The amyloid hypothesis is not dead, but it may be gasping for breath

The amyloid hypothesis has been the dominant theory for the cause of Alzheimer’s disease for over 20 years. In brief, the theory holds that Alzheimer’s disease is caused by accumulation of beta-amyloid that damages nerve cells in the brain.  Beta-amyloid is cleaved from the large amyloid precursor protein (APP) into two main fragments which are released outside the cell, beta-amyloid 40 and beta-amyloid 42 containing respectively 40 and 42 amino acids. In normal people, these peptides are rapidly removed, but in people with Alzheimer’s disease the metabolic ability to degrade them is decreased, the peptides accumulate, form fibrils and ultimately amyloid plaques that in turn damage neurons and trigger formation of abnormal tau pathology causing neuronal death. There certainly is no doubt that amyloid plaques and tau-containing neurofibrillary tangles exist in Alzheimer’s disease. The question has been are they causative or are they simply the debris left over as the disease progresses. Soluble beta-amyloid is neurotoxic in mouse models of Alzheimer’s, but there is increasing unease that things are not as simple as the amyloid hypothesis would suggest.  Many studies costing billions of dollars have so far failed to show a significant reversal or even slowing of disease progression in Alzheimer’s patients treated with anti-amyloid drugs. Some of these drugs have been very effective in removing amyloid plaques, but none have demonstrated clinically significant slowing of cognitive deterioration. The failure of these drugs to date has been blamed on giving them too late.  The idea is that once cognitive impairment has started, there has already been death of neurons that cannot be repaired. This has led to several trials of anti-amyloid dugs in subjects with presymptomatic Alzheimer’s disease. They have positive biomarkers for Alzheimer’s such as amyloid and/or tau PET scans or blood tests but they do not yet have any cognitive impairment. One of these, the trial of crenezumab in the Colombian kindred with the autosomal dominant presenilin 1 mutation, has already failed (see my post). The results of the other studies will start to appear over the next few years.

The frustration at failing to find any effective anti-amyloid medications for Alzheimer’s received a bombshell this week with the publication in Science of a six-month investigative study that alleges falsified data in a 2006 study published in Nature that had been an important lynchpin for the amyloid hypothesis. The authors of that Nature paper claimed to have isolated a 56-kilodalton soluble beta-amyloid assembly from the brains of cognitively impaired, transgenic mice carrying human DNA coding for human amyloid precursor protein. They called this isolate Aβ*56.  When they injected purified Aβ*56 into the brains of young, healthy rats, the rats became cognitively impaired. The authors proposed that“…Aβ*56 impairs memory independently of plaques or neuronal loss, and may contribute to cognitive deficits associated with Alzheimer’s disease.” This appeared to be very strong evidence that a soluble beta-amyloid fragment could by itself cause the neuronal damage found in Alzheimer’s disease. According to experts in the field, working with amyloid fragments is technically difficult as they can be very unstable, so it was not particularly alarming that this work could not be replicated by others. The investigation done for Science found evidence of data falsification in the Nature paper calling into question the validity of other papers based on these results. 

I have not totally given up on the amyloid hypothesis.  There still is evidence that amyloid can be neurotoxic. However I suspect that the causes of neuronal loss and dementia in Alzheimer’s will turn out to be more complicated. According to Grace Stutzmann’s comment after an Alzforum article this week about the Aβ*56 fiasco (one of 23 comments): “we also see the rise of alternative mechanisms independent of amyloid pathology, including neuroinflammatory cascades, synaptic pathophysiology, calcium mishandling, and mitochondrial dysfunction—none of which are mutually exclusive.” Clearly we have much more to learn before we fully understand the neurobiology of Alzheimer’s disease.

Thanks to my daughter Susannah Gibbs for bringing the Science article to my attention.