Move over α, β, η, and γ. There’s a shiny new secretase in town, and it may sink its teeth into APP before the others get a bite. A new study published in Nature Communications on November 9 claims that asparagine endopeptidase (AEP)—a protease that switches on in acidic conditions—is a δ-secretase. Like a doting mother, AEP cuts APP into bite-sized portions for toddler BACE1 to chew on, facilitating an increase in Aβ production. The researchers, led by Keqiang Ye at Emory University School of Medicine in Atlanta, reported that AD mice lacking AEP made fewer Aβ plaques, skirted synaptic dysfunction, and remembered much better than counterparts expressing the δ secretase. The researchers found APP fragments allegedly produced by the protease in the cerebrospinal fluid of AD patients, and the protease’s expression and activity in the brain increased with age. Given that Ye had previously reported that AEP cleaves tau to generate toxic fragments, the researchers proposed that the δ secretase represents a prime target for AD therapeutics. However, as some commentators pointed out, much is left to learn about how and where AEP inflicts its damage.

“This enzyme might be a crucial trigger for neurodegeneration during the aging process, so we think inhibiting it could provide a substantial therapeutic benefit,” Ye told Alzforum. Ye and colleagues are screening small-molecule inhibitors of the enzyme and investigating the potential of AEP-generated APP fragments as biomarkers for AD.

AEP, also known as legumain, is a lysosomal cysteine protease that cleaves peptide bonds after asparagine residues. In the brain, AEP expression and activity ramps up in response to hypoxia and to excitotoxicity, and the enzyme can hasten neuronal cell death. In Ye’s lab, the protease presented itself unexpectedly when a postdoc accidentally swapped a neutral buffer for an acidic one during an experiment, and the researchers subsequently found that the protease, AEP, cleaves SET, a protein that prevents DNA degradation (see Liu et al., 2008). The group subsequently reported that AEP promoted tau hyperphosphorylation by cleaving, and thus activating, an inhibitor of the phosphatase PP2A, which normally keeps such rampant phosphorylation in check (see Basurto-Islas et al., 2013). They went on to report that the protease cleaves tau in a way that promotes neurofibrillary tangles, and also snips TDP-43—a key player in frontotemporal dementia and amyotrophic lateral sclerosis (see Oct 2014 newsHerskowitz et al., 2012). 

First author Zhentao Zhang and colleagues wondered if AEP could also process APP. The researchers mixed kidney or brain extracts with APP, and found that lysates from the wild-type mice, but not AEP knockouts, generated two major APP fragments. Importantly, these fragments only appeared when the mixture was kept at an acidic pH, when AEP was active. Treating the lysates with an AEP inhibitor or with antibodies that block AEP function prevented this cleavage. The two APP fragments weighed in at 80kDa and 130kDa, and mass spectrometry revealed that they were N-terminal fragments cleaved at asparagine 373 and 585 in APP, respectively. Mutating these residues prevented AEP processing of APP.

Might AEP activity influence Aβ production? To find out, the researchers first tested whether AEP affected the cleavage of APP by BACE1, the rate-limiting step in Aβ production, or by the α-secretase ADAM10, which diverts APP processing away from amyloid production. They found that BACE1 more readily cleaved the shorter C-terminal AEP-generated fragment APP585-695 than it did full-length APP, while ADAM10 processed these proteins at the same rate. Both α- and β-secretases processed the longer AEP-generated C-terminal fragment (APP374-695) as well as they did full-length APP. Zhang reported that neuronal cultures from AEP KO mice produced much less Aβ40/42 than did wild-type neurons. Conversely, overexpression of the APP585-695 fragment in HEK293 cells promoted Aβ production, while knocking down AEP in these cells reduced production. Together, these findings indicated that AEP cleavage of APP at asparagine 585 promoted BACE1 cleavage, which occurs just 11 amino acids away at residue 596.

Given that AEP and BACE1 seemed to cooperate in Aβ production, the researchers next investigated where the two enzymes got together in the cell. They probed fractionated brain extracts with antibodies specific for APP, BACE1, AEP, and several endolysosomal markers. In normal five-month-old mice, the researchers found minimal overlap between the proteins, as BACE1 and APP appeared predominantly in endosomal and Golgi fractions, while AEP was relegated to the lysosome. However, in brain extracts from five-month-old 5xFAD mice, P301S tau mice, or aged wild-type mice, AEP spread out among both endosomal and lysosomal fractions and overlapped significantly with BACE1 and APP. Ye speculated that this change in AEP localization was due primarily to increased abundance of the protein in both aged and diseased brains.

The researchers used confocal microscopy to track APP and AEP in primary neurons. They observed AEP in both lysosomal (LAMP1+) and endosomal (EEA-1+) compartments in neurons from 5xFAD mice, as well as in wild-type neurons treated with pre-aggregated Aβ, which reportedly affects membrane trafficking. APP appeared to mingle with AEP in the cells.

Ralph Nixon of the Nathan Kline Institute in Orangeburg, New York, pointed out that the brain fractionation and confocal co-localization experiments were difficult to interpret, as such experiments produce notoriously variable results and were not adequately quantified in the paper. However, he added that the findings certainly highlight the importance of endosomal compartments in the pathogenesis of AD. “This adds new players to a location already known to be an AD hotspot, namely the endosomal pathway,” he told Alzforum. “The findings potentially bring together even more pathobiology in that location.”

Interestingly, the researchers employed a suite of fragment-specific antibodies to detect levels of APP1-373, APP1-585, and APP585-695 in human CSF. Concentrations of all three, and AEP, were higher in AD patients than in healthy controls. These fragments and the protease also increased in the brains of mice as they aged.

To zero in on AEP’s potential role in synaptic and cognitive symptoms of AD, the researchers crossed AEP KO mice with 5xFAD mice. Without AEP, mice retained more dendritic spines and synapses, had less deficits in long-term potentiation, and preserved their spatial memory, as assessed by their ability to learn the location of a hidden platform in the Morris water maze. In an injection model, APP harboring the Swedish, London, and Arctic mutations was rendered less toxic by ablating the AEP cleavage sites. Together, these findings suggested that AEP cleavage of APP promoted synaptic dysfunction and memory problems.

What about the soluble APP fragments released by AEP cleavage? Ye and colleagues also reported that the N-terminal APP1-373 fragment was toxic to neurons and promoted apoptosis, while the APP1-585 fragment did not. Ye speculated that synaptic dysfunction mediated by AEP boiled down to enhanced Aβ production through BACE1 cleavage of APP585-695, and to the release of this toxic N-terminal APP1-373 peptide fragment. Interestingly, the N-terminal fragment of APP has been implicated in neuronal apoptosis before, through its binding of Death Receptor 6 (DR6), though that does not seem to drive AD pathology (see Feb 2009 conference news; May 2014 news). 

Nixon acknowledged the potential importance of both of these APP products in neuronal dysfunction, but noted that the researchers did not examine the role of β-CTF, another product of BACE1 cleavage that may cripple endosomal trafficking (see Jul 2015 news). He added that the AEP products join the pantheon of APP fragments researchers must now grapple with to understand AD. This includes products of the recently described η-secretase cleavage (see Aug 2015 news). Ye speculated that AEP could also be responsible for generating these fragments, as they were cleaved at asparagine residues, albeit different ones than Ye detected due to AEP cleavage. Why Ye detected different fragments than those generated by η-secretase cleavage remains unknown, but Ye chalked it up to the use of different methods.

Kaj Blennow of the University of Gothenburg in Sweden commented that whiffs of potential AEP products of APP have long been in the air. A cleavage site just one amino acid away from the N585 site was attributed to a “δ secretase” nearly two decades ago (see Simons et al., 1996). Blennow added that many other fragment possibilities arising from AEP cleavage, such as APP585-596, which would be released following AEP and BACE1 cleavage, might be detected in CSF if they are there. “Such APP or APP/Aβ peptides may serve as biomarkers to study AEP processing in AD patients or as theragnostic markers in trials targeting AEP,” he said. Blennow and Ye are currently collaborating to track such markers. Beyond biomarkers, Blennow wrote that AEP stands out as an attractive therapeutic target.—Jessica Shugart 


Make a Comment

Comments on this content

  1. Zhang et al. nicely show that amyloid precursor protein (APP) is processed at new δ-secretase sites by asparagine endoprotease (AEP) using AEP gene knockout (AEP KO) mice expressing human APP containing the Swedish mutant β-secretase site sequence (APPSwe). The AEP KO APPSwe mice have lower brain Aβ levels and reduced memory deficits relative to their APPSwe counterparts, which express normal levels of AEP. The authors conclude that AEP may be a new drug target for AD. The paper thus adds to the growing list of proteases, which cleave APP, that affect neurotoxic APP fragment production and are potential AD therapeutic targets.

    The AEP KO may have different effects on Aβ in transgenic mice expressing APP containing the wild-type β-secretase site sequence (APPwt) than in APPSwe mice. This is because AEP KO increases cathepsin B levels and activity relative to AEP-sufficient mice (Miller et al., 2011; Shirahama-Noda et al., 2003) and cathepsin B overexpression raises brain Aβ1-40/42 and pyroglutamate Aβ-3-40/42 levels in transgenic mice expressing APPwt (Hook et al., 2014). Further, cathepsin B KO reduces Aβ peptide levels in mice expressing APPwt but not in mice expressing APPSwe (Kindy et al., 2012). Thus, the AEP KO increases levels of cathepsin B that may increase Aβ in APPwt, but not in APPSwe mice, consistent with the results of Zhang et al.

    While AEP and cathepsin B are both cysteine proteases, they are structurally very different and thus different compounds may inhibit each. For example, the compound E64, which is a potent cathepsin B inhibitor, does not inhibit AEP (Chen et al., 1997). Thus, AEP and cathepsin B inhibitors are likely to be different compounds even though both are cysteine proteases.


    . Asparagine endopeptidase is required for normal kidney physiology and homeostasis. FASEB J. 2011 May;25(5):1606-17. Epub 2011 Feb 3 PubMed.

    . Biosynthetic processing of cathepsins and lysosomal degradation are abolished in asparaginyl endopeptidase-deficient mice. J Biol Chem. 2003 Aug 29;278(35):33194-9. Epub 2003 May 29 PubMed.

    . Brain pyroglutamate amyloid-β is produced by cathepsin B and is reduced by the cysteine protease inhibitor E64d, representing a potential Alzheimer's disease therapeutic. J Alzheimers Dis. 2014;41(1):129-49. PubMed.

    . Deletion of the cathepsin B gene improves memory deficits in a transgenic ALZHeimer's disease mouse model expressing AβPP containing the wild-type β-secretase site sequence. J Alzheimers Dis. 2012;29(4):827-40. PubMed.

    . Cloning, isolation, and characterization of mammalian legumain, an asparaginyl endopeptidase. J Biol Chem. 1997 Mar 21;272(12):8090-8. PubMed.

Make a Comment

To make a comment you must login or register.


News Citations

  1. Does Novel Tau Protease Promote Pathology?
  2. Keystone: Death Receptor Ligand—New Role for APP, New Model for AD?
  3. Reports Weaken Death Receptor Link to Alzheimer's Disease
  4. Partners in Crime: APP Fragment and Endosomal Protein Impair Endocytosis
  5. Enter Aη: Alternative APP Cleavage Creates Synaptotoxic Peptide

Research Models Citations

  1. 5xFAD
  2. hTau.P301S

Paper Citations

  1. . Neuroprotective actions of PIKE-L by inhibition of SET proteolytic degradation by asparagine endopeptidase. Mol Cell. 2008 Mar 28;29(6):665-78. PubMed.
  2. . Activation of asparaginyl endopeptidase leads to tau hyperphosphorylation in Alzheimer disease. J Biol Chem. 2013 Jun 14;288(24):17495-507. PubMed.
  3. . Asparaginyl endopeptidase cleaves TDP-43 in brain. Proteomics. 2012 Aug;12(15-16):2455-63. PubMed.
  4. . Amyloidogenic processing of the human amyloid precursor protein in primary cultures of rat hippocampal neurons. J Neurosci. 1996 Feb 1;16(3):899-908. PubMed.

Other Citations

  1. APP

Further Reading

Primary Papers

  1. . Delta-secretase cleaves amyloid precursor protein and regulates the pathogenesis in Alzheimer's disease. Nat Commun. 2015 Nov 9;6:8762. PubMed.