The protein alpha-synuclein (ASYN) is thought to be involved in the development of dementia with Lewy bodies (DLB). Overexpression of ASYN has been linked to cellular toxicity and human disease, and in experimental models, chaperones such as heat shock proteins (HSPs) are protective against ASYN toxicity. We have assessed the abundance of mRNA for ASYN and chaperones and the abundance and solubility of the encoded proteins in temporal cortex from sporadic human DLB. We found a reduction of ASYN mRNA in DLB (44.9% of control). The abundance of the Triton-soluble fraction (bioavailable protein) was not altered, but there was an increase of the Triton-insoluble component (likely representing aggregates). We evaluated 3 chaperones: HSP70, HSP90, and HDJ1. HSP70 mRNA was increased in DLB, whereas the mRNAs for HSP90 and HDJ1 were unchanged. HSP70 accumulated in the Triton-soluble fraction, whereas HSP90 and HDJ1 proteins accumulated in the Triton-insoluble fraction. These observations suggest that sporadic DLB is not associated with overexpression of ASYN. Rather, the persistence of normal soluble ASYN protein levels, despite the reduction of its mRNA, suggests a primary defect in clearance of the protein. However, this reduced clearance cannot be attributed to a failure of chaperone expression, because their mRNA is unchanged or increased in the DLB brain.
Dementia with Lewy bodies (DLB) is increasingly recognized as an important cause of dementia and cognitive impairment in the elderly (1-3). Pathologically, DLB is characterized by widespread appearance of Lewy Bodies (LBs) in cortical regions and other areas of the brain, as well as degenerating neurites in the hippocampus and other regions (1-3). Both the clinical spectrum and pathologic features of DLB overlap to some degree with Alzheimer disease (AD). Indeed, some senile plaques are observed in most cases of DLB, and cortical LBs may be found in some cases of pathologically definite AD. Despite the overlap, the disorders do appear to be distinct (4).
Although the cause of DLB remains unknown, recent evidence suggests that mechanism likely involves the protein alpha-synuclein (ASYN). This same protein has also been implicated in the pathobiology of Parkinson disease (PD) as well as multiple system atrophy (MSA), and led to the classification of all 3 disorders as “synucleinopathies” (5-7). Alpha-synuclein is a small, abundant brain protein of uncertain function (2, 8-11). It is a major component of LBs, and it is found in aggregated form in degenerating neurites in DLB (2, 10). Mutations in the coding sequence of ASYN cause autosomal-dominant forms of synucleinopathy (5, 12-23). Recently, it has been shown that duplication or triplication of the normal ASYN gene also causes autosomal-dominant forms of synuclein-related disease, including both parkinsonism and dementia (24-31). These observations imply that overexpression of normal ASYN is sufficient to cause disease and suggest that the more common nonfamilial cases of PD and DLB may also be triggered by alterations in ASYN production. Indeed, recent studies have reported that polymorphisms in the ASYN promoter are a risk factor for sporadic disease (32-34). However, several studies aimed at determining ASYN mRNA expression levels in sporadic synucleopathies have yielded contradictory results. A study by Rockenstein et al suggested that the brain of patients affected by DLB had a significant increase in ASYN mRNA relative to other members of the synuclein family of proteins (35). No difference was observed between mRNA levels of ASYN in MSA patients and control patients as determined with quantitative polymerase chain reaction (PCR) (36), but a study using a ribonuclease protection assay showed that the ASYN message was actually diminished in the substantia nigra but not in the cortex of patients with PD compared with control (37). Similar results were obtained by in situ hybridization by Kingsbury et al (38) and Nichols et al (39). Other studies suggest that overexpression of ASYN is actually protective (40) and that reduced expression levels are associated with synuclein-related diseases (41).
Although overexpression of ASYN is clearly capable of inducing the pathologic features of synucleinopathy, there is also evidence to suggest that the disease state may be caused by defects in ASYN clearance. Alpha-synuclein is a naturally unfolded protein (42, 43), but when present in sufficiently high concentration, it can be induced to fold and aggregate both in vitro and in vivo (44). In DLB and other synucleinopathies, these aggregated forms can be detected by the appearance of detergent-insoluble high-molecular-weight aggregates. These misfolded and aggregated forms are believed to be responsible for the toxic affects of ASYN overexpression (45). Refolding or clearance of misfolded or aggregated proteins requires the actions of protein chaperones such as the heat shock protein (HSP) as well as the proteosome system (44, 46, 47). Recently, it has been shown that systemic impairment of proteosome function can lead to accumulation of ASYN and loss of dopaminergic neurons in the brain, replicating many features of PD (48-51). In contrast, augmentation of the chaperone system can reduce the toxicity of ASYN. In a Drosophila model of synucleinopathy, overexpression of HSP70 protects against the loss of dopaminergic neurons, whereas inhibition of endogenous HSP70 by a dominant negative protein enhances toxicity (52, 53). HSP70 is also protective in cellular models of ASYN toxicity and is present together with ASYN in human LBs (44, 47). The relationship between molecular chaperone systems in human brain and the pathogenesis of DLB is uncertain. It is possible that the pathology of DLB could arise from a failure of the chaperone system without any alteration in the expression of ASYN.
To investigate the relationship among ASYN, chaperone proteins, and the pathologic features of DLB, we have used quantitative methods to assess the ASYN mRNA and the abundance and solubility of ASYN protein in samples of temporal cortex from sporadic cases of human DLB. We have also examined the mRNA and protein levels of several of the chaperones linked to the processing of ASYN in the same cases.
Materials and Methods
All human brain tissues were obtained through the Massachusetts Alzheimer Disease Center (ADRC) brain bank, the Harvard Brain Tissue Resource Center, or the University of Maryland brain bank. The pathologic diagnosis of DLB was established using the criteria of McKeith et al established in 1996 (1). A comparison set of AD brains was diagnosed using the criteria of the Consortium to Establish a Registry for Alzheimer's Disease (CERAD) (54-57) and the Neuropathological Staging of Alzheimer-Related Changes (58). All of the control brains were neuropathologically normal. In addition to these diagnostic classifications, all of the cases were evaluated for the severity of the Alzheimer-type pathology and classified according to Braak and Braak criteria (58) for density and distribution of neurofibrillary tangles. All of the samples consisted of frozen blocks of temporal cortex.
Because of the large number of independent variables to be tested, we used a 2-part design for the studies of mRNA abundance. We first examined the abundance of all the candidate genes of interest in an exploratory group of samples of human postmortem temporal cortex. This group included 8 control cases and 8 DLB cases. We also included in this initial experiment a set of 14 AD cases along with a second independent group of 16 controls matched to the age and postmortem interval (PMI) of the AD cases (n = 16). We evaluated the relative abundance of the mRNAs of 4 heat shock proteins (HDJ-1 HSP90α, HSP90β, and HSP70), ASYN, the chaperone interacting protein CHIP, synaptophysin, and two housekeeping genes (neuron-specific enolase [NSE] and glyceraldehyde-3-phosphate dehydrogenase [GAPDH]). In this part of the study, standard curves for PCR were constructed using dilutions of mRNA isolated from normal cortex. The data obtained were normalized to GAPDH, and the abundance of each transcript was expressed as percent of the mean of the corresponding control group. For each of the transcripts, we evaluated the statistical difference between the DLB cases and their own matched controls and between the AD cases and their matched controls with a 2-tailed Student t-test for unequal variances.
To avoid a type II statistical error, transcripts of interest identified in the exploratory group (HDJ1 HSP90α, HSP70, and ASYN) were reexamined in a second group of DLB and control cases. This was the second independent group of samples with no cases shared with the first group of control and DLB cases. Using the results of the exploratory study, we performed a power analysis and determined that a sample size of at least 10 would provide a statistical power of 0.70 or greater with alpha of 0.05 for each of the variables. Thirteen control cases and 10 DLB cases, matched for age and PMI, were selected to be part of this second experimental group. In this set of experiments, we generated standards for each assay by subcloning PCR products containing the target genes. This allowed us to express the results as number of copies of each transcript per nanogram of input RNA. The statistical difference between controls and DLB for each target gene was calculated with a one-tailed Student t-test for unequal variances. All the cases used in both the exploratory study and in the confirmatory study were subsequently used for the protein analysis.
RNA Extraction and First-Strand Synthesis
Total RNA was extracted from the human brains using Trizol (Sigma-Aldrich, St. Louis, MO) according to the manufacturer's specifications. Briefly, approximately 100 mg of brain tissue was homogenized in 1,000 μL of Tri-reagent and 200 μL of chloroform was added to the samples. After a brief centrifugation to separate the aqueous phase, 500 μL of isopropanol was added and the mixture was placed on dry ice for 1 hour. The samples were then centrifuged at 4°C for 40 minutes at 21,000 × g. This precipitated total RNA into a pellet that was subsequently washed twice with 75% ethanol. After the second wash, the pellet was dried and solubilized in RNAse and DNAse-free water (Eppendorf, Westbury, NY). Five to 8 μg of total RNA was treated with 5 U RNase-free DNase I (RQ1; Promega, Madison, WI) for 15 minutes at 37°C. After the DNAse treatment, the RNA was reextracted with phenol/chloroform-isoamyl alcohol and quantified using a spectrophotometer (Eppendorf). The quality of the RNA sample was assessed using an Agilent 2100 Bioanalyzer (Agilent Technologies, Palo Alto, CA). All cases used for quantitative PCR had RNAs that yielded 2 sharp ribosomal 18S and 28S bands.
First-strand cDNA synthesis was carried out on 4 μg of the total RNA from each sample with an Invitrogen (Carlsbad, CA) Superscript first-strand synthesis kit for reverse transcriptase (RT)-PCR according to the manufacturer's specifications. All the samples were equilibrated so that the final concentration of cDNA yielded by the first strand synthesis reaction was 30 ng/μL. This stock solution of cDNA for each of the cases used in this study was further diluted into 3 ng/μL and 0.3 ng/μL working solution and preserved at -20°C until the next step.
Quantitative Polymerase Chain Reaction
Quantitative PCR (QPCR) reactions were carried out in a 96-well plate using an iCycler (BioRad, Hercules, CA) and SYBR Green PCR Master Mix (Applied Biosystems, Foster City, CA). Short, synthetic (18-22 mer) PCR primers were designed using the tools available through the MIT Whitehead Institute web page (http:www-genome.wi.mit.edu/cgi-bin/primer/primer3.cgi). Primer sets were selected to amplify small (100-300 bp) amplicons for each candidate mRNA and to have high PCR efficiencies (preferably 80-95%). The specificity of each candidate PCR amplicon was evaluated by melting curve (Tm) analysis, ethidium bromide gel electrophoresis, and direct sequencing. We carried out a concentration curve using known concentrations of reference cDNA or cDNA clones for each primer set used. This curve was used to both calculate the primer set efficiency for each particular experiment and to quantitate products. Each transcript of interest was quantified using a modification of the ΔCt method (59) that interpolates the Ct back into the standard curve with an algorithm embedded in the software provided with the iCycler. All the primer sets used in this study are illustrated in Table 1, along with their respective optimal annealing temperatures, average efficiencies, and optimal final concentration in the reaction mix.
For protein studies, brain tissue was mechanically homogenized in 10 vol (w/v) of cold lysis buffer (Tris/HCl 50 mM pH 7.4, NaCl 175 mM, EDTA 5 mM pH 8.0, protease inhibitor cocktail [Roche, Basel, CH]) and sonicated for 10 seconds to generate total cell lysates. These lysates were divided into Triton X-100 soluble and insoluble fractions by adding Triton X-100 to total cell lysates (final concentration 1%) and incubating for 30 minutes on ice followed by centrifugation (15,000 × g, 60 minutes, 4°C). The supernatant was designated as Triton X-100 soluble fraction. The pellet was dissolved in lysis buffer containing 2% SDS and sonicated for 10 seconds. This was designated the Triton X-100 insoluble fraction. The protein concentration in each fraction was determined using a Lowry protein assay (BioRad) according to the manufacturer's specifications. Twenty to 40 μg of each cell lysate was loaded onto 10% to 20% Tris-Glycine gels (Novex, San Diego, CA) for Western blot analysis. Protein was transferred to Immobilon-P membrane (Millipore, Bedford, MA) and blocked in blocking buffer (Licor, Lincoln, NE) for 1 hour before the addition of the primary antibody (anti-Hsp70/SPA830, anti-Hsp90/SPA812, or anti-HDJ-1/SPA400, Stressgen, Victoria, BC, Canada; anti-ASYN/Syn-1, BD Transduction Lab, San Jose, CA) at room temperature for 1 to 2 hours or overnight at 4°C. The blots were washed 3 times in Tris-buffered saline with 0.2% Tween (TBS-T, pH 7.4) and incubated at room temperature for 1 hour in fluorescent-labeled secondary antibodies (IRDye 800 anti-rabbit or anti-mouse, Rockland Immunochemicals, Gilbertsville, PA; 1:3000 or Alexa-680 anti-rabbit or anti-mouse, Molecular Probes, Eugene, OR, 1:3000). After three times washing in TBS-T, the immunoblots were processed and quantified using the Odyssey infrared imaging system (Lycor). Electrophoresis gels were stained for total protein with Coomassie blue and quantified as loading controls using the Odyssey infrared imaging system.
Characteristics of the Subjects
Each of the 2 independent groups of DLB and control specimens selected for the RNA studies were closely matched with respect to age at the time of death and postmortem interval (Table 2). An additional analysis of the extent of AD-type pathology in these cases revealed that in the first DLB group, 25% of the cases had no detectable AD pathology, 25% of the cases were classified as entorhinal (Braak stage I-II), 25% of the cases were classified as limbic (III-IV), and 25% of the cases were classified as neocortical (V-VI). In the second DLB group, 40% of the cases had an entorhinal distribution of AD pathology (I-II), 40% of the cases a limbic distribution, and 20% a neocortical distribution of NFT. A chi-squared test of the frequencies of each Braak stage did not reveal any differences between the 2 groups.
Exploratory RNA Study
The exploratory study was conducted to evaluate the relative abundance of the mRNAs for our 10 target genes in the human postmortem temporal cortex. Each of the targets was assayed using the primers described in Table 1, and the abundance in each sample was normalized to the abundance of GAPDH in the same sample. This study revealed that 3 of the 10 targets examined were expressed at significantly lower levels in DLB cases: both of the isoforms of HSP90 (HSP90α and HSP90β) and ASYN (Table 3). No significant difference in the expression of HDJ1 or CHIP was detected. Markers of neuronal (NSE) and synaptic (synaptophysin) structure were modestly reduced (30% and 35%, respectively), but neither of these changes reached statistical significance. The only mRNA that was substantially increased was HSP70, which exhibited a fairly large increase in mean abundance (173% of control), but this trend did not reach statistical significance (p = 0.10).
A different pattern was observed in the comparison sample of AD cases. In this group, HSP90α and HSP90β were unchanged with respect to the controls, whereas ASYN was significantly decreased in AD samples as well. HSP70 was substantially increased in the AD samples (211% of control, p < 0.003). In the AD samples, both NSE and synaptophysin were consistently and significantly reduced in abundance, perhaps reflecting the greater neuronal loss observed in the temporal cortex of the Alzheimer brain as compared with the DLB brain.
Confirmatory RNA Study
Based on the results of the exploratory study, 4 transcripts were selected and studied in a second independent cohort of DLB and control cases. In this set of experiments, we used DNA standards generated by PCR, which allowed us to calculate absolute concentrations of each transcript (as copy number per nanogram of input RNA). The values for each of the transcripts were normalized to the abundance of GADPH in each sample, which was similar for the 2 groups (control 4208 ± 754.7 copies/ng RNA; DLB 3743 ± 800.3; p = 0.430).
Examination of this confirmatory cohort revealed that as in the first group, the abundance of ASYN was markedly reduced in the DLB cases (Table 4). In this set of samples, the increase in HSP70 mRNA levels was even more striking (294% in control), and the difference between DLB and control group was statistically significant (p < 0.024). The only result of the first study that failed to replicate was the difference in the abundance of HSP90α; in the confirmatory cohort, the abundance of this transcript did not differ between the DLB and control groups. The expression of HDJ1 did not differ between the control and DLB cases. Regression analysis did not show a significant relationship between ASYN mRNA levels and the Alzheimer Braak stage or between HSP70 mRNA levels and the Alzheimer Braak stage, indicating that the changes seen are not dependent on the level of Alzheimer-type pathology present in the DLB brains analyzed.
The abundance of proteins was studied separately in the total brain lysates and in the 1% Triton X-100-soluble and -insoluble fractions of the same lysates. The Triton-soluble fraction is thought to represent bioavailable protein, whereas the Triton-insoluble component may represent proteins that are aggregated or sequestered (9). In the total cell lysates, we observed an increase in the abundance of the Hsp90 protein in DLB brains as compared with controls (p < 0.05), whereas Hsp70, HDJ1, and ASYN were not different between the 2 groups of samples (Fig. 1, top panel).
Separation of the crude homogenates into Triton X-100-soluble and -insoluble fractions revealed substantial differences between the DLB and control groups (Fig. 1, bottom panel, and Fig. 2). The most striking alterations were in the detergent-insoluble fraction. We have previously reported that in DLB, there is increased ASYN in the detergent-insoluble fraction (44) and observed a similar enhancement in this study. The mean abundance of insoluble ASYN was increased more than 3-fold, although there was substantial variability, with some cases exhibiting marked elevation whereas others overlapped with the normal range (Fig. 2). We attempted to correlate the magnitude of change in insoluble ASYN with variables, including age, postmortem interval, AD Braak staging (58), or PD Braak staging (60), but no relationship was apparent. The increase in insoluble ASYN was accompanied by a marked increase in the abundance of insoluble HSP90 protein (more than 5-fold), as well as a significant increase in insoluble HDJ1 (Fig. 2). HSP70 was not altered in the insoluble fraction, although it is known to be present together with HSP90 in LBs (47, 52, 61, 62). Within the Triton X-100-soluble fraction, we found a modest reduction in HSP90 and HDJ1 and an increase in HSP70. Triton-soluble ASYN was not altered, consistent with our previous studies (44).
In this study, we have examined the expression of ASYN in normal and DLB postmortem temporal cortex to evaluate the hypothesis that elevated expression of ASYN contributes to ASYN aggregation and toxicity in this disease. Contrary to the result predicted by this hypothesis, we found that there was a striking reduction in the mRNA for ASYN in DLB. Despite this reduced expression of ASYN mRNA, the overall abundance of the protein and the abundance in the Triton-soluble fraction were not altered, and in many cases, there was an increase of ASYN protein in the Triton-insoluble component. We also evaluated 3 protein chaperones (HSP70, HSP90, and HDJ1) thought to interact with ASYN to test the hypothesis that chaperone downregulation could contribute to the accumulation of insoluble forms of ASYN. We did not find evidence for downregulation of the mRNA for any of these chaperones, and there was a consistent enhancement of the mRNA for HSP70 in DLB (for a summary of mRNA results, see Fig. 3). Similarly, the total protein abundance of these chaperones was not reduced, and in the case of HSP90, the total protein abundance was somewhat increased. There was an alteration in the solubility of the chaperones with an accumulation of HSP90 and HDJ1 in the Triton-insoluble component. In light of these results, we cannot attribute the pathology of DLB to overexpression of ASYN. Interestingly, all the DLB cases show either normal or elevated insoluble ASYN and normal levels of total and soluble ASYN comparable to the normal cases, despite a consistent marked decrease in the levels of alpha-synuclein mRNA. If mRNA levels are a reflection of the rate of production of the protein, these results suggest instead that the primary mechanism may be impaired clearance of ASYN so that the protein abundance is maintained in the presence of a large reduction in ASYN mRNA. Such a reduction in clearance might result from an impairment of the function of specific degradative mechanisms or, alternatively, from the difficulty of clearing proteins that have already aggregated. However, this impaired clearance does not appear to be the result of a failure of expression of chaperone proteins, at least of those that we have directly examined.
Several previous studies aimed at determining ASYN mRNA expression levels in sporadic synucleopathies have yielded contradictory results. A study by Rockenstein et al comparing ASYN mRNA levels in postmortem temporal cortex of DLB and AD and other diseases reported a significant increase in ASYN mRNA in DLB and no change in AD when compared with control patients (35), whereas we found the levels of ASYN mRNA to be decreased in both DLB and AD postmortem temporal cortex. An important difference between the 2 studies is that Rockenstein et al used a semiquantitative method (ribonuclease protection assay [RPA]) to compare the relative levels of ASYN to those of beta and gamma-synuclein, and the alteration reported is an increase in the proportion of ASYN as a percentage of the sum of alpha, beta, and gamma synuclein mRNA signal (35). In contrast, we used a quantitative PCR method with synthetic standards that allowed us to directly calculate concentrations of the target mRNAs. Thus, our result reflects a true alteration in ASYN mRNA, independent of any changes that may occur in the abundance of other members of the synuclein family. Interestingly, another study using RPA but using a direct measure of the signal for ASYN did indeed show a reduction in ASYN mRNA in the substantia nigra in PD (37), but a limitation of this study is that it may reflect the extensive cell loss that occurs in PD rather than a change in expression. Because cell loss is not prominent in DLB, it is unlikely to contribute to the results we obtained. Our findings are also in agreement with the in situ hybridization studies of Kingsbury et al (38) and Nichols et al (39).
Another strength of our study is that to examine the abundance of mRNAs, we used a study design in which we evaluated 2 independent cohorts of control and DLB samples. This approach was necessary because of the large number of different variables we wanted to examine and the substantial intersubject variance in mRNA to be expected in a study of human postmortem brain. We used careful analysis of the pathologic features of the cases as well as matching of the ages and extent of concurrent AD-type pathology. We obtained clear replication of the results for ASYN showing a marked reduction and HDJ1 showing no change in each of the 2 datasets. HSP70 increased in both datasets, although this did not quite reach statistical significance in the first set. One of the observations in the exploratory dataset, an observed reduction in HSP90 mRNA in DLB, failed to replicate. We therefore regard the initial result as uncertain and conclude that there is no convincing evidence for a change in HSP90 mRNA in our study. We believe these observations demonstrate the value of independent replication of data in human studies in which large variance around the mean is expected.
Despite our finding that ASYN mRNA is markedly reduced in DLB, we did not find a reduction in the abundance of the encoded protein. Indeed, although the concentration of soluble ASYN was normal, the abundance of Triton-insoluble forms was markedly enhanced. Similar results have been reported by others (63, 64) and presumably reflect the aggregation of ASYN together with other proteins in LBs and other inclusions as well as within Lewy neurites. It is certainly true that measures of mRNA abundance are not a direct measure of the rate of production of the encoded proteins, and there may be divergence between mRNA and protein levels. In addition, it is important to note that we have analyzed postmortem specimens that represent the end stage of the disease, and these results may not be reflective of the state earlier in the disease process. Also, some mRNA species may be more unstable than others in the postmortem state. This should not alter the validity of the differences we have detected, but might obscure additional differences in species that are very labile after death. Nevertheless, the striking disparity between the reduction in mRNA and increase in insoluble ASYN that we observed leads us to hypothesize that the primary defect in sporadic cases of DLB is not overproduction of ASYN. Instead, we propose that the primary defect may be a disorder of clearance, leading to accumulation of insoluble forms. Indeed, the reduced mRNA abundance may be a normal homeostatic response to a reduced rate of protein clearance.
We did examine in detail one component of the clearance process for ASYN, the protein chaperones HSP70, HSP90, and HDJ1. There is considerable experimental evidence to suggest that these chaperones play an important role in the clearance of ASYN. In particular, these chaperones as well as the related protein torsinA can suppress toxicity of ASYN in a cellular model (47, 65), and HSP70 markedly reduces the formation of high-molecular-weight aggregates of ASYN in a transgenic mouse model (44, 66). HSP70 is also protective in a Drosophila model of ASYN toxicity (52). In our studies, we failed to find any convincing evidence that failure of chaperone expression could account for the development of sporadic DLB. In fact, HSP90 and HDJ1 are unchanged in the DLB brain, and HSP70 is expressed at higher levels in the DLB brain as well as AD, which may represent a response to disease-induced stress.
There are several limitations to our study of chaperones in DLB. The absence of reductions in mRNA or protein abundance does not rule out abnormal chaperone function. In particular, it is possible that alterations in the stoichiometry of the different HSPs could affect their function. For example, the interactions among HSP70, HSP90, the cochaperone HOP, and their substrates is mediated by the ATP-dependent formation of heterocomplexes with a specific stoichiometry of 2 HSP90:1 HOP:1 HSP70 (67). If HSP90 is decreased in the pool of proteins that are available for biologic processes (Triton X-100-soluble fraction) and HSP70 is increased, the specific stoichiometry between these 2 proteins may result in chaperone dysfunction. In addition, our survey of chaperones was necessarily not comprehensive. In addition to HSP90, HSP70, and HDJ-1, which we analyzed, there are a number of other potential protein chaperones such as torsinA, which is known to accumulate in LBs (68). Finally, this study does not address the issue of cellular localization of the chaperone proteins analyzed and of the changes observed. It is possible that the changes observed are neuronal in nature or that they occur in glial cells and reflect the effect of gliosis.
It would be interesting to analyze the behavior of alpha-synuclein and molecular chaperone in the brain of cases with alpha-synuclein locus duplication and triplication. There have been few autopsy studies in these pedigrees, and only one study has addressed the issue of alpha-synuclein levels in the brain of carries of these genetic alterations. Miller et al analyzed alpha-synuclein mRNA and total protein levels in 2 cases with alpha-synuclein locus triplication and showed that the patients affected by the triplication had increased mRNA levels and increased high-molecular-weight and insoluble alpha-synuclein (39). The elevated mRNA in these cases is expected because of the underlying genetic abnormality but differs from the reduction in ASYN mRNA we have observed in sporadic DLB. This difference supports our hypothesis that the mechanism of disease in sporadic DLB is different from that in the duplication and triplication-related forms and is not related to overproduction of ASYN. The sporadic and triplication forms of the disease do, however, seem to share similar accumulation of insoluble ASYN protein.
In summary, although it is clear that ASYN-related disease can be triggered by overexpression of ASYN, our data suggest that this is not likely to be the mechanism responsible for most sporadic cases of DLB. Rather, our data favor a model in which ASYN is transcriptionally downregulated and the clearance of ASYN is reduced. We believe that it is of interest that the abundance of some of the chaperones is increased in the insoluble fractions, perhaps pointing to a primary defect in the proteosome itself. It is also important to note that even if the function of the chaperones is not defective in DLB, it is still possible that therapies that augment chaperones might be beneficial in the disease.
The authors thank Christine E. Keller-McGandy, Zane R. Hollingsworth, and Dr. Charles R. Vanderburg for their invaluable technical help.