New Alzheimer’s Approach Suggested by Data Mining

By Allison Proffitt 

October 1, 2024 | In a paper published last month in Brain (DOI: 10.1093/brain/awae216), researchers explored data from 24 past Alzheimer’s Disease trials. The University of Cincinnati team posits a new theory for why the most recent drugs have shown some success against the disease and how to build on that success.  

The drivers of Alzheimer’s Disease have long been debated. Amyloid plaques and Tau tangles are suspected culprits, but for years, anti-amyloid-b monoclonal antibodies have significantly reduced brain amyloid while not resulting in cognitive or clinical benefits for patients. Two more recent anti-amyloid-b (Ab) monoclonal antibodies—lecanemab (2023, Eisai and Biogen) and donanemab (2024, Eli Lilly)—seem to have turned the tide, both reducing brain amyloid and showing cognitive benefits.  

“We’ve had essentially not only just failures, but situations in which people treated with anti-amyloid treatments were worse compared to placebo,” explains Alberto Espay, professor of neurology in the University of Cincinnati College of Medicine and senior author on the study. “Removing amyloid is quite a toxic enterprise,” he notes. “One in three people develop brain swelling or bleeding or both,” a side effect called ARIA (amyloid-related imaging abnormalities).  

“What we wanted to do was try to explain why it is that there seems to be an exception here.”  

The difference, Espay suggests, is that the newer mABs both decrease brain amyloid plaques and increase levels of the 42-amino acid isoform of Ab when tested in cerebral spinal fluid.  

Ab42 is a common protein, known to exert strong antioxidant and neuroprotective effects with roles in enhancing synaptic plasticity and memory, the authors wrote in the paper. This is fascinating, says Espay, because while the brains of elderly patients both with and without Alzheimer’s Disease show amyloid plaques—making it an unreliable marker, low levels of Ab42 are universally associated with the disease. And Ab42 seems to fall to abnormally low levels before any plaques are visible on imaging and before cognition suffers.  

“The increase in Ab42 explains the cognitive endpoints at least as well as the decreases in amyloid [plaques],” Espay argues. He theorizes that Ab42 is the normal configuration of the same amyloid protein that becomes plaques. In some cases, the body can keep healthy levels of Ab42, even as some of the protein loses its function locked in clumpy plaques. Those would be individuals who show plaques on brain imaging but have no neurodegeneration symptoms.  

For others, total Ab42 levels fall in response to environmental stressors. “The proteins that we have react to things in the environment to protect the brain. If that exposure continues unabated, then the proteins become depleted,” Espay says. Meanwhile, plaques are forming, so not only do patients have amyloid locked in plaques, but there is not enough normal Ab42 left for normal brain functioning. These patients have Alzheimer’s Disease.  

“It turns out that there was no correlation whatsoever between how much of the amyloid plaques you would have and how severe your Alzheimer’s would be,” Espay says. “Once [amyloid proteins] adopt a configuration that’s abnormal, which is known as plaques, they just cannot work. It’s just a loss of their function; it is not a gain of toxicity.”  

Espay acknowledges that this idea is a paradigm shift in how we view amyloid plaques and how Alzheimer’s drugs should be targeted. He recommends looking for therapies that proactively increase Ab42, not just try to remove plaques. The drugs that are effective, he believes, work “not by how much it removes amyloid—because so many other drugs have removed amyloid brilliantly—it’s by how impressive the increase in Ab42 appears to be,” he says.  

Data Access Question 

Espay and his team discovered this correlation by digging deep into all anti-Ab drug trials in PubMed with patients with early to mild-to-moderate AD published between January 1985 and November 2023. The team looked at controlled trials of Ab-targeting drugs with follow-up of at least one year, looking for reports of both Ab plaques visible in PET imaging and Ab42 levels in cerebral spinal fluid.  

They found 24 trials of 10 anti-Ab drugs and 37 datasets comprised 25,996 AD study participants.  

Lecanemab, the first FDA-approved treatment of Alzheimer’s, increases Ab42 “very, very markedly,” Espay says. He theorizes that the increase in Ab42 was significant enough to outweigh the toxicity of removing plaques.  

But you wouldn’t know that is you just read the paper. “That data was reported in the supplementary materials, so that if you read the New England Journal of Medicine article for lecanemab (DOI: 10.1056/NEJMoa2212948), you would never see any text that [reports] that this therapy increases the level of the [Ab42] proteins,” Espay said.   

In fact, Espay and the other authors highlight that the Eisai/Biogen partnership in the lecanemab trial dictated that no data sharing was permitting. “We could only get group-level data because that’s the level of data that they published. We need to get individual-level data,” Espay said, to be able to look closer. The authors call out those limitations to data access as a main limitation to their work. 

Espay hopes that the companies doing research in anti-amyloid-b monoclonal antibodies will reconsider their individual-level data in light of these findings and do a much more refined study. But he also hopes for more transparency overall.  

“It is not adequate in this day and age to hold data from the public. In this age of transparency, it is an anachronism by these companies to say, well that data is ours and we’re not going to share it.”