Not on the menu

Abstract

A common feature of neurodegenerative diseases is the accumulation of disease-specific, aggregated protein species in the nervous system. Transmissible spongiform encephalopathies are universally fatal neurodegenerative diseases involving the transconformation and aggregation of prion proteins. At the cellular level macroautophagy has been identified as an efficient pathway for the clearance of these toxic protein aggregates. Hence, recent research has focused on the pharmacological manipulation of autophagy as a potential treatment for neurodegenerative diseases. Independent of their effects on the estrogen receptor, tamoxifen and its metabolite 4-hydroxytamoxifen are well known inducers of autophagy. However, we recently reported that the ability of 4-hydroxytamoxifen to clear prion infection is independent of autophagy. In contrast, we provide a model whereby perturbation of cholesterol metabolism, and not autophagy, is the main mechanism whereby 4-hydroxytamoxifen is able to exert its anti-prion effects. Thus, while tamoxifen, a widely available pharmaceutical, may have applications in prion therapy, prions may also represent a special case and may require different pharmacological interventions than other proteinopathies.

Keywords: :

• prion
• tamoxifen
• 4-hydroxytamoxifen
• cholesterol
• autophagy
• neurodegenerative
• aggregates
• proteinopathies
• U18666A
• trafficking

This article refers to:

The Role of Protein Aggregates and Autophagy in Neurodegenerative Disease

A near universal feature of neurodegenerative diseases is the accumulation of aggregated proteins in the nervous system.^1,2 Parkinson Disease, Huntington’s Disease, Alzheimer Disease, frontal-temporal lobe dementia, and prion diseases are universally fatal and involve the cruel psycho-motor deterioration of sufferers. Post-mortem pathological examination reveals that their brains are burdened with aggregated protein species unique to each disease. At the cellular level, aggregated proteins pose a particular problem to cells insofar as their disposal is difficult. Indeed, the ubiquitin-proteasome system (UPS) for degrading cytosolic proteins appears to be ineffective for detangling large protein aggregates and unfolding individual proteins for degradation.³ This problem is particularly dire for the cells of non-renewable tissues, such as neurons, which cannot dilute the aggregated protein burden by cell division, nor undergo apoptosis and be replaced.^1,2 However, there is one pathway that seems to be able to efficiently dispose of large protein aggregates - macroautophagy. Macroautophagy (hereafter referred to as autophagy) is the process whereby portions of the cytosol and damaged organelles are engulfed in autophagosomes; subsequently, these structures fuse with lysosomes (forming autolysosomes) to degrade their contents and recycle essential cell building blocks. The reliance of neurons on this pathway is borne out by experiments demonstrating that genetic knockout of essential autophagy genes in mice, atg5 or atg7, results in the accumulation of protein aggregates in the nervous system accompanied by neurodegeneration and death within months of birth.^4,5 The autophagy pathway is thought to have evolved in eukaryotic organisms in order to recycle building blocks and salvage energy during times of nutrient deprivation. In higher organisms it may serve the additional function of enabling non-renewable tissues to rid themselves of the toxic protein aggregates that accumulate as the result of metabolic stress.¹ It may be the case that evolution has taken advantage of the chronic nutrient deprivation inevitably suffered by our ancestors to perform housekeeping functions, such as the clearance of toxic protein aggregates, as a by-product of salvaging nutrients and building blocks. The existence of large segments of society in the developed world that have never suffered chronic nutrient deprivation might be a contributing factor to the rise of Alzheimer Disease and other dementias.⁶ With a looming wave of neurodegenerative diseases on the horizon, researchers have begun to wonder if autophagy could be exploited pharmacologically to stem the tide. Hence, a flurry of research activity has recently been undertaken to explore the potential of chemical agents to clear disease-specific aggregated protein species by stimulating autophagy.

Prion Protein Aggregates—A Special Case?

Transmissible spongiform encephalopathies are infectious, neurodegenerative diseases involving the abnormal folding of the cellular prion protein PrP^C into a pathological conformer, PrP^Sc.⁷ This malconformed state also leads to its accumulation as aggregates in both the cytoplasm of affected neurons and the interstitial spaces within the brains of afflicted individuals. It has been proposed that the amount of protease-resistant, aggregated prion protein is maintained by a balance between PrP^Sc formation by conversion of native PrP^C and its destruction in lysosomes, which could be mediated by the autophagic pathway, as shown in the case of other neurodegenerative diseases of protein aggregation.⁸ However, the role of autophagy in prion diseases in highly debated.⁹ The appearance of multi-vesicular bodies and autophagic vacuoles in both prion-infected, neuronal cells in culture¹⁰ and in brain biopsies from prion-infected patients^11,12 suggested a protective role in the disease. Alternatively, it was proposed that autophagy may contribute to the spongiform changes that are a pathological hallmark of prion-affected brains, and may be activated by apoptosis.^11-13

We have recently explored the question of the biological role of autophagy in prion infection and disease and looked more closely into the molecular interplay between autophagy, prion propagation, trafficking and clearance using a tandem fluorescent reporter, LC3 (tf-LC3), containing both GFP and RFP.¹⁴ With tf-LC3 it is possible to distinguish autophagosomes from autolysosomes on the basis of the differential sensitivity of the green and red fluorophores to the acidic conditions in the latter compartment.¹⁵ Using this tool we found that chronic scrapie infection enhances autophagic flux¹⁴ but not autophagy per se, which is already active in neuronal cell culture as previously shown in primary neurons.^16,17 Despite the increase in autophagic flux, we observed that PrP^Sc was mostly absent from autophagic vesicles, even when lysosomal degradation was impaired. This raised the possibility that PrP^Sc was not processed by the autophagic machinery.

To further investigate the effects of autophagy on prion infection we tested tamoxifen along with a panel of other autophagy inducers, including rapamycin and lithium chloride. Surprisingly, tamoxifen was the only compound to consistently cause a substantial decrease in proteinase-K-resistant PrP^Sc. Although it has pleiotropic effects, tamoxifen is a well-known inducer of autophagy.^18,19 Puzzled by these results we decided to test the active metabolite of tamoxifen, 4-hydroxytamoxifen (OHT), which had been reported to be a somewhat weaker inducer of autophagy based on its reduced induction of Beclin-1.²⁰ Contrary to expectations, we consistently found OHT to be a more potent anti-prion agent than tamoxifen insofar as its ability to clear proteinase-K-resistant prion infection from cultured cells. This result suggested that autophagy was not responsible for the anti-prion effects of tamoxifen and OHT.¹⁴ We tested this hypothesis directly using a gene-silencing approach. Inhibition of autophagy using anti-atg7 siRNA did not affect the ability of OHT to reduce PrP^Sc content in infected cells, suggesting that autophagy was not involved in this process. Furthermore, prolonged knockdown of Atg7 did not lead to an accumulation of PrP^Sc above the levels observed in untreated cells, as would be expected if it were being degraded by autophagy.

Having ruled out autophagy we began to investigate other possible mechanisms to explain the anti-prion effects of tamoxifen and OHT. Interestingly, we observed that prion degradation seemed to depend upon degradation in lysosomes. However, it is important to note that delivery to lysosomes did not appear to occur by the autophagic route given that no increase in co-localization between autophagosomes and prion protein occurred upon induction by OHT. The localization of prions in lysosomes is not surprising in and of itself, for this is a well known site for prion distribution.²¹ However, enrichment of PrP in lysosomes under OHT treatment coincided with the accumulation of cholesterol in these organelles (Fig. 1). This is interesting for a number of reasons.

Figure 1. Schematic presentation of PrP trafficking in infected cells and upon OHT treatment. In infected cells PrP^C and PrP^Sc interact at the plasma membrane in cholesterol-rich lipid domains called lipid rafts. Upon internalization, both PrP^C and PrP^Sc can recycle via the endosomal recycling compartment (ERC) or can be routed for degradation in lysosomes. Subcellular cholesterol distribution influences PrP^Sc trafficking in the endocytic pathway. In untreated, infected cells, the majority of PrP recycles through the cholesterol-rich ERC, supporting conversion of PrP^C to PrP^Sc. Treatment with 4- hydroxytamoxifen (OHT) induces cholesterol accumulation in enlarged late endosomes (LE). PrP^Sc production and degradation defines the cellular load of infectious prions. We propose that 4-hydroxytamoxifen-induced changes in PrP^Sc trafficking favor PrP^Sc degradation. EE, early endosomes; LYS, lysosomes. (Adapted with permission from Marzo et al., 2013).¹⁴

First, in contrast to other neurodegenerative diseases the prion protein is not exposed to the cytosol²²; it is a GPI-anchored, extrinsic membrane protein found on the exofacial leaflet of membranes. For this reason it may be able to evade detection by the autophagic machinery in the cytosol. Under normal circumstances it can be distributed to the Golgi, the endosomal recycling compartment (ERC), lysosomes and the plasma membrane.²³ When we treated cells with rapamycin, the classic autophagy-inducing agent, we did observe a slight reduction of proteinase-K-resistant PrP^Sc. It could be that a population of PrP molecules was degraded due to autophagy by virtue of being in the wrong subcellular locale at the wrong time. However the majority of PrP molecules may evade the autophagic machinery and escape degradation explaining the weak effect compared with tamoxifen and OHT.

Second, attenuation of prion protein expression/generation by pharmacological manipulation of cholesterol and glycosphingolipid metabolism has been well established.^24-26 For example, inhibition of cholesterol synthesis with statins has been shown to delay the onset of symptoms and prolong survival times in scrapie-infected mice.²⁷ Furthermore, treatment with U18666A perturbs cholesterol trafficking causing a Niemann-Pick-like phenotype resulting in lysosomal accumulation of cholesterol and depletion of prion protein.²⁵ Tamoxifen/OHT treatment shares many features with U18666A treatment. Like U18666A, both tamoxifen and OHT inhibit cholesterol synthesis and perturb cholesterol trafficking²⁰ (Fig. 2). Therefore, it seems the anti-prion effects of tamoxifen/OHT are due to a classic mechanism of interference with cholesterol metabolism and not to autophagy as was first hypothesized.

Figure 2. Possible mechanisms of prion clearance by tamoxifen and OHT. Above dashed line: Tamoxifen and OHT may impair trafficking of cholesterol out of lysosomes. The high cholesterol environment favors PrP^Sc degradation and/or impairs PrP^Sc transconformation from PrP^C substrate molecules (left lysosome). In contrast, without drug treatment normal cholesterol trafficking supports an environment that prevents PrP^Sc degradation and/or favors PrP^Sc formation (right lysosome). Below dashed line: Tamoxifen and OHT inhibit cholesterol biosynthesis at different steps. While tamoxifen treatment leads to zymosterol accumulation, OHT leads to desmosterol accumulation.²⁰ Desmosterol may be a more potent inhibitor of PrP^Sc formation or promoter of its degradation than zymosterol as indicated by the differently weighted lines impacting on PrP^Sc formation.

As described above, one of the properties shared by tamoxifen and OHT is the ability to inhibit cholesterol biosynthesis.²⁰ Interestingly, these closely related compounds induce the accumulation of distinct biosynthetic intermediates of cholesterol in treated cells (Fig. 2). Specifically, while tamoxifen leads to the preferential accumulation of zymosterol, OHT treatment causes the preferential accumulation of desmosterol. Interestingly, U18666A, a potent anti-prion compound²³ also results in desmosterol accumulation by way of desmosterol reductase inhibition.²⁸ We used the polyene, macrolide filipin to detect the subcellular location of cholesterol by fluorescent microscopy.¹⁴ Since filipin has been shown to bind sterols promiscuously,²⁹ it is unclear whether the sterol content we detected in lysosomes was cholesterol or one of these other biosynthetic intermediates. Therefore, it could be that an abundance of these intermediates in lysosomal membranes, in place of cholesterol, could perturb prion synthesis or promote its degradation. This possibility is supported by our results demonstrating that although inhibition of lysosomal proteases by NH₄Cl treatment rescued prions from degradation induced by OHT treatment, the effect was not complete. This result might indicate that in addition to promoting prion degradation, tamoxifen and OHT may also foster a sterol-rich lysosomal membrane environment that inhibits the conversion of PrP^C to PrP^Sc.

To further complicate the picture, cholesterol depletion itself has also been shown to induce autophagy.²⁰ Thus, the major anti-prion effects of tamoxifen and OHT are likely due to their inhibition of cholesterol biosynthesis and perturbation of lipid trafficking; in contrast, minor anti-prion activity may be due to autophagy, induced as a by-product of interference with cholesterol biosynthesis. Unlike other neurodegenerative diseases of protein aggregation, prions diseases may represent a special case, responding to perturbation of cholesterol metabolism rather than to autophagy. That being said, tamoxifen has also been shown to play a protective role in Alzheimer Disease and this effect was shown to be partly due to its effects on cholesterol biosynthesis.³⁰ In light of these data, it would be interesting to investigate the effect of tamoxifen and OHT on other proteinopathies.

Conclusion

We consistently observed tamoxifen and especially its metabolite OHT to be superior to known autophagy inducing agents (including rapamycin and lithium) in their ability to clear cellular prion infection. While autophagy induction may be a reliable means of degrading aggregated proteins of cytosolic origin such as huntingtin, tau, Aβ, or α-synuclein it may be less effective for prions, which are sequestered from the cytosol. We found the ability of tamoxifen and OHT to reduce prion burden to be independent of autophagy. Furthermore, we demonstrated that the anti-prion effect was due in part to degradation in lysosomes and that these lysosomes were laden with sterols similar to what was observed for another anti-prion compound: U18666A. Thus, while autophagy may be a worthy avenue for discovering therapies for other neurodegenerative diseases of protein aggregation, prions may represent a unique case. As the search continues for a prion therapeutic, it may be more fruitful to investigate the cholesterol biosynthesis and trafficking networks as classic pathways known to be involved in the generation and degradation of prion protein aggregates rather than those of autophagy.

Acknowledgments

We would like to thank all the members of the Zurzolo Lab for helpful discussion. DB was funded by a post-doctoral fellowship from the Canadian Louis Pasteur Foundation. The work in CZ lab was supported by European Union FP7 (Priority, Grant 222887), by ANR (Blanc, Priontraf 2009) and (DISCover, 2009 NEUR 00203) and by Pasteur-Weizmann Foundation (2010–2012).

Submitted

06/25/13

Revised

07/15/13

Accepted

07/18/13

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.