Alzheimer hypothesis stuck in intellectual cul-de-sac

Dementia has become Australia’s latest national health priority. John/Flickr

The size and growing magnitude of the dementia epidemic has forced a welcome decision by health ministers to name the disorder as Australia’s latest national health priority area.

Yet for researchers searching for a dementia cure, a recent large negative clinical trial based directly on the amyloid hypothesis of Alzheimer’s disease underlines the idea’s growing irrelevance and even negative influence.

A brief history

Alois Alzheimer described his famous case of Auguste D in 1906. This patient was a 50-year-old woman with complex psychiatric symptoms and memory loss. On autopsy, a dense distribution of deposits or plaques were found throughout her brain. These days we wouldn’t diagnose her with garden variety Alzheimer’s disease – but so started the long and complex saga that is this illness.

In the 1980s, several groups (one led by Australian professor Colin Masters) isolated the main protein in these “senile plaques” as fibrillar amyloid: long, twisted and insoluble proteins assumed to be the prime culprit in Alzheimer’s disease. For the next 20 years, hundreds of science papers were written about the toxic effect of fibrillar amyloid and how the build up of these molecules in the form of senile plaques was antecedent to dementia, and the basis for the accompanying neuronal loss. Remember that dementia is the clinical syndrome of progressive mental dysfunction that causes patients to present to their doctor.

Cracks in the hypothesis

In the last two decades, three inconvenient truths have emerged for the amyloid hypothesis. First, it’s now clear that about 30% of individuals with moderate to severe senile plaques at death were never demented in life. So the plaques (and their component fibrillary amyloid) don’t necessarily lead to dementia. It’s also now clear that the majority of people diagnosed with dementia have a mix of two pathologies in their brain – fibrillar plaques and vascular disease. And finally, that amyloid is a natural physiological protein in the brain, vital to synaptic function.

But during this time, some smart scientists came up with the idea of developing a vaccine against amyloid. The argument was that if the body could better eliminate it then the terrible mental deterioration experienced by those with dementia should halt, if not reverse.

First failed trial

Mice results were marvellous – in fact, we have since cured Alzheimer’s in mice many times over. But human studies have been spectacular failures. The first incarnation of the vaccine was based on active immunisation, relying on the body to raise antibodies to a foreign amyloid challenge. After 6% of participants developed severe brain inflammation, the trial was called off. In retrospect, this was perhaps unsurprising given amyloid is a physiological protein.

Researchers have cured Alzheimer’s in mice many times over. Kessa Ligerro

Just as troublesome was the observation that in those patients who mounted a strong immune response upon vaccination, autopsy showed there was indeed evidence of removal of senile plaques from their brain (hurray) – but their dementia progressed unabated (boo).

So, in living, breathing humans, the clinical disorder of dementia decouples from the disease state of fibrillar amyloid plaques. These results also raise a deeply troublesome issue for basic scientists and those in the drug business. Mice models of Alzheimer’s disease are only nominal, mere reflections of our theories; if the theory is weak then the model is worthless.

History repeated

The drug company Elan, which developed the original amyloid vaccine was taken over by Pfizer and Johnson & Johnson in 2008. The new team concluded that the problem didn’t lie with the amyloid hypothesis, but rather that a passive vaccine was needed.

Fast forward four years, a $1bn investment and a series of clinical trials of bapineuzumab involving 2400 patients. Last week, the results of Phase III trials were announced: no evidence of clinical efficacy on any cognitive or functional outcome (even after data mining), and all future trials of the drug halted.

These kinds of results should have triggered a profound reconsideration of the amyloid hypothesis. So is the field now finally moving on? Perhaps and perhaps not. Investors are running for the door (see these hard-hitting commentaries from Forbes, here and here), but the die-hards have thrown two counter-punches.

New orthodoxies

The first tenet of the new orthodoxy is that the pathological process starts 20 to 30 years before symptoms appear. While probably true, there are two fatal flaws with the argument’s extension – that we should be starting anti-amyloid therapy in our 40s and 50s.

The cross-section of the brain of an Alzheimer’s sufferer. AJ Cann

First, as now conceded, 30% of those with significant amyloid plaques (the overt disease) will never dement (the disorder), but die of other causes with intact mental function. So taking a potentially dangerous and expensive anti-amyloid drug for 20 to 30 years would prove futile in those who otherwise would never have developed the disease, and redundant in the one-in-three who do. What’s more, we need to ask again what exactly is the disease process of interest? There are now serious questions about whether fibrillar amyloid has a causal relationship with neuronal loss and related mental dysfunction. So any drug attempting to rid the brain of it is suspect.

The second new orthodox view rejects the assumption of the last 100 years that fibrillar amyloid is the villain. Rather, it considers that the amyloid you can’t see under the microscope is the real problem. Small ensembles (oligomeric) of soluble amyloid are thought to be toxic, not the twisted insoluble stuff first seen by Alois Alzheimer.

And what of that mass of research concluding that fibrillar amyloid was bad? That was based on barnyard concentrations of the stuff, the kind of non-physiological doses that would kill off anything. In fact, if we test amyloid at a more natural physiological range, it seems to help neuronal function.

In fact, one of the founders of the amyloid hypothesis has gone to the extent of speculating that fibrillar amyloid may be a good thing. That it may be the body’s way of locking away the harmful oligomeric amyloid in relatively innocuous plaques. What was once evil is now blessed, and the toxic culprit invisible, hard to measure and in an unknown dynamic relationship to “good” amyloid.

To consummate confusion in the area is the current fad of looking for “biomarkers” – biological tests to tell us the state of amyloid in the brain. The most fancy and expensive of these is PIB imaging, a molecular technique for visualizing plaques in the live brain. Several companies and research groups around the world are trying to champion it as a method of better diagnosing Alzheimer’s disease and monitoring treatment.

And what type of amyloid does it visualize? Fibrillar – the same type the new orthodoxy rejects as pathologically irrelevant.

Oligomeric amyloid may or may not turn out to be one of the causal factors behind dementia, but let us not repeat the same mistakes of the last 100 years. We must understand its physiological function, test it at realistic concentrations, and if vaccines against this form of the protein also fail, then the amyloid hypothesis should be buried once and for all.

After all, there’s nothing more destructive to the progress of medical science than an unfalsifiable belief.

Found this article useful? A tax-deductible gift of $30/month helps deliver knowledge-based, ethical journalism.