine p<0.01). Linear regression analysis with dummy variables did not show any statistical difference in the inhibitory activities of MK-801 or memantine under either conditions of excitotoxic glutamate or QUIN.

32Astrocyte ionotropic glutamate receptors (iGluRs) and in particular NMDARs may play an important role in facilitating glial signalling in the CNS. We have shown that adult and foetal human primary astrocytes express all known NMDAR subunits including NR1. NR2A, 2B, 2C, 2D and NR3A, 3B. These results are in accordance with previous immunohistochemical studies showing that NR1 and NR2A/B are expressed by the rodent and human astrocytic processes.

33We also found that astrocyte calcium concentrations can be significantly increased when stimulated with glutamate (p<0.05) and that the mechanism of calcium elevation is at least partially attributable to activation of the NMDAR pathway. These results are supported by the real-time fluorescent microscopy data, revealing the presence of calcium entry following astrocyte stimulation with a selective NMDAR agonist (Figure 3 and 4). Results from the fluorometry studies (Figure 3) also substantiate this by showing the inhibition of calcium influx following the addition of selective NMDAR antagonists (p<0.05). Furthermore, the similar NMDAR expression profiles between adult and foetal specimens suggest that studies on foetal astrocytes may be applied to adult models. We previously compared simian adult astrocytes and human foetal astrocytes for the ability to produce chemokines and chemokines receptors and we did not found any differences between adult and foetal cells. However further research is required to assess any potential differences between adult and foetal NMDARs functions of individual NMDAR subunits.

34The finding that NMDAR agonists such as glutamate can cause a rise in intracellular calcium levels within astrocytes may help to explain the mechanism of glial signalling within the CNS. Whilst in neurons, sodium channels are the main receptors for initiating depolarisation, it is now known that a different form of signalling occurs in glial cells, which involves the generation of calcium waves. Astrocytic α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptors (AMPARs) and metabotropic glutamate receptors (mGluRs) have been suggested to fulfil this role of facilitating rises in intracellular calcium, however the results of this study reveal that NMDAR activation may also be an important mechanism for initiating calcium influx given that these ion channels traditionally have much higher calcium permeability when compared to AMPARs.

35The glial signalling pathways triggered by this rise in intracellular calcium are uncertain. Glutamate release has been shown to occur following an increase of cytosolic calcium in astrocytes. Previously supported mechanisms of astrocytic neurotransmitter efflux were centred upon cytosolic calcium release from internal stores through the IP3 pathway following the activation of G-protein coupled mGluRs. This study however, suggests an alternative pathway involving NMDAR activation to achieve astrocyte intracellular calcium elevation, subsequently facilitating extracellular glutamate release. Nonetheless, it is likely that both mGluRs and iGluRs are involved in glial signalling and further research would be required to determine any unique roles of each subtype.

36Whilst physiological concentrations of glutamatergic neurotransmitters may be involved in glial communication in the CNS, pathological conditions can also result from prolonged exposure to excitotoxic levels of glutamate and other NMDAR agonists such as QUIN. Concerning AD, β-amyloid (Aβ) plaques are characteristic of this condition and their presence during diseased states has been well documented. These neurotoxic lesions are often found surrounded by extensive areas of cellular necrosis. The mechanism by which glial excitotoxicity is caused by Aβ aggregation in the brain is still unclear. One study by Guillemin showed that Aβ plaques induced the expression of IDO-1 in surrounding macrophages and microglia, resulting in the increased production of QUIN from these cells. This finding was supported by subsequent studies revealing that the concentration of QUIN was raised in brain sections from AD patients. QUIN is also able to increase tau phosphorylation leading to the formation of neurofibrillary tangles. Memantine is able to reverse this effects and.

37From the results of the LDH cytotoxicity experiments, the ability of selective NMDAR antagonists to reduce the level of astrocyte cell death in the presence of high concentrations of QUIN suggests that the excitotoxic effects of QUIN are at least partially mediated through their actions on glial NMDARs (Figure 3). Although it is known that QUIN is an agonist of neuronal NMDARs, this study also postulates that the neurotoxin acts via a similar mechanism in astrocytes to cause glial cell death. Furthermore, it has been shown that QUIN not only over-activates NMDARs but it can also inhibit astrocytic glutamate uptake and glutamine synthetase expression, compounding the excitotoxic environment in the CNS during neuroinflammatory conditions,. Therefore, the results of this study suggest that astrocytic NMDARs may be responsible for the glial pathology found in AD and other neurological diseases due to their involvement in facilitating glutamate and quinolinic acid excitotoxicity.

38In conclusion, the present study is the first to conclusively show the functional expression of NMDARs in human primary astrocytes. This discovery has far-reaching implications, particularly in redefining the role of astrocytes in both physiological progresses and pathological conditions.