Introduction

Virtually all age-associated neurodegenerative diseases are linked to the abnormal conformation and assembly of proteins in the nervous system [46, 48]. The hallmark of these disorders is the transformation of a normally soluble peptide or protein into insoluble fibrils that are deposited in the brain tissue [6, 9, 25, 26]. The cerebral proteopathies include a range of devastating neurological disorders such as Alzheimer’s disease (AD), Parkinson’s disease (PD), Huntington’s disease and Creutzfeldt–Jakob disease (CJD). Age is a factor in the emergence of all of these maladies, although the means by which age influences their expression remains unresolved. While AD is characterized histopathologically by large numbers of senile plaques and neurofibrillary tangles, such lesions also occur, in much smaller numbers, in many non-demented elderly humans [49]. Senile plaques and cerebral β-amyloid angiopathy (CAA) consist primarily of deposits of the protein fragment Aβ, which most often is 40 (Aβ40) or 42 (Aβ42) amino acids in length. In aging humans, parenchymal Aβ42 deposits usually precede the appearance of dense, Aβ40-rich aggregates [47].

Creutzfeldt–Jakob disease belongs to the transmissible spongiform encephalopathies (TSE), or prion diseases, which are characterized by the accumulation of an abnormal isoform of the prion protein (PrPSc). Histopathologically, CJD is diagnosed by the classical triad of spongiform change, neuronal loss, and gliosis, as well as the presence of aggregated PrPSc [5, 12, 14, 43]. In the brains of older CJD patients, Alzheimer-like β-amyloid (Aβ) deposits sometimes can be observed [5, 21], but the presence of such lesions appears to be a non-specific effect of age. However, in these cases, PrP often decorates the periphery of Aβ-plaques, suggesting that Aβ may influence the deposition of PrP [21].

Despite many differences between CJD and AD, both diseases share important pathogenetic similarities that could lead to similar therapeutic strategies [2]. Three broad therapeutic approaches to abrogating the proteopathic cascade are suggested by the commonalities among these disorders: (1) reduce the production of the aggregation-prone proteins, (2) prevent their self-assembly and toxicity, or (3) promote their degradation and removal [46, 48].

In recent years, increasing attention has been devoted to the myriad functions of a class of transport proteins—the so-called ATP-binding cassette transport proteins (ABC transporters)—in the central nervous system. In this regard, P-glycoprotein (P-gp, ABCB1, MDR1) appears to be particularly important. P-gp acts as an efflux pump for several endogenous and exogenous substances and thus plays a critical role in the protection of tissues with excretory and/or barrier function, such as intestine, kidney, and liver. P-gp also is active at the blood-brain barrier (BBB) and blood-cerebrospinal fluid (CSF) barrier [17, 18, 37]. In normal brain, P-gp is expressed by the endothelial cells of blood vessels, especially capillaries [1, 18], and by epithelial cells of the choroid plexus [15].

P-glycoprotein has been shown to actively transport Aβ in vitro [28] and in vivo [11], and deficits in P-gp may be involved in the pathogenesis of Alzheimer’s disease and CAA [44, 45]. Importantly, cerebrovascular P-gp expression decreases with age [45], and could thereby impair the ability of the brain to expel excess proteins. We hypothesize that the resulting accumulation of proteins in cells overwhelms cellular ubiquitin-proteasomal degradation, and thus facilitates the emergence of proteopathic disorders of the nervous system. In the present study, we investigated the relationship between P-gp and prion disease, and found that CJD, like AD and CAA, is associated with a decrease in the expression of cerebrovascular P-gp compared to age-matched control cases. Since P-gp expression can be induced by several drugs, selective enhancement of brain P-gp function might be employed to treat or prevent the abnormal accrual of proteins that characterizes many neurodegenerative disorders.

Materials and methods

Tissue samples of occipital and temporal lobe were obtained at autopsy from ten subjects with histologically, biochemically, and genetically confirmed CJD who died between the ages of 59 and 82 years (mean age 70.3 years), as well as ten age-matched control cases (mean age 70.2 years). Sex and age distribution are shown in Table 1. Brain tissue from all cases was fixed similarly in 4% buffered formalin, inactivated 1 h with 98% formic acid and embedded in paraffin according to the standard guidelines for processing CJD tissues. Sections were subjected to conventional staining and to immunostaining for PrP (3F4) upon hydrolytic autoclaving. Genetic analysis and determination of glycotype were performed as described previously [20].

Table 1 Age, sex distribution, and diagnosis of cases analyzed (M Male, F Female, CJD Creutzfeldt–Jakob disease)
Full size table

Immunohistochemistry

For immunohistochemistry, 6 μm-thick tissue sections were cut, mounted on slides and dried overnight at 60°C. Sections were deparaffinized in xylene, rehydrated in a graded series of ethanol and water, and then treated with 10 mM citrate buffer at pH 6.0. Endogenous peroxidase was blocked with 0.3% H2O2. For staining, the biotin-streptavidin immunoperoxidase method with the LSAB®HRP detection system (DAKO, Hamburg, Germany) was used. P-gp was immunolabeled with monoclonal antibody JSB-1 (Alexis, Grünberg, Germany). Aβ was labeled using end-specific monoclonal antibodies to Aβ40 or Aβ42 (Chemicon, Planegg-Muenchen, Germany).

For quantification of P-gp, five consecutive digital pictures were taken of neocortex, white matter and leptomeninges (magnification ×200) from identical regions of each case; for quantification of Aβ40 and Aβ42, four consecutive digital pictures of cortex (magnification ×100) and five consecutive digital pictures of leptomeninges (magnification ×200) were taken (3CCD color camera, Hitachi HV-C20 M, Hitachi Denshi Ltd., Japan, and Axioskop, Zeiss, Jena, Germany). Subsequently, the optical density and the area occupied by immunopositive structures were determined in each photograph using KSRun software (KSRun Version 3.0, Zeiss), revealing a final middle optical density (mod) value for each case.

In selected cases, immunohistochemical double-labeling with antibodies to Aβ (DAKO, clone 6F/3D, dilution 1:50) and PrP (DAKO, clone 3F4, dilution 1:25) was performed. First, Aβ was stained using the EnVision Dual Link System (DAKO) and diaminobenzidine (DAB) as chromogen, yielding a brown color for Aβ. Subsequently, PrP immunostaining was performed with the Chem Mate Detection Kit, Alkaline Phosphatase/RED (DAKO) and fast red as chromogen, staining the PrP deposits red.

Statistical analysis

Statistical analysis of the data was performed using the non-parametric Mann–Whitney U-Test with a two-tailed significance threshold (P<0.05).

Results

The data for CJD cases regarding the pattern of PrP-immunostaining, PrPSc-type and the methionine/valine polymorphism at codon 129 of the PRNP gene are shown in Table 2. The PrP immunostaining pattern was described as punctate, patchy or plaque-like, respectively. Vascular deposition of PrP was noted in only one case (below). PrPSc-subtypes were divided into two groups (I or II) based on variation at codon 129 of the PrP gene [33]. The cohort included a representative selection of geno- and PrPSc-types in order to minimize potential bias in the analysis.

Table 2 Distribution and immunohistochemical pattern of PrP deposition (Punc Punctate, Pa Patchy, Pl Plaque-like), PrPSc-type, methionine/valine polymorphism at codon 129 of the PRNP gene, and Aβ deposition in ten Creutzfeldt–Jakob cases
Full size table

P-glycoprotein was expressed strongly in endothelial cells of cerebral blood vessels, especially capillaries. Quantitatively, CJD cases had significantly lower P-gp labeling within the cortex (P=0.005), leptomeninges (P=0.003) and white matter (P=0.037) when compared to controls (Fig. 1).

Fig. 1 
figure 1

Quantitative immunohistochemical expression of P-gp. CJD cases show significantly lower P-gp expression than controls within the cortex, leptomeninges, (*P<0.01) and white matter (**P<0.05). Mod Middle optical density

Full size image

Five out of ten CJD cases had age-associated Aβ42-immunoreactive plaques within the cortex, as well as Aβ42 deposits in the walls of leptomeningeal blood vessels. Two cases also had Aβ42-CAA within the brain parenchyma. Four of the five CJD cases with Aβ42 lesions also showed Aβ40-positive plaques within the cortex, including two cases with intracerebral Aβ40-CAA involvement as well as three cases with leptomeningeal Aβ40-CAA. Regarding the extent of the Aβ burden there were three cases with isolated Aβ plaques only, as is often the case in the non-demented elderly. In contrast, the two cases with intracerebral CAA also showed numerous Aβ plaques, comparable to the lesion density in AD brains. No amyloid deposits were observed in control cases (except one case with a single diffuse plaque).

There was no colocalization of P-gp expression and β-amyloid deposition within the same vessel, i.e., vessels with Aβ deposition in their walls showed no endothelial P-gp labeling, and vice versa (Fig. 2). Generally, prominent vascular Aβ deposition was observed in arterioles and small arteries. There was only one case with Aβ deposits within capillaries, and the absence of P-gp expression was evident in these afflicted vessels as well. Interestingly, this case with capillary β-amyloid angiopathy showed the lowest P-gp expression of all cases analyzed.

Fig. 2 
figure 2

Immunohistochemical expression of β-amyloid in a case of CJD (a) and a consecutive section with P-gp immunostaining (b). Note Aβ deposition as plaque and amyloid angiopathy, and the negative expression of P-gp. A control case with strongly detectable P-gp is shown in (c). Bar=50 μm

Full size image

Because five of the CJD cases also had cerebral Aβ deposition, and because P-gp expression is significantly reduced in blood vessels that manifest CAA, we asked whether the deficits in P-gp expression in CJD cases are influenced by the simultaneous presence of Aβ40 or Aβ42. Comparing CJD cases with or without Aβ40/42 deposition, respectively, there were no differences in P-gp expression (Table 3).

Table 3 Mean middle optic density (mod) value of P-gp expression in Aβ-positive and negative CJD cases
Full size table

Double immunolabeling often detected colocalization of Aβ and PrP within the same lesions, but the distribution of the two proteins differed. Some senile plaques had an Aβ core and an accumulated rim of PrP (Fig. 3a), whereas some dense PrP plaques were rimmed by Aβ (Fig. 3b). We also observed protein layering in certain plaques, such that a layer of PrP was interposed between an inner core and outer shell of Aβ (Fig. 3c). Interestingly, although PrP was generally absent in vessels, one case with severe leptomenigeal CAA showed occasional small arteries with a colocalisation of Aβ and PrP within the vessel wall (Fig. 3d).

Fig. 3 
figure 3

Immunohistochemical double-labeling of Aβ (brown) and PrP (red) in Creutzfeldt–Jakob cases. a Plaque with an Aβ core and a PrP shell. b Plaque with a PrP core and an Aβ shell. c Plaque with central Aβ deposition, a layer of PrP and an outer accumulation of Aβ. d Vessel wall with colocalisation of Aβ and PrP (arrow). Bar=50 μm

Full size image

Discussion

It has been suggested that decreased clearance of peptides (especially Aβ) from the brain at the BBB could contribute to the buildup of proteins, leading to AD and possibly to other neurodegenerative disorders such as CJD [13, 39, 48]. In previous studies we found an inverse relationship between the expression of cerebrovascular P-gp and the deposition of Aβ in the brains of elderly humans, suggesting that a diminution of P-gp function might be involved in the early pathogenesis of AD [44, 45]. Furthermore, P-gp expression declines in the cerebral blood vessels of older humans even in the absence of CAA [45], and could thereby increase the probability of aberant protein accumulation with age. Since there is mounting evidence for pathogenic commonalities in the accumulation of Aβ and PrPSc in diseased tissues [2, 3], we hypothesized that the expression of P-gp might be diminished in cases with idiopathic CJD.

Our present results show that P-gp expression within endothelial cells of brain vessels is significantly lower in cases with CJD than in age-matched control cases. Unlike Aβ, the accumulation of PrPSc in the cerebral vasculature is quite rare [except in a Gerstmann–Straeussler–Scheinker syndrome (GSS) pedigree with a premature stop codon in the PrP gene] [36]. Accordingly, there was almost no immunohistochemically detectable PrP in the blood vessels of the idiopathic CJD cases that we examined. The paucity of demonstrable vascular PrP raises the question of whether down-regulation of the P-gp transporter in the endothelium is the cause or the consequence of the prion disease process. Half of the CJD cases that we analyzed also displayed parenchymal and/or vascular Aβ deposits, as expected in a cohort of older subjects [49]. As in our previous study [45], the presence of Aβ within a particular vessel wall was associated with a deficiency of P-gp within the affected segment. Although there was no statistically significant difference in overall P-gp expression in CJD cases with or without cerebral β-amyloidosis, one CJD case manifesting Aβ-CAA in neocortical capillaries had the lowest P-gp expression of all cases analyzed. Taken together, these findings support the view that CJD is associated with a reduction of P-gp expression independent of the co-existence of Aβ deposits, but further studies are required to determine whether the co-presence of PrP and Aβ is linked to an even greater diminution of transporter expression and function.

There are several potential means by which P-gp function might be related to prion disease. One possibility is that P-gp could be down-regulated in the brain as a result of the disease process, i.e., that P-gp expression loss is a consequence rather than a cause of PrP accumulation. Longitudinal studies of P-gp expression in PrP-transgenic mice, which do not deposit Aβ with age, would help to establish whether P-gp function declines before, or after, the onset of disease phenotype.

Another possibility is that the loss of P-gp is a causative factor in the accumulation of PrP in brain. The age-associated loss of P-gp function [45] suggests one means whereby age might influence the risk of CJD. To date, there is no evidence that PrP itself is a substrate for P-gp, but there are two general, alternative pathways through which P-gp reduction might stimulate the accumulation and subsequent toxicity of PrPSc.

First, the loss of P-gp function could facilitate the buildup of Aβ, which then acts as a heterologous seed for the aggregation of PrPSc. Aβ deposition frequently is found in TSE, including CJD [8], although at present it is uncertain whether Aβ accumulation is pathogenically linked to prion disease, or whether it is an independent (e.g., age-related) process. Interestingly, the levels of Aβ42 are decreased in the CSF of patients with CJD, similar to the reduction seen in patients with AD [32, 50]. Other studies have reported a colocalization of Aβ and PrP in the same plaques in CJD cases, suggesting that preexisting β-amyloid might augment PrP accumulation by promoting the aggregation of one amyloidogenic protein (PrPSc) onto a core composed of the other (Aβ) [21, 27]. Our results confirm that PrP can decorate the periphery of Aβ plaques. We also observed the converse, i.e., that Aβ can deposit on PrP plaque cores. Occasionally, there were even multilayered plaques, with central Aβ deposition, a rim of PrP and an outer accumulation of Aβ. Furthermore, there were rare blood vessels with mural colocalisation of Aβ and PrP. These results suggest that both proteins interact complementarily, and that each may act as a seed for the accumulation of the other. This mechanism could be evaluated in dual PrP-βAPP transgenic mice.

A second possibility is that the accumulation of aberrant proteins (including Aβ) due to diminished P-gp activity could overwhelm the protein-degrading enzymes of the ubiquitin-proteasomal system (UPS) and thereby facilitate the accumulation and conversion of PrP. The UPS plays a fundamental part in many basic cellular processes by degrading a wide range of specific cellular proteins, including mutated and misfolded proteins [10]. Dysfunction of the UPS contributes to the accumulation of proteins in neurodegenerative diseases, such as α-synuclein in PD [34], or Aβ [29] and tau [31] in AD. PrP also is degraded via the UPS [35, 52]. Inhibition of the UPS, e.g., due to aging or drug treatment, causes an accumulation of normal PrPC in the cytoplasm, where it can be spontaneously converted into a PrPSc-like species because it is not promptly degraded by the UPS [30]. Interestingly, there are recent hints that P-gp interacts actively with the proteasome complex [4].

It cannot be excluded that, in CJD, there are additional mechanisms that damage the integrity of the BBB and, secondarily, reduce the expression of P-gp. However, since P-gp acts as a critical detoxifying system in the brain [42], we favor the hypothesis that the age-associated decline in P-gp function elevates the levels of toxic proteins such as PrP in brain, thereby increasing the likelihood that they will accumulate to pathogenic levels via permissive templating [22].

Additionally, P-gp might play a part in the pathogenesis of neurodegenerative diseases by one or more indirect mechanisms. Diminished P-gp expression could cause a pronounced influx of exogenous, neurotoxic compounds, leading to the damage and loss of neurons that is believed to promote the development of some cases of idiopathic parkinsonism [16, 19]. What is more, P-gp could act as a neuroprotective factor by suppressing the activation of caspases [40, 41] involved in apoptosis.

Similarities in the pathogenesis of aberrant protein deposits in CJD and AD have been suggested in the context of the concept of protein misfolding diseases, or proteopathies. Since these neurodegenerative disorders may be amenable to similar therapeutic principles [2, 48], selectively augmenting cerebral P-gp expression represents a novel therapeutic strategy to forestall the accumulation of insoluble proteins in the brain. P-gp activity can be modulated by a variety of substances such as verapamil or cyclosporin A (inhibition) [51], as well as rifampin or St. Johns wort (induction) [17, 23]. Because P-gp plays an important part in a variety of tissues, the challenge for drug discovery will be to identify an agent that selectively enhances P-gp function in the brain.

Finally, it has been shown that the MDR1 gene, which codes for P-gp, is highly polymorphic [7, 24], and that P-gp expression can be influenced by these polymorphisms [24, 38]. Thus, it is conceivable that variations in the MDR1 gene might influence the risk of developing certain neurodegenerative proteopathies, including CJD, an issue that warrants further research.

Conclusion

The results of our study show that the cerebrovascular expression of the multifaceted transporter P-gp is down-regulated in CJD. We propose that diminution of P-gp expression with age reduces the expulsion of toxic proteins by the BBB. The buildup of these proteins within cells overwhelms the protein degradation machinery, thereby further promoting the accumulation of aberrant proteins such as PrP. Our findings suggest that selectively augmenting cerebrovascular P-gp function could open new therapeutic pathways for the prevention and/or treatment of prionoses and other proteopathic disorders of the central nervous system.