The excitotoxin quinolinic acid induces tau phosphorylation in human neurons

Abstract

Some of the tryptophan catabolites produced through the kynurenine pathway (KP), and more particularly the excitotoxin quinolinic acid (QA), are likely to play a role in the pathogenesis of Alzheimer's disease (AD). We have previously shown that the KP is over activated in AD brain and that QA accumulates in amyloid plaques and within dystrophic neurons. We hypothesized that QA in pathophysiological concentrations affects tau phosphorylation. Using immunohistochemistry, we found that QA is co-localized with hyperphosphorylated tau (HPT) within cortical neurons in AD brain. We then investigated in vitro the effects of QA at various pathophysiological concentrations on tau phosphorylation in primary cultures of human neurons. Using western blot, we found that QA treatment increased the phosphorylation of tau at serine 199/202, threonine 231 and serine 396/404 in a dose dependent manner. Increased accumulation of phosphorylated tau was also confirmed by immunocytochemistry. This increase in tau phosphorylation was paralleled by a substantial decrease in the total protein phosphatase activity. A substantial decrease in PP2A expression and modest decrease in PP1 expression were observed in neuronal cultures treated with QA. These data clearly demonstrate that QA can induce tau phosphorylation at residues present in the PHF in the AD brain. To induce tau phosphorylation, QA appears to act through NMDA receptor activation similar to other agonists, glutamate and NMDA. The QA effect was abrogated by the NMDA receptor antagonist memantine. Using PCR arrays, we found that QA significantly induces 10 genes in human neurons all known to be associated with AD pathology. Of these 10 genes, 6 belong to pathways involved in tau phosphorylation and 4 of them in neuroprotection. Altogether these results indicate a likely role of QA in the AD pathology through promotion of tau phosphorylation. Understanding the mechanism of the neurotoxic effects of QA is essential in developing novel therapeutic strategies for AD.

Figure 1. Simplified diagram of the kynurenine pathway.

Figure 2. Immunodetection of tau and QA in AD sections.

QA (red; Alexa 647) and tau (AT8; green; Alexa 594) immunolabelling in control and AD cingulate cortex (a–e) and hippocampus (f–g). a) Neuritic plaque in the cingulate cortex of a late-stage AD showing the AT8-positive neurites surrounding by QA-positive cells. The majority of this labelling can be seen in vesicle-like structures within microglial cells, however QA+ neurons in the plaque vicinity are also present (asterix). b) Co-localisation of an intracellular neurofibrillary tangle and QA. This tangle containing neuron is contacted by QA+ non-neuronal cells. See also supplementary figure S3 for additional views through this neuron. c) QA-positive vesicles within a microglia and d) neuron. e) Punctate tau-positivity in a neuron also positive for QA. f) 73-year old neurologically normal control hippocampal CA1 pyramidal cells with very faint immunopositivity for QA. This case had small numbers of entrohinal tangles, but otherwise no other AD-type lesions were detected. In non-AD cases, QA+ neurons rarely occurred outside the hippocampal-entorhinal sector. g) Hippocampus of a mild AD case (Braak stage IV) with QA co-localisation with AT8+ intraneuronal tangles. Scale bars 10 µm a–e; 50 µm f–g.

Figure 3. Immunodetection of Tau Phosphorylation in primary cultures of human neurons.

Primary cultures of human neurons were immunostained for tau phosphorylation with three different antibodies: Line A, staining of total tau using mAb Tau 5 (Abcam); line B, staining of phosphorylated tau using mAb Tau 180 (Innogenetics); and line C, staining of phosphorylated tau using mAb Tau 8 (Innogenetics). The left column corresponds to untreated neurons, the middle column to QA 500 nM treated cells and the right column to QA 1200 nM treated cells. The histograms represent the semi-quantification of the immunostaining intensity. Intensity has been quantified using ImageJ 10.2. This experiment has been performed in triplicate with primary cultures of neurons prepared from three different brains. Three individual microscopic fields were analysed for each treatment and the SEM for the data was determined to be <5%.

Figure 4. QA induces tau phosphorylation at multiple serine and threonine residues:

Primary human foetal neurons were cultured at a density of 1×10⁶ cells per well in a 12-well plate in the absence or presence various concentrations of QA. 100 nM Okadaic acid was used as positive control for tau phosphorylation. After 72 hours, cells were lysed in RIFA buffer. The lysate protein resolved in 10% SDS-PAGE and Western blot performed with the antibodies indicated. Antibodies tested are: A, AT8 (1∶1000 dilution); B, AT180 (1∶1000); C, AT270 (1∶2000); D, PHF-1 (1∶1000). Signal for total tau was determined with the antibody Tau at 1∶1000 dilution. Blots were quantified using the Quantity-One software (Biorad). Signal for phospho tau epitopes was normalized by the signal for total tau and the data expressed as percent of the normalized signal in control cultures. Data presented are the mean (±SD) of five different experiments with neurons from five different brains. Each graph is accompanied with a representative blot in which lane 1 is control; lane 2, 100 nM QA; lane 3, 350 nM QA; lane 4, 500 nM QA; lane 5, 1200 nM QA; lane 6, 100 nM Okadaic acid.

Figure 5. QA treatment induced the aggregation of tau:

A: Representative blot showing the presence of a high molecular weight (∼165 kDa) band recognized by the PHF-specific antibody AT270 in response to QA treatment. Treatments are the same as shown in Figure 4. B: Quantitative estimation of the high molecular weight tau band normalized for total tau. The data presented are the average (±SD) of four experiments from four different brains and are expressed as the ratio of signal for AT270 to that of signal for total tau.

Figure 6. QA treatment reduces total phosphatase and PP2A activity in cultured neurons:

Neuronal cultures were treated with the concentrations of QA shown, for 72 hours. Cells were lysed in the phosphatase activity buffer (50 mM Tris, pH 7.5, 150 mM NaCl, 1% NP-40, 1 mM PMSF, 1 mM Benzamidine, 10 mM β-mercaptoethanol (βME), 2 µg/ml aprotinin, 2 µg/ml pepstatin and5 µg/ml leupeptin. A: Total phosphatase activity was determined using pNPP (10 mM) as a substrate in 50 mM Tris-HCl, pH 7.4, 0.1 mg/ml BSA, 3 mM CaCl₂, 1 µM calmodulin and 3 mM MnCl₂. The reaction mixture was incubated at 30°C for 30 min and the absorbance measured at 405 nm. Data shown are the average (±SD) of three experiments. B: OA-inhibited (PP2A) in the lysate of neurons treated with QA. Total phosphatase activity towards pNPP was determined as above in the presence or absence of 10 nM OA. The difference in the activity lost due the addition of 10 nM OA was calculated as PP2A activity. C: Activity of immunoprecipitated PP2A towards pNPP. PP2A was immunoprecipitated from 400 µg lysate protein with anti-PP2A (clone 1D6). 20 µl of bead were resuspended in 200 µl of phosphatase activity buffer containing 50 mM Tris-HCl, pH 7.4, 3 mM MnCl₂ and 10 mM pNPP. The reaction mixture was incubated at 30°C for 30 min with occasional mixing. Beads were removed and the OD of the clear supernatant was measured at 405 nm. Data from three separate experiments are shown as mean (±SD).

Figure 7. QA treatment decreases the expression of PP1, PP2 and PP5 but increases the expression of PP2B:

20 µg lysate protein (or 20 µl PP2A-IP beads for Figure 7C) was resolved on 10% SDS-PAGE, protein transferred on to PVDF membrane and probed with respective antibodies for each phosphatase. For loading control, the same membranes were stripped and re-probed with anti-actin antibody. For PP2A-IP the band of IgG heavy chain was used as loading control. Graph are mean (±SD) of three experiments (from three separate brains) and are accompanied by a representative blots. Each bar of the quantification graph represents the corresponding band above for each phosphatase.

Figure 8. QA-induced tau phosphorylation occurs through the activation of NMDA receptor, and memantine inhibits QA-induced tau phosphorylation:

Primary cultures of human neurons were treated with 500 nM each of three NMDA receptor agonists (QA, glutamate and NMDA) and three antagonists (memantine, MK-801 and AP-5). One well (termed as cocktail) was treated with a combination of all the three antagonists (500 nM each). Control wells were cultured in medium alone. Cells were cultured for 24 hours and were then lysed in RIFA buffer. 20 µg of protein lysate was resolved on 10% SDS PAGE, transferred onto PVDF membrane and probed with phosphor-specific antibodies AT8 and AT180 (1∶100 dilution). Antibody Tau (DAKO, 1∶10000) was used for total tau. Quantization data is presented as ratios of AT8 and AT-180 signal to total tau signal and is the average (±SD) of three experiments. A representative blot is for each graph is shown in the insert.

Figure 9. Determination of the genes induced by QA in primary cultures of human neurons using PCR arrays:

The 3D histogram shows 84 genes tested in the Pathfinder kit®. Only 10 genes (in red) were significantly increased (expression>2x the control). No genes were significantly decreased.

Figure 10. hypothetical model of QA effects on Tau phosphorylation in primary human neurons.