Brain endothelial cells synthesize neurotoxic thrombin in Alzheimer's disease

Abstract

Alzheimer's disease (AD) is characterized by neuronal death; thus, identifying neurotoxic proteins and their source is central to understanding and treating AD. The multifunctional protease thrombin is neurotoxic and found in AD senile plaques. The objective of this study was to determine whether brain endothelial cells can synthesize thrombin and thus be a source of this neurotoxin in AD brains. Microvessels were isolated from AD patient brains and from age-matched controls. Reverse transcription-PCR demonstrated that thrombin message was highly expressed in microvessels from AD brains but was not detectable in control vessels. Similarly, Western blot analysis of microvessels showed that the thrombin protein was highly expressed in AD- but not control-derived microvessels. In addition, high levels of thrombin were detected in cerebrospinal fluid obtained from AD but not control patients, and sections from AD brains showed reactivity to thrombin antibody in blood vessel walls but not in vessels from controls. Finally, we examined the ability of brain endothelial cells in culture to synthesize thrombin and showed that oxidative stress or cell signaling perturbations led to increased expression of thrombin mRNA in these cells. The results demonstrate, for the first time, that brain endothelial cells can synthesize thrombin, and suggest that novel therapeutics targeting vascular stabilization that prevent or decrease release of thrombin could prove useful in treating this neurodegenerative disease.

Figure 1

Cultured rat brain endothelial cells were exposed to either serum-free media containing 0.1% bovine serum albumin (control), media plus 100 μmol/L H₂O₂, or media plus 1 μmol/L PKC inhibitor bisindolymalemide (BIM) for 24 hours. RNA was extracted, reverse transcribed, and amplified using primers specific for rat thrombin and the housekeeping gene β-actin. The intensity of the bands is graphically shown below RT-PCR bands. The experiment was performed three times and a representative RT-PCR is shown. **P < 0.01 vs. control.

Figure 2

A: Proteins from control- (C) and AD-derived microvessel lysates were separated by SDS-PAGE, transferred to a polyvinylidene difluoride membrane, and immunoblotted for human thrombin (37 kDa). Loading equivalency was confirmed using GAPDH; n = 4. B: RNA from control (C) and AD-derived microvessels was extracted, reverse transcribed, and amplified using primers specific for human thrombin and the housekeeping gene GAPDH; n = 4. C: Proteins from control- (C) and AD-derived CSF were separated by SDS-PAGE, transferred to a polyvinylidene difluoride membrane, and immunoblotted for human thrombin (37 kDa). Loading equivalency was confirmed using albumin; n = 4.

Figure 3

Alzheimer’s disease (A–D) and control (E–H) brain sections were examined by immunofluorescence for the presence of the endothelial cell marker Von Willebrand factor and for thrombin (×20). 4,6-Diamidino-2-phenylindole staining for nuclei was comparable in both AD (C) and control (G) sections. Staining for Von Willebrand factor was detectable in both AD (B) and control (F) tissues. In contrast, reactivity to the thrombin antibody (green immunofluorescence) was clearly present in AD (A) but barely detectable in control (E) vessels. Similarly, yellow/white immunofluorescence, which reflects a merged image of Von Willebrand and thrombin staining, was strong in AD sections (D) but not discernible in controls (H). Scale bar = 15 μm.