Elusive Gamma-Secretase Identified

Model of an inhibitor targeted to g-secretase interacting with presenilin. The background shows an Alzheimer's brain that carried a presenilin mutation, with immunohistochemistry revealing abundant amyloid plaques (red-orange patches). For years researchers have been perplexed by the identity of g-secretase, an enzyme that cuts amyloid precursor protein (APP) into amyloid ß (Aß) fragments that form telltale plaques in the brains of Alzheimer's patients. Now researchers from Merck Rese

Nadia Halim
Jun 25, 2000

Model of an inhibitor targeted to g-secretase interacting with presenilin. The background shows an Alzheimer's brain that carried a presenilin mutation, with immunohistochemistry revealing abundant amyloid plaques (red-orange patches).
For years researchers have been perplexed by the identity of g-secretase, an enzyme that cuts amyloid precursor protein (APP) into amyloid ß (Aß) fragments that form telltale plaques in the brains of Alzheimer's patients. Now researchers from Merck Research Laboratories and Harvard Medical School have independently solved the mystery, bringing the field one step closer to a therapeutic approach against Alzheimer's disease (AD). "All of the new and exciting information coming out is in agreement. It looks like presenilin is the long-sought g-secretase," says Dennis Selkoe, professor of neurology at Harvard Medical School.

Presenilin 1 (PS1) has been linked to AD since PS1 was first identified and cloned in 1995.1 PS1 mutations cause a rare form of early-onset familial AD. Bart De Strooper, group leader at the Flanders Interuniversitary Institute for Biotechnology, Leuven, Belgium, tightened the link by showing that neuronal cells derived from PS1-deficient mouse embryos exhibit significantly decreased g-secretase activity.2

A few years later, the labs of Selkoe and collaborator Michael Wolfe, associate professor of neurology at Harvard Medical School, presented some unexpected findings at a neuroscience meeting in Los Angeles and went on to publish them in April 1999.3 The team showed that presenilin is tightly linked to Aß production, and therefore g-secretase activity, in cells. Furthermore, g-secretase has the properties of an aspartyl protease. Upon closer examination the Selkoe and Wolfe labs discovered that presenilin has two aspartates in two of its eight transmembrane domains. When the aspartates are mutated, g-secretase activity significantly decreases. Together, the findings raised the specter that presenilin might itself be g-secretase.

As intriguing as those results were, Selkoe and Wolfe faced considerable skepticism. Researchers wanted to see the biochemical evidence to support the conclusion drawn from genetic and cellular studies.

Steve Gardell, director of biological chemistry at Merck Research Laboratories, and his colleagues took up the challenge. The Merck group used a g-secretase inhibitor identified through whole-cell assays to attack the problem. The researchers concluded that the inhibitor was binding to the active site of g-secretase based on the observed changes in inhibitory potency when the structure of the compound was altered. The inhibitor, which was structurally similar to aspartyl protease inhibitors, was converted into a photoaffinity probe. Such probes are used to label and identify their molecular targets. The Harvard researchers independently took a similar approach with compounds designed by Wolfe to interact specifically with the diaspartyl active site of g-secretase. Both groups identified PS1 as the protein that directly binds to the g-secretase inhibitors.4,5 Interestingly, the inhibitors did not bind to the intact single-chain form of PS1, but specifically labeled the cleaved heterodimer, which is associated with bioactivity.

Charles Glabe, professor of molecular biology and biochemistry at the University of California, Irvine, who is working in Alzheimer's research, agrees that the evidence is compelling. However, some still may be skeptical. The experiment that would prove without a doubt that the two molecules are the same involves making new Aß from APP using pure presenilin. However, this has not worked yet in the hands of many researchers. Other proteins in the cell may be necessary for presenilin to function as a protease. "Right now it is impossible to do the killer experiment because we haven't found the partners of presenilin," explains Selkoe.

Dennis Selkoe
Researchers may also be skeptical because presenilin is like no other known protease, which is the main reason it has taken so long to identify presenilin as g-secretase. The enzyme appears to be a novel aspartyl protease with an intramembranous catalytic machine. "The g-secretase/presenilin story is exciting because this is an unprecedented type of protease that cuts its substrate within the phospholipid bilayer," says Selkoe. It doesn't cut in the aqueous environment of the cytosol or the extracellular fluid, where virtually all other proteases work.

Moving Forward

More than likely researchers will be using presenilin to go beyond the whole-cell assays of the past. "Assays now can be geared toward increasing the number and diversity of inhibitors that we find. We will then use medicinal chemistry to elucidate structure-activity relationships and optimize the properties of these compounds," says Gardell. In fact, Merck has already devised an assay for soluble g-secretase activity.6

Researchers may also test known g-secretase inhibitors for presenilin binding. Inhibitors that don't bind to presenilin might be binding to a molecule that associates with presenilin. Such an inhibitor could then be used to fish out the identity of partner molecules. Selkoe and Gardell indicate that they are looking for the partners that may make up a g-secretase complex. Disrupting such a complex may be another way of inhibiting Aß production.

Secretases are attractive drug targets because they generate Aß peptides. Thus strategies to block the formation of the Aß may in turn block the generation of the plaques and perhaps interfere with the onset and progression of the disease. "In light of recent findings that g-secretase inhibitors interfere with the formation of amyloid plaques in animal AD models, it would be surprising if these compounds did not have a similar effect in man," says Gardell.

Last year researchers at Amgen Inc. of Thousand Oaks, Calif., cloned ß-secretase,7 the other enzyme involved in producing Aß and set the stage for an intensive ß-secretase inhibitor search. Surprisingly, the whole-cell screening assays run over the past several years identified many g- secretase inhibitors, but failed to find many ß-secretase inhibitors. It may be that the ß-secretase inhibitors have trouble crossing the cell membrane or are toxic to the cells. On the other hand, the structure of g-secretase inhibitors may be more amenable to uptake into the cells. Selkoe mentions that he hasn't heard of any ß-secretase inhibitors that are close to clinical trials yet.

Uncertain Outcome

Many researchers believe that Aß buildup occurs years or decades before the first symptoms of Alzheimer's and is essentially the cause of disease. The most powerful evidence comes from the discovery of Alzheimer's genes, which when mutated increase amounts of Aß. Transgenic animal modeling has also supported this theory. "We think that all the cases of Alzheimer's are due to increased levels of Aß in the brain. The vast majority aren't due to presenilin mutations, but there must be other reasons why their Aß builds up," says Selkoe.

However, it is still not clear that lowering Aß levels will control the symptoms of AD. Another camp of researchers thinks that amyloid buildup is not the cause of AD, but just a marker of the disease. If this is the case, g- and ß-secretase inhibitors will not fulfill their therapeutic purpose.

"The first generation of antiamyloid drugs are essentially in hand already," says Selkoe, referring to the g-secretase inhibitors. In fact, Bristol-Myers Squibb, another company with a strong research program in this area, has indicated at meetings that it started a clinical trial in March using a g-secretase inhibitor to treat people with mild Alzheimer's.

The answer to the debate about whether amyloid plaques cause the neurodegeneration in AD or not lies with ß- and g-secretase inhibitors developed for clinical trials. "If such agents block plaque formation once they can get into the brain, as I suspect they will, we may finally be able to [get some answers]," concludes Gardell.S

Nadia S. Halim can be contacted at nhalim@the-scientist.com


1. R. Sherrington et al., "Cloning of a gene bearing missense mutations in early-onset familial Alzheimer's disease," Nature, 375:754-60, 1995.

2. B. De Strooper et al., "Deficiency of presenilin-1 inhibits the normal cleavage of amyloid precursor protein," Nature, 391:387-90, 1998.

3. M.S. Wolfe et al., "Two transmembrane aspartates in presenilin-1 required for presenilin endoproteolysis and g-secretase activity," Nature, 398: 513-7, 1999.

4. Y-M. Li et al., "Photoactivated g-secretase inhibitors directed to the active site covalently label presenilin 1," Nature, 405:689-94, June 8, 2000.

5. W.P. Esler et al., "Transition-state analogue inhibitors of g-secretase bind directly to presenilin-1," Nature Cell Biology, in press.

6. Y-M. Li et al., "Presenilin 1 is linked with g- secretase activity in the detergent solubilized state," Proceedings of the National Academy of Sciences, 97:6138-43, May 23, 2000.

7. R. Vassar et al., "ß-secretase cleavage of Alzheimer's amyloid precursor protein by transmembrane aspartic protease BACE," Science, 286:735-41, 1999.