Neurofibrillary tangles, which contain a mixture of hyperphosphorylated tau fragments, are a hallmark of tauopathies such as Alzheimer's and frontotemporal dementia. Scientists debate which of many tau proteases contribute most to pathology. Now, researchers led by Keqiang Ye, Emory University School of Medicine, Atlanta, add another to consider. In the October 19 Nature Medicine, they report that asparagine endopeptidase (AEP) cuts tau in several places. Knocking out this cysteine protease partially rescued both tau pathology and behavioral deficits in transgenic mice expressing mutant forms of human tau. What’s more, the researchers found AEP-generated tau fragments in the brains of people with the disease. “This study suggests AEP is a major player in Alzheimer’s pathology,” Ye told Alzforum.

Now you see it, now you don’t.

Neurons in the hippocampus of tau-transgenic mice stain positive (red) for the N368 fragment (left), but not when AEP is knocked out (right). Nuclei are stained blue. [Image courtesy of Zhang et al., Nature Medicine.]

Also known as legumain, as it was first found in a bean, AEP normally resides in the lysosome. In 2008, Ye and colleagues discovered that under acidic conditions in the brain brought on by stroke, AEP wanders into the cytoplasm and nucleus, where it cuts and deactivates an enzyme that prevents DNA degradation. Loss of DNA then leads to neuron death (see Liu et al., 2008). Khalid Iqbal and colleagues at the New York State Institute for Basic Research in Developmental Disabilities, Staten Island, found that during normal aging, and in Alzheimer’s disease, AEP also enters the cytoplasm, where it cleaves an inhibitor of a tau phosphatase, leading to tau hyperphosphorylation (see Basurto-Islas et al., 2013). Could AEP cleave tau as well?

To find out, first author Zhentao Zhang and colleagues compared tau protease activity in wild-type mice with that in AEP-knockout mice (see Shirahama-Noda et al., 2003). Because AEP is the only mammalian enzyme known to cleave peptide bonds that follow an asparagine (N), they looked for fragments of tau ending in that amino acid. In vitro, kidney extracts from the wild-type mice cleaved tau in many places, resulting in a mixture of fragments. Two, 1-N255 and 1-N368, were most abundant. Extracts from AEP-knockout mice produced neither of these two fragments, indicating the endopeptidase was responsible them (see image above). A series of in vitro experiments revealed that these cleavages occurred independently of other tau proteases, which include caspases, calpains, and thrombin (see Wang et al., 2010). Mass spectroscopy also found fragments ending in N368 in postmortem human brains of AD patients, suggesting the human protein cleaves tau there as well.

To see if AEP cleavage of tau changes as mice grow older, the researchers examined brain extracts taken from wild-type mice at different ages using an anti-tau N368 antibody they had just developed. Older mice had more N368 tau. The antibody also detected more of the tau epitope in human AD brain samples than in samples from age-matched controls. It often bound neurofibrillary tangles. Interestingly, Aβ dose-dependently elevated N368 in rat primary neuron cultures, suggesting that Aβ promotes AEP degradation of tau.

Does AEP-cleaved tau affect neurons? In vitro, the tau peptides assembled into paired helical fragments faster than the full-length protein. They also reduced polymerization of tubulin, the building block of microtubules. When the researchers overexpressed the AEP-generated tau fragments in primary neurons, they were robustly phosphorylated, axons shrank, and apoptosis ramped up. Knocking out AEP in P301S tau-transgenic mice reduced hyperphosphorylation of tau by 30 percent. Compared to P301S controls, AEP knockouts had more dendritic spines and synapses, and better synaptic function. Knocking out the protease also reversed learning deficits on the Morris water maze, and restored memory on the cued-fear and contextual-fear assays.

Altogether, the data suggest that AEP mediates some downstream effects of tau cleavage and of P301S mutant tau. The authors hypothesize that as mice and people get older, more AEP is released from lysosomes into the cytoplasm, where it cleaves tau, leading to its hyperphosphorylation and aggregation.  

The findings fit with Iqbal's results. “Ye's work is a nice addition to the story,” he told Alzforum. “[Cleavage of tau] is another means by which the asparaginyl endopeptidase could promote tau pathology.” Like the authors, he suspects other disease-related proteins could be cleaved by AEP as well. Ye previously reported that AEP cuts TDP-43 in people with frontotemporal lobar degeneration (see Herskowitz et al., 2012). 

The work also jibes with findings from Ralph Nixon's lab at New York University, Langone Medical Center. He found that reducing tau cleavage by calpain, another protease, could treat tau-related neurodegenerative diseases (see Jul 2014 news story). Since some tau is cut and aggregates even in AEP knockouts, calpain inhibition could prove complementary to AEP inhibition, Nixon guessed. “This is an impressive set of observations that resonates with existing work,” he told Alzforum. “It reinforces that AEP is a legitimate target for anyone considering tau modulation and modification as a therapeutic avenue.” Ye is testing an AEP inhibitor, which he says penetrates the brain, in mouse models of AD and other diseases.

“The results are interesting and provide new insight into the mechanistic role of proteases in AD neuropathology,” said Robert Rissman, University of California, San Diego. He cautioned that inhibiting AEP and disrupting its normal function could pose problems for a number of required biological processes, as would be the case for inhibiting other proteases. Erik Portelius, University of Gothenburg, Sweden, wrote in an email to Alzforum that it would be interesting to know if these AEP-generated fragments turn up in CSF and if so, how they compare between Alzheimer's disease patients and controls (see full comment below).—Gwyneth Dickey Zakaib


  1. This is a very comprehensive study where the authors have used various techniques and models to show that a specific endopeptidase (AEP) degrades tau. It has been known for some years that tau is degraded and fragments of tau have previously been detected in both human brain tissue and cerebrospinal fluid (CSF) (see, for example, Zilka et al., 2012; Meredith et al., 2013). However, the exact molecular nature of the tau fragments has so far been difficult to determine. In the study by Zhang and colleagues, they show the exact cleavage sites for AEP. From a biomarker perspective, it would have been interesting to know if these AEP-generated fragments are present in CSF and if so, how would these tau fragments perform in an Alzheimer's disease and control study.

    View all comments by Erik Portelius
  2. The study unequivocally shows a predominent role of asparagine endopeptidase (AEP)/legumain in the degradation of tau, and precisely determines the cleavage site of the asparagine-specific protease. The absence of AEP-derived tau fragments in AEP-knockout mice proves the in vitro data. However, a central question related to the topology of molecular events mediated by this protease in the paper from Zhang et al. remains unanswered: If AEP is a lysosomal protease, how does tau fragmentation occur mainly in the cytosol?

    The authors speculate that AEP is translocated to the cytosol under specific conditions, where it might act on tau. However, they also present clear data that AEP is inactive at cytosolic pH and is only able to cleave tau under acidic conditions. Previous studies had shown, moreover, that AEP is irreversibly inactivated and denatured under neutral pH conditions (Chen et al., 1997). Taken together, cytosolic tau degradation by AEP seems unlikely.

    Tau was previously shown to be degraded at least in part by autophagy, with both chaperone-mediated autophagy and macroautophagy playing a role (for detailed reviews see Wang et al., 2010; Wang and Mandelkow 2012). Very recently, phosphatidylinositol-binding clathrin assembly protein (PICALM) was shown to modulate autophagy and thereby tau accumulation (Moreau et al., 2014). Finally, overexpression of the transcription factor EB (TFEB) was recently shown to mediate the clearance of toxic tau species in the rTg4510 mouse model of tauopathy (Polito et al., 2014).

    It remains to be clarified if AEP proteolysis of tau occurs in the cytosol or in autolysosomes after autophagic engulfment. This question has a major impact on possible therapeutic interventions aimed at AEP as a target. 


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

    . Synergy and antagonism of macroautophagy and chaperone-mediated autophagy in a cell model of pathological tau aggregation. Autophagy. 2010 Jan;6(1):182-3. PubMed.

    . Degradation of tau protein by autophagy and proteasomal pathways. Biochem Soc Trans. 2012 Aug;40(4):644-52. PubMed.

    . PICALM modulates autophagy activity and tau accumulation. Nat Commun. 2014 Sep 22;5:4998. PubMed.

    . Selective clearance of aberrant tau proteins and rescue of neurotoxicity by transcription factor EB. EMBO Mol Med. 2014 Jul 28;6(9):1142-60. PubMed.

Make a Comment

To make a comment you must login or register.


News Citations

  1. Reining in Calpain Tempers Tau and Delays Neurodegeneration

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. . 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.
  4. . Proteolytic processing of tau. Biochem Soc Trans. 2010 Aug;38(4):955-61. PubMed.
  5. . Asparaginyl endopeptidase cleaves TDP-43 in brain. Proteomics. 2012 Aug;12(15-16):2455-63. PubMed.

Further Reading


  1. . The asparaginyl endopeptidase legumain is essential for functional recovery after spinal cord injury in adult zebrafish. PLoS One. 2014;9(4):e95098. Epub 2014 Apr 18 PubMed.
  2. . Structural analysis of asparaginyl endopeptidase reveals the activation mechanism and a reversible intermediate maturation stage. Cell Res. 2014 Mar;24(3):344-58. Epub 2014 Jan 10 PubMed.
  3. . Animal Models of the Sporadic Form of Alzheimer's Disease: Focus on the Disease and Not Just the Lesions. J Alzheimers Dis. 2013;37(3):469-74. PubMed.

Primary Papers

  1. . Cleavage of tau by asparagine endopeptidase mediates the neurofibrillary pathology in Alzheimer's disease. Nat Med. 2014 Nov;20(11):1254-62. Epub 2014 Oct 19 PubMed.