It’s more than a century since Alois Alzheimer first noticed and reported sticky plaques in the post-mortem brain of a patient with what is now known to have been Alzheimer’s disease, yet there is still disagreement on whether the plaques actually cause the disease (New Scientist, 2016). The role of plaque in the disease has been questioned following several high profile failures of drugs that eliminate plaque already existing in the brain or that seek to stop it being produced there. The only plaque clearing drug to have shown any clinical promise so far, aducanumab, also caused quite severe side effects, such as brain swelling. Many of the plaque blocking drugs that have been tried also caused severe side effects, including degeneration of nerves, the brain and the retina, and growths in the pancreas and reductions in blood sugar levels. Drugs called gamma secretase inhibitors designed to block plaque production by another route caused severe gut problems and even cancers to develop.

In a study published in Science Translational Medicine, a team from Merck Research Laboratories has reported results of early human and animal trials of a drug called verubecestat, which targets the production of protein plaques associated with Alzheimer’s (Scientific American, 2016). Team leader Matthew Kennedy said “It’s a summary of the discovery and early-stage profiling of what we hope is going to be a new therapeutic for Alzheimer’s. It represents well over a decade of investment in this project by many, many scientists.” Definitive conclusions will have to await the results of larger, ongoing phase III clinical trials to assess their efficacy, effectiveness and safety, but experts say the results are promising.

Verubecestat is a BACE1 inhibitor. BACE1 (Beta-site Amyloid precursor protein Cleaving Enzyme 1, also known as beta-secretase 1) is an enzyme involved in producing amyloid beta, a protein that clumps together, eventually forming the plaques surrounding neurons that are Alzheimer’s key hallmark. The amyloid hypothesis of Alzheimer’s proposes that the accumulation of amyloid beta aggregates in the brain drives a cascade of biological events leading to neurodegeneration. By blocking BACE1, the hope is this approach could prevent the buildup of these clumps in the first place. But until now, development of these drugs has been hindered by problems finding molecules with the right characteristics, and concerns over theoretical and actual side effects.

Amyloid is formed when amyloid precursor protein (APP) is cleaved into pieces by BACE1 and another enzyme called gamma-secretase. APP protrudes from cell membranes into the space between cells, where the enzymes can cut it. Production of amyloid beta involves two snips. First, BACE1 cleaves it some distance from the cell (producing fragments called sAPP beta) then gamma-secretase cuts the remaining stub off at the cell membrane. The fragment released by this cut is amyloid beta. BACE1 inhibitors work by attaching to the enzyme and preventing it from cleaving APP, thereby decreasing production of amyloid.

BACE1 was discovered in 1999 by a team led by molecular biologist Robert Vassar, now at Northwestern University, who was not part of this study. Researchers have been studying its function using mice engineered to lack the BACE1 gene. These studies have revealed numerous consequences including problems with insulation and guidance of neural wiring, retinal pathology and neurodegeneration, raising concerns that BACE1 inhibitor drugs might have side effects. Another challenge was developing molecules big enough to attach to BACE1 but still able to cross the blood–brain barrier. Several candidate drugs have now been developed, but a recent clinical trial was halted due to evidence of liver toxicity.

The Merck team has developed a molecule that appears to overcome these challenges. They tested the drug on animals and found it significantly reduced levels of both amyloid and sAPP beta in the blood, cerebrospinal fluid and brain in a dose-dependent manner. There were no signs of toxicity, even after treatment of up to six months in rats and nine months in monkeys. The only obvious side effect was reduced fur pigmentation in mice and rabbits, although this wasn’t seen in monkeys. The researchers then moved on to small, early-stage human trials to assess safety and tolerability and inform the choice of suitable doses for later trials.

The tests involved 32 people with Alzheimer’s and over 100 healthy volunteers. Verubecestat reduced amyloid and sAPP beta in the cerebrospinal fluid of healthy adults who took the drug for two weeks and patients with mild to moderate Alzheimer’s who took it for one week. At the lowest doses, the drug reduced amyloid beta by a half to three quarters, rising to 90% at higher doses. No one in the human trials experienced side effects. “This is the first detailed report of what a BACE inhibitor does in humans,” said Dennis Selkoe of Harvard University, a leading Alzheimer’s researcher who was not involved in the work. He said “The good news is they didn’t see evidence so far of any of the side effects we’re concerned about with BACE inhibition.” Robert Vassar said this is probably because the doses used did not fully inhibit BACE1 activity. He said “It might be you only need a little bit of BACE active in the brain and body to prevent side effects.” Another possibility is that some of the consequences for mice lacking BACE1 from birth are developmental effects that don’t apply when the enzyme’s activity is lowered in adults.

These results helped propel testing to full-blown clinical trials, making verubecestat the first BACE1 inhibitor to reach phase III trials. Matthew Kennedy said “It’s really the first molecule of its kind to combine [amyloid]-lowering potency with a very positive safety profile that allows us to treat patients for the time needed to determine if there will be clinical benefits on cognition.” Two trials testing long-term outcomes in patients are ongoing. The first involves roughly 2,000 patients with mild to moderate Alzheimer’s for 18 months. The second has around 1,500 participants with early signs of Alzheimer’s (as indicated by amyloid plaques revealed in positron emission tomography brain scans), for two years. Robert Vassar said “The big issue is: What will the long-term safety of these drugs be? People may have to take these drugs for the rest of their lives, the trials are two years at most; what happens beyond that, as people get older, we have no idea.”

The second trial could prove a crucial test, because a treatment that limits production of amyloid is likely to work best at the earliest stages of the disease. Plaques may start building two decades before symptoms appear, so by the time a diagnosis is given it may be too late for this approach to help. Researchers await results (expected in 2017 and 2019, respectively) eagerly. If participants show slowed decline of cognitive functions together with reduced amyloid, it would provide strong support for the hypothesis that the protein clumps cause Alzheimer’s.

Another approach is antibody therapies. One of these drugs, aducanumab, dramatically reduced amyloid in the brains of patients with mild stages of Alzheimer’s in a small trial reported in the September issue of Nature. Some participants also showed slower cognitive decline, although this too awaits confirmation from a larger, ongoing clinical trial. Dennis Selkoe said “I’m also encouraged by the aducanumab data. These are all good shots on goal.” Researchers believe antibody therapies work by “mopping up” existing amyloid aggregates whereas BACE1 inhibitors prevent the protein from being produced, so the two could prove complementary, Robert Vassar said “There may be hope for healing the brain with such an approach,” he adds. “We can’t bring back dead neurons but we might be able to heal the ones that are still alive.”

Other treatments under investigation involve modifying gamma-secretase enzymes, tackling inflammation or targeting the tau protein tangles that occur in Alzheimer’s. Dennis Selkoe said “Different approaches are necessary, and we envision patients will get multiple treatments once they’re approved. For now, none have made it across the finish line.”

Other Alzheimer’s researchers have welcomed Merck’s results, but are more cautious about whether verubecestat will have an impact on symptoms. “We have to hope that this excellent animal data really does translate into something useful in the real, human disease,” said the originator of the plaque hypothesis, John Hardy of University College London.

“While this study is promising, more work needs to be done to know if this drug will be an effective treatment for Alzheimer’s disease,” said Tara Spires-Jones of the University of Edinburgh. She said despite strong evidence that production of amyloid beta kick-starts the disease process, several large clinical trials have previously shown that lowering amyloid beta failed to significantly help patients.

Others were worried about potential long-term effects of taking verubecestat. “My concern is that amyloid beta must be blocked for over a decade or so before symptoms start appearing, so any drug that does this could become like a ‘statin’ for Alzheimer’s disease,” said Rudolph Tanzi at Harvard Medical School. He said if people have to take verubecestat for decades, it could have unanticipated longer-term effects in the brain and other parts of the body by disrupting other substances that interact with BACE1. He said he has developed a drug that interferes with but doesn’t block gamma secretase, another enzyme that helps to create amyloid beta. Pfizer has developed a similar drug, and will be presenting results on how well it’s worked so far in December at a meeting in San Diego, California. Rudolph Tanzi said “I’m more optimistic about these, in terms of drugs that might need to be used for decades.”