What now for the amyloid hypothesis?

In July 2022, a bombshell dropped on the Alzheimer's disease research field. For years, researchers had searched for something that caused the disease’s telltale amyloid plaques — complex tangles of a protein called amyloid-beta (Aβ) frequently found in the brains of patients with neurodegenerative disorders. A series of studies published starting in the mid-2000s reported the discovery of a toxic form of Aβ in the brains of mouse models of Alzheimer's disease called Aβ*56. Researchers hoped that Aβ*56 was the protein that snowballed into those amyloid plaques. But a team of sleuths found that many of the papers describing Aβ*56 were fraudulent and contained an array of faked images and blots. The fraud seemed to call the entire idea of amyloids causing Alzheimer’s disease into question. 

Given the thunderous news of the summer, the path forward for Alzheimer’s disease might seem uncertain. But news of the amyloid hypothesis’s death may have been exaggerated.

“Science is guilty of misinterpreting the importance of Aβ*56. It was not foundational,” said Jeffrey Cummings, a neurobiologist and long-time Alzheimer’s disease researcher at the University of Nevada, Las Vegas. “It was proven to be false within about two years ” (1). 

How amyloids in general, and Aβ in particular, became the central figures of Alzheimer’s disease when other players have been known for years is a complex story. Some researchers mistook Alzheimer’s disease for a simple condition of tangled protein when the modern view is that it is actually a broader syndrome with many contributing factors. A large part of this mistake comes down to the disease’s sheer complexity, its sometimes subtle symptoms, and its decades-long development time. But much of it may also come down to simple human instincts, error, and even financial gain. 

Promising beginnings

“The amyloid hypothesis” describes how a variety of neurodegenerative disorders, including Alzheimer's disease, result from proteins misfolding from a healthy structure into an alternative shape that forms visible clumps. Through an unknown mechanism, these clumps kill neurons in the brain.

An amyloid hypothesis for Alzheimer’s disease can be traced to a 1984 paper by George Glenner and Caine Wong at the University of California, San Diego describing Aβ as the misfolded protein that drives Alzheimer’s disease pathogenesis (2). But there are a host of other amyloid diseases such as Creutzfeldt-Jakob disease and Bovine Spongiform Encephalopathy, and each has its own biology, symptoms, and cast of molecular characters that contribute to disease, often routed through the central misfolded protein.

For example, the gamma secretases Presenilin 1 and 2 (PSEN1/2) produce Aβ peptides in the brain. Mutations in these gamma secretases are thought to cause increased production of Aβ and predispose a patient to form amyloid clumps (3,4). 

“The reason that the amyloid hypothesis started out was pretty strong,” said Becky Carlyle, a neurobiologist who studies neurodegeneration at the University of Oxford. “They found out that lots of people with dementia had these plaques. We were able to actually look at what proteins were in them; there was an absolute ton of amyloid. And then that combined with the fact that we have familial early onset Alzheimer’s disease, which involves mutations in the enzymes that produce amyloid peptides, meant that that was a pretty good starting point, I think, for the theory and certainly merited a lot of attention.”

Amyloids are the most visible and striking markers of Alzheimer’s disease. In stained slices of the brain, they form unmistakable dark plaques. But over the last decade, researchers have begun to view amyloid plaques less as a pernicious agent of destruction and more as a simple and sometimes even misleading biomarker.  

“To me, there are two major things that really cast doubt upon the future [of amyloids]. One is the fact that once we started doing really well thought out general population cohort studies of neuropathology, we did find that about one third of the people in the population that, when you look at their brains post-death, look like they should have Alzheimer's disease, like they should have dementia, when you see the plaque load in their brains. But they actually don't. They're completely cognitively unimpaired,” said Carlyle. 

This is a fundamental disconnect between the hallmark symptom of amyloid diseases. Visible amyloid clumps had been thought to be the cause of these neurodegenerative diseases, killing neurons en masse until neurological symptoms set in. But it’s clear now that plaques in the brain do not predict whether a patient will have dementia. On the other hand, neither does their absence. Neurodegeneration occurs in Alzheimer’s disease, prion diseases, and other amyloid diseases with and without plaque formation.

“That suggests to me that, okay, if you do have amyloid, you are more likely to have dementia than someone who doesn’t, but it’s not a one-to-one relationship. There are other things that have to be mediating the presence of dementia in these people,” said Carlyle.

Nevertheless, until recently, research on treatments for Alzheimer’s disease has focused almost exclusively on amyloid plaques. The only new treatment in recent memory for the disease is aducanumab, a monoclonal antibody that clears plaques. Aducanumab earned FDA approval in the summer of 2021 despite only having a marginal ability to slow cognitive impairment. It may have passed review because within the parameters of the amyloid hypothesis, it’s quite successful. Aducanumab almost completely clears patient brains of amyloid plaques (5).

“And then the second one, of course is aducanamb,” said Carlyle. “We now have a drug that pretty much clears amyloid plaques from the brain. And potentially there's a tiny, tiny incremental change in cognitive performance, but not something that an individual person is going to notice from taking the drug.

“It's probably fair for them to be saying maybe if you took it earlier, then maybe it would have some positive effects. But the reality is, it's a drug with a ton of dangerous side effects. And we don't want to be giving it to people in their 40s at the off chance that they might develop dementia,” said Carlyle.

Just as concerning is that it’s difficult to find a recurring genetic signature that consistently reappears within families with a history of Alzheimer’s disease and points towards a single cause. Only 10% of patients with Alzheimer’s disease have a family history, while the other 90% are sporadic and have no known cause (6). 

Even still, genetics is considered by far the most dominant risk factor for developing Alzheimer’s disease. There is one genetic marker that has a strong correlation with developing Alzheimer’s disease, called apolipoprotein E (APOE). A healthy version of this gene, APOE3, is functional. A single amino acid substitution, cysteine 112 to arginine 112, creates the APOE4 variant, which is completely nonfunctional. Inheriting one copy of APOE4 doubles a patient’s chances of developing Alzheimer’s disease while two copies increases it eight- to 12-fold. APOE4 was identified decades ago, but researchers only identified a causal role in neuronal demyelination and death in 2022 (7). There are currently no therapeutics in development that target APOE4 (8). 

Never heard of it

This disconnect of amyloids, genetics, and disease, the absence of real therapeutics, and allegations of fraud have created noise in the media, and while researchers in the field have concerns, the uproar over Aβ*56 does not match what researchers in the field felt.

“The amyloid hypothesis has numerous reasons why it's been followed for a long time,” said Carlyle. “A lot of those were very, very strong scientific reasons. I think there's good reason that it was the focus for a very long time. But I joined the field in 2017, and this particular peptide — I had never heard of it.”

The focus on the Aβ*56 fraud revealed a popular view that only amyloids and Aβ caused Alzheimer’s disease. Another protein called tau has been known for decades to form its own protein clumps within neurons (called neurofibrillary tangles), and to form those tangles specifically in response to Aβ misfolding (9). Nevertheless, for years, the focus remained on amyloids and Aβ.

“People have confused or assumed that diagnosis and treatment are the same thing,” said Kwasi Mawuenyega, a protein biologist at MilliporeSigma who spent 15 years studying Alzheimer’s disease at Washington University in St. Louis. While the presence of amyloids and Aβ fragments was the sole diagnostic biomarker and treatment target for Alzheimer’s disease for years, tau tangles are becoming a better diagnostic tool and indicator of the disease’s complexity. Focusing exclusively on amyloids was simply a case of tunnel vision, according to Mawuenyega. 

“A little bit of people's infatuation, or I will say attachment to one specific biomarker, maybe blurred their view of the big picture,” he said. Since many biopharma companies had focused for so long on amyloids, it became a self-perpetuating cycle where scientists funded by industry also looked at amyloids, according to Mawuenyega. “When I went to Alzheimer's disease meetings and conferences, people had a lot of conflicts of interest,” he said. “They are scientific advisors to many of the drug companies that are supporting their research. It got to a point where if you go to a conference, and you're listening to somebody give a talk, they’ll show their disclosures, and it's two, three slides of just that.” 

Compared to the heart

As of January 2022, there were 172 clinical trials testing 143 different drugs that target Alzheimer’s disease. A scant 20 of them target amyloids, a few target Aβ, and none target Aβ*56 (8).

Despite a reputation for chasing amyloids’ tail, the field is moving on to other targets such as gamma-secretases and tau. Instead of treating Alzheimer’s disease and neurodegeneration like a bacterial infection, something that can be cleared away with the right pill, researchers are reframing Alzheimer’s disease as a multi-faceted condition with no singular cause. Treating it will require a full lifetime of primary care, management, lifestyle changes, and medication.

 “If we look to some other fields like cardiovascular medicine or renal medicine, those diseases are much simpler than these brain disorders,” said Carlyle. “But they still don't have one diagnostic test. They still don't have one treatment route. They have different diagnostic tests; they have full diagnostic pathways; and they have different treatments to treat different aspects of these diseases.”

“Yes, it's probably a bit of amyloid; it's probably a bit of tau,” Carlyle continued. “But we also have to think about inflammation; we have to think about reactive oxygen, stress, oxidative stress, all of these different things probably come together along with your lifetime of potential stressors and your genetic susceptibility to dementia.”

The ecology of the disease

“This is a pivotal year in Alzheimer's disease therapeutics,” said Cummings. He pointed to a number of ongoing Alzheimer’s disease clinical trials with targets outside the amyloid hypothesis. Gantenerumab, a monoclonal antibody targeting toxic Aβ,  produced negative phase III clinical data at the end of 2022. Lecanemab, another monoclonal antibody that targets and clears toxic soluble Aβ aggregates before they form amyloids, slowed cognitive decline by 27% in a phase III trial in patients with Alzheimer’s disease, leading to its FDA approval in January 2023. Lecanemab also reduced the appearance of amyloid plaques in the brain, a critical sign of effectiveness for researchers and for government regulators.

Unfortunately, that 27% improvement might be misleading. Researchers score dementia progression using the Clinical Dementia Rating sum of boxes (CDR-SB), a scale that rates patients on neurological function markers like memory, orientation, and personal care. On that scale, lecanemab’s improvement was 27% better than placebo but in absolute units was a scant 0.45 point improvement out of 18 total points. Other studies had set a minimum of a 0.98 point improvement as clinically meaningful for just mild cognitive impairment and as much as 1.63 points for fully developed Alzheimer's disease (10). At the same time, two patients in lecanemab’s clinical trial treatment arm died of strokes leading to brain hemorrhages that scientists suspect to be side effects from the treatment (11). 

Other therapeutics that target the broader ecology of Alzheimer’s disease are not far off. Baricitinib, a Janus-kinase inhibitor that reduces neuroinflammation should return clinical data by October 2023. AL002, a monoclonal antibody that helps microglia clear Aβ, is expected to produce data by August 2023. Even a vaccine against Alzheimer’s disease-related forms of tau, termed ACI-35, should give a readout by October 2023. 

Although drug developers are turning away from amyloids and researchers are recognizing other causes of Alzheimer’s disease, scientists won’t minimize amyloids’ importance or throw them out altogether. “People said, ‘Well, you can have amyloid in the brain and not have dementia.’ That's absolutely true, said Cummings. “However, amyloid seems like it establishes the ecology in which other processes can occur that result in cell dysfunction and dementia."

Cummings sees the most mystery and promise in the molecular ecology of the brain, the ecosystem that draws together individual toxic proteins like Aβ or tau and the amyloids and neurofibrillary tangles. Other researchers agree and are now looking for future therapeutic targets in secretases, inflammation, and microglia and the immune system. 

“We can measure monomers [of Aβ and tau] very well,” said Cummings. “We can measure plaques very well. And everything else is invisible to us.” 

References

1. Shankar, G. M. et al. Amyloid-β protein dimers isolated directly from Alzheimer’s brains impair synaptic plasticity and memory. Nat Med  14, 837–842 (2008).
2. Glenner, G. G. & Wong, C. W. Alzheimer’s disease: Initial report of the purification and characterization of a novel cerebrovascular amyloid protein. Biochemical and Biophysical Research Communications  120, 885–890 (1984).
3. Kelleher, R. J. & Shen, J. Presenilin-1 mutations and Alzheimer’s disease. Proc. Natl. Acad. Sci. U.S.A.  114, 629–631 (2017).
4. Lanoiselée, H.-M. et al. APP, PSEN1, and PSEN2 mutations in early-onset Alzheimer disease: A genetic screening study of familial and sporadic cases. PLoS Med  14, e1002270 (2017).
5. Sevigny, J. et al. The antibody aducanumab reduces Aβ plaques in Alzheimer’s disease. Nature  537, 50–56 (2016).
6. Gribkoff, V. K. & Kaczmarek, L. K. The need for new approaches in CNS drug discovery: Why drugs have failed, and what can be done to improve outcomes. Neuropharmacology  120, 11–19 (2017).
7. Blanchard, J. W. et al. APOE4 impairs myelination via cholesterol dysregulation in oligodendrocytes. Nature  611, 769–779 (2022).
8. Cummings, J. et al. Alzheimer’s disease drug development pipeline: 2022. A&D Transl Res & Clin Interv 8, (2022).
9. Götz, J., Chen, F., van Dorpe, J. & Nitsch, R. M. Formation of Neurofibrillary Tangles in P301L Tau Transgenic Mice Induced by Aβ42 Fibrils. Science  293, 1491–1495 (2001).
10. The Lancet. Lecanemab for Alzheimer’s disease: tempering hype and hope. The Lancet  400, 1899 (2022).
11. Second death linked to potential antibody treatment for Alzheimer’s disease. (2022). doi:10.1126/science.adf9701