A cat might be the last thing a sickly AD mouse needs, particularly if it is CatB (cathepsin B), a papain-family lysosomal cysteine protease suspected of contributing to the production of amyloid-β (Aβ). But in an ironic twist, a new paper reveals an entirely different CatB/mouse relationship. Li Gan and colleagues at the Gladstone Institute of Neurological Disease and the University of California at San Francisco report that CatB not only degrades soluble Aβ, but attacks fibrillar forms as well, a rare talent among the proteases that might clear Aβ in vivo. The work appears in this week’s Neuron.

The original idea that cathepsin B (CatB) might be a secretase stems from its location in endosomes and lysosomes where Aβ is produced. In addition, CatB has been found in enlarged endosomes and neuritic amyloid plaques in AD brain. To test the secretase hypothesis, first author Sarah Mueller-Steiner and colleagues crossed their human APP mouse with CatB knockouts (see ARF related news story). If CatB were a secretase, they reasoned, the hAPP/CatB-/- mice should have fewer plaques than their cathepsin-replete littermates. Instead, the researchers found the precise opposite—young mice lacking CatB displayed more, not fewer, amyloid plaques in their hippocampus and cortex. The increased plaque load was accompanied by a small increase in total Aβ and Aβ1-42 levels which did not reach statistical significance. However, the relative ratio of Aβ1-42 compared to total Aβ was significantly elevated. The mice also had increased neuropathology, as indicated by decreased staining for the synaptic calcium binding protein calbindin: the same group previously showed that calbindin loss correlates tightly with cognitive decline in this strain of mice (see ARF related news story). When the authors suppressed CatB in cultured primary mouse neurons, they found that levels of both Aβ1-42 and total Aβ increased, while boosting the protease had the opposite effect. No effect was seen on APP processing, suggesting that CatB influenced Aβ degradation.

Cathepsin B has been reported to localize to plaques (reviewed in Nixon and Cataldo, 2006), so the researchers looked for the enzyme in the brains of hAPP mice. They found that in older hAPP mice, cathepsin was associated with 20 percent of all plaques and 70 percent of mature neuritic plaques, which contained fibrillar Aβ, as indicated by thioflavin S staining. In younger mice, most plaques, both diffuse and neuritic, were cathepsin-positive. They also found the enzyme in both neurons and activated glia surrounding the plaques, and that all three cell types—neurons, astrocytes and microglial cells—secreted the enzyme in culture. In neurons, CatB localized to the lysosomal/endosomal compartment.

Could induction of CatB represent a protective mechanism against Aβ toxicity? Aβ1-42 induces CatB mRNA in microglia (Gan et al., 2004), and when the authors added Aβ1-42 to N2A neuronal cells in culture, they also observed an increase in CatB mRNA and activity. Aβ1-40, however, had no effect on neuronal CatB unless it was pre-aggregated. Consistent with these in vitro results, the investigators observed higher CatB activity in the hippocampus of young and middle-aged hAPP mice. This induction might fail with aging, as they did not see upregulation of CatB in older hAPP mice. Their results are consistent with the idea that defects in Aβ clearance with aging might underlie amyloid accumulation.

In vitro studies showed that CatB clips Aβ1-42 at its C-terminus, leaving the less amyloidogenic, and therefore less toxic, Aβ1-40, Aβ1-38, and Aβ1-33 peptides as products. The enzyme appeared to cleave Aβ1-42 sequentially, first to Aβ1-40, then to Aβ1-38. Higher concentrations of enzyme yielded Aβ1-33, but did not completely degrade the peptide. Interestingly, CatB also cleaved aggregated Aβ into less fibrillogenic species, eliminating the high molecular mass assemblies that are detected on polyacrylamide gels. Cleavage by CatB was higher at pH 6, the conditions inside lysosomes where it might encounter Aβ. But the enzyme also degraded Aβ at pH 7, suggesting it could work outside cells as well.

To ask the all-important question of whether CatB can reduce established plaques in hAPP mice, the researchers used lentiviral vectors to express the enzyme in the brains of 12-15-month-old hAPP mice. After 3 weeks, they detected a boost in CatB activity, and reduction in plaques. CatB was as effective at reducing plaques as a neprilysin construct introduced in the same way, and was more effective than neprilysin at removing thioflavin S-positive plaques, suggesting its superior ability to remove aggregated Aβ. Recently, matrix metalloprotease 9 was also shown to degrade Aβ fibrils in vitro (Yan et al., 2006); this capacity is not shared by Aβ degraders—insulin-degrading enzyme, neprilysin, and plasmin.

In an accompanying commentary, Greg Cole and Sally Frautschy of the University of California at Los Angeles write that the identification of a major protease responsible for intracellular and extracellular clearance of Aβ assemblies is full of promise, and they discuss the possibility of using this information to therapeutic advantage. “For example, the authors observed that microglia secrete abundant CatB. This may be regulated by their activation state, which, for example, can be upregulated by the amyloid vaccine or anti-amyloid antibodies, currently one of the most effective methods of reducing preformed thioflavin S-labeled amyloid plaques,” they write. But they also caution that CatB might contribute to neurotoxic effects of microglia. The first step must be finding out if CatB plays a similar role in humans as it does in the mouse.—Pat McCaffrey


  1. The study by Mueller-Steiner and colleagues firmly establishes cathepsin B as a biologically relevant Aβ-degrading enzyme. The results add to mounting evidence that the major pathways to the lysosome, endocytosis, and autophagy are important in APP processing for Aβ generation and degradation, and that dysfunction in these pathways early in Alzheimer disease promotes β-amyloidogenesis and neurodegeneration (Nixon and Cataldo, 2006). The authors propose that insufficient CatB activity may contribute to AD, although expression of CatB and other lysosomal hydrolases, including another potential Aβ-degrading cathepsin (CatD), increases in AD and AD models according to their data and earlier work. CatB function could be insufficient, however, if enough of the protease fails to reach Aβ-generating compartments of cells, or its action is impeded once it arrives. This may well be the case in AD brain where autophagy is markedly impaired. Autophagic vacuoles (AVs) accumulate in huge numbers within dystrophic neurites, reflecting their incomplete clearance from neurons. This clearance normally occurs when autophagosomes fuse with lysosomes, and cathepsins efficiently digest autophagic substrates.

    We have shown that Aβ is generated in AVs during autophagy (Yu et al., 2005). Autophagy-generated Aβ is normally degraded efficiently in lysosomes, but in AD, this Aβ accumulates within AVs that collect massively in the AD brain. If, as we suspect, autophagosomal-lysosomal fusion, or subsequent steps in autolysosome maturation, are impaired/slowed in this condition, even elevated levels of CatB would be rendered functionally “insufficient.” Elevating CatB levels further, as proposed, might overcome the degradative impasse, although it is worth noting that cathepsin overexpression in some systems promotes cell death after injury. Also, modulating one cathepsin significantly affects the levels of other hydrolases in the lysosome. Strategies to reverse the autophagy block, while more challenging a prospect at present, may have additional merit.

    As a sidebar, the data in the report do not formally exclude an "APP secretase" role for CatB. They do show that CatB modulations in vivo have a net effect in favor of degradation of Aβ over any (putative) formation process, but it does not follow from the data that degradation is the only action of CatB on APP metabolism or even the only relevant one. CatD, for example, is reported to have both Aβ-degrading activity and "APP secretase" activity in vitro.


    . Macroautophagy--a novel Beta-amyloid peptide-generating pathway activated in Alzheimer's disease. J Cell Biol. 2005 Oct 10;171(1):87-98. PubMed.

    . Lysosomal system pathways: genes to neurodegeneration in Alzheimer's disease. J Alzheimers Dis. 2006;9(3 Suppl):277-89. PubMed.

  2. One highlight of this past year in Alzheimer disease research has been the flood of interest and activity around Aβ proteolysis. Several new Aβ-degrading enzymes have been discovered, and others already known have stood tests of increased experimental scrutiny. This report by Sarah Mueller-Steiner and colleagues presents cathepsin B (CatB) as a new member of the Aβ-degrading enzyme family, demonstrating that CatB is capable of degrading fibrillar synthetic Aβ as well as clearing Aβ plaques in vivo.

    Previous investigation has linked CatB to Alzheimer disease (AD) by several mechanisms. Extracellular CatB immunoreactivity and enzymatic activity, not present in normal brains, have been demonstrated at the sites of amyloid plaque deposition in human brain (Cataldo et al.). This new paper similarly identifies CatB in association with plaques, specifically enriched at thioflavin S-positive plaques, and demonstrates CatB expression by multiple cell types in the brain. An APP transgenic combined with CatB knockout mouse had an impressive two- to fourfold elevation in plaque deposition, though no change in absolute levels of Aβ as determined by ELISA. The ability of CatB to promote plaque clearance was further demonstrated by viral CatB overexpression in APP transgenic mice.

    There has been some controversy over the role of CatB in APP processing, as some studies have found CatB to directly cleave APP, functioning like β-secretase to elevate Aβ levels. By showing that CatB deficiency in vivo produces no change in full-length APP, APP C-terminal fragments, and APPs, Mueller-Steiner and colleagues provide convincing evidence that CatB is not a prominent APP processing enzyme.

    In relation to other human CNS diseases, there has been significant interest in CatB for its upregulation in brain tumors (Rempel et al., 1994), likely contributing to metastasis. CatB activity has also been shown to be elevated in the CSF of patients with inflammatory neurologic diseases (Nagai et al., 2000). It is unclear if induction of CatB expression is a protective or toxic response in the disease process, but several studies propose that CatB is required for microglial-mediated neurotoxicity (Gan et al., 2004), and that even CatB by itself is neurotoxic (Kingham et al., 2001). Certainly, more work at the basic research level should be done before promoting CatB as a potential therapeutic.

    The findings of this report raise some interesting questions and possibilities. First, why does the significant plaque elevation in APP-overexpressing and CatB-deficient mice not change absolute Aβ levels? Is CatB acting by sequestering the most aggregation-prone Aβ, and does preventing plaque deposition raise levels of soluble oligomers? Second, what mechanism targets secreted CatB to thioflavin S plaques? Is it simply the age of the deposit or the actual structure of the fibrils that matters? Third, this is another example of existing plaques being removed through a proteolytic mechanism. Some evidence suggests that plaques cannot be cleared by the brain after suppressing APP transgene expression (Jankowsky et al., 2005), which may indicate that elevating Aβ clearance mechanisms by proteolytic, transport, antibody, or other means must be enlisted to therapeutically lower Aβ in the brain. Fourth, this finding may increase interest in the analysis of the CatB genetic locus for association to AD, as happened for IDE. Finally, can CatB be used as a therapeutic tool itself? Some of the biological details of CatB’s regulation are known, and if CatB upregulation is not itself toxic, perhaps these data could be exploited as a therapeutic approach to Alzheimer disease.


    . Cathepsin B expression and localization in glioma progression and invasion. Cancer Res. 1994 Dec 1;54(23):6027-31. PubMed.

    . Cystatin C and cathepsin B in CSF from patients with inflammatory neurologic diseases. Neurology. 2000 Dec 26;55(12):1828-32. PubMed.

    . Identification of cathepsin B as a mediator of neuronal death induced by Abeta-activated microglial cells using a functional genomics approach. J Biol Chem. 2004 Feb 13;279(7):5565-72. PubMed.

    . Microglial secreted cathepsin B induces neuronal apoptosis. J Neurochem. 2001 Mar;76(5):1475-84. PubMed.

    . Persistent amyloidosis following suppression of Abeta production in a transgenic model of Alzheimer disease. PLoS Med. 2005 Dec;2(12):e355. Epub 2005 Nov 15 PubMed.

  3. This is an extensive and very carefully controlled study in which the authors demonstrate that cathepsin B (CatB) can readily degrade Aβ in vivo and in vitro. At pH 6.0, CatB similarly degrades both non-aggregated and aggregated synthetic Aβ1-42, while at pH 7.0, CatB effectively degrades aggregated but not non-aggregated Aβ. At both pH 6.0 and 7.0, CatB acts as a carboxypeptidase, trimming Aβ1-42 first to Aβ1-40 and then Aβ1-38. But CatB can also act as an endopeptidase cleaving Aβ to generate Aβ1-33. In vivo viral expression of CatB caused a substantial decrease in amyloid burden, with lentiviral expression of CatB having an effect comparable to lentiviral NEP, but with the former much more effective at decreasing the number of compacted plaques.

    Consistent with this selectivity towards aggregated Aβ, over 70 percent of compacted (thioflavin S-positive) plaques from old (16-20-month) APP transgenic mice co-stain for CatB, and in cell culture, application of aggregated Aβ caused an increase in CatB transcription and activity. Conversely, knockout of CatB in an APP overexpressing background increased amyloid burden as assessed by immunostaining with 3D6, thioflavin S staining, and ELISA of guanidine-HCl-extracted Aβ. Interestingly, while the level of Aβ1-42 was significantly decreased in CatB Kos, the level of total Aβ was not significantly decreased. The latter reflects accumulation of some of the partially degraded Aβ1-42, that is, Aβ1-40, Aβ1-38, and Aβ1-33.

    Together, these results indicate that CatB plays an important role in the regulation of aggregated Aβ levels and suggests that a deficit in CatB activity may facilitate Aβ accumulation. However, in Alzheimer disease it is not apparent that there is a deficit in CatB activity. Clearly, this report will prompt further investigations to address this point, and may encourage new studies to look at ways of stimulating CatB activity in vivo. But in regard to the upregulation of CatB, further studies would be well advised to consider what effect newly solubilized Aβ1-33, 1-38, and 1-40 might have on cognition and on vascular deposition. As with everything in life, the utility of upregulating CatB will largely depend on timing. Increased CatB activity would be most beneficial at intervals proceeding profound amyloid deposition, whereas upregulation at later time points could be counter-productive.

  4. The conclusion of this paper, that cathepsin B does not have β-secretase activity for human APP (hAPP), is misleading. The transgenic mice used in this study express hAPP containing Swedish mutations at the β-secretase site (hAPPswe). hAPPswe is an exceedingly rare form of hAPP. The vast majority of Alzheimer patients express hAPP containing the wild-type β-secretase site (hAPPwt). The altered β-secretase site in the Swedish mutant form can change the β-secretase activity compared to that utilized for the wild-type β-secretase site. Thus, in agreement with our findings, the data in this paper support the narrow conclusion that cathepsin B does not have β-secretase activity for the rare APPswe. Unfortunately, the failure to make this point clearly in the paper has led some readers to erroneously conclude that cathepsin B has no β-secretase activity for hAPP generally.

    In fact, cathepsin B has excellent β-secretase activity for the wild-type β-secretase site required for production of β-amyloid (Aβ). We recently showed that cathepsin B inhibitors reduce both Aβ production and β-secretase activity in guinea pigs expressing APP containing the human wild-type β-secretase site, reported by Hook et al. (2007). Cathepsin B inhibitors resulted in substantial reduction of Aβ peptide levels in brain by 50 to 70 percent. These inhibitors decreased brain levels of CTFβ derived from APP by processing at the β-secretase site, which indicated that the cathepsin B inhibitors reduced β-secretase. It has been reported that cathepsin B activity in Aβ-containing secretory vesicles does not cleave the Swedish mutant site (Hook et al., 2002). Therefore, the finding that cathepsin B is not involved in processing Swedish mutant APP, as reported by Mueller-Steiner et al. (2006), was to be expected. Importantly, however, results by Hook et al. (2007) show that cathepsin B remains a critical parameter for generating neurotoxic Aβ from wild-type APP that is relevant to the majority of Alzheimer disease. Thus, cathepsin B should continue to be considered in the regulation of Aβ production in Alzheimer disease research.


    . Cysteine protease inhibitors effectively reduce in vivo levels of brain beta-amyloid related to Alzheimer's disease. Biol Chem. 2007 Feb;388(2):247-52. PubMed.

    . Beta-amyloid peptide in regulated secretory vesicles of chromaffin cells: evidence for multiple cysteine proteolytic activities in distinct pathways for beta-secretase activity in chromaffin vesicles. J Neurochem. 2002 Apr;81(2):237-56. PubMed.

    View all comments by Greg Hook
  5. Dr. Greg Hook commented that the Swedish mutation at the β-secretase site in hAPP may have prevented β-secretase activity of cathepsin B (CatB). We addressed this possibility in our original article (1). Primary cortical neurons from CatB–/– mice and CatB+/+ littermate controls were transduced with an adenoviral vector expressing wild-type hAPP. Aβ42 levels were significantly higher in the supernatants from CatB–/– neurons than CatB+/+ neurons (Figure 3). If CatB acted as β-secretase for wild-type hAPP, CatB–/– neurons would have produced less, not more, Aβ42. Therefore, our results suggest that neuronal CatB is unlikely to exert β-secretase activity regardless of whether the Swedish mutation is present or not. This conclusion is perfectly consistent with the fact that no Aβ has been detected in the brains of BACE1-deficient mice (2).

    Recently, Hook et al. (3) reported that the cysteine protease inhibitor E64d or CA074Me significantly reduced Aβ levels in the brains of guinea pigs. However, neither inhibitor is specific for cathepsin B. Previous studies showed that CA074Me inactivates both CatB and cathepsin L within murine fibroblasts, demonstrating that it is not a specific inhibitor of CatB in living cells (4). E64d is a potent and irreversible inhibitor of calpains and cathepsins (B and L) (5). Since E64d and CA074Me are non-specific, their Aβ-reducing effects cannot be specifically attributed to CatB.

    In our study, to specifically increase or decrease CatB activities, we used a variety of genetic approaches, including genetic deletion in knockout mouse models, RNA interference in primary neurons, and lentiviral-mediated overexpression in both primary neurons and in mouse models (1). With reassuring consistency, all data obtained with these approaches support the conclusion that CatB reduces human Aβ levels in vivo.

    Lastly, I would like to note that Dr. Hook co-founded American Life Science Pharmaceuticals, Inc., a company that has been awarded an SBIR to develop inhibitors of CatB as a potential therapy for AD (3).


    . Antiamyloidogenic and neuroprotective functions of cathepsin B: implications for Alzheimer's disease. Neuron. 2006 Sep 21;51(6):703-14. PubMed.

    . Mice deficient in BACE1, the Alzheimer's beta-secretase, have normal phenotype and abolished beta-amyloid generation. Nat Neurosci. 2001 Mar;4(3):231-2. PubMed.

    . Cysteine protease inhibitors effectively reduce in vivo levels of brain beta-amyloid related to Alzheimer's disease. Biol Chem. 2007 Feb;388(2):247-52. PubMed.

    . CA-074, but not its methyl ester CA-074Me, is a selective inhibitor of cathepsin B within living cells. Biol Chem. 2002 Jul-Aug;383(7-8):1305-8. PubMed.

    . Irreversible inhibitors of serine, cysteine, and threonine proteases. Chem Rev. 2002 Dec;102(12):4639-750. PubMed.

    View all comments by Li Gan

Make a Comment

To make a comment you must login or register.


News Citations

  1. The Limits of Compensation: Double Loss of Proteases Causes Massive Neuronal Death
  2. Calbindin Study: Is Calcium the Molecular Handle on Dysfunction in AD?

Paper Citations

  1. . Lysosomal system pathways: genes to neurodegeneration in Alzheimer's disease. J Alzheimers Dis. 2006;9(3 Suppl):277-89. PubMed.
  2. . Identification of cathepsin B as a mediator of neuronal death induced by Abeta-activated microglial cells using a functional genomics approach. J Biol Chem. 2004 Feb 13;279(7):5565-72. PubMed.
  3. . Matrix metalloproteinase-9 degrades amyloid-beta fibrils in vitro and compact plaques in situ. J Biol Chem. 2006 Aug 25;281(34):24566-74. PubMed.

Further Reading

Primary Papers

  1. . Antiamyloidogenic and neuroprotective functions of cathepsin B: implications for Alzheimer's disease. Neuron. 2006 Sep 21;51(6):703-14. PubMed.
  2. . Cat and mouse. Neuron. 2006 Sep 21;51(6):671-2. PubMed.