Alzheimer’s: How Not to Study a Disease

Yves here. As KLG explains in detail below, received wisdom on Alzheimer’s disease is a classic example of at least two very damaging yet widespread failures in what passes for scientific reasoning. The first is foundational myths, that persuasive-seeming early theories that become well accepted are remarkably difficult to dislodge. One classic example is ulcers. As recently as 20 years ago, a full 1/3 of GI specialists were treating acid as the cause of ulcers, not bacteria, as had been established in 1982. The second is that correlation is not causation. Mind you, KLG has other items in his bill of particulars, such as bad incentives like greed.

KLG’s overview:

Alzheimer’s disease (AD) has remained refractory to meaningful intervention despite intensive research, especially during the past 30-40 years. It may be that no cure can be found to AD. But the almost total lack of success could be due to inadequate theories of AD and faulty assumptions that are embodied in the Amyloid Cascade Hypothesis (ACH), which can be viewed as the Central Dogma of Alzheimer’s Disease. The foundation of the ACH was constructed more than a hundred years ago, shortly after Dr. Alois Alzheimer identified his eponymous plaques in the brain of his first patient after she died. The ACH has ruled modern research on AD since then. This history is deep in the archives of basic and clinical research on AD but is covered exceedingly well in How Not to Study a Disease: The Story of Alzheimer’s (2022) by Karl Herrup of the University of Pittsburgh. In addition to showing where the research community went wrong on AD, this history is an object lesson in the perils of epistemic closure in any field. It also seems likely that these wrong turns have led to some of the more egregious failures of science that have recently come to light. There are also lessons here for the science and the political economy of science as emerging diseases cause ever more damage the world over. We, scientists and citizens alike, should heed them well.

By KLG, who has held research and academic positions in three US medical schools since 1995 and is currently Professor of Biochemistry and Associate Dean. He has performed and directed research on protein structure, function, and evolution; cell adhesion and motility; the mechanism of viral fusion proteins; and assembly of the vertebrate heart. He has served on national review panels of both public and private funding agencies, and his research and that of his students has been funded by the American Heart Association, American Cancer Society, and National Institutes of Health

We discussed Alzheimer’s disease (AD) here two years ago and that led me to read much of the literature in the history of AD research.  My questions were  “How was the amyloid cascade hypothesis (ACH) developed, and why has it dominated research on AD for the past 30+ years despite leading to no interventions that are effective in preventing or arresting the horrific and inexorable course of AD?”  As treated at length in the very accessible How Not to Study a Disease: The Story of Alzheimer’s [1] by Karl Herrup of the University of Pittsburgh, it turns out the ACH was baked into the cake in the first decade of the twentieth century, shortly Dr. Alois Alzheimer published the case report of his first patient with an odd form of dementia, Auguste D, in 1907: About a peculiar formation of the cerebral cortex.

This case report is just that, a presentation based on one patient.  Case reports are essential to modern medicine, but they are not “proof” of anything.  Nevertheless, Alois Alzheimer had become a member of the scientific circle of Emil Kraepelin, in whose laboratory he identified plaques in postmortem samples from the brain of Auguste D after she died in 1906.  Kraepelin was an early psychiatrist, probably the first, to put forth the theory that mental/psychiatric illnesses are caused by biological changes in the brain of those so afflicted.  His argument with Sigmund Freud about the nature of mental illness is a recurring theme in Mind the Science by Jonathan N. Stea, which was discussed here recently.  Dr. Kraepelin subsequently included “Alzheimer’s Disease” in the 1910 edition of Psychiatrie.  As noted by Herrup, this was a “bold and reckless” choice by Kraepelin because once a disease is named in a textbook (or the DSM), it becomes much more real than a mere hypothesis in a case report or scientific paper.

The clear implication from the beginning was that plaques cause AD.  But, and this was not considered a hundred years ago and has barely been considered up to the present, something else could have caused both plaques and the disease.  Yes, Alois Alzheimer identified his now eponymous plaques in the brains of several patients postmortem.  However, he never did the control experiment, which would have been to examine the brains of the elderly who died without showing evidence of dementia (more on this below).

According to Herrup, this demonstrates the differences between science and medicine.  The response of medical establishment to AD goes something like this: “We have millions of people with AD.  We must do something, now!  So, let’s start clinical trials based on what we know, and what we know is that amyloid plaques cause AD.”  On the other hand, a biomedical scientist would respond, “Yes, we should do what we can and closely watch what happens, but we know next to nothing about the biological causes of AD.  Therefore, while we could get lucky in the short term, we cannot begin to respond intelligently to this disease until we know a lot more.”  The biomedical scientist was ignored.

But there were and are many potentially promising theories of AD other than the ACH.  Over the past 25 years at least 30 genes have been linked to AD.  One of them is similar to an ancient, evolutionarily conserved protein from amoebas to humans that has been the primary focus of my research for the past 25+ years.  The complex world of normal and pathological cell biology is a crowded space that is not generally acknowledged by clinician and layperson alike.  These mechanisms include but are not limited to the following.

Inflammation. A large subset of these AD-linked genes (I do not distinguish between a gene and its protein product here) are involved in the inflammatory response.  It has been shown that microglial cells, which are the brain’s version of the macrophages of the immune system that cause your skin to get red, swollen, and warm to the touch when your immune system is fighting off a bacterial infection, are altered in the brains of AD patients.  People who have taken anti-inflammatory drugs long-term for other conditions are less likely to be diagnosed with AD.

Lipid/Cholesterol Metabolism.  APOE (pronounced A-po-E, all long vowels) is an essential protein on the surface of the lipoprotein particles that transport lipids and cholesterol (essential nutrients and components of every membrane in every cell in the body) through the circulation.  There are several APOE genes, and APOE4 is the number one risk factor for sporadic (i.e., non-familial) AD. [2]  The connection between APOE4 and AD remains to be determined.  Too much information, but a description is here (scroll down to Alzheimer Disease 2).

Vesicle trafficking and Energy Metabolism.  Cargoes are transported in our cells in membrane-bounded vesicles. Several putative AD genes are involved in this process, which is ubiquitous in eukaryotic cells and particularly essential for neuronal function.  Other cellular processes implicated in AD include energy metabolism.  For example, mitochondria (the powerhouses of the cell, as we all learned in high school biology) are disrupted in the brains of AD patients.  This could lead to an energy deficit and oxidative damage in specific areas of the brain, with inflammation to follow.

Infection.  Bacteria or viruses may also be involved in the etiology and progression of AD.  Early research was promising and a current review in an authoritative journal outlines the potential involvement of herpes viruses in AD.  Viral infection would lead to neuroinflammation, and my gut feeling, for whatever it may be worth, is that an infectious agent is involved in early stages of AD in many if not all patients.

Neurotransmission.  And finally, data point to dysregulation of neurotransmission in AD.  The AD drug Aricept (donepezil) increases the level of the neurotransmitter acetylcholine in the brain.  Aricept improves cognition at the margin in some patients by some measures but does not prevent progression of disease.

Thus, many biological mechanisms could contribute to the origin(s) and progression of AD.  But concomitant with the development of modern molecular biology (i.e., gene cloning), the ACH

 became dominant and virtually exclusive when the amyloid precursor protein (APP) was cloned in the mid-1980s.  Here was a signal triumph of modern biomedical science, at a time when an analog of Kraepelin’s thesis (i.e., mental illness is caused by a biological/structural problem in the brain) became dominant: Disease is caused by mutant genes, and when we know the gene, we can find the cure in short order. [3]  Thus did the ACH become the Central Dogma of Alzheimer’s Disease: APP makes amyloid makes Alzheimer’s disease.  The Central Dogma of Molecular Biology, dating to the late-1950s (Francis Crick), is DNA makes RNA makes Protein.  With a minor modification for the existence of retroviruses such as HIV that integrate into the human genome (viral RNA makes DNA makes RNA makes Protein), this Central Dogma provided the theoretical basis for modern molecular biology and genetics.

Scientific politics and institutional imperatives, but not scientific research, soon confirmed the Central Dogma of Alzheimer’s Disease.  The National Institute on Aging (NIA) was established in 1974.  Since we all hope to grow up without growing too old and decrepit to enjoy maturity, NIA was a natural expansion of an NIH that now has 27 institutes and centers.  Without going into the details, a marketing strategy to develop support for the NIA “required a new reshaping of the narrative.  Alzheimer’s disease needed to become recognized as the most common form of dementia.”  Yes, scientists market their brands, contrary to their self-image.  This worked.  In 2021, 70% of the NIA budget was devoted to AD research.  Virtually all this funding was in support of the Central Dogma of Alzheimer’s Disease. [4]  In Herrup’s memorable formulation “The amyloid flea is wagging the Alzheimer’s tail which is wagging the NIA dog…This is not how you study a human disease.”

This is where Big Pharma comes into the picture.  Aricept works at the margin for some patients but does not halt or reverse progression of AD.  Aricept is a cheap chemical, not a blockbuster drug.  On the other hand, a cure for AD would be worth billions.  And this explains why over the past 25 years a series of expensive Phase III clinical trials have failed to make any substantial progress in confirming the Central Dogma of Alzheimer’s Disease.

A brief narrative would include the following:  Soon after amyloid was identified a mouse model that expressed human amyloid was constructed.  A “vaccine” against this amyloid resolved the plaques.  This was tried in human subjects.  Those who were successfully immunized against amyloid nevertheless continued to decline at a rate virtually indistinguishable from patients who were not immunized.  This intervention was clearly not working in humans as it did in mice.  In other models, mice do not exhibit the progressive loss of function like that in humans.  They have defects but they do not get worse as the mice age.  If these model mice are vaccinated against amyloid these defects go away.  It goes without saying, but this is apparently not appreciated, that if these triumphs of modern molecular genetics were a good model of human AD, even the best result would not restore function that had been lost.  AD progression in humans is inexorable and irreversible.

So, why has Big Pharma persisted up to now?  Herrup proposes three reasons that are not mutually exclusive: (1) Stubbornness – associated with the myth of the omniscient (and very rich) CEO of the modern transnational Big Pharma firm, (2) Greed –  throwing good money after bad will lead to that billion-dollar breakthrough, and (3) Bad advice – in this Herrup is direct to the point of rudeness.  But I have been following AD research closely for about 25 years, since I worked with colleagues who were very good scientists enthralled with the ACH.  One was a physical biochemist whose research was far beyond my ken.  He could expound at length on the structure of soluble (inconsequential) versus insoluble (pathognomic) amyloid but almost nothing about the disease itself.  Here is Herrup’s take on bad advice to Big Pharma:

What began as maneuvers of academic politics – inflation of the meaning of Alzheimer’s disease and the suppression of nonamyloid research – ended up creating a bevy of inbred (Ouch!) experts who were repeatedly and exclusively called upon as advisors to the industry.  So certain were these experts if their own models of Alzheimer’s disease that they repeatedly reassured the CEOs they were on exactly the right course.

This is the magical thinking leading to Aduhelm (aducanamab), the brand (keyword) that did not work, despite the wishes of Biogen (one of the first Biotech companies, established in 1978 before biotech even existed; founders Walter Gilbert and Philip Sharp were awarded Nobel Prizes) and the Food and Drug Administration of the United States.

What is the current state of the ACH/Central Dogma of Alzheimer’s Disease?  Remember the controls that Alois Alzheimer did not do?  They have been done.  Upon postmortem examination of brains of healthy people who died in old age but showed no signs of even mild cognitive impairment or dementia, up to one-third of these brains have enough amyloid plaques to support a diagnosis of Alzheimer’s disease they did not have.  Therefore, plaques (and Tau tangles, ignored here but no more likely than amyloid plaques to cause AD) are not the cause of AD.  In the mouse model of familial AD, which is probably what the 51-year-old patient zero Auguste D had, the mouse disease does not resemble human disease.  These mice develop plaques but have few other health problems.  They are also hard to distinguish from the wild-type controls in their behavior (grooming, feeding, social activity).  Contrast this with human patients in memory care units with their contemporaries who do not have AD.  Moreover, when AD mice are successfully vaccinated against amyloid, they are restored to “normal” within days.  Contrast this with any AD treatment to date in human patients.  Aricept may marginally slow cognitive decline in AD patients, but it does not reverse it.

The Central Dogma of Alzheimer’s Disease was only that, dogma.  It has been an intuitive cul de sac.  Unfounded dogma leads nowhere.  It is past time, 100+ years after the first description of Alzheimer’s disease, for an approach that focuses on the basic biological mechanisms (plural) of AD.  This is the only way forward.  Before the biological bases of cancer and metastasis were understood, largely based on research within the last 50 years and accelerating in the past 25 years, clinical oncology was also limited to what may be called intuitive approaches such as the radical mastectomy– cut it all out – first performed by William Stewart Halstead (one of the four physicians who founded Johns Hopkins Hospital) in 1882.  The operation became even more radical in time, even after realization that it often did not prevent recurrence at a secondary site (metastasis) while mutilating the patient.  All of us have had friends or relatives treated in this wayb.  All of us also have friends whose breast cancer has been cured by modern, biologically based clinical oncology.

Alzheimer’s is a brain disease of aging.  Thus, the answer(s) to AD are likely to come from a deeper understanding of the biology of aging and the local and systemic effects of aging on the brain.  This will require a complete rebalancing of the research portfolio of the National Institutes of Health and other funding agencies so that the molecular, cellular, and organismal understanding of aging in the brain reach the levels of the extensive (but still sometimes insufficient and sometimes misdirected) knowledge we have of heart disease and cancer.  This rebalancing can be done only if the rewards of research are changed, and the Central Dogma of Alzheimer’s Disease is displaced as the ruling paradigm of AD research.

Basic biomedical research takes time and patience.  It is tedious and incremental and not amenable to fads such as the Central Dogma of Alzheimer’s Disease.  Very few useful scientific advances can be rushed.  Nevertheless, the tenure clock ticks inexorably for the young scientist in a research university or medical school or NIH, so one must go with the flow to be successful in a dogfight that often turns nasty.  This is one major reason that biomedical research can lead to dogmatic dead ends, some of which are filled with untoward behavior (covered here before).  And this recently revealed outrage involving research on Alzheimer’s and Parkinson’s disease, which is unbelievable but probably true! [5]

Finally, as Karl Herrup puts it toward the end of his book, with authority:

It is precisely because we have relied so heavily on this hurry-up-and-do-something approach that we are in the situation we find ourselves in today – no cure and few clues…We must do something…(but)..there is nothing ethical about endlessly repeating expensive trials based on old, unreliable disease models such as the amyloid cascade hypothesis.

The “hurry-up-and-do-something approach” in biological research has a dismal history, one that most recently has failed to lead to any real solution to a worldwide pandemic. [6]  That particular dogma was “vaccination is the one true path to success for one novel disease.”  Yes, vaccines are an essential tool in the control of infectious disease.  But not always, and especially not recently.  This was known regarding coronaviruses as early as the 1950s.  Unfortunately, the importance of vaccination in preventing infectious disease has been called into question by the institutional and political responses to COVID-19.  This has the makings of catastrophe.  Dogma can be deleterious.

The Amyloid Cascade Hypothesis told us that the definition of Alzheimer’s disease is the presence of amyloid plaques in the brain.  This is not true.  Many of us of a certain age have enough plaques in our brains to support a diagnosis of AD but without dementia or even mild cognitive impairment.  Going forward the definition of Alzheimer’s disease must be based on the clinical symptoms of the patient, not on protein deposits and tangles remaining after cells die in the patient’s brain.  The symptoms of AD cannot be understood until research on AD is based on a biological model of the disease.  The health and wellbeing of millions depend on this.  The likely paths (not path) have been outlined above.  Other candidates will emerge, perhaps from a completely unexpected direction, which would include the function of APP itself.

This is how you study a human disease.

Notes

[1] Much of what follows is based on this book, which is an astonishing read that illustrates everything that must not be done when studying a human disease.  I recommend it highly, as a good read and a source for those who want to dig deeper.  I can also imagine responses from the usual suspects to this book, which I have not read but plan to get to soon in another context.

[2] Sporadic, i.e., non-familial, disease.  Familial AD accounts for relatively few cases of AD.  It was thought originally that cleavage of Amyloid Precursor Protein (APP) to produce amyloid plaques would reveal the secret of AD.  This research has revealed much about amyloidogenic fragments but less about the disease itself.  Similarly, Familial Amyotrophic Lateral Sclerosis (FALS) is associated with the enzyme superoxide dismutase (SOD1), which converts the highly reactive superoxide anion to hydrogen peroxide, which is then split by the enzyme catalase to produce water.  FALS represents only 5% of ALS cases.  The connections between SOD1 and ALS remain obscure, but the courses of disease of familial and sporadic ALS are similar.

[3] I was a research laboratory technician/coordinator when the Cystic Fibrosis Transmembrane Conductance Regulator (CFTR) was cloned in 1989.  A drug combination has recently shown promise in treating the most common form of cystic fibrosis.  It was essential to know the gene/protein associated with CF to reach this point, but these drugs are simple and expensive compounds that help misfolded mutant CFTR find its way to the cell membrane of cells in many tissues other than the lung and function properly when it arrives.  Molecular genetics per se has nothing to do with this and there is no prospect of gene therapy for cystic fibrosis.

[4] In the 1990s Karl Herrup was told that if his group was not studying amyloid, they were not studying AD.  This rings true.  Ten years later I was told that I was wasting my time with a proposal on the cellular function of a protein that interacted with huntingtin (the protein associated with Huntington disease, HD).  All support was instead directed at huntingtin itself.  This struck me as odd, but I was (not so) young but (still) naïve.  It turns out that “my” protein is probably not directly related to HD, but there was no way to know at the time.  The protein is involved in intracellular vesicle trafficking, which has been associated with AD.

[5] To cut to the chase: “After Science brought initial concerns about (Eliezer) Masliah’s work to their (University of California-San Diego and NIA) attention…a 300-page dossier…revealed…a steady stream of suspect images between 1997 and 2023 in 132 of his published research papers.”  The analysts, who were not paid by Science noted “In our opinion, this pattern of anomalous data raises a credible concern for research misconduct and calls into question a remarkably large body of scientific work.”  As illustrated in the table in the news article, Masliah is among the world’s top 10 scientists in certain subfields, including mouse models of AD.  Scientists at the end of the piece “worry about (this apparent catastrophe) giving science a further black eye, just as the public’s confidence in science is sinking to new depths.”  My reaction is unprintable.  Another link is here if the Science link is paywalled.

[6] A comparative analysis of our collective responses to Alzheimer’s disease, AIDS, and COVID-19 will provide useful lessons for the future, but only if the entire research community is reorganized for biomedical science instead of “Bioscience” that could make a lot of money.  The scientists and the infrastructure are in place and can be repurposed to the benefit of all.  We need only the wit and the will to discard Neoliberalism – “The market is the measure of all things, even those that cannot be measured.”