Insoluble cellular prion protein and its association with prion and Alzheimer diseases

Abstract

The soluble cellular prion protein (PrP^C) is best known for its association with prion disease (PrD) through its conversion to a pathogenic insoluble isoform (PrP^Sc). However, its deleterious effects independent of PrP^Sc have recently been observed not only in PrD but also in Alzheimer disease (AD), two diseases which mainly affect cognition. At the same time, PrP^C itself seems to have broad physiologic functions including involvement in cognitive processes. The PrP^C that is believed to be soluble and monomeric has so far been the only PrP conformer observed in the uninfected brain. In 2006, we identified an insoluble PrP^C conformer (termed iPrP^C) in uninfected human and animal brains. Remarkably, the PrP^Sc-like iPrPC shares the immunoreactivity behavior and fragmentation with a newly-identified PrP^Sc species in a novel human PrD termed variably protease-sensitive prionopathy. Moreover, iPrP^C has been observed as the major PrP species that interacts with amyloid β (Aβ) in AD. This article highlights evidence of PrP involvement in two putatively beneficial and deleterious PrP-implicated pathways in cognition, and hypothesizes first, that beneficial and deleterious effects of PrPC are attributable to the chameleon-like conformation of the protein and second, that the iPrP^C conformer is associated with PrD and AD.

Introduction

Prion diseases (PrDs) are a group of fatal neurodegenerative disorders in which cognitive decline and dementia constitute the early predominant clinical manifestations.1 In contrast to Alzheimer disease (AD), which is the most common cause of dementia in adults, and characterized by the deposition of noninfectious amyloid-β (Aβ) plaques consisting of a peptide containing 42 amino acids (termed Aβ42) generated from the amyloid precursor protein,2 PrDs are associated with the deposition in the brain of an infectious prion protein conformer (PrP^Sc) that is derived from its cellular prion protein (PrP^C) by a structural transition.3

PrP^C and PrP^Sc are the two major PrP conformers studied so far. They share the same primary structure but have distinct secondary structures,4–6 which explains the distinction in their physicochemical and biological features, physiology and pathophysiology. PrP^C is monomeric, rich in α-helical structure, sensitive to proteinase K (PK)-digestion, soluble in non-denaturing detergents, non-infectious and present in both uninfected and scrapie-infected brains. PrP^Sc, on the other hand, is oligomeric or aggregate, rich in β-sheet structure, partially resistant to PK, insoluble in detergents, infectious and present in infected brains only. Although PrP^C is the physiologic form of the protein, it has been well documented that the co-existence of PrP^C and PrP^Sc constitutes the prerequisite for the emergence of PrDs. The distinct prion strains and phenotypes of PrDs which have been identified in animals and humans,7,8 are probably associated with variable conformations of PrP^Sc.9–12 Recent identification of the insoluble cellular PrP (iPrP^C) in the uninfected human and animal brain raises the possibility that additional PrP^C conformers in the brain may be implicated in the beneficial or deleterious effect of PrP^C and that they may play a role in the pathogenesis of PrDs and other neurodegenerative disorders.13

PrP^C is Physiologically Involved in Human Cognitive Processes

PrP^C may play a role in human cognitive processes by interacting with other proteins in synapses.14 Long-term memory formation is believed to begin at synapses located on dendritic spines.15 PrP^C is highly concentrated at the synapse,16 presynaptically and postsynaptically as well.16–19 PrP^C significantly affected synapse formation, and the simple incubation of cultured hippocampal neurons with recombinant PrP (rPrP) in vitro induced rapid elaboration of axons and dendrites, along with increases in the number of synaptic contacts.20 In fact, a dominant synaptic impairment is often observed in PrDs,21 impairment which may result from the conversion of PrP^C into PrP^Sc.

PrP deletion impaired long-term potentiation (LTP) and decreased a fast inhibition involving GABA-A receptors.22,23 Also, it has been shown that PrP-knockout mice developed either an age-dependent impairment in memory consolidation24 or in hippocampus-dependent spatial learning.25 These deficits were reversed by re-expression of PrP^C.26 On the other hand, mice overexpressing PrP exhibited supra-physiologic responses, and a positive correlation was evident between the expression level of PrP^C and the overall strength of glutamatergic transmission in the hippocampus.26 These observations strongly suggest that PrP^C constitutes the components of cognitive processes in synapses. While the detailed molecular events underlying the effect of PrP^C on cognitive processes remain unclear, it has been observed that blocking the association of PrP^C with stress-inducible protein 1 and with laminin by antibodies adversely influences memory.27,28 Therefore, under normal conditions, the intact interaction of PrP^C with other proteins in the synapses may be necessary to maintain normal cognitive functions through a beneficial PrP^C-implicating pathway. Finally, recent findings that overexpression of PrP^C inhibits β-secretase cleavage of APP and Aβ formation, and that Aβ formation increases in manipulated PrP-deficient cells and in PrP-knockout mouse brains,29 also suggest that PrP^C plays a protective role against cognitive impairment by AD.

PrP^C is Pathologically Involved in Human Cognitive Processes

At the same time, however, several lines of evidence indicate that PrP^C may also be implicated in a pathway which adversely affects cognitive performance. Depletion of PrP^C in mice with early prion infection reversed spongiform change and prevented the emergence of clinical symptoms and neuronal loss.30 Moreover, the early hippocampal pathology, including impaired synaptic responses, preceded the accumulation of PrP^Sc deposits and the pathologic changes rapidly reversed after PrP^C depletion.31 Similarly, blocking certain PrP^C areas by anti-PrP antibodies, or by reducing and alkylating reagents, to diminish the availability of PrP^C, prevented PrP conversion and even cured prion infection in vitro and in vivo.32–34 Cumulatively, these observations are entirely consistent with the remarkable finding that PrP^C-knockout mice are resistant to prion infection.35 Therefore, PrP^Sc as well as PrP^C are necessary for the pathogenesis of PrDs.

The detrimental face of PrP^C also has been reported in other abnormal conditions. For example, its potential role in the pathogenesis of AD has recently come to light. PrP^C polymorphism, either methionine (M) or valine (V) at residue 129 (129M or 129V), may modulate the number of Aβ deposits during aging.36 Remarkably, of the 225,000 proteins that were screened in a cell model, only PrP-expressing cells were found to strongly support the binding of soluble Aβ42 oligomers, the neurotoxic species; moreover, in mouse brain slices, Aβ42 oligomers suppressed LTP in CA1 hippocampal neurons in a PrP^C-dependent manner.37 This suppression was completely abolished by deletion of murine PrP residues 32–106 or significantly reduced by an antibody against PrP93-109. AD transgenic mice with intact PrP^C expression exhibited deficits in spatial learning and memory, whereas no impairment of spatial learning and memory was detected in mice lacking PrP^C.38 These findings were not reproduced in different animal models by other laboratories.39–41 However, Strittmatter and co-workers' findings were consistent with studies in which the anti-PrP antibodies administrated peritoneally or intracerebroventricularly either improved AD mouse deficits in cognitive learning or prevented the inhibition of LTP by Aβ oligomers.42,43 Strikingly, PrP^C was observed to mediate neurotoxic signaling of β-sheet-rich conformers including not only PrP^Sc and Aβ oligomers but also yeast prions and designed β-peptides in cell models.44 The PrP^C-mediated neurotoxic signaling arises from the binding of these β-sheet-rich conformers to the intrinsically disordered N-terminal PrP domain between residues 27 and 89. PrP 27–89 includes the octapeptide repeat region that contains the Aβ42-specific binding sites.45 In addition, several new studies also suggest that PrP may be involved in AD beneficially or deleteriously.45–49

Accordingly, PrP^C may be implicated in the impairment of memory and cognition via an unknown pathway which is triggered by abnormal protein aggregation such as PrP^Sc in PrDs and Aβ oligomers in Alzheimer disease (Fig. 1).50 As a result, under pathologic conditions, blocking or depleting PrP^C becomes an important, if not exclusive strategy for restoring cognitive functioning. In PrDs, some or all of PrP^C converts to PrP^Sc, which may not only damage the favorable pathway by a decrease in PrP^C but also trigger the unfavorable pathway by the newly formed PrP^Sc. However, the molecular mechanisms underlying prion-related synaptic changes remain unclear. Nevertheless, the above evidence suggests that an unfavorable PrP^C-implicated pathway exists in human cognitive processes, especially in PrDs and Alzheimer disease. Therefore, it would be important to determine which proteins are involved in the downstream of the Aβ-PrP^C or PrP^Sc-PrP^C pathway (Fig. 1).

The Beneficial and Deleterious Effects of PrP^C and its Chameleon-Like Conformation

What is the molecular basis of the beneficial and deleterious effects of PrP^C? One striking structural feature of PrP is its chameleon-like conformation,11,12 which may be structurally associated with beneficial and deleterious effects of the protein.50 In aqueous solutions, rPrP formed a variety of conformations including pH-dependent α-helical conformations and thermodynamically more stable conformations rich in β-sheet.51 The existence of PrP folding intermediates was also indicated by hydrogen exchange experiments,52 and by high pressure NMR and fluorescence spectroscopy.53,54 In addition to a monomer, a β-oligomer and an amyloid fibril,55–59 two additional polymeric transient intermediates were also identified during fibrillogenesis of rPrP in vitro.60

The tendency of PrP to form multiple non-native isoforms rich in β-sheets in vitro, as demonstrated by biophysical studies on rPrP, may represent a unique intrinsic feature of this protein. In other words, the multiple conformational forms of rPrP may represent an intrinsic molecular spectrum of PrP^C in vivo. If so, it should be expected that more PrP conformers would be present in the brain in addition to the two major conformers PrP^C and PrP^Sc. We did indeed confirm this when we identified novel conformers which form insoluble cellular PrP aggregates and protease-resistant PrP species in uninfected human and animal brains.13 Moreover, by using gel filtration we revealed that PrP in uninfected human brains is present not only in monomers but also in oligomers and large aggregates.13 These new PrP conformers, which we termed iPrP^C, account for ∼5–25% of total PrP including full-length and N-terminally truncated forms; and a portion of iPrP^C is resistant to PK-digestion even at 50 µg/ml.13 These conformers exhibit high affinity for the gene 5 protein (g5p), a single-stranded DNA-binding protein which can also specifically bind to PrP^Sc but not to PrP^C.13,61 Notably, cytosolic PrP aggregates were also observed in pancreatic beta-cells of uninfected rats in response to hyperglycemia.62 By differential SDS solubility assay, PrP^C species with either lower or higher solubility were also differentiated in the brain homogenates of non-infected humans, sheep and cattle.63 Although the iPrP^C molecule possesses PrP^Sc-like physicochemical properties—for instance, insolubility in non-denaturing detergents, a strong tendency to form aggregates and resistance to PK-treatment—it is unclear whether these small amounts PrP^C aggregates acquire infectivity in the normal human brain. Indeed, like other PrP^Sc-directed antibodies,64–66 g5p captures not only infectious PrP^Sc but also non-infectious PrP aggregates or acid-denatured PrP species (unpublished data). In our laboratory we are investigating the infectivity of iPrP^C using both animal and cell models. Finally, the heterogeneity of PrP has been observed in its glycosylation and endogenous fragmentations,67,68 which may also contribute to the various functions of the protein.

Insoluble PrP^C and Prion Disease

Since iPrP^C possesses PrP^Sc-like physicochemical properties, there is a possibility that iPrP^C is an intermediate between PrP^C and PrP^Sc (Fig. 1) or silent PrP^Sc.13 The observation that the brains of bigenic mice are capable of clearing prions led to the proposition that the normal brain contains low levels of PrP^Sc, which play a role in cellular metabolism.69 Therefore, it is possible that under normal circumstances, despite the presence of a small amount of PrP^Sc, the brain can nevertheless maintain equilibrium between the formation and clearance of this PrP^Sc. Since this small amount of PrP^Sc does not induce a neurodegenerative disorder, presumably it is in a silent state. However, significant increases in levels of the silent prions induced by infection, PrP mutation, or unknown causes may trigger PrDs (Fig. 1). Fascinatingly, using protein misfolding cyclic amplification (PMCA), Barria and co-workers generated a new infectious prion without adding exogenous PrP^Sc seeds,70 which indicates the possibility that PMCA replicated a PrP^Sc intermediate existing in the brain homogenate, or that the silent prion was activated by the sonication-incubation cycles involved in the PMCA process.

An alternative hypothesis is that iPrP^C is a conformer which, when it increases induces an atypical form of PrDs. In fact, Westaway et al. discovered a novel neurologic syndrome in Tg mice overexpressing wild-type PrP, in which degeneration of skeletal muscle, peripheral nerves and the central nervous system was dependent on transgene dosage.71 The increased amounts of wild-type PrP^C might form aggregates that induce degeneration. Indeed, Chiesa et al. revealed that homozygous Tg mice overexpressing wild-type PrP at ∼10-fold but not hemizygous mice overexpressing wild-type PrP at ∼5-fold developed a spontaneous neurodegenerative disorder manifesting tremor and paresis.72 Nevertheless, punctate PrP deposits and abnormally enlarged synaptic terminals with a dramatic proliferation of membranous structures were observed in both types of mice. Interestingly, the overexpressed PrP assembled into insoluble aggregates with mild PK-resistance but acquired no infectivity.72 By using transgenic flies, Fernandez-Funez et al. demonstrated that misfolding and neurotoxicity of wild-type PrP are sequence-dependent: Hamster PrP formed larger amounts of PrP aggregates and induced spongiform degeneration, whereas rabbit PrP formed only very smaller amounts of PrP aggregates and did not induce spongiform degeneration.73 Remarkably, the same study also found that although small amounts of PrP aggregates were also detected in young flies (day 1) expressing hamster PrP, spongiform degeneration was not evident. Therefore, the small amounts of PrP aggregates were unable to induce spongiform degeneration. Spongiform degeneration only occurred in older flies (day 30) when the levels of PrP aggregates increased.

The small amount of PK-resistant PrP we identified in uninfected human brains exhibited a peculiar immunoreactivity behavior: higher affinity for 1E4 but poor affinity for 3F4.13 The two antibodies have adjacent epitopes on PrP.74,75 The same immunoreactivity behavior has also been observed in a new PrP^Sc species which we recently identified in a novel human PrD termed variably protease-sensitive prionopathy (VPSPr).76,77 VPSPr involves an abnormal PrP with peculiar glycosylation and fragmentation and exhibits clinical features similar to those of non-Alzheimer dementias: in particular, frontotemporal dementia, diffuse Lewis body disease and normal pressure hydrocephalus.76–78 The molecular hallmark of VPSPr is 1E4-detected pathogenetic PK-resistant PrP^Sc with a ladder-like electrophoretic profile.76,77 PrP^Sc from VPSPr exhibits immunoreactivity behavior and three PK-resistant core fragments similar to those of iPrP^C (Fig. 2). These similarities suggest that they also share a common molecular metabolic pathway (Fig. 1). Like sCJD, VPSPr affects all patients regardless of three PrP genotypes defined by 129 MV-polymorphism; however, the allelic prevalence is distinct in the two diseases.77 Interestingly, the amounts of PK-resistant PrP^Sc in VPSPr are greatly affected by the polymorphism, which has not been observed in sCJD.77 Moreover, the infectivity of PrP^Sc from VPSPr seems to be much lower compared to that of PrP^Sc from sCJD. Preliminary data revealed no clinical phenotype during the normal life span of the transgenic mice expressing human PrP-129V at 6-fold inoculated with brain homogenates from cases of VPSPr-129VV.79 Nevertheless, 30% of the mice exhibited peculiar PrP plaques with a distinctive topography and minimal or no spongiform degeneration. Most of these mice also had the PK-resistant PrP^Sc whose profile exhibited the ladder-like electrophoresis detected by 1E4. Therefore, VPSPr characterized by the deposition in the brain of iPrP^C-like PrP^Sc represents a PrD which is distinct from classical PrDs, bearing more resemblance to other neurodegenerative diseases such as Alzheimer disease and tauopathies.78 Because of the similarities between iPrP^C and PrP^Sc from VPSPr, it would be important to investigate the possibility that VPSPr results from an increase in the amount of iPrP^C (Fig. 1).

Insoluble PrP^C and Alzheimer Disease

We recently demonstrated that iPrP^C is the main species that interacts with Aβ in AD.80 Because of the specific or direct binding of Aβ42 to PrP, as indicated by previous studies in reference 37, 38 and 81 as well as our peptide membrane array and co-immunoprecipitation of soluble PrP and Aβ,80 it is most likely that PrP and Aβ42 bind directly to each other within insoluble complexes. Histologically, PrP deposits often accompany Aβ-positive plaques in AD brains.82–84 Although the exact biological relevance of the interaction between iPrP^C and Aβ is unclear at present, it has been revealed that aggregation of one protein may facilitate aggregation of the other.85 Synergistic interactions between other amyloidogenic proteins associated with neurodegeneration also have been reported to promote each other's fibrillization, amyloid deposition and formation of filamentous inclusions in transgenic mice.86,87 Morales et al. recently observed that an increase in the efficiency of Aβ42 aggregation in vitro was dependent on PrP^Sc dosage. Moreover, insoluble PrP^Sc aggregates also seem to facilitate Aβ42 aggregation in vivo and AD mice developed a strikingly higher load of cerebral amyloid plaques that appeared much faster in prion-infected than in uninfected mice.85 Our finding that Aβ42 binds to iPrP suggests that iPrP (the PrP^Sc-like forms in uninfected human brains) may facilitate fibrillization of Aβ42 in AD (Fig. 1). If this is the case, the possibility must be considered that a significant increase in the total number of Aβ plaques observed in bigenic mice overexpressing PrP86 might result from an increase in the formation of iPrP. Since the less toxic insoluble Aβ42 aggregates constitute the end products of highly toxic soluble Aβ42 oligomers, formation of the large aggregates facilitated by iPrP^C may reduce the amount of Aβ42 oligomers. The decrease in the levels of toxic Aβ42 oligomers would then attenuate the cognitive impairment induced by Aβ42 oligomers in AD. As a result, iPrP^C may play a protective role in AD. Given that iPrP^C interacts with insoluble Aβ42, whereas soluble PrP^C binds soluble Aβ42 in vivo, it is possible that distinct PrP conformers binding to different Aβ42 species thereby function either as receptors for soluble Aβ42 oligomers or as modulators of insoluble Aβ42 deposition. If this hypothesis could be proved, it would establish that the chameleon-like conformation of PrP^C is coupled with its beneficial and deleterious effects. This hypothesis could be tested by intracerebrally injecting anti-PrP antibodies against either soluble or insoluble PrP species in AD animal models.

Insoluble PrP^C and Long-Term Memory Storage

The iPrP^C species that might be of a conformation different from PrP^C may have a physiological function. A neuronal isoform of the cytoplasmic polyadenylation element binding protein (CPEB) involved in long-term memory has been demonstrated to regulate local protein synthesis by activating translationally dormant mRNA. Moreover, these events stabilized synapse-specific long-term facilitation in Aplysia.88 Interestingly, Si and colleagues further demonstrated that Aplysia CPEB exhibited self-perpetuating prion-like properties in sensory neurons or in yeast.89,90 These studies suggest that prion-like conformational changes are indispensable for the maintenance of structural synaptic changes required for long-term memory.91,92 It is reasonable to assume that the conversion of soluble PrP^C monomers into insoluble PrP oligomers or aggregates is associated with long-term memory storage in the normal human brain (Fig. 1). In contrast to PrP^C, the iPrP^C molecule has been demonstrated to bind to g5p, the single-stranded DNA binding protein.13 Accordingly, the possibility cannot be ruled out that iPrP^C binds to mRNA in vivo. The PrP gene is believed to be genetically associated with human long-term memory performance, as evidenced in the observation that 24 hours after a word list-learning task, carriers of either the polymorphism methionine/methionine (M/M) at residue 129 (129MM) or M/valine (V) (129 MV) genotype recalled 17% more information than did 129VV carriers.91 Therefore, the association between PrP and human memory may be mediated by M129V polymorphism, and the 129M allele may have a beneficial effect on long-term memory. It was proposed that the impact of a putative PrP conformation, rather than the pathologic PrP^Sc, on long-term memory in healthy humans was related to physiologically occurring conformational changes.91,93 It would therefore be intriguing for us to test our hypothesis that iPrP^C is the PrP species that plays a beneficial role in human cognitive processes.

Conclusions

The modern study of PrD following the discovery of PrP^Sc and PrP^C opened an extraordinary chapter in the history of the life science, which made it understandable that PrD may be composed of both transmissible and non-transmissible forms,94 possibly because of the chameleon-like conformation of PrP^Sc.11 Similarly, it is most likely that PrP^C also possesses a chameleon-like feature which is reflected not only in its conformation but also in its function. Subsequent discovery of iPrP^C and the demonstration of its potential association with atypical PrD and AD may open even newer avenues in the exploration of the pathogenesis of PrD and AD, thereby facilitating identification of new targets for the therapeutics of both diseases.

Abbreviations

AD	=	Alzheimer disease
CPEB	=	cytoplasmic polyadenylation element binding protein
g5p	=	gene 5 protein
iPrPC	=	insoluble cellular prion protein
PrD	=	prion disease
PrP	=	prion protein
PrPC	=	cellular prion protein
PrPSc	=	pathologic conformer of PrPc
rPrP	=	recombinant prion protein
VPSPr	=	variably protease-sensitive prionopathy

Figures and Tables

Figure 1 Diagram of the pathway involved in prion- or Aβ-induced dysfunction of memory and cognition as well as cell death. PrP^Sc or Aβ oligomers impair memory and cognition, and induce cell death by binding to PrP^C at the synapse but the downstream molecular events (the black box with a question mark) of the PrP^Sc-PrP^C or PrP^C-Aβ complex are unknown. The iPrP^C species derived from the soluble PrP^C might play a role in one or more of the following events: (1) long-term memory storage; (2) PrP^Sc formation in classic CJD; (3) initiating VPSPr; and (4) facilitating formation of Aβ42 fibrils in AD.

Figure 2 Comparison of PK-resistant PrP core fragments from non-CJD, VPSPr and sCJD. PrP captured with g5p from brain homogenates of three non-CJD subjects was treated with PK and PNGase F prior to SDS-PAGE and immunoblotting with anti-PrP antibody 1E4. Fifty µl of insoluble fraction (P2) equivalent to 500 µl of supernatant (S1) from a low-speed centrifuge was used for each non-CJD subject. In these highly concentrated samples from non-CJD subjects, three PK-resistant PrP (Pr^Pres) fragments migrating at ∼20 kDa, ∼17–18 kDa and ∼6–7 kDa were detected with 1E4. But these Pr^Pres fragments were not detectable by 3F4 (data not shown). In contrast, three PrP fragments with similar gel mobility were also detected in only 5 µl of S1 preparation from a VPSPr case (129MV) by 1E4 after treatment with PK and PNGase F. However, only one PrP band was detected in sCJD samples after PK and PNGase F treatment.

Acknowledgments

The authors are grateful to Dr. Pedro Fernandez-Funez for helpful comments. This work was supported by the National Institutes of Health R01NS062787, the University Center on Aging and Health with the support of the McGregor Foundation and the President's Discretionary Fund (Case Western Reserve University), the Alliance BioSecure and the CJD Foundation.