By Deborah Borfitz 

August 18, 2022 | Research in Alzheimer’s disease is a “supertanker” of work focused on a handful of pathologies such as an abnormal accumulation of tau and beta-amyloid proteins in the brain. A comprehensive survey of the scientific literature has shown that certain biological pathways, especially immune and metabolic ones, have been the most studied over the past 30 years despite hundreds of others being associated with the mind-robbing condition, says Winston A. Hide, Ph.D., director of the Precision RNA Medicine Core Facility at Beth Israel Deaconess Medical Center (BIDMC) and an associate professor of medicine at Harvard Medical School. 

Changing the course and speed of that supertanker will take enormous effort, but the need is inescapable, says Hide. “What is striking to me is that although the same 25 or 30 pathways have been intensively studied for decades, those pathways aren’t significantly enriched with genetic or functional findings. They are enriched with people studying them.” 

Hide is part of the team behind a study published recently in Frontiers in Aging Neuroscience (DOI: 10.3389/fnagi.2022.84690) that is advocating for a more data-driven approach to Alzheimer’s research where sets of biological pathways get grouped with mechanisms the body itself uses to ward off the disease. The relentless quest for a single-gene or single-pathway solution has turned up few effective treatments and no cure, he says, highlighting the need to start thinking about the relationship between sets of pathways and the phenotypes they produce.  

That Alzheimer’s disease will become a “neurological epidemic” is a worrisome possibility, says Hide. “I believe treatments can be developed that can accommodate the fact that the pathways don’t work alone. Choosing the [right] pathways to go after in the very earliest stages before the disease manifests... is the way to get ahead.” 

Surprising Signal 

For the study, researchers conducted a systematic assessment of more than 200,000 scientific publications published since 1990 to understand the breadth and diversity of biological pathways associated with Alzheimer’s disease. Sarah Morgan, a postdoctoral researcher at BIDMC, used Microsoft Academic Graph and AMiner to find associations in text based on a rigorous set of manual criteria, Hide says, which eliminated most of the spurious associations. 

The ubiquity of Alzheimer’s association pathways came as a complete and discouraging surprise to the research team, he adds. Nearly all of the pathways (91%) were associated with Alzheimer’s disease in at least five publications each and in more than 100 scientific papers nearly half of the pathways were linked. 

To ensure the initial findings weren’t a fluke, the group performed a number of rigorous analyses to assess if pathway distributions were being affected by other variables, such as how many times an article was cited or a journal’s citation rate. “We looked at all sorts of funky stuff, but the signal just wouldn’t go away no matter what we did,” says Hide.  

It is unsettling to see papers being published where the findings have no genetic or functional association with an Alzheimer’s phenotype in a dish, Hide says, pointing to an ever-growing body of evidence of limited value. Very few gene-based Alzheimer’s disease studies are even available for perusal. The genetic approach itself needs refining to start looking at rare variants that are not part of the processes usually found in Alzheimer’s disease studies where common variants (e.g., APOE4 and APP) are being examined. 

Again and again, bile acid secretion—the end product of cholesterol metabolism—comes up as an Alzheimer's-associated process, he notes, but how is that related to neurodegeneration in the brain? “These are the sorts of questions we need to be asking.” 

‘Anti-Alzheimer’s Pathways’ 

The underlying mechanisms of Alzheimer’s disease could be targeted in vitro in groups of redundantly acting pathways. “The most studied pathways are functionally related to each other, either by sharing genes or by sharing their activity,” Hide cites as an example. “They are all working together in some way.” Grouping those pathways and mechanisms together could therefore be revelatory. 

Patients could be segmented into subgroups based on the various “anti-Alzheimer’s pathways” in their body, he continues. It’s clearly not all about measuring levels of plaque in the brain. “There is a group of people who have very bad Alzheimer’s pathology in the brain... and they are not sick. They can think completely normally.” 

Most recently, in a case report appearing in Acta Neuropathologica (DOI: 10.1007/s00401-022-02467-8), researchers describe a woman with a family history of early-onset Alzheimer’s disease who lived dementia-free into her 70s, seemingly protected by an APOE3 mutation that spared critical brain regions from characteristic tau pathology. In the APOE family of genes, APOE3 is the most common variant but not typically associated with either reduced or increased risk for Alzheimer’s. 

Others have Alzheimer’s disease pathology in their brain but appear to have no significant loss of cognition. A recent study has identified biological processes that appear to be related to conferring resilience to disease pathology, including overall maintenance of synapse structure and molecular mediators such as REST and NRN1 protein levels (Trends in Neurosciences, DOI: 10.1016/j.tins.2022.02.005).   

Individuals like this are successfully warding off the disease, says Hide. “So, instead of going after the proteins that can [be removed]... by providing antibody therapy, why don’t we look more carefully for the mechanisms that the body uses itself to address Alzheimer’s?” 

Hard Truths 

The approach now being taken by Hide and his colleagues is to use each pathway as a proxy for a unit of function, in lieu of using a gene as a unit of inheritance, to get at answers. The top-ranked pathways of Alzheimer’s by textual reference aren’t the most important ones, so “that is a great problem to solve,” he says. 

In fact, they are soon to publish a new report where they have consistently found sub-clusters of co-activated pathways in Alzheimer’s disease, says Hide. Most intriguingly, the team has identified about 20 pathways that are always activated in patients suffering from Alzheimer’s disease irrespective of the pathological hallmarks in their brain. 

The novel, data-driven approach would be useful in choosing the best pathways to target, he says. “That is driving us to form a more black-and-white picture of the functions of Alzheimer’s where we can be categorical about it.” 

The pervasiveness of pathways is by no means unique to Alzheimer’s, says Hide, citing the same paradigm exists with Parkinson’s disease as well as diabetes and cancer. Therefore, even scientists focused on an important mechanism against complex diseases should be forced to address the basis of it when conducting a drug trial. 

Research is being stymied by targets that get graded from a podium rather than experiments in the lab and predictions from computational assessment of what is happening, says Hide. That might not change without new regulations to ensure studies include a forecast on how a drug is going to work and transparency about its weaknesses. 

“We need to instill a healthy skepticism for dogma in the next generation of Alzheimer’s researchers because [it] is misleading us,” Hide says. “We must be more accommodating of the hard truths of Alzheimer’s to move the supertanker so that it can start sailing to Port Cure.”