Monomeric C-reactive protein--a key molecule driving development of Alzheimer's disease associated with brain ischaemia?

Abstract

Alzheimer's disease (AD) increases dramatically in patients with ischaemic stroke. Monomeric C-reactive protein (mCRP) appears in the ECM of ischaemic tissue after stroke, associating with microvasculature, neurons and AD-plaques, Aβ, also, being able to dissociate native-CRP into inflammatory, mCRP in vivo. Here, mCRP injected into the hippocampal region of mice was retained within the retrosplenial tract of the dorsal 3rd ventrical and surrounding major vessels. Mice developed behavioural/cognitive deficits within 1 month, concomitant with mCRP staining within abnormal looking neurons expressing p-tau and in beta-amyloid 1-42-plaque positive regions. mCRP co-localised with CD105 in microvessels suggesting angiogenesis. Phospho-arrays/Western blotting identified signalling activation in endothelial cells and neurons through p-IRS-1, p-Tau and p-ERK1/2-which was blocked following pre-incubation with mCRP-antibody. mCRP increased vascular monolayer permeability and gap junctions, increased NCAM expression and produced haemorrhagic angiogenesis in mouse matrigel implants. mCRP induced tau244-372 aggregation and assembly in vitro. IHC study of human AD/stroke patients revealed co-localization of mCRP with Aβ plaques, tau-like fibrils and IRS-1/P-Tau positive neurons and high mCRP-levels spreading from infarcted core regions matched reduced expression of Aβ/Tau. mCRP may be responsible for promoting dementia after ischaemia and mCRP clearance could inform therapeutic avenues to reduce the risk of future dementia.

Figure 1. Kinexus Western phospho-microarray analysis and Western blotting of mCRP-induced signalling in BAEC.

A shows quantitative Kinexus phospho-protein screening array carried out on BAEC after exposure to mCRP (8 minutes) demonstrated up-regulation of several potentially important proteins that may be implicated in AD pathology including Tau (2.3 fold) Focal adhesion kinase and IRS-1 (3.4 fold). IRS-1 was investigated in more detail in our in vitro studies Fig. 1B shows by Western blotting in the same samples, that mCRP induced approximately a fourfold increase in p-IRS expression compared with control untreated cells (bar chart). P-Tau was also increased by approximately 5-fold (C) and in addition, we showed that the cellular content of Aβ1–42 increased 3-fold over 24 h whilst PEN-2 also increased (2-fold; 8 minutes). These experiments were carried out at least twice and a representative example is shown.

Figure 2. Effects of siRNA knock-down of IRS-1 on BAEC angiogenesis and cell signalling:

BAEC were subjected to siRNA knock-down of IRS-1as described in the Materials and Methods section of this article. After 48 h treatment, approximately an 85% reduction in IRS-1 gene expression was noted (A). In contrast, NC siRNA had no effect on IRS-1 gene expression. Knock-down was tested for each experiment and found to be similar and the figure shows a representative example. (B) a reduction (50%) in mCRP-induced tube-like-structure formation was seen in siRNA-treated cells. The bar chart shows significant reduction in tube formation in the presence of IRS-1 siRNA (**p < 0.01; *p < 0.05). Pre-incubation with our characterised antibody specific for mCRP (4 h; 1 μg/ml) was able to inhibit both p-ERK1/2 and p-IRS-1 expression in BAEC (C) and also, significantly, tube-like-structure formation (D; *p < 0.05). Note nCRP was not tested in these assays as we have previously shown that it has no pro-angiogenic activity These experiments were carried out three times and where statistical analysis was performed, results represent the mean ± S.D of these experiments.

Figure 3. Characterization of mCRP-induced vascular activation:

(A) BAEC spheroids were generated to examine the effect of mCRP on sprout structure and formation in a 3-dimensional system. In normal culture conditions, sprouting was slower, sprouts had a thicker appearance and cell-cell junctions were maintained (left panel). In the presence of mCRP, sprouts formed more quickly, were notably thinner in appearance and the inter-cellular gaps between cells was notably larger (right panel). (B) Dorsal matrigel implants containing mCRP (10 μg/ml; 72 h) produced strong and visible haemorrhagic angiogenesis (iv; arrows) compared with a typical, normal looking vascular response seen in the presence of VEGF (ii; 25 ng/ml), whilst nCRP (10 μg/ml) produced very little angiogenic response (p < 0.05 increase in the presence of mCRP and VEGF compared with control implants) (iii). In (C) the graph shows a significant increase in monolayer permeability in the presence of mCRP (10 μg/ml; 8 h) using a Millipore-based filter assay, similar to that produced by 10% DMSO (p < 0.01 increase in FITC dextran penetrating the monolayer in the presence of either mCRP or the positive control DMSO), and lighter regions in the images shown indicate areas of increased permeability. (D) Expression of adhesion molecules was examined in BAEC treated with mCRP (10 μg/ml; 24 h). NCAM expression was increased by approximately 2.8 fold whilst VCAM, ICAM and integrins were not affected (data not shown). β-tubulin was used as the house keeping control (gel and bar chart shown). These experiments were repeated at least twice and a representative example is shown.

Figure 4. Behavioural changes observed following mouse-hippocampal injection of mCRP.

Stereotactic injection of mCRP (50 μg) directly into the CA1 hippocampal region of mice were examined 3–4 weeks after operation. A significant reduction in balance and grip was see in non-transgenic animals in the presence of mCRP (wire hang test; (A) whilst head dipping latency in the Boissier’s hole-board test was significantly increased in 3xTg mice and further increased when mCRP was present (B). In the same test, the latency to entry into 4 holes was significantly increased in both mCRP-treated non-transgenic and 3xTg mice. Statistical analysis was done using two-way ANOVA (*P < 0.05).

Figure 5. Cognitive effects of mCRP on mice following hippocampal injection.

(A–C) shows results for novel object recognition. Whilst no effects were seen at time zero, a trend was seen after 2 h and after 24 h, visual discrimination ratio was significantly decreased in the presence of mCRP in both non-transgenic and 3xTg mice (C). Similarly, in the water maze test (D–F), mCRP-treated mice showed a significant increase in distance covered (D) and also reduction of % time in the target quadrant (E,F). Statistical analysis was done using two-way ANOVA (*P < 0.05).

Figure 6. IHC of mouse brain tissue sections following mCRP hippocampal injection.

Example staining results from histological and ICH analysis of mouse brain tissue sections (5 μ). (A) (i–iii), shows mCRP-positive ventricles (i), neurons close to the injection site (ii) and positive neuronal staining around cortical ventricular tracts (iii). (iv) There are numerous positively (peri-nuclear) stained irregular hypertrophic looking cortical neurons. In v, CA1 hippocampal neurons show strong mCRP-positivity whilst in vi, distant staining was observed in neurons of the hypothalamic region. In vii-viii, cortical microvessels are clearly stained for mCRP. ix shows non-transgenic control brain hippocampal neurons negative for mCRP staining and (x) shows cortical neurons also negatively stained for mCRP. (B) P-Tau positivity was increased in sections of cortical neurons. (i) shows an example of negative staining in hippocampal neurons of a control mouse whilst (ii) shows p-tau staining in cortical neurons from a 3xTg mouse and (iii) an equivalent section from an mCRP-injected animal. (iv) shows notable hippocampal (peri-nuclear) staining whilst (v) shows positive axonal cortical neuronal staining (v). Aβ staining is shown in (C) (i) shows negative staining in a normal non-injected mouse, ii and iii show plaque –like element staining and neuronal axons with peri-nuclear staining respectively, whilst iv and v show sections of 3xTg mice demonstrating a similar staining pattern. (D) shows p-IRS-1 and mCRP staining in cortical neurons and plaques (i-ii; serial sections) and matching areas of serial sections showing co-localization in ventricular tracts local to the injection site (iii-iv), CA1 hippocampal neurons (v-vi). vii-viii, shows the presence of p-IRS-1-labelled plaque-like mCRP-positive structures in the same region as mCRP positive areas. (ix) shows non-transgenic mouse cortical region negatively stained for p-IRS-1. (E) shows double immunofluorescent micrographs demonstrating Ca1 neuronal co-localization of mCRP (TRITC) and p-Tau (i-iii; plus insert), and in cortical microvessels (iv; CD31 = FITC). Areas of direct co-localization appeared yellow. Due to antibody binding restrictions, active angiogenic vessels were labelled with CD105 in serial sections and compared with mCRP staining. V-vi show mCRP and CD105 staining in the same vessels of serial sections respectively. No positively stained microvessels were observed in the normal non-transgenic mouse cortex (data not shown). Magnification bars 2.5 mm = × 400).

Figure 7. Western blotting showing mCRP-neuronal protein phosphorylation.

Rat cortical neurons cultured in basal medium showed approximately an 8.5 fold increase in p-Tau expression and a 5-fold increase in p-ERK1/2 by Western blotting after 8 minutes treatment with mCRP (I-ii respectively; 10 μg/ml). In addition, (ii) mCRP induced increased phosphorylation of p-IRS-1 (5 fold), p-Akt (2.5 fold) and p-APP (2 fold). (iii) shows that pre-incubation with our anti-mCRP blocking antibody was able to inhibit mCRP signalling through p-ERK1/2 and p-Tau. These experiments were carried out at least twice and a representative example is shown. (iv) Kinexus phospho-protein Western array carried out on control neurons versus mCRP-treated cells (8 minutes; 10 μg/ml) revealed further proteins that could be involved in neuronal-mCRP signalling including focal adhesion kinase (2.2 fold increase) and p-53 (1.7 fold increase). Magnification bars 2.5 mm = ×400).

Figure 8. Tau fibrilization assay.

In vitro assay showing Tau 244–372 aggregation induced by mCRP (10 μg/ml; 24 h) (c), with a similar profile to that produced by the positive control arachidonic acid (150 μM). (a) shows the control Tau incubated with all other buffer component’s minus CRP or arachidonic acid (b). These experiments were carried out at least twice and a representative example is shown. Magnification bars10 mm = 1 μM.