Targeting Amyloid in Alzheimer’s Disease: No Longer Enough?

by Jim Schnabel

October 3, 2008

A postmortem study of Alzheimer’s patients given an amyloid vaccine eight years ago shows that the vaccine cleared amyloid plaques from their brains but failed to reduce the progress of their dementia.

Alzheimer’s patients whose brains were largely cleared of beta-amyloid “plaques” by a vaccine still experienced worsening dementia, according to a study reported in the July 19 issue of the Lancet by Clive Holmes and colleagues at the University of Southampton, U.K.

“These findings imply that progressive neurodegeneration can occur in Alzheimer's disease despite removal of plaques,” the authors noted.

The study was based on a postmortem examination of patients who had received an amyloid-clearing vaccine in an ill-fated clinical trial organized by Elan Pharmaceuticals in 2000. Although the study is based on a relatively small number of samples, its results are consistent with a growing body of evidence suggesting that targeting beta-amyloid plaques in Alzheimer’s patients is at best an insufficient strategy for preventing or treating the disease.

The amyloid vaccine

The Elan trial, which eventually enrolled about 300 people with Alzheimer’s, was halted in 2002 after the “AN-1792” vaccine provoked a dangerous inflammatory condition in the brains and spinal cords of a small percentage of recipients. Elan discontinued development of that vaccine, and the general strategy of clearing amyloid by provoking the immune system into an anti-amyloid antibody response is now questionable after yet another Elan amyloid vaccine, “ACC-001,” was terminated last April after causing blood vessel inflammation in one patient.

These trials are potentially valuable, however, because they offer an answer to the question: Does the removal of amyloid do anything to stop the progress of dementia, as has been assumed? Towards finding an answer, Elan researchers followed the AN-1792 patients after the trial was halted, performing cognitive tests on surviving patients as well as postmortem exams of the brains of some who died.

Such exams are usually complicated by the difficulties of recovering brain tissue before it is damaged by postmortem decay, and in many cases patients or their families do not provide consent for autopsies. But Elan announced in 2004 that it had managed to examine the brains of three deceased patients from the AN-1792 study and had found evidence of significant beta-amyloid clearance by the vaccine.

That was portrayed as good news at the time. Less emphasized was the fact that Elan had also analyzed the vaccine recipients with cognitive tests established as part of the original trial protocol. According to those criteria there was no evidence that the vaccine had slowed the progress of their dementia.

From only three autopsies, any analysis could hardly be conclusive. The new Lancet study, though, extends this data set. Holmes and his colleagues obtained consent for postmortem brain analysis from 80 patients who had been in the AN-1792 trial and performed postmortems on the brains of eight of 20 patients who died during their study window. Again they found evidence for significant plaque clearance, and the statistical trends suggested that higher antibody responses had resulted in more extensive plaque clearance.

But again there was no evidence that the patients who had lost their beta-amyloid plaques had experienced a slowing of their dementia. “All but one of the individuals who died during the follow-up phase had clear end-stage dementia before death, including the two individuals with highest mean [levels of] antibodies to beta-amyloid and almost complete elimination of plaques,” Holmes and his colleagues noted in their Lancet paper.

Re-thinking the amyloid hypothesis

This is a time of crisis for the “amyloid hypothesis” of Alzheimer’s, which has absorbed the most research dollars and by far the most drug development funds since it emerged as the leading theory for the disease in the mid-1990s. In the past year, two promising drugs known to reduce the accumulation of amyloid in “Alzheimer’s mice” have failed to significantly slow the disease in large scale trials in humans. In June it was announced that a third drug, Elan’s monoclonal-antibody “passive vaccine” designed to clear amyloid directly, also failed to show a significant effect in slowing the disease according to the analytical criteria originally established for the trial. (A “post-hoc analysis” claimed evidence for an effect in a subset of patients.) [See Dana news story on bapineuzumab, June 19.]

For many researchers, these results confirm other findings suggesting that beta-amyloid shouldn’t be the sole target for disease therapies, but that another culprit, the protein known as tau, should be cleared as well.

Tangled fibrils of tau, clumped in dying neurons, form the second characteristic pathological sign of Alzheimer’s in humans—beta-amyloid plaques being the first characteristic sign. Tau tangles have often been dismissed as mere markers of the Alzheimer’s disease process, rather than factors that actually damage neurons and worsen dementia. They also appear in other dementia-associated illnesses, including a set of disorders known as frontotemporal dementias. In recent years, however, evidence has emerged that mutations in the gene for tau can lead to hereditary neurodegenerative disorders. Transgenic mice that produce mutant tau also show tau-based tangles and neuronal degeneration with similarities to those seen in Alzheimer’s.

In 2001, a team from the Mayo Clinic in Jacksonville reported in Science that the addition of beta-amyloid to the brains of such mice makes the tau-based tangles more intense and widespread. “There’s good evidence that plaques and tangles are on the same pathway, with plaques actually causing tangles,” says Sam Gandy, a neurologist at Mount Sinai School of Medicine who chairs the Alzheimer's Association’s Medical and Scientific Advisory Council.

Karen Duff, an Alzheimer’s researcher at Columbia University who has worked with amyloid- and tau-producing mouse models, notes that there is also evidence to suggest that the process can continue on its own, once harmful forms of tau have appeared. “If tau is not reduced in the amyloid targeting approach, tangles may increase,” she says.

Mouse studies suggest that early clearance of amyloid can interrupt this apparent cascade of neuronal destruction enough to reduce levels of tau. But many researchers now regard an Alzheimer’s diagnosis as a relatively late event in a decades-long disease process, which means that even aggressive amyloid clearance in newly diagnosed patients may come too late—as the pathology results from Holmes and his colleagues appear to confirm.

Gandy suggests that the best strategy may be a drug “cocktail,” in which an amyloid-clearing therapy is augmented with a therapy targeted against harmful forms of tau. “An anti-amyloid plus anti-tangle combination will be very exciting to put through a clinical trial,” he says.