Amyloid Toxicity: Myths, Contradictions & Failures

For over a hundred years, since Alois Alzheimer published his famous plaque and tangle pictures, it has been taken for granted that these plaques and tangles are actively killing the neurons by being toxic. There was no proof needed of why or how they are toxic, the details and mechanisms were to come eventually over time to prove what we already knew all along, these structures are toxic. Observations and experiments were not performed to test the toxic gain-of-function (GOF) theory of amyloid pathology, but just to verify what we already know, they are toxic.

Here, and exactly here, is where the pathology (as a scientific discipline) got its own disease; confirmation bias. No amount of counterevidence was ever enough to challenge GOF. Any counterevidence was either accommodated somehow or simply dismissed, leading to a group of myths and contradictory arguments that are recited almost religiously in papers, reviews, conferences, and media interactions to defend GOF. A smokescreen of decisive claims that carefully hides the weaknesses of GOF and buries any serious challenges to its claims in obscurity. However, while this was successful for GOF dominance over the field, it never led to a successful treatment (and this includes aducanumab), because as Richard Feynman once said, "For a successful technology, reality must take precedence over public relations, for Nature cannot be fooled."

Here, I will discuss the major myths and contradictory arguments that are usually weaved together in acrobatic and confusing statements to give us the epistemological mess of a theory that is GOF. But first, I will discuss why well-meaning scientists stick by a clearly weak and ineffective theory. Why??

Why the GOF?

The answer is one word: genetics. It is the firm belief that there is no other way but GOF to interpret the genetic evidence that links familial forms of neurodegenerative diseases to the proteins and peptides found in the amyloid plaques. In the absence of a better alternative, GOF proponents assert, this is the best we have got, and we must keep trying intellectually and experimentally to make sense of it until we succeed. Sounds reasonable, only if it were true, but it is not. The root conviction that genetics must mean GOF is simply wrong. The genetic evidence clearly indicates the causal involvement of these proteins in the respective diseases, but it does not indicate how they are involved. Genetics points out the cause, not the mechanism, and GOF is not the only mechanism by which these proteins can cause harm.

When a protein aggregates, two things happen simultaneously (like in an hourglass, Fig. 1); as the aggregated species accumulate (GOF), the soluble species gets depleted, and its function is lost (loss-of-function - LOF). The genetic evidence is essentially agnostic and neutral towards the question of whether GOF or LOF is the pathogenic mechanism. So why? Why LOF is ignored despite being equally compatible with the genetic evidence, while GOF gets diehard support? The answer lies in a group of myths that are repeated without actual questioning or validation to dismiss LOF out of hand without any serious consideration. These myths include the following.

Figure 1. Protein aggregation and soluble protein depletion happen simultaneously.  

1. The “no phenotype” argument: Knock out/down animals of the amyloidogenic proteins show no phenotype

This is just a blatant fallacy. However, it is usually the first argument that is used to dismiss LOF as irrelevant and inconsequential. While it is easily provable to be just a myth, it also flies in the face of common sense by assuming that the tens of proteins that are involved in amyloid pathologies, which are evolutionary conserved and produced normally throughout life, are without function and their depletion has no consequence whatsoever. Moreover, there are literally tens of papers showing the devastating effects of knocking out/down Aβ, Tau, α-synuclein, PrPC, SOD-1, TDP-43, and other proteins that are involved in neurodegenerative diseases. We cite many of these papers in our review (1 , Box 1) and they are cited in several other reviews as well (2–4). The “no phenotype argument” does not make biological sense in the first place, besides being demonstrably false.

2. OK, they have functions, but we don’t “fully understand” their functions

Well, we don’t “fully understand” anything, and science never claims that it will ever do. We can only understand things to a certain extent and increase our understanding incrementally over time. We don’t “fully understand” why the corona virus affects certain people differently and we don’t “fully understand” all the causes of diabetes, but this still hasn’t paralyzed us from developing successful vaccines and replacement therapies based on what we know in virology and endocrinology. Moreover, we don't “fully understand” the mechanisms of toxicity (if any – (see below) ) of amyloids, but we still spend billions developing drugs targeting them. So, the mythical and elusive ideal of “fully understanding” is applied selectively to LOF but not GOF.

Additionally, there is plenty of literature demonstrating the important functions Aβ and α-synuclein for example (for Aβ roles in neurogenesis: 10 refs here and synaptic plasticity: 10 refs here, for α-synuclein check ref. (3)).  Despite all this knowledge, there is no point that can be reached where we can claim that we “fully understand” their functions. However, if we compare LOF and GOF to a more realistic and rational standard of how they relatively compare to each other in terms of what we know about both, LOF easily wins. In addition to what we know about the important functions of these proteins in neurogenesis and synaptic plasticity for example, LOF is more compatible with the rest of our biological knowledge. Evolutionary conserved and highly expressed proteins (such as Aβ and α-synuclein) are supposed to have important functions, and protein depletion, if not compensated, is never good nor inconsequential. GOF on the other hand struggles with its basic concepts and mechanisms, and fails to answer simple questions like: what is the toxic species? What are the physicochemical mechanisms of toxicity or prion “protein-only replication”? Thus, in realistic standards, we understand much more about protein function and the detrimental effects of protein depletion than the hypothetical GOF species and mechanisms.  

3. Patients, especially with mutations, have high toxic protein levels, don’t they?

This is probably the biggest myth of all, entirely false, yet very common. Clinical evidence consistently shows that the soluble levels of the proteins involved in neurodegenerative diseases are LOWER in patients compared to controls. Patients with familial forms of Alzheimer’s disease, both due to mutations (5) or gene duplication (Down syndrome (6)), have lower levels of Aβ42 in the CSF. The same is true for familial forms of Parkinson’s disease and Creutzfeldt-Jakob disease, where the levels α-synuclein and PrPC are lower in patients (7, 8). This is also the case in the sporadic forms of Alzheimer's, Parkinson's, and Creutzfeldt-Jakob diseases (9–11). You expect a toxic protein to be higher in patients, but the opposite is always the case. In fact, a lower soluble level of the protein is a much better predictor of disease status and progression than the amount of aggregates (12). Additionally, Aβ42 depletion is associated with cognitive impairment and dementia in many neurodegenerative diseases that do not specifically harbor Aβ amyloid plaques. These diseases include: frontotemporal dementia, amyotrophic lateral sclerosis, Parkinson’s disease, Parkinson’s disease dementia, dementia with Lewy bodies, multiple system atrophy, progressive supranuclear palsy, and corticobasal degeneration (13). Moreover, we have recently found that high levels of soluble Aβ42 in the brain are associated with normal cognition in plaque-positive individuals, no matter how high the plaque burden is in their brains (14) (Fig. 2).

Figure. 2 from Sturchio et al 2021, showing that high levels of soluble Aβ42 preserve cognition regardless of the amyloid plaque burden. 

The abovementioned myths are examples of the easily demonstrably false statements that are used to defend GOF at the expense of LOF. However, GOF is not only inconsistent with the evidence, it is also inconsistent with itself. This can be demonstrated by how contradictory and confusing some of GOF’s basic assumptions are, such as:  

1. The most basic GOF premise is that the plaques are the hallmarks of disease, but are they? 25-35% percent of cognitively normal elderly have amyloid plaques in their brains (12). How can something be called “hallmark” if it cannot distinguish between healthy and affected individuals in a quarter to a third of the cases? It is like finding an Eiffel tower in every third or fourth city you visit. At some point, it seizes to be considered a landmark of Paris.

2. GOF was supposed to be based on the toxicity of the hallmark (which is not quite a hallmark), but is it toxic? According to the founders of the two major GOF hypotheses (the amyloid cascade hypothesis and the prion hypothesis), plaques are not toxic, but might actually be protective (Fig. 3).

Figure. 3

3. So, a non-hallmark-hallmark that was supposed to be toxic turned out to be protective, probably. Why are the GOF proponents still advising the industry to target the non-toxic, probably-protective plaques?

4. Also, if the plaques that are composed of amyloid fibrils are not toxic, even protective, why is the shape/polymorph/strain of amyloid fibrils important in the pathology if it is not pathological to start with? How would a particular polymorph/strain of fibrils be important in propagating a particular type of pathology if it is protective and non-toxic?

5. If toxicity is not mediated by the non-hallmark-hallmark that turned out to be protective and not toxic, but mediated by an elusive soluble substrate that is called oligomers, how come the patients have LOWER levels of the soluble proteins, the fuel of the supposedly-toxic soluble oligomers?

These are not open scientific questions; these are blatant contradictions and epic epistemological failures that should not be present in any theory that claims to be scientific. Taken together with the demonstrably false myths that are used to keep it going (see above), GOF is just a terribly bad theory that is neither consistent with available evidence nor with itself. Moreover, GOF’s mechanismless and contradictory assumptions are not even consistent with basic physicochemical and biological principles. In light of established physicochemical principles, amyloids are too thermodynamically stable (non-reactive) to be toxic themselves, and oligomers are too thermodynamically transient to be relevant. Furthermore, there is no physicochemical basis for the so-called “protein-only replication”, a process that can otherwise be fully explained by the well-established principles of phase-transition and nucleation (see review (1) for more details). In terms of established biological principles, proteins evolve to perform a function, not to be toxic, and uncompensated depletion of soluble functional proteins is never harmless. Thus, the powerful GOF emperor that has ruled the field for decades is naked of any internal or external consistency. The only thing that the GOF theory is consistent at is its failure to produce successful therapies, which is of course a very natural result given how weak and contradictory it is. Problem-solving is a function of the conceptual framework we use to solve them. Wrong, weak, or unfounded concepts inevitably lead to wrong, weak or unfounded solutions, in this case, therapeutics.

Better alternative?

There are many things that are fundamentally wrong about GOF, and it is time to run some epistemological clean-up and start over. We can start from the points we agree on, which are:

1. Genetics says proteins such as Aβ are linked to the pathology

2. Plaques are not toxic

3. The soluble levels of these proteins are lower in patients compared to controls

4. Proteins have functions and protein depletion is never inconsequential

Then LOF emerges as a logical necessity, and not a crazy or controversial idea. It is in fact the only way to save the genetic evidence from the convoluted, contradictory, and confusing argumentation of GOF. LOF is compatible with well-established concepts in physical chemistry and biology, and with data from animal models and clinical studies. The only thing remaining to support LOF even further is to actively deplete the soluble levels of the protein in the brain of healthy human subjects and see that it induces cognitive impairment, but who would do such an experiment?!

Unfortunately, this experiment has already been done, justified by GOF that depleting a functional protein in healthy human subjects is a good idea. The “successful” depletion of Aβ led to a smaller brain volume and worse cognition in the treated group compared to the placebo group (15, 16) (Fig. 4). Many other clinical trials targeting Aβ have led to a similar outcome, worse cognition, and more atrophy in the treatment group (17).

Figure 4. from Egan et. al 2019, showing that pharmacological depletion of soluble Aβ induces cognitive impairment in the treatment group compared to placebo.  

Why so harsh?

After the aducanumab scandal that threatens to bankrupt the healthcare system for a GOF-based drug that does not work, it is now more important than ever to be uncompromisingly clear and vocal about the root causes of failure. Science is not always about finding a middle ground if the data clearly points otherwise. We cannot keep coming with excuses for a clearly failed research program at the expense of patients’ time and money. Also, we cannot only blame Biogen for selecting data while the field has been mired in selectivity, myths, and contradictions for so long. The field is dominated by a singular unshakable assumption: amyloids are directly responsible for neurodegeneration via a GOF toxicity mechanism, and protein depletion, while present, is completely and absolutely inconsequential and irrelevant. This unitary position is not only dogmatic and irrational, it is also weak and inconsistent with many biological and clinical facts. To hide that, GOF was protected by a group of myths and confusing self-contradictory arguments that kept it alive for so long and kept any serious challenges to its authority at bay. Aducanumab approval is just the bitter fruit of a bad tree.   

It is time to end GOF monopoly and expose its flaws with the clearest and strongest terms, since it has now escaped the realm of academic discussion and poses real-life harm to the health and finances of millions of people. We need to be clear that the genetic evidence should not be used to extort people into accepting GOF, since LOF is equally compatible with the genetic evidence. Additionally, claims such as “knock-out animals do not have phenotype”, “we don’t fully understand the function”, and “patients have higher levels of the protein” are factually incorrect (myths), and should be called out as BS whenever mentioned. Moreover, we understand much more about protein function than about toxicity mechanisms, and the hallmark of the disease is protein depletion, irrespective of the plaque load.

LOF is not a perfect theory, no scientific theory is, but it is a far more compelling and consistent theory than GOF. It follows logically and naturally from what we know about physical chemistry, thermodynamics, evolutionary biology, and protein function and it is compatible with much of the animal and clinical evidence. It also offers a new way to think about the pathogenesis and novel path to develop therapeutics based on non-aggregating protein analogues as replacement therapy to restore the normal levels and functions of the depleted proteins. The replacement approach already works for diabetes using insulin and an amylin analogue (pramlintide), why not for neurodegenerative diseases?

GOF has managed to push itself to the forefront of research based on a group of contradictory and mythical assumptions. Now, it is pushing itself even further, from journal pages and conference halls into patients’ veins and pockets. This requires more than just rage against aducanumab, it requires exposing the whole research program for what it is, a total mess, and requires fighting for a better alternative. It requires a whole new LOFolution!

 

References

1. Malmberg,M., Malm,T., Gustafsson,O., Wright,A., Andaloussi,S. El and Ezzat,K. (2020) Disentangling the Amyloid Pathways : A Mechanistic Approach to Etiology. Front. Neurosci., 14, 1–11.

2. Kent,S.A., Spires-Jones,T.L. and Durrant,C.S. (2020) The physiological roles of tau and Aβ: implications for Alzheimer’s disease pathology and therapeutics Springer Berlin Heidelberg.

3. Benskey,M.J., Perez,R.G. and Manfredsson,F.P. (2016) The contribution of alpha synuclein to neuronal survival and function - Implications for Parkinson’s disease. J. Neurochem., 137, 331–359.

4. Saccon,R.A., Bunton-Stasyshyn,R.K.A., Fisher,E.M.C. and Fratta,P. (2013) Is SOD1 loss of function involved in amyotrophic lateral sclerosis? Brain, 136, 2342–2358.

5. Bateman,R.J., Xiong,C., Benzinger,T.L.S., Fagan,A.M., Goate,A., Fox,N.C., Marcus,D.S., Cairns,N.J., Xie,X., Blazey,T.M., et al. (2012) Clinical and biomarker changes in dominantly inherited Alzheimer’s disease. N. Engl. J. Med., 367, 795–804.

6. Portelius,E., Hölttä,M., Soininen,H., Bjerke,M., Zetterberg,H., Westerlund,A., Herukka,S.K., Blennow,K. and Mattsson,N. (2014) Altered cerebrospinal fluid levels of amyloid β and amyloid precursor-like protein 1 peptides in Down’s syndrome. NeuroMolecular Med., 16, 510–516.

7. Kasuga,K., Tokutake,T., Ishikawa,A., Uchiyama,T., Tokuda,T., Onodera,O., Nishizawa,M. and Ikeuchi,T. (2010) Differential levels of α-synuclein, β-amyloid42 and tau in CSF between patients with dementia with Lewy bodies and Alzheimer’s disease. J. Neurol. Neurosurg. Psychiatry, 81, 608–610.

8. Dorey,A., Tholance,Y., Vighetto,A., Perret-Liaudet,A., Lachman,I., Krolak-Salmon,P., Wagner,U., Struyfs,H., De Deyn,P.P., El-Moualij,B., et al. (2015) Association of cerebrospinal fluid prion protein levels and the distinction between Alzheimer disease and Creutzfeldt-Jakob disease. JAMA Neurol., 72, 267–275.

9. Meyne,F., Gloeckner,S.F., Ciesielczyk,B., Heinemann,U., Krasnianski,A., Meissner,B. and Zerr,I. (2009) Total prion protein levels in the cerebrospinal fluid are reduced in patients with various neurological disorders. J. Alzheimer’s Dis., 17, 863–873.

10. Stomrud,E., Minthon,L., Zetterberg,H., Blennow,K. and Hansson,O. (2015) Longitudinal cerebrospinal fluid biomarker measurements in preclinical sporadic Alzheimer ’ s disease : A prospective 9-year study. 1, 403–411.

11. Parnetti,L., Paciotti,S., Farotti,L., Bellomo,G., Sepe,F.N. and Eusebi,P. (2019) Parkinson’s and Lewy body dementia CSF biomarkers. Clin. Chim. Acta, 495, 318–325.

12. Villemagne,V.L. and Chételat,G. (2016) Neuroimaging biomarkers in Alzheimer’s disease and other dementias. Ageing Res. Rev., 30, 4–16.

13. Mollenhauer,B., Parnetti,L., Rektorova,I., Kramberger,M.G., Pikkarainen,M., Schulz-Schaeffer,W.J., Aarsland,D., Svenningsson,P., Farotti,L., Verbeek,M.M., et al. (2016) Biological confounders for the values of cerebrospinal fluid proteins in Parkinson’s disease and related disorders. J. Neurochem., 139, 290–317.

14. Sturchio,A., Dwivedi,A.K., Young,C.B., Malm,T., Marsili,L., Sharma,J.S., Mahajan,A., Hill,E.J., Andaloussi,S. EL, Poston,K.L., et al. (2021) High cerebrospinal amyloid-β 42 is associated with normal cognition in individuals with brain amyloidosis. EClinicalMedicine, 000, 100988.

15. Egan,M.F., Kost,J., Voss,T., Mukai,Y., Aisen,P.S., Cummings,J.L., Tariot,P.N., Vellas,B., Van Dyck,C.H., Boada,M., et al. (2019) Randomized trial of verubecestat for prodromal Alzheimer’s disease. N. Engl. J. Med., 380, 1408–1420.

16. Sur,C., Kost,J., Scott,D., Adamczuk,K., Fox,N.C., Cummings,J.L., Tariot,P.N., Aisen,P.S., Vellas,B., Voss,T., et al. (2020) BACE inhibition causes rapid, regional, and non-progressive volume reduction in Alzheimer’s disease brain. Brain, 10.1093/brain/awaa332.

17. Espay,A.J., Sturchio,A., Schneider,L.S. and Ezzat,K. (2021) Soluble Amyloid-β Consumption in Alzheimer’s Disease. J. Alzheimer’s Dis., 10.3233/jad-210415.