The secretase BACE1 kicks off the process of snipping Aβ from amyloid precursor protein (APP), but what else does this protease do? The question is not merely academic. After a decade of frustration, BACE1 inhibitors are at last moving into clinical trials for Alzheimer’s disease. If scientists knew more about the biology of the secretase, they might know what adverse effects to watch out for in trial participants. Two new papers now provide some clues. In the June 22 EMBO Journal online, researchers led by Stefan Lichtenthaler at the German Center for Neurodegenerative Diseases, Munich, Germany, describe a novel, sensitive method for identifying proteins cleaved by BACE1. Using this protocol, they found some two dozen novel candidates, and validated four.

With a different methodology, researchers led by Bart De Strooper at the University of Leuven, Belgium, came up with a largely overlapping list of likely substrates, as reported online in the June 14 Journal of Biological Chemistry. De Strooper’s group went on to extensively characterize and confirm two of the top hits, the neural cell adhesion molecules L1 and CHL1. Notably, the majority of the proteins on both lists have known roles in cell adhesion or axon guidance. The results suggest that BACE1, which is most highly expressed soon after birth, may play a role in forming connections in the brain, and perhaps even in synaptic plasticity and learning.

“BACE1 has a much broader role in the brain than we anticipated,” Lichtenthaler said. “This gives us an idea of what kind of potential side effects we may want to look for in a clinical trial, and it raises the possibility that we can use secreted proteins as novel biomarkers.”

For many years, it appeared that BACE1 might not be a druggable target, as several pharmacological programs stalled out. However, a new generation of inhibitors now shows good drug properties (see ARF related news story; ARF news story). Currently, pharmaceutical company Eli Lilly, Indianapolis, Indiana, has a compound in a Phase 2 trial, while Merck, Whitehouse Station, New Jersey, has one in Phase 1 (see ARF related news story). AstraZeneca, Wilmington, Delaware, also has a compound in Phase 1. So far, these drugs have passed toxicology and safety studies, but it is always possible that subtle problems may crop up with long-term use.

Previous studies looked for BACE1 substrates by a candidate approach or by screening overexpression systems that may not reflect physiological conditions (see ARF related news story on Hemming et al., 2009). To screen using primary neuronal cells, Lichtenthaler’s group developed a novel proteomics method called secretome protein enrichment with click sugars (SPECS). These are chemically modified sugars that behave like the real thing, and “click” together with a target molecule to simplify their detection and purification. Because the majority of secreted and transmembrane proteins are glycosylated, the researchers were able to capture fragments that incorporated those sugars and were shed into culture media. They then identified proteins in this “secretome” by mass spectrometry. The advantage of this method is that secreted proteins can be easily separated from the more abundant serum proteins present in the media, which can otherwise overwhelm the signal, Lichtenthaler said.

To narrow down the field to BACE1 substrates, first author Peer-Hendrik Kuhn compared the secretomes of wild-type mouse neuronal cultures incubated with and without a BACE1 inhibitor. Kuhn and colleagues identified 34 candidate substrates, of which 23 are novel. For in-vivo follow-up, the authors chose four proteins: L1, CHL1, the adhesion molecule contactin-2, and seizure protein 6, which has been genetically linked to epilepsy. They verified that BACE1 knockout mice and mice treated with a BACE1 inhibitor poorly cleave these proteins compared to control animals. In future work, Lichtenthaler plans to validate more of the candidates and look more closely at their function, he told Alzforum.

In the JBC paper, first author Lujia Zhou also used primary neuronal cultures. Zhou and colleagues got around the problem of serum proteins by incubating the cultures for four hours in serum-free media, with or without a BACE1 inhibitor. (The disadvantage of serum-free media, however, is that it can stress neurons.) After putting shed and secreted proteins through a mass spectrometer, the authors identified 13 candidate substrates, 10 of them novel, and most in agreement with Lichtenthaler’s list. They chose the neural cell adhesion molecules (NCAMs) L1 and CHL1 for further analysis. BACE1 knockout mice and mice treated with a BACE1 inhibitor did not cleave these proteins as well as did control mice, the authors report. The authors also determined the BACE1 cleavage sites in both proteins. Coauthor Soraia Barão told Alzforum they plan to do more functional studies of these NCAMs, for example, examining the part they play in producing the phenotype of the BACE1 knockouts.

What role does BACE1 cleavage of these proteins play in normal physiology? BACE1 knockout mice have synaptic problems, learning deficits, and are prone to epileptic seizures (see ARF related news story; ARF news story). Intriguingly, De Strooper and colleagues showed that L1 and CHL1 localize to synapses, putting them in the right place to affect plasticity. In addition, two recent functional studies showed that BACE1 activity is required for axon guidance in the olfactory system (see ARF related news story on Rajapaksha et al., 2011; Cao et al., 2012). The new proteomics screens propose candidate molecules that might mediate these processes.

The data suggest that BACE1 cleavage of L1 and CHL1 could play a role in synaptic plasticity and learning, said Ciaran Regan at University College Dublin, Ireland, who studies NCAMs. These adhesion molecules are known to affect synaptic plasticity (see ARF related news story; Dityatev et al., 2000). Regan noted that endocytosis of cell adhesion molecules occurs about two to six hours after training (see Bailey et al., 1992), just before new synapses form. “It suggested to us that perhaps this endocytosis is loosening the network, allowing it to become more plastic,” he said. Cleavage of adhesion molecules may accomplish the same thing, he speculated. “You could imagine it as freeing up the system, allowing more movement because those anchors are gone.” In this way, cells can remodel synaptic contacts in response to stimuli. To nail down a relationship between the secretase and learning, the next step might be to look for modulation of BACE1 activity or expression in the early post-training period, Regan suggested.

Whether companies developing BACE1 inhibitors are concerned about these new substrates cropping up is unclear. Barão noted that the new screens open the door for functional studies, and may help scientists determine how far they can safely inhibit the protease. Most pharmaceutical agents only partially inhibit the secretase, and do not produce all the side effects seen in BACE1 knockouts, Lichtenthaler pointed out. That may be because other proteases compensate for the loss of BACE1, at least for some substrates, he suggested. It is also not clear how essential the protease is in adult brain, as it appears to be most active during early development. However, Lichtenthaler’s group verified that their four validated candidates were cleaved equally well in adult brain as in early postnatal.—Madolyn Bowman Rogers


  1. I really think these are interesting data, as they provide another possible neuroplastic mechanism to potentially explain the neurocognitive deficits associated with AD. In essence, it is proposed that BACE cleaves cell surface cell adhesion molecules, and in so doing, allows modification of the neural circuitry.

    Most previous considerations [of synaptic changes in plasticity] have been related to endocytosis. Eric Kandel demonstrated this in vivo in the early 1990s, and later on we demonstrated the same in vivo. How endocytosis relates to Aβ-induced cognitive deficits is, of course, the next crucial step.

    In this regard, we know that APP is internalized at two to six hours following a learning paradigm, and that this precedes the time when Aβ exerts its synaptotoxic effects on the newly formed learning-associated synapses. It may very well be that endocytosis or proteolytic release, as shown in the present paper, of cell adhesion molecules may facilitate a loosening of synaptic connections in the neural network and allow further remodeling of the circuitry—for good or for bad.

    The next steps forward have to be most interesting.

    View all comments by Ciaran Regan

Make a Comment

To make a comment you must login or register.


News Citations

  1. Paper Alert: First BACE Trial Published—New Inhibitor On Deck
  2. Getting to First BACE: BACE1 Inhibition Takes a Step Forward
  3. Keystone: Therapies Around ApoE—Has Their Time Come?
  4. Dual Secretase Approach Shows Promise in AD Mice
  5. BACE Knockouts Show More Synaptic Troubles
  6. Off BACE: Lack of Secretase Causes Seizures in Knockout Mice
  7. Not the Usual Suspects: Tracking BACE Inhibition, Axon Role
  8. Peptide Sparks Synaptic Plasticity, Improves Memory in Rodents

Paper Citations

  1. . Identification of beta-secretase (BACE1) substrates using quantitative proteomics. PLoS One. 2009;4(12):e8477. PubMed.
  2. . The Alzheimer's β-secretase enzyme BACE1 is required for accurate axon guidance of olfactory sensory neurons and normal glomerulus formation in the olfactory bulb. Mol Neurodegener. 2011;6:88. PubMed.
  3. . The precision of axon targeting of mouse olfactory sensory neurons requires the BACE1 protease. Sci Rep. 2012;2:231. PubMed.
  4. . Synaptic strength as a function of post- versus presynaptic expression of the neural cell adhesion molecule NCAM. Neuron. 2000 Apr;26(1):207-17. PubMed.
  5. . Serotonin-mediated endocytosis of apCAM: an early step of learning-related synaptic growth in Aplysia. Science. 1992 May 1;256(5057):645-9. PubMed.

External Citations

  1. Phase 2 trial
  2. Phase 1
  3. Phase 1

Further Reading

Primary Papers

  1. . Secretome protein enrichment identifies physiological BACE1 protease substrates in neurons. EMBO J. 2012 Jul 18;31(14):3157-68. PubMed.
  2. . The neural cell adhesion molecules L1 and CHL1 are cleaved by BACE1 protease in vivo. J Biol Chem. 2012 Jul 27;287(31):25927-40. PubMed.