27 July 2006. This is part 1 of our 3-part series. Also see part 2 and part 3, or download PDF.
Advances in understanding BACE1, the β-secretase enzyme relevant to Alzheimer disease, stood out as a notable trend at the 10th International Conference on Alzheimer’s Disease and Related Disorders (ICAD), held from July 15 to 20 in Madrid. At ICAD, bits and pieces of news were rustling up a fresh breeze in the air that came as a welcome change after a doldrums of sorts. Hope that this enzyme would serve as a new drug target first rose when researchers led by Martin Citron cloned it in 1999. At first blush, BACE1 looked like a safer target than its big brother γ-secretase, because knockout mice generated by Robert Vassar and several other independent groups all appeared largely normal. When Jordan Tang’s group solved the BACE1 crystal structure a year later—why, it seemed that all that was left to do was for clever drug designers to get busy and, presto, serve up a suitable small molecule drug. But the going got tough when BACE1 proved to be a recalcitrant drug target. What’s more, basic scientists began to whisper that BACE1 might not be as straightforward a target as initially thought. In Madrid, researchers for the first time presented a potent BACE1 inhibitor, fledgling immunotherapy approaches, and new data on its biology and potential as a biomarker. Read on for summaries of a plenary lecture and some of the 48 other presentations on BACE1. As always, Alzforum encourages presenters and attendees to amend our selected notes with their own.
In the plenary reviewing current knowledge on BACE, Citron, of Amgen in Thousand Oaks, California, first recapped that BACE1 and 2 are single transmembrane aspartyl proteases. They are related to the HIV retropepsin, which is a thoroughly studied drug target. One reason why BACE1 is less well understood, besides having been known for only six years, is that it undergoes numerous post-translational modifications that influence its activity in still-mysterious ways, Citron noted. Some things are known, however. BACE2 appears to play little, if any, role in AD pathogenesis. Cell biologists have pieced together that BACE1 traffics through the secretory pathway, moving from the trans-Golgi network to the plasma membrane, where it becomes pinched off into to endosomes and from there is retrieved again for further transport. BACE1 is thought to cleave APP most readily in endosomes and the trans-Golgi network, said Citron. It forms homodimers, and appears to do its work in lipid rafts.
One of the hottest questions in BACE research these days is whether BACE1 is upregulated in AD, and whether this upregulation comes as an epiphenomenon in late-stage AD or plays an early role and contributes to pathogenesis. Numerous reports have found that BACE1 activity increases with age and even more so in AD. Yet no familial AD loci containing BACE1 polymorphisms, much less AD-causing mutations in the BACE1 gene, have been found. This raises the underlying question of what regulates BACE1 expression. Many interactions of BACE1 and other proteins are on the map, including with reticulons, GGA proteins, and sorLa, but which ones participate in AD pathogenesis remains a puzzle. Other research has implicated BACE1 in an inflammatory feed-forward loop, and energy depletion as occurs in an atherosclerotic, underperfused brain is also thought to trigger BACE1.
Tang’s BACE1 crystal structure, and Amgen’s, too, showed that the active site comprises eight subsites, and that it would be difficult for a single small molecule drug to touch them all. Studying which of these sites a drug needs to hit has taken up much of the intervening time since 2000, Citron said. Only clinical trials will show whether BACE1 can be inhibited safely. In the interim, basic research has put potential concerns to watch for on the drug developers’ radar screen. Potential risks include that interfering with APP metabolism could narrow the therapeutic wiggle room if indeed Aβ turns out to perform an essential biological function, for example, in synaptic activity. Moreover, BACE1 has proven to cleave other substrates more readily than APP, and any physiological consequences of inhibiting these reactions remain unclear at present. The list of published substrates includes ST6Gal I, Psgl-1, LRP, and neuregulin-1 (see ARF related Madrid story).
BACE1 knockout mice are fertile, viable, and appear to age normally, but little is known about how they fare when stressed while aging. Some studies have identified subtle memory deficits, though this issue remains controversial, and some BACE1/2 double knockout mice tend to die early. In Madrid, Alex Harper and colleagues from GlaxoSmithKline in Harlow, Great Britain, reported that BACE1 knockouts had trouble gaining weight with age.
Removing BACE1 protected the mice against the weight gain usually seen on a high-fat diet. Lack of BACE1 also appeared to increase the mice's insulin sensitivity in the face of a glucose challenge test, pointing to some still-mysterious metabolic role for BACE1. The BACE1 knockout mice also tended to die earlier than did wild-type controls. On the plus side, however, a different safety concern that has been raised about inhibiting APP cleavage by either BACE or its downstream successor γ-secretase appears less worrisome upon further inspection. It concerns a loss of physiological gene expression signaled by the intracellular tail of APP, aka AICD. A few genes, including neprilysin, KAI1, APP itself, or GSK3β, had been implicated as AICD target genes. Yet subsequent studies in different labs have struggled to reproduce these findings, and in Madrid, Sebastian Hébert in Bart de Strooper’s group in Leuven, Belgium, reported that in their hands, too, reducing AICD through secretase inhibition had no major effect on any of those genes (see also Hébert et al., 2006).—Gabrielle Strobel.