The words “tangle” and “Alzheimer’s disease” (AD) often occur together in articles about the disease. There, “tangle” usually refers to a specific structural aberration found in the brains of AD patients. It denotes the mesh-like bundles of protein threads, composed of the brain protein called “tau”, often abundant in these brains though not in normal ones. They are termed neurofibrillary tangles (NFT) and they are distinct from but frequently accompany the deposits, termed “plaques”, of a second group of proteins always found in the brains of AD sufferers, namely amyloid-beta (Aß) proteins. In the title of this piece, however, “tangle” refers to the confusing set of ideas and interpretations about the nature of AD itself and the disputes about current AD therapies. These two aspects of AD – the scientific controversies and the arguments about therapies – are, of course, closely connected. Furthermore, these matters have become entangled with larger issues that go beyond AD itself, involving not-infrequent fraud in clinical research and the fact that major findings in it are too often not found to be repeatable. Those controversies then feed back into the basic scientific questions about the conditions being investigated. That has happened in this case. In the year 2024, the science of AD is indeed a tangled mess.
Before going into the details, however, let us note an irony of the situation. AD is, indeed, a major health crisis for humanity, afflicting by one estimate about 50 million people in the world today, a figure that is probably an underestimate. Yet, this disease burden is an indirect consequence of the major successes of public health measures of the last two centuries or so. As a result average human life expectancy has greatly lengthened, by decades, from the late 19th century in much of the world. The specific improvements have been in: sanitation generally, water and air quality, maternal care, new or improved therapies (vaccines, antibiotics, new drugs), and more reliable food supplies and better nutrition. Because of the resulting expansion of human lifetimes, the numbers of people reaching old age have also greatly increased. Correspondingly, since AD is a disease of old age, the numbers and percentage of people experiencing AD have ballooned. This particular situation illustrates a larger truth: when one problem or set of problems is solved, the solution(s) often generate new ones. (One sees this all the time in the realm of social and political changes.)
The fact of senile dementia has been recognized as long as human civilizations have existed and cases probably existed even in prehistory, in earlier hunter-gatherer societies although these presumably would have had far fewer elderly individuals, hence many fewer cases. For millennia, the diseases of old age, especially senile dementia, remained just a regrettable fact of life.
A crucial insight was provided, in the early 20th century, by a German doctor with a strong interest in diseases of the nervous system. His name was Alois Alzheimer (1864-1915). One of his early patients, in 1901, was a woman patient showing symptoms of senile dementia at age 51, hence well before her mid60s, the more typical times of onset for this condition. As with other cases of senile dementia, her earliest symptom was increasing loss of short-term memory, which soon changed into a more general memory deficit, then loss of other cognitive capacities, then the loss of still more basic capacities for looking after herself, and controlling her bodily functions. When Alzheimer did a post-mortem examination of her brain, it was found to be not only reduced in size, relative to brains of people her age not suffering dementia, but larded with deposits of a fatty substance, the “plaques” mentioned above. This substance was later found to be the substance now known as amyloid beta protein or Aß. There were also dense tangles of a thread-like structure, the neurofibrillary tangles or NFT.
The “Eureka” moment for Alzheimer was the discovery, in the post-mortem examination of the brains of older people who had had typical senile dementia was that these showed similar characteristics: reduced brain size, many plaques of Aß throughout the brain and many neurofibrillary tangles. In some of these brains, the NFT were even a more prominent feature than the plaques but both were features of the brains of those who had died with senile dementia. Before this discovery, it had been assumed that people with “early onset” dementia, like Alzeimer’s first patient, had a different condition than those who succumbed to senile dementia in old age. After this finding, it was realised that the end-pathology was essentially the same in both forms of senile dementia but that there must be different initiating triggers. As is appreciated today, a particular form of biological dysfunction can be the outcome of different initiating conditions. An example is chronic fatigue syndrome (ME/CFS) and “long covid”, whose main symptoms are the same but which begin differently. (see prior article)
A further difference between early-onset and late-onset senile dementia was a marked familial inheritance pattern in the early-onset condition that was not associated with the more typical cases of dementia amongst the elderly. Later it would be found that mutations in certain genes were strong risk factors in early-onset dementia but that these mutant genes were not involved in more typical, late onset, cases of what had become known as Alzeimer’s disease. (The disease was named “Alzheimer’s disease” in 1910 by Emil Kraepelin, a psychiatrist colleague and friend of Alois Alzheimer.)
When one condition, let us call it A, is found in association with another, B, the question arises as to the nature of the connection. Is it just a coincidence and unimportant correlation? Or is it an indication of one causing the other, especially if A always precedes B in time? In this particular case, the question was: were the plaques and the NFT that slowly accumulate in the brains of people with AD just a side- accompaniment of the disease or were they actually the cause of it?
The latter – that the Aß plaques and the NFT were actually causative of neural damage in senile dementia – was formally proposed in several influential papers, one published in 1984, a second set (from three different labs) in 1991. By the late 1980s, the evidence was beginning to be much stronger that the accumulation of Aß plaques and NFT were both damaging to brain cells and specifically to the neurons of the brain. This evidence involved even tighter correlations in humans, various genetic and biochemical findings in human brains, and both experiments in mice and in in vitro experiments on neurons in culture. A key idea in these early papers was that the reduced size of brains in Alzheimer disease sufferers reflected massive neuronal death in those brains, triggered in particular by the lethal effects of the Aß plaques. Of course, such cell death could and would almost certainly also affect the neural “wiring” in such brains, further affecting cognitive function.
There were complications in the picture of course, not least some evidence that the NFT might themselves be an accelerating factor in plaque accumulation, at least in some patients. Hence, Aß proteins are probably not the sole cause of the neuronal damage in AD. Another complexity is that some brains in people who showed no signs of AD were found to have significant amounts of Aß plaques. Hence, while Aß plaques might be the main source of neuronal damage, there is no simple quantitative relationship, at least no universal one, between its accumulation and the extent of neuronal damage.
Not least, there was a growing list of “risk factors”, many of them variant genes, not all of whose effects could be directly or easily tied to Aß. By the late 1980s, at least 20 risk factors had been identified. Today, the list of known risk factors exceeds 60 and may be considerably more. (See no. 10 in this series on “risk factors”.) They include: longevity itself, gender (AD is much more frequent in women than men), obesity, exposure to certain heavy metals, insufficient sleep, herpes virus infection and various forms of malnutrition. This plethora of risk factors raised further questions. Could all of these risk factors be ones that were simply increasing the rate of Aß accumulation or its amounts? Or were the Aß plaques just one of several kinds of damage to the brain in AD?
These questions in turn are, of course, highly relevant to questions about therapies for AD. If Aß proteins were not the sole cause of senile dementia, but perhaps only a contributing cause, should it be the sole or even primary focus of attempted therapies? If not, what other elements or processes in AD should researchers concentrate on?
Nevertheless, the evidence in favour of the Aß hypothesis on AD causation grew from the late 1980s onwards. And in 2006, a particularly influential paper involving experiments in mice was published that seemed to demonstrate that a particular form of Aß protein was directly toxic to neurons and perhaps was the main agent of the pathogenesis associated with plaques. The approach involved turning on synthesis of these proteins, by a sophisticated genetic trick, and the work involved a proper cause-effect kind of test, not merely a correlative study.
Of course, the biotech and pharmaceutical corporations were taking note of these developments. Increasingly, it seemed a good bet that therapies aimed at reducing Aß accumulation might be the best way forward in treating AD. Yet, the problems in designing such therapies were formidable. This was reflected in the slow progress of such work. Indeed, it could be argued that there was essentially no progress in AD therapy for more than three decades. (In contrast, there were major therapeutic advances in other difficult diseases during this period.)
That changed this year, 2024, with the launch of two drugs. These were both monoclonal antibodies against Aß proteins and these were delivered by injection into the bloodstream. A critical problem in this approach is the blood-brain barrier, the physiological mechanism that the mammalian brain has for restricting soluble proteins in the blood from reaching the brain. Yet, with large enough amounts of these proteins, sufficient quantities were reaching the brains of the patients to create a measurable reduction in the deposition of the Aß plaques. The clinical trials of these two drugs, over a period of many years, had suggested that the cognitive decline in AD patients was also measurably, significantly, slowed by this treatment.
This was the best result achieved, with the reduction in such decline on the order of 30%. This was neither a cure nor a complete prevention of mental decline, yet it was significant and constituted a genuine advance in fighting AD though still far short of what is needed. It was, however, very dependent on the particular brain physiology of the individual patient. In particular, the drugs were found to be most effective in those with larger amounts of NFT than those with smaller amounts. This suggested an interesting interplay between these two proteins in creating the loss of neuronal cells and cognitive decline. Furthermore, there was no significant reduction in mental decline in AD sufferers who were more seriously affected. In other words, the therapeutic effect was highly context dependent for these two treatments.
Needless to say, evaluations of the benefits of the new therapies, were, and remain, highly controversial. First, there was the question of whether this incomplete and partial protection was sufficient to justify heavy reliance on them. Second, they are highly expensive. Even within the primary target group of AD patients in North America and Europe, those who are the most likely to receive them, there will be many who simply cannot afford them. Nor is it clear that their production could be rapidly ramped up to supply that group even if cost was not a consideration. As for people in the poorer countries outside of the traditional West, there will be very few indeed who will benefit from them in the near term.
Another problem that is little discussed in the popular press is that once a therapy for a difficult medical condition has been developed, usually with a considerable investment by the companies involved, that tends to inhibit developing new therapies, some of which might, in the long term, prove to be more effective. The company or companies involved in the initial developments, quite naturally, want to expand the market for their treatments. In effect, a partial advance might slow down the development of more efficacious ones. Whether this potential problem becomes an actual one in the case of the present monoclonal antibody therapies for Aß proteins remains to be seen but it is one that needs discussion.
If these complications were not enough, another one developed this year. Recall the landmark paper published in 2006 noted earlier. For the past few years, it has come under scrutiny. It was claimed by whistle-blowers, who had studied it carefully, that some of the critical images had been manipulated to create a false impression. In effect, this was a charge of fraud. If validated, this is the worst thing that can befall a paper. And a few months ago, it was retracted by the authors. Hence, a critical piece of evidence supporting the Aß hypothesis has been withdrawn, and at just the point where therapies, based on that hypothesis, had become available. The retraction does not, of course, falsify the hypothesis but it weakens the grounds for believing in it. And if the Aß hypothesis is not the whole story of the causes of AD, this raises questions about how exclusively therapies should be based on it.
Let us try to summarize where matters stand in understanding the science of AD. First of all, the picture of the causation of AD remains somewhat cloudy. Over all, the evidence is good that Aß accumulation in the brain is a causative effect of damage to human brains. Yet, it is still not clear that it is the sole cause. Indeed, that seems unlikely at this point. The retraction of the 2006 paper has certainly removed one major prop of this idea though other observations, as noted above, contribute to the doubts. Secondly, the new therapies, whatever their benefits should not be accepted as the end-point for new treatments. They are only a palliative, not a cure. For the sake of the prospective many 10s of future AD sufferers, not to mention the already 50 or so million present ones, there needs to be space for new therapies.
Perhaps a small step forward would be to embrace the full complexity of the problem though this seems counter-intuitive. Normally, in science, one looks for the simplest hypothesis accounting for the facts – so called “Occam’s razor” – and then tries to validate or falsify it. The Aß hypothesis was such a hypothesis: Aß proteins were proposed to cause neuronal death and their accumulation in the brain the source of the pathogenesis. It made initial sense to concentrate on reducing its accumulation in the brains of affected people.
However, given the plethora and diversity of the risk factors, we need to see the pathology as the result of a long sequence of events, many of which are still little known, but which lead to Aß accumulation and NFT, and perhaps other changes that contribute to neuronal death. If there is actually a sequence of causative events that precede those harmful consequences, then interfering with those earlier steps might prove to be the key. To simplify, picture the process of pathogenesis as a sequence of molecular/cellular events, each triggering the next, let us call it a pathway of pathogenesis;
A → B → C → D → ……………………………. → Aß, NFT, plus other pathogenic substances(?)
Of course, there may be not one but two or more distinct pathways that end up with the same “bad products”. Nevertheless, interfering with any of the early steps might be just as efficacious as trying to stop things near the late steps.
That of course would require identifying those early steps. Yet, the known risk factors provide possible hints to their nature. Perhaps, far more study using mouse experimentation, and looking at those variables, could unearth some valuable information about the earlier steps. If so, those identifications might suggest new biochemical or molecular targets for intervention. This is only a suggestion with no guarantee that it would succeed but it is worth considering.
In the meantime, there could be a massive public health program focussed on the known risk factors and reducing human exposure to them. Expensive? Yes, probably. Intrinsically hard? No. We have an example of such an approach: the campaign to reduce cigarette smoking, focussed on a known major risk factor of lung disease, has undoubtedly saved millions of lives. The fight against AD is more complex and would be more difficult but difficulty should not be grounds for passivity. I would say that we owe it to our families, friends and our collective human posterity to try this public health approach while the search for new therapies proceeds.
Supplementary reading:
Selkoe, D.J., Hardy, J. (2016). The amyloid hypothesis of Alzheimer’s disease at 25 years. EMBO Molecular Medicine 8: 595-608.
A thorough review but ending with a basically positive evaluation, of the amyloid beta hypothesis of Alzheimer’s disease.Armstrong, R.A. (2019). Risk factors for Alzheimer’s disease. Folia Neuropathologica 57: 87-105
A good review and discussion of the many risk factors for AD that have been identified and an argument for taking into account the importance of risk factors known to be important in ageing.Couzin-Frankel, J. (2024). New Alzheimer’s drug clears FDA advisory vote despite unknowns. Science 384: 1164-1165.
This news article covers the approval by the FDA of the latest monoclonal antibody against Aß and the questions surrounding this new therapy.Piller, C. (2024). All the Alzheimer’s research we did not do. The New York Times, July 7, 2024, op-ed section.
A NYT piece on how the stress on the Aß hypothesis has perhaps shifted attention away from other hypotheses, possibilities, and approaches.