Proteases and their Inhibitors in Neurodegenerative Disease

Roles of proteolysis and lipid rafts in the processing of the amyloid precursor protein and prion protein

N.M. Hooper


In the amyloidogenic pathway, the APP (amyloid precursor protein) is proteolytically processed by the β- and γ-secretases to release the Aβ (amyloid-β) peptide that is neurotoxic and aggregates in the brains of patients suffering from Alzheimer's disease. In the non-amyloidogenic pathway, APP is cleaved by α-secretase within the Aβ domain, precluding deposition of intact Aβ peptide. The cellular form of the PrPC (prion protein) undergoes reactive oxygen species-mediated β-cleavage within the copper-binding octapeptide repeats or, alternatively, α-cleavage within the central hydrophobic neurotoxic domain. In addition, PrPC is shed from the membrane by the action of a zinc metalloprotease. Members of the ADAM (a disintegrin and metalloproteinase) family of zinc metalloproteases, notably ADAM10 and TACE (ADAM17) display α-secretase activity towards APP and appear to be responsible for the α-cleavage of PrPC. The amyloidogenic cleavage of APP by the β- and γ-secretases appears to occur preferentially in cholesterol-rich lipid rafts, while the conversion of PrPC into the infectious form PrPSc also appears to occur in these membrane domains.

  • Alzheimer's disease
  • amyloid precursor protein
  • Creutzfeldt–Jakob disease
  • GPI anchor
  • lipid raft
  • prion protein


AD (Alzheimer's disease) is the major neurodegenerative disease of the aging brain, which is extremely complex and ill understood but which poses an ever-expanding burden on the Health Service in an aging population. By 2010, it is estimated that there will be half a million AD sufferers in the U.K. In contrast, TSEs (transmissible spongiform encephalopathies), such as CJD (Creutzfeldt–Jakob disease), are relatively rare but have received much attention in recent years because of the potential number of individuals in the U.K. affected following the epidemic of BSE (bovine spongiform encephalopathy). A major issue of current concern is that blood donors may be subclinically infected with the BSE agent and could potentially transmit variant CJD through blood transfusion to healthy individuals.

AD is characterized by studying the deposition of senile plaques consisting predominantly of the Aβ (amyloid-β) peptide of 40–42 amino acids in the brain [1]. Aβ, particularly the longer Aβ1-42, causes either directly or indirectly the neurodegeneration seen in AD. Aβ is derived by proteolytic cleavage of the APP (amyloid precursor protein), a type I integral membrane glycoprotein. TSEs or prion diseases are a group of neurodegenerative disorders that include CJD, Gerstmann–Straussler–Scheinker disease, fatal familial insomnia and kuru in humans, scrapie in sheep and BSE in cattle. The key event in the pathogenesis of these diseases is the conformational conversion of the normal cellular form of the prion protein, PrPC, into a pathogenic form PrPSc that aggregates in the brains of affected individuals and animals [2]. PrPC is anchored in the cell membrane through a C-terminal GPI (glycosylphosphatidylinositol) moiety.

In addition to both APP and PrP being critically involved in neurodegenerative disorders, there are strong similarities in the proteolytic processing and membrane localization of the two proteins and it is these areas that this review focuses on.

Proteolytic cleavage of APP

In the amyloidogenic pathway (Figure 1), the β- and γ-secretases cleave APP at the N- and C-termini of the Aβ peptide respectively. β-Secretase has been identified as a membrane-bound aspartic protease termed BACE, Asp2 or memapsin [3], whereas γ-secretase appears to be a complex of at least four proteins, presenilin, nicastrin, Aph-1 and Pen-2 [4]. Two transmembrane aspartate residues within presenilin probably constitute the active site of this protease. In the non-amyloidogenic pathway (Figure 1), APP is cleaved within the Aβ domain by α-secretase. This proteolytic cleavage prevents the deposition of intact Aβ peptide and results in the release of a large soluble ectodomain, sAPPα, from the cell that has neuroprotective and memory-enhancing effects [5,6]. Under normal circumstances, the majority of APP is cleaved by α-secretase with minimal processing by β-secretase. As the α- and β-secretases compete for the same substrate, an increase in Aβ peptide production correlates with a decrease in sAPPα levels and vice versa [7,8]. Consistent with this, AD patients have decreased levels of sAPPα in their cerebrospinal fluid [9,10]. However, the mechanisms regulating the distribution of APP between the amyloidogenic and non-amyloidogenic pathways remain unclear, and identifying these regulatory mechanisms is central to our understanding of the pathogenesis of AD.

Figure 1 Schematic diagram of the proteolytic processing of APP

APP is a type I integral membrane protein (white rectangle) with the Aβ domain shown as the cross-hatched region. The membrane is represented by the grey-shaded rectangle. In the non-amyloidogenic pathway, APP is cleaved within the Aβ domain by α-secretase, predominantly ADAM10 and TACE, to release the soluble ectodomain fragment sAPPα. In the amyloidogenic pathway, APP is cleaved at the N-terminus of the Aβ domain by β-secretase, BACE, with the release of the soluble ectodomain fragment sAPPβ. The resulting membrane-anchored C-terminal fragment is then cleaved at the C-terminus of the Aβ domain by γ-secretase, a multi-component complex containing the presenilins.

Members of the ADAM (a disintegrin and metalloproteinase) family of proteases, namely ADAM9, ADAM10 and TACE (tumour necrosis factor-α converting enzyme; ADAM17), have been shown to possess α-secretase activity (reviewed in [11]). A consensus view appears to be emerging that there is a team of metalloproteases capable of cleaving APP at the α-secretase site. In different cell types, and possibly under particular cellular conditions, different members of this team contribute to a greater or lesser extent to the α-secretase cleavage of APP. For example, in the human neuroblastoma SH-SY5Y cell line we have shown, using selective inhibitors [12,13] and an antisense-oligonucleotide approach [14], that ADAM10 has a major role in the constitutive cleavage of APP with TACE having a minor role. However, TACE may have a more significant role to play in the protein kinase C-stimulated cleavage of APP [15]. Cholesterol-lowering drugs, metal chelators, steroid hormones and non-steroidal anti-inflammatory drugs have all been shown to inhibit the production of Aβ, and some of these agents are currently being evaluated in clinical trials as treatments for AD [16]. A few of these agents have also been shown to increase the non-amyloidogenic processing of APP. In the case of the cholesterol-lowering drugs, this effect may, in part, be due to alteration in the expression of ADAM10 [17]; however, whether the other compounds affect the non-amyloidogenic processing of APP, and if so their mechanism of action, is not known.

Proteolytic processing of PrP

The C-terminal half of PrPC has a compact domain structure consisting of three α-helices and a two-stranded anti-parallel β-sheet, whereas the N-terminal half is less structured [18]. Within this N-terminal domain are four copies of an octapeptide repeat sequence PHGG(G/S)WGQ that bind Cu2+ ions. This has led to proposals that PrP is involved in cellular copper metabolism and has anti-oxidant properties [19]. Within the centre of the polypeptide chain is a highly conserved hydrophobic region (residues 106–126) that, at least as an isolated peptide, has neurotoxic properties [20]. The ‘normal’ cleavage of PrPC, in the brain and in cultured cells, occurs inside the 106–126 neurotoxic region of the protein, leading to the formation of a soluble N-terminal fragment (N1) and a GPI-anchored C-terminal fragment (C1 or PrP-II) that can no longer be converted into PrPSc (Figure 2) [21,22]. Similar to the non-amyloidogenic cleavage of APP, this ‘normal’ cleavage of PrP is stimulated by agonists of the protein kinase C pathway [23]. Overexpression or deletion of the gene for ADAM10 increased or decreased the formation of N1 respectively and similar transfection and knock-out approaches demonstrated that TACE contributed to the protein kinase C-mediated production of N1 [24]. Thus, both APP and PrPC undergo proteolytic cleavage in their ‘toxic’ domains by members of the ADAMs family [25].

Figure 2 Schematic diagram of the proteolytic processing of the cellular form of the PrP

PrPC is a GPI-anchored protein (white rectangle) with copper-binding octapeptide repeats (grey shaded) and central hydrophobic and neurotoxic region (residues 106–126) (cross-hatched). PrPC can be shed from the membrane through the action of a zinc metalloprotease, resulting in a soluble form of the protein. PrPC can also be subject to α-cleavage by members of the ADAM family within the central hydrophobic and neurotoxic region to generate the soluble N1 fragment and the membrane-anchored C1 fragment. Alternatively, PrPC can be subject to reactive oxygen species-mediated β-cleavage within the octapeptide repeats to generate the soluble N2 fragment and the membrane-anchored C2 fragment.

PrPC can also be cleaved within or adjacent to the octapeptide repeats to generate a 21 kDa C-terminal fragment C2 and the corresponding N-terminal fragment N2 (Figure 2) [22,26,27]. This cleavage event appears to be mediated by reactive oxygen species [28] (N.T. Watt and N.M. Hooper, unpublished work) and has been termed β-cleavage [29]. Recently, the processing of PrPSc in scrapie-infected brain and cells to a C2-like fragment, that has the same molecular mass as the protease-resistant core PrP-(27–30), by the Ca2+-dependent calpain proteases has been reported [30].

A soluble, shed form of PrP has been observed in both the medium of cultured cells and in human CSF (cerebrospinal fluid) [21,31,32], although it has not been established whether this is due to cleavage within the GPI region or to loss of the GPI anchor owing to proteolysis near the C-terminus. Recently, we observed that PrP is proteolytically shed from the cell surface (Figure 2) [33]. The shed form of PrP was recognized by an antibody generated to an epitope near the C-terminus of the polypeptide chain. Surprisingly, the inhibition profile with a range of hydroxamate-based zinc metalloprotease inhibitors was identical to the α-secretase shedding of APP [13], implicating an ADAM protease, possibly ADAM10, in this process. It is not clear what role soluble forms of PrP play in prion diseases. On the one hand, GPI-minus forms of PrPC may be required for the initial steps of infection [34] and thus an increase in the shedding of PrP may enhance the infectious process. On the other hand, increased shedding of PrPC would reduce the amount of membrane-bound proteins available for subsequent conversion into PrPSc [35]. Thus understanding the mechanism of shedding of PrP and the role of ADAM proteases in its ‘normal’ metabolism is central to the pathogenesis of prion diseases and may provide novel therapeutic approaches.

Role of cholesterol-rich lipid rafts in regulating the processing of APP and PrP

Lipid rafts, domains of the plasma membrane enriched in cholesterol, glycosphingolipids, sphingomyelin and acylated proteins have been implicated in a range of biological processes, including intracellular trafficking, transmembrane signalling, lipid and protein sorting, viral uptake and regulated proteolysis [36,37]. Rafts are characterized biochemically by their relative insolubility at 4°C in certain detergents such as Triton X-100.

It has been hypothesized that the amyloidogenic processing of APP occurs primarily in rafts, whereas the non-amyloidogenic processing occurs in non-raft regions of the membrane [38]. Although only a minor proportion of APP, presenilin and BACE are localized in rafts [3941], experimental evidence does support the occurrence of amyloidogenic processing in rafts [41,42]. Significantly, a decrease in cellular cholesterol, which disrupts the structure and function of rafts, results in a decrease in Aβ production and an increase in sAPPα formation [17,43]. Such an alteration in the processing of APP may, in part, account for the beneficial effect of cholesterol-lowering drugs, statins, in lowering the prevalence of AD [44]. Recently, we have investigated the significance of rafts in APP processing using the novel approach of targeting the β-secretase BACE exclusively to rafts by replacing its transmembrane and cytosolic domains with a GPI anchor [45]. Expression of this GPI-anchored form of BACE substantially up-regulated the secretion of both sAPPβ and Aβ, above the levels observed from cells overexpressing wild-type BACE. This effect was reversed when rafts were disrupted by depleting cellular cholesterol. These results suggest that processing of APP to Aβ occurs predominantly in rafts and are in agreement with another report showing that antibody cross-linking induced co-patching of APP, BACE and raft marker proteins and increased Aβ production [46].

Lipid rafts also appear to play a critical role in the conversion of PrPC into PrPSc, as depletion of cellular cholesterol decreased the formation of PrPSc [47]. Although the GPI anchor on PrP mediates its raft association, there is evidence that the interaction of PrP with model membranes, including SCRLs (sphingolipid-cholesterol-rich raft-like liposomes), can occur in a GPI-independent manner [34,48]. Utilizing alternatively anchored forms of PrP expressed in SH-SY5Y cells, we observed that a GPI anchor is not obligatory for the raft association of PrP, which occurred irrespective of both the mode and topology of membrane anchorage [49]. Deletion mutagenesis demonstrated that the N-terminal region of the PrP ectodomain (residues 23–90) was necessary for this raft association and that this region was sufficient for raft association when fused to the ectodomain of a non-raft protein. The 23–90 region of the PrP ectodomain may function as a raft-targeting determinant, either by association with another raft protein or by directly interacting with raft-associated lipids. Although PrP has been reported to bind to several proteins, including neuronal cell adhesion molecules, plasminogen, the 37/67 kDa laminin receptor and stress-inducible protein 1, binding in all cases involved regions C-terminal to residue 90 of PrP. An interaction of PrP with raft lipids is supported by the report that hamster PrP lacking a GPI anchor can bind to SCRLs and that this binding is markedly lowered by deletion of the 34–94 region of the protein [34]. The raft-targeting determinant in the ectodomain of PrP may play a critical role during the process of prion infection. In support of this, the conversion of PrPC-like proteinase K-sensitive PrP (PrP-sen) to PrPSc-like proteinase K-resistant PrP (PrP-res) by exogenous PrP-res required a GPI-independent, rather than a GPI-directed, interaction of PrP-sen with SCRLs [34]. In addition, raft-bound PrP-sen resisted conversion into PrP-res until PrP-sen was released from rafts by phospholipase digestion, the PrP-res was inserted into contiguous membranes or cell contact occurred [35,50]. Clearly identifying the molecule(s) in rafts that interact with the N-terminal region of PrP has implications both for the mechanism of prion infection and for the normal metabolism of PrPC.


Like many proteins, APP and PrP are subject to a variety of proteolytic cleavage events that modulate their biological functions. In some cases, the role of these cleavages and the function of the fragments generated are well-documented, in other cases these remain to be determined. The amyloidogenic processing of APP appears to occur within cholesterol-rich lipid rafts, which are also the site of conversion of PrPC into PrPSc. Inhibition of the proteolyic cleavage of APP by the β- and γ-secretases are potential therapeutic strategies for the treatment of AD. Whether inhibition of one of the proteolytic cleavages of PrP would be a potential therapeutic target for the treatment of TSEs remains to be determined.


The financial support of the Medical Research Council, the Biotechnology and Biological Sciences Research Council and the European Union is gratefully acknowledged. The contributions of past and present members of the author's laboratory are also gratefully acknowledged.


  • Proteases and their Inhibitors in Neurodegenerative Disease: Focused Meeting held at the Royal Institute of British Architects, London, U.K., 17 November 2004. Organized by and Edited by H. Nagase and J. Saklatvala (Imperial College London, U.K.).

Abbreviations: AD, Alzheimer's disease; ADAM, a disintegrin and metalloproteinase; APP, amyloid precursor protein; Aβ peptide, amyloid-β peptide; BSE, bovine spongiform encephalopathy; CJD, Creutzfeldt–Jakob disease; GPI, glycosylphosphatidylinositol; PrP, prion protein; PrP-res, proteinase K-resistant PrP; PrP-sen, proteinase K-sensitive PrP; SCRL, sphingolipid-cholesterol-rich raft-like liposome; TACE, tumour necrosis factor-α converting enzyme; TSE, transmissible spongiform encephalopathy


View Abstract