Homodimerization as a molecular switch between low and high efficiency PrP^C cell surface delivery and neuroprotective activity

Abstract

PrP^C is associated with a variety of functions, and its ability to interact with a multitude of partners, including itself, may largely explain PrP multifunctionality and the lack of consensus on the genuine physiological function of the protein in vivo. In contrast, there is a consensus in the literature that alterations in PrP^C trafficking and intracellular retention result in neuronal degeneration. In addition, a proteolytic modification in the late secretory pathway termed the α-cleavage induces the secretion of PrPN1, a PrP^C-derived metabolite with fascinating neuroprotective activity against toxic oligomeric Aβ molecules implicated in Alzheimer disease. Thus, studies focusing on understanding the regulation of PrP^C trafficking to the cell surface and the modulation of α-cleavage are essential. The objective of this commentary is to highlight recent evidences that PrP^C homodimerization stimulates trafficking of the protein to the cell surface and results in high levels of PrPN1 secretion. We also discuss a hypothetical model for these results and comment on future challenges and opportunities.

Keywords: :

• Alzheimer disease
• dimerization
• neuroprotection
• prion
• protein secretion
• β-sheet oligomers

This article refers to:

PrP Homodimerization

PrP dimers were experimentally detected in bovine, mouse and hamster brain homogenates,^1-3 and in N2a cells expressing hamster PrP^C or endogenous PrP^{C 4, 5}. Predictions and experimental evidences clearly showed that the hydrophobic domain (residues 112-MAGAAAAGAVVGGLGGYMLGSA-133) mediates PrP^C dimerization.^5,6

In the context of prion propagation, PrP^C dimerization is generally perceived as an important molecular step toward conformational change of the soluble α-helical PrP^C to the insoluble β-sheet-rich PrP^{Sc 7–10}. Also, the connection between dimerization and toxicity has been supported for several years by a model suggesting in vivo noxiousness of cross-linked PrP^C at neurons surface.¹¹ However, this model was recently clearly invalidated, and passive immunotherapy with monoclonal antibodies against PrP^C is seriously considered.^12-14

Results from two different studies led to the conclusion that dimerization of PrP^C is actually a positive mechanism in physiological conditions. First, in experimental settings where PrP^C displays a stress-protective signaling activity, dimerization correlates with neuroprotection.⁵ In contrast, impaired dimerization in prion-infected cells correlates with exacerbated sensitivity to stress. In a second approach, we used an inducible dimerization strategy to control PrP^C dimerization in cell culture. Using a permeable dimerizer, we discovered that PrP^C dimerization increases PrP^C trafficking to the cell surface. This resulted in a very large increase of all extracellular PrP^C species, including PrPN1 and shed PrP^C (Fig. 1), within a few hours.¹⁵ In this experiment, the burst in secretion of PrPN1 after PrP^C dimerization resulted in levels sufficient to inhibit Aβ oligomers-mediated cell death compared with basal levels of PrPN1 that were clearly inadequate.¹⁵ In this experimental setting, dimerization is a molecular switch between cell death and survival. Enforced dimerization of PrP^C also resulted in large increase of PrPC1 and shed PrP^{C 15}. Other studies demonstrated that PrPC1 acts as a GPI-anchored dominant negative inhibitor of PrP^Sc formation,¹⁶ and that soluble dimeric PrP binds PrP^Sc and can antagonize prion propagation.¹⁷ Thus, increased trafficking of PrP^C may be an effective mechanism to deliver protective molecules at the cell surface and in the extracellular space.

Figure 1. Enforced dimerization increased shedding of PrP^C. Left panel, western blot analysis of PrP^C immunoprecipitated from the cultured medium of cells expressing Fv-PrP, without (control) or with enforced dimerization (dimerizer). Right panel, Densitometric analysis of four independent experiments showed that addition of the dimerizer caused a 3.6 fold increase of shed PrP^C compared with control cells. In this experiment, the dimerizer binds to the Fv domain.¹⁵

Overall, these studies strongly support the proposition that physiological PrP^C dimers are neuroprotective.

Mechanism and Regulation of Dimerization-Induced PrP^C Cell Surface Delivery

We did not elucidate the detailed mechanism of how enforced dimerization stimulates PrP^C delivery at the plasma membrane. However, our observations provided insights for two characteristics. First, this mechanism is membrane anchor-independent. Thus, the interaction with a membrane partner or the localization in lipid rafts is not essential for dimerization-induced trafficking. Second, levels of total PrP species after dimerization were noticeably increased after dimerization.¹⁵ Since dimerization did not change PrP^C stability (Fig. 2), we hypothesize that it improves maturation efficiency and passage through protein folding quality control checkpoints. A third observation from previous reports indicates that deletion of the natural dimerization domain does not prevent cell membrane delivery of the protein.¹⁸ From these data, we propose a model in which two secretion mechanisms regulate PrP^C trafficking to the cell surface (Fig. 3). A constitutive delivery mechanism independent from the presence of the dimerization domain controls basal levels of PrP^C at the plasma membrane. In addition, a dimerization-regulated secretion mechanism allows very large increases of extracellular PrP^C species in a short period of time. Compared with the constitutive pathway, the efficiency of the regulated pathway is remarkable. By definition, our model is hypothetical, and the use of an artificial dimerization strategy to enforce PrP^C homodimerization may have exaggerated a normally occurring mechanism. Thus, the large increase in PrP^C trafficking after dimerization may not be as dramatic in physiological conditions when the natural hydrophobic domain mediates PrP^C dimerization.

Figure 2. Enforced dimerization does not increase the stability of PrP^C dimers. Fv-PrP^C expressing cells were treated with vehicle (control) or the dimerizer for 20 h. Cells were subsequently treated with cycloheximide (0h) and total cellular Fv-PrP^C levels were estimated by densitometric analyses over time. Levels of Fv-PrP^C at 8 and 24 h post-treatment were normalized to 0h. Dimerization did not significantly change the degradation rate of Fv-PrP^C. For experimental details, see Béland et al.¹⁵

Figure 3. Acute neuroprotection induced by dimerization of PrP^C. Constitutive secretion of PrP^C provides cells with basal levels of neuroprotective species (Left). After dimerization, PrP^C secretion increases sharply and causes a strong rise of full length PrP^C (FL) and PrPC1 at the cell membrane and of PrPN1 and shed PrP^C in the extracellular medium. PrPN1 and shed PrP^C are able to bind and neutralize toxic β-sheet oligomers (Right) and provide a strong neuroprotective response. More efficient trafficking of PrP^C to the cell surface also suggests that the risk of potential intracellular accumulation of toxic PrP^C species is reduced.

That PrP^C may have evolved a well conserved domain to stimulate the efficiency of its own trafficking through the secretory pathway is possible. Indeed, PrP^C signal sequence is slightly inefficient and acute stress in the endoplasmic reticulum results in intracellular accumulation and neuronal toxicity in cultured cells and in vivo.^19-21 In addition, PrP^C familial or artificial mutants that accumulate intracellularly are also neurotoxic.^22,23 Dimerization may help preventing fatal accumulation of intracellular PrP^C.

According to our model, we predict that promoting PrP^C dimerization in the secretory pathway may help fighting neurotoxic insults, including the neutralisation of β-sheet oligomers (Fig. 3). In order to modulate PrP^C dimerization as a possible therapeutic intervention, an important aim is to elucidate its regulation. It is tempting to speculate that endogenous molecules may bind to the dimerization domain and modulate the dimerization process. This domain binds several ligands, including lipids, heparin sulfate glycosaminoglycans, the secreted glycoprotein vitronectin and the secreted heat shock protein Sti1.²⁴ However, these mostly extracellular ligands are unlikely to modulate PrP^C dimerization in the secretory pathway. Currently, the modulation of endogenous PrP^C dimerization is not possible and finding a strategy to stimulate this mechanism remains an important challenge. Interestingly, the model that dimerization improves the efficiency of delivery at the cell surface is reminiscent of other membrane receptors, including some GPCR proteins.^25-29 In this field, targeting molecular chaperones that regulate GPCRs dimerization and trafficking is a clear novel therapeutic avenue.²⁹

The Elusive Connection between Dimerization, α-Cleavage and Membrane Topology

During translocation into the ER, 80–90% of the total PrP^C chains pass completely the ER membrane and get into the lumen.^19,30 Additionally, 2% of total PrP^C inserts either with their N terminus in the ER lumen (^NtmPrP) or in the opposite orientation (^CtmPrP).²¹ In these type I and type II single-pass transmembrane forms, respectively, the dimerization domain acts as a transmembrane domain. Experimental mutations that increase the transmembrane domain hydrophobicity result in higher levels of ^CtmPrP and increased neurodegeneration in mice.^31-33 Interestingly, five disease-associated mutations in human PRNP (Pro105Leu, Gly114Val, Ala117Val, Gly131Val and Ala133Val) increase the hydrophobicity of the transmembrane domain. Thus, ^CtmPrP may be the neurotoxic molecule in the corresponding familial forms of human prion disease.³⁴

Several artificial deletions of different size in this domain outside of the α-cleavage site partially or completely inhibit the α-cleavage,^35-37 indicating that the dimerization domain is also essential for this cleavage. Since α-cleavage occurs in the late secretory pathway,³⁸ PrP^C mutants unable to reach the cell surface should be resistant to α-cleavage. However, some mutants with deletions that do not affect the α-cleavage site in the dimerization domain-still traffic to the plasma membrane but have a reduced α-cleavage activity.³⁶ These results indicate that the dimerization domain is a key feature recognized by the enzyme responsible for this cleavage.

It is certainly possible that the same domain is involved in dimerization, α-cleavage and membrane integration of minor topological forms, three apparently independent functions. PrP^C is a multifunctional protein with multiple topological forms which interacts with a large variety of ligands. Thus, a PrP^C domain of 22 amino acids able to display these three functions and also bind different substrates would not be surprising for researchers in the prion field. Alternatively, these three functions may be connected if dimerization were a requirement for α-cleavage. This issue will likely require the identification of the α-protease in order to be appropriately addressed.

PrP^C-Based Therapeutic Opportunities for Neurodegenerative Disorders

Levels of secreted PrPN1 achieved after dimerization are impressive.¹⁵ This burst of secretion is particularly interesting since PrPN1 neutralizes β-sheet-rich oligomers toxicity, including soluble Aβ oligomers.^15,39,40 Specifically, Alzheimer disease is set to surge in the next future, and any potential therapeutic strategy disserves in depth investigations. In prion diseases, the N-terminal polybasic residues of PrP^C bind PrP^Sc and are required for efficient prion propagation.⁴¹ Increasing extracellular PrPN1 levels would be expected to prevent PrP^Sc binding to the neuronal surface and inhibit prion conversion and induction of intracellular toxic signals. However, several issues should be addressed prior to envisaging a therapeutic avenue based on PrPN1. First, biochemical characterizations of PrPN1 is lacking such as whether it is secreted as monomeric or oligomeric species and in association with other proteins. Second, the mechanism of how PrPN1 neutralizes β-sheet-rich oligomers remains to be determined. A direct interaction is the most straightforward hypothesis, but this interaction has not been validated yet in vivo. Third, it will be important to determine the fate of neutralized β-sheet-rich oligomers complexes and the mechanism of clearance of these fatal species. Last but not least, provided that all previous issues are resolved and the bioactivity of PrPN1 against β-sheet-rich oligomers is well characterized, the delivery mode of PrPN1 through increased dimerization or direct administration will also have to be addressed. Manipulating PrP^C trafficking or using a PrP^C metabolite as a possible therapeutic molecule may sound peculiar. Yet, there is a large body of evidence that PrP^C is neuroprotective,^5,42-44 and the translation of this characteristic in medical settings in the future might not be science fiction.⁴⁵

Acknowledgments

This research was supported by a grant from the Canadian Institutes for Health Research. We thank Julie Motard for critical reading of the manuscript.

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.