Clusterin in brain and cerebrospinal fluidin Alzheimer s disease

Sammanfattning: The apolipoprotein clusterin is suggested to be involved in the pathogenesis of Alzheimer¡¦s disease (AD). At the start of these studies, results had indicated that clusterin was elevated in AD brain, compared with controls. Since the brain extracellular space is in direct contact with the cerebrospinal fluid (CSF), it may reflect brain processes, and thus clusterin was an interesting target to examine in both brain tissue and CSF. We studied brain tissue from patients with AD, vascular dementia (VaD) and controls. Using quantitative gel electrophoresis and Western blot the clusterin level was found to be increased in AD related brain regions (hippocampus, frontal and parietal cortex), but unchanged in the cerebellum, a region less affected by AD pathology. In VaD, there were no statistically significant changes in any brain regions. Using immunohistochemistry, there was intense staining for clusterin in both neuronal and astroglial cells, while only a portion of the senile plaques (SP) showed positive staining in AD. Thus, since we found no correlation between clusterin and the number of SP, we suggested that there is no link between clusterin and early beta-amyloid (Abeta) deposition and/or formation of SP in AD brain. Clusterin was then quantified by ELISA in CSF from patients suffering from AD, VaD, and Parkinson¡¦s disease, and compared to controls. No significant change was found between AD or VaD and controls. However, in Parkinson's disease patients the clusterin level was significantly lower than in controls. Clusterin was also quantified in a CSF material from acute stroke patients, where it was compared with levels of glial fibrillary acidic protein (GFAp). Despite an elevation of GFAp from 1-2 days until 1 week after stroke, indicating reactive astrocytes, the clusterin level was unchanged. Clusterin glycosylation in CSF was also analysed, since we speculated wether the glycosylation could inhibit the quantification of clusterin. Using enzymatic deglycosylation, we found the carbohydrates on clusterin in CSF to be of N-linked type and sialic acids. No O-linked carbohydrates were found. Using a sandwich ELISA, clusterin was then re-quantified in CSF in both its native and deglycosylated forms, first in a pilot study, then in a larger material of AD and control cases. Deglycosylation increased detection of clusterin, thus supporting our theory of steric hindrance in the antibody binding, and showing that quantification of post-translationally modified proteins may be affected by methodology. The larger CSF material showed clusterin to be significantly increased in both native and deglycosylated samples in AD. However, since there is an overlap of clusterin levels between AD and controls, clusterin is not valid as a diagnostic marker for AD. However, here we suggest that comparing the groups of native control samples and deglycosylated AD samples, may discriminate ~83% of AD cases from controls. To summarise, we show that clusterin is affected by AD pathology in the brain, and has an elevated level in the CSF of AD patients. However, so far, CSF clusterin is not valid as a diagnostic marker for AD, due to overlap of the mean levels between AD and controls.