A lively platter of slide talks on APP processing and β-secretase (BACE) activities yielded a few interesting tidbits on the localization and regulation of β cleavage at this year’s annual meeting of the Society for Neuroscience in Atlanta, Georgia. The presentations also left the impression that cathepsin B, most recently in the news for digesting fibrillar Aβ (see ARF related news story), may yet have a role to play in the production of Aβ, as a β-secretase candidate protease.
Two talks focused on the localization and fluctuations of the BACE protein in brain. In the first, Matthew Kennedy and colleagues at Schering-Plough Research Institute, Kenilworth, New Jersey, described a radio-labeled small molecule BACE inhibitor that specifically binds BACE in mouse tissues—there was no binding to brains from BACE knockout mice. In brain sections of 6-week-old mice, they detected inhibitor binding throughout the CNS, with higher levels in the CA2/3 regions of the hippocampus and in the olfactory bulb. Using transgenic CRND8-APPswe/ind AD mice that show very early plaque deposition (10-12 weeks of age), they found that BACE levels increase after plaque deposition. In the mice, they saw a 50 percent increase in BACE inhibitor binding only after 24-30 weeks. They also saw a change in localization of the protein to high-staining foci in the cortex, which appear to colocalize with plaques.
The delay in upregulation of BACE is at odds with data presented by Jie Zhao in Robert Vassar’s lab on a new monoclonal antibody they developed to BACE. (See our ARF Madrid story for Vassar’s talk on the antibody, which the lab produced by immunizing BACE knockout mice.) Unlike all commercial antibodies tested, which either recognized multiple proteins on Western blots, or stained the brains of BACE knockout just as well as wild-type mice, the new antibody was monospecific for BACE. Zhao’s immunostaining results revealed increased BACE in the 5XFAD APP/S1 (see ARF related news story) and Tg2576 transgenic mice around amyloid plaques. The BACE elevation occurred in parallel with the rise in amyloid burden in 5XFAD mice. Increases were also seen in brain of human AD patients.
Consistent with the labeled inhibitor experiments of the Schering-Plough group, Zhao showed colocalization of BACE around amyloid plaques. Confocal views revealed BACE in an outer ring of Aβ40 that surrounded the inner, Aβ42-rich plaque. From these and other recent results, Vassar and colleagues hypothesize that energy insufficiency stress, caused, for example, by vascular hypoperfusion, induces an increase in BACE. The subsequent amyloid deposition may itself further activate BACE. Peri-plaque BACE was found in association with neuronal markers.
In a related presentation, Sarah Cole, also at Northwestern, showed some early results from crossing BACE knockouts with the 5XFAD APP/S1 strain. When Cole and colleagues looked at 4-month-old BACE-heterozygous mice, they saw a normalization of early markers of neuronal stress, namely p25 protein and phospho p38 MAP kinase. Until the mice get older, the researchers won’t know if the pathology is actually prevented, or just delayed.
Is BACE the One and Only β-secretase?
The observation that knocking out BACE1 shuts down the constitutive production of Aβ from cultured neurons and the accumulation of Aβ in FAD mouse models has been taken as proof that BACE1 is the predominant enzyme responsible for β cleavage of APP. But is BACE really the only β-secretase? Work presented by Vivian Hook, University of California, San Diego, made a strong case for cathepsin B as an additional β-secretase with a role to play in the activity-regulated secretion of Aβ.
Hook and colleagues have proposed previously that the majority of Aβ is released from cells via regulated secretion, as in the activity-dependent production and release of Aβ at synapses. According to their hypothesis, only a small amount of extracellular Aβ would derive from constitutive secretion, the BACE-dependent pathway through which Aβ makes its way into the conditioned medium of cultured neurons. The researchers, who have long studied the processing and release of neuropeptides, proposed this model based on their studies of adrenal chromaffin cells, which are widely used as a model of regulated secretion of neurotransmitters. In these cells, they have identified cathepsin B as the major β-secretase activity in secretory vesicles. BACE1 is present in the vesicles but accounts for only a small part of the total processing activity during KCl or nicotine-stimulated secretion. From these results, Hook and colleagues have proposed that cathepsin B could be the main β-secretase in neurons (Hook et al., 2002; Hook and Reisine, 2003; Hook et al., 2005).
One reason why cathepsin B has been overshadowed by BACE could be that, as Hook and colleagues demonstrated several years ago, the enzyme prefers to cleave wild-type APP, and displays much lower activity toward the altered site created by the Swedish mutation. The mutant sequence is preferred by BACE, and pops up in many of the FAD animal models which have been used to study the role of BACE in vivo. For studying β-cleavage inhibitors, Hook and coworkers favor the guinea pig model, which has a wild-type APP sequence that matches the β-secretase cleavage site in humans.
Hook used that animal to look at the role of cathepsin B in APP processing in the CNS. In her talk, she showed that infusion of selective cathepsin B inhibitor CA-074Me into the brains of guinea pigs reduced Aβ40 and Aβ42 levels by approximately one-half. In a second talk, Hook showed that another cathepsin B inhibitor, loxistatin, had the same effect after direct intracerebroventricular infusion. She saw decreased production of Aβ and the C-terminal β-secretase fragment of APP in the brain and in isolated synaptosomes. These results suggest that cathepsin B does participate in Aβ production in vivo as a β-secretase, and that inhibitors may have some potential as Aβ-lowering agents. This work is now in press in Biological Chemistry, Hook said.
A listener was quick to ask how Hook could reconcile her results with the recent paper from Li Gan showing that AD mice with the cathepsin B gene knocked out have worse plaque pathology than mice with the enzyme (see ARF related news story). In that study, the researchers found no evidence that cathepsin B acted as a secretase, but in fact found that it was able to degrade Aβ fibrils.
Hook pointed out that the AD mouse model used in that paper (the hAPP J20 line) incorporated the Swedish mutation, which could explain why the researchers saw no evidence for APP processing by cathepsin B. In animals with wild-type human APP, there could be a different outcome. Clearly, there is a lot more work to be done, but the data so far suggest that where Aβ production is concerned, BACE may be just one part of the story.—Pat McCaffrey.