A Science-Based Falsifiability Test for the Amyloid Hypothesis (AHyp)

The identification of Aβ peptides as the major components of amyloid plaque and the discovery in familial cases of Alzheimer’s diseases (AD) of multiple mutations in the genes implicated in their production led to the formalization of the Amyloid Hypothesis (AHyp) [1]. The AHyp asserts that Aβ peptides misfold into toxic aggregates that initiate and drive the pathological mechanisms leading to neurodegeneration and AD. AHyp has been the dominant working hypothesis in the AD field and the primary paradigm directing the development of drug-based therapies for the last three decades [2].

However, despite thousands of studies intended to validate the AHyp, and after thousands of other studies presenting alternative AD explanations refuting the AHyp, the etiology of AD remains unsettled. Indeed, the authors of the recent NIA-AA research framework, which surprisingly redefined AD based on biological markers, stated that we are right where Alois Alzheimer, Oskar Fischer, and their colleagues left the field a century ago: “We do not know what causes AD, and we have no effective treatments. Although Aβ plaques and pathologic tau are the hallmark pathologic changes, we have limited understanding of how they came about. In this setting, it is important to examine all possible mechanisms” [3]. How can this be? How many dozens, hundreds, thousands, or even tens of thousands of studies does it take to prove or disprove a hypothesis?

According to their authors and supporters, each AD hypothesis is supported by strong evidence and persuasive arguments, as if there were a bottomless pit of findings and arguments that scientists can cherry pick (often with the help of a versatile reference manager) to support almost any idea. It is not clear, though, whether this disturbing reality has to do with science, in the sense that AD and other enigmatic neurodegenerative diseases, including Parkinson’s disease (PD), amyotrophic lateral sclerosis (ALS), and Creutzfeldt-Jakob disease (CJD), just happen to be very difficult to study and to understand, or whether we already have answers (see [4] for a glimpse of hope), but the problem is with the way the scientific enterprise operates, which precludes a resolution.

Although it would make sense to have hypotheses scientifically proven before they are used to direct very expensive clinical trials involving thousands of innocent patients, in the case of AHyp this rationale was turned upside down, and the clinical trials were projected to represent AHyp falsifiability tests. However, after each unsuccessful clinical trial, even after treatments that reached their objective of completely removing the amyloid, the AHyp falsifiability goalposts were moved (discussed in [5]).

One approach to advance the AD field would be to embrace a broader and more inclusive multifactorial perspective on AD etiology and on the biological role of Aβ in AD, as suggested two decades ago by Mark Smith, George Perry, and their colleagues [6-8]. Such an approach would advance the field by increasing the explanatory power of all reproducible data and observations, which currently are only regarded as significant within the narrow context of specific hypotheses and cannot be readily explained in the context of AHyp [5-11]. Very importantly, this multifactorial perspective would open AD to a myriad of immediate prevention and therapeutic approaches, which, unlike the amyloid-based therapies promoted by the AHyp and the NIA-AA research framework, would help not only with AD but with many other human health issues and diseases. Interestingly, the U. S. National Plan to address AD and related dementias was just updated [12] to include actions for promoting healthy aging by reducing risk factors, including physical inactivity, hypertension, smoking or excessive alcohol drinking, unhealthy diet, diabetes, infectious diseases, exposure to toxins, physical brain trauma, depression, and low cognitive/social/educational attainments, among many other damaging factors and conditions associated with aging. These sensible interventions should also relieve some of the anxiety/desperation in persons with dementia and their families (as well as in some of their supporting groups), who have been emphatically primed (or self-primed) to expect (or even demand) pharmaceutical solutions (discussed in [13, 14]).

Another way to advance the AD field would be to replace the current clinical-based falsifiability test of AHyp with a science-based test. Ideally, this test should be applied to other hypotheses in the AD field, as well as to those that address the etiology of PD, ALS, CJD, and other enigmatic neurodegenerative diseases. Furthermore, if this aspiration were not already high enough, the results of this test should also guide the way forward in virtually all basic, biomedical, and public health fields. This objective might appear to be part of a science fiction scenario, but it is based on a single, relatively straightforward genetic test that addresses all human biological traits and conditions by determining and analyzing the extent of human genetic variation. Fortunately, the genome sequencing technologies for generating this data are well established and routinely used in hundreds of laboratories around the world, so the only obstacle that might stand in the way of implementing this extraordinary project has nothing to do with science but rather with a poorly run scientific process.

The AHyp falsifiability test proposed here is straightforward: (i) individuals who do not produce the Aβ peptides (denoted here with the annotation Aβ -/-) due to certain mutations (i.e., genetic variation) in the genes involved in their production are expected to be completely resistant to AD-like dementia (in other words, “No Aβ, No AD”; not to be confused with “No Aβ amyloid/plaques, No AD”—the futile therapeutic approach associated with the therapeutic arm of AHyp), and (ii) individuals with reduced production of Aβ peptides (Aβ +/-) are expected to be at least partially, if not fully, resistant to AD-like dementia. The same rationale applies also to all the other proteins implicated in neurodegenerative diseases, including tau, α-synuclein, prion protein (PrP), and TAR DNA-binding protein 43 (TDP-43).

The major appeal of the AHyp was its rationalization in the context of genetic findings and analysis, which, when compared to biological and pathological findings, are usually more informative. However, the genetic data loses some of its clarity when, in conjunction with environmental factors, it is translated into a phenotype, both in health and disease. For example, although the early genetic data indicated that the Aβ peptides/assemblies play a casual role in familial AD, and possibly in sporadic AD, these genetic findings were neutral with regard to the disease mechanism—gain of (toxic) function (GOF) versus loss of (biological) function (LOF)—both of which were thought to be the result of a protein misfolding process leading to amyloid formation. So, why did the AHyp embrace the GOF paradigm? Was it due to a poor scientific judgment?

Not when judged in the context of two well-established paradigms in protein chemistry and biology, the protein misfolding dogma [15] and the Nobel Prize winning prion hypothesis [16], both of which consider the amylogenic process to be incompatible with physiological functions but conducive to pathogenic mechanisms leading to disease. However, the protein misfolding dogma and the prion hypothesis are both questionable and, likely, have confused or misled the thinking on the role of Aβ, tau, α-synuclein, TDP-43, PrP, and other amyloidogenic proteins in the disease process [17-23]. Nevertheless, even if we recognize that the advocates of the AHyp and its surrogate concepts, such as the notion of self-propagating pathogenic protein aggregates [24], have embraced the protein misfolding dogma and the prion hypothesis in good faith, on account of their (near) universal acceptance by the scientific community, the lack of attention to the biological function of Aβ and the other amyloidogenic proteins implicated in neurodegenerative diseases, which would enable scientists to assess the merit of the GOF and LOF paradigms, is difficult to excuse, especially considering that it was not for lack of warnings. For example, six years after the Nobel Prize for the prion hypothesis, Kurt Wunthrich (himself a Nobel laureate for his work on the three-dimensional structure of biological macromolecules) submitted that: “we must understand the function of the normal prion protein before we can understand prion diseases” (quoted in [25]). Or, consider the apparent distress of John Hardy, one of the AHyp founders, when critically reappraising his own hypothesis: “A major concern about the amyloid hypothesis is that we have very little idea as to the functions of APP or the possible function of Aβ” [26].

If the GOF model is indeed problematic, then the alternative paradigm, the LOF (see above), must be correct. Correct? Not so (see below). Nevertheless, unlike the GOF-based AHyp and its supporting paradigms (i.e., the protein misfolding dogma and the prion hypothesis), the LOF hypothesis brings the biological function of all these proteins front and center in explaining the etiology of neurodegenerative diseases, as it prompts the critical question: what kind of biological function, when lost, would lead to neurodegeneration?

Although the LOF paradigm (i.e., loss of biological function due to protein misfolding) in the causation of AD, PD, ALS, and CJD has recently been revived [27], the current genetic data does not have the power to differentiate between GOF and LOF mechanisms, except perhaps in the case of CJD and other transmissible spongiform encephalopathies (TSEs). In 2012, the first natural genetic variation, a nonsense mutation in the PrP gene, which leads to a PrP -/- phenotype, was discovered among healthy homozygous dairy goats in Norway [28]. These goats did not develop any pathology/disease analogous to TSE, nor did the numerous transgenic PrP knockouts generated in mice, sheep, goats, cows, and other animal models. This also appears to be the case with many Aβ, tau, α-synuclein, and TDP-43 transgenic knockouts, but the findings need to be corroborated by natural knockouts.

As I proposed previously, the aggregation of PrP, Aβ, tau, α-synuclein, and TDP-43 does not induce a LOF, nor a toxic GOF, but rather is an activity associated with their evolutionary selected biological function in innate immunity [22, 23]. The innate immunity function of these proteins confers on them both protective and destructive properties, a characteristic common to many other members of the immune system.

The major strength of the innate immunity hypothesis is that it can explain the results of many studies that have been used to support the other hypotheses in the field, including the majority of the findings believed to support the AHyp. The immune reactivity paradigm is gaining increased support (see [29, 30]), but it needs to be fully evaluated. The human genetic variation study proposed here would also go a long way toward deciphering the biological function of all proteins implicated in AD, as well as the function of all other human proteins. Moreover, by identifying the human genetic variation, including heterozygous and, possibly, homozygous knockouts, which would generate the Aβ +/- and Aβ -/- phenotype, respectively, the results of this extraordinary study would also provide a scientific falsifiability test for the AHyp.

A human genetic variation project of this magnitude might appear daunting. Although the logistics and financing of a study of this magnitude are challenging, they are nevertheless feasible and realistic. For example, a recent exome sequencing and analysis study of 454,787 UK Biobank participants indicates that sequencing five million individuals would facilitate the identification of 500+ heterozygous LOF carriers for the majority of human genes [31]. Considering that the heterozygous or even homozygous knockouts of the genes coding for Aβ, tau, α-synuclein, TDP-43, and PrP are not lethal or highly deleterious, they are likely to be found at expected frequency in outbred human populations and, possibly, at higher frequency in consanguineous populations.

Moreover, a coordinated major project of this nature will increase the productivity and standardization of the sequencing and analysis processes while dramatically reducing the cost, which would be a fraction of the cost of clinical trials. Furthermore, there are already dozens of ongoing or planned human genome sequencing projects throughout the world (e.g., [32, 33]), covering millions of people, a number that at the current rate will soon be in the tens of millions.

In light of the newly proposed Advanced Research Projects Agency for Health (ARPA-H) and its intent “to develop breakthroughs—to prevent, detect, and treat diseases like Alzheimer's, diabetes, and cancer” [34], the scope and the timing for analyzing the extent of human genetic variation are in full concordance with ARPA-H’s overall mission “to accelerate the pace of breakthroughs to transform medicine and health” [34].

Claudiu I. Bandea, PhD

[1] Selkoe DJ, Hardy J (2016) The amyloid hypothesis of Alzheimer's disease at 25 years. EMBO Mol Med 8, 595-608.
[2] Hampel H, Hardy J, Blennow K, Chen C, Perry G, Kim SH, Villemagne VL, Aisen P, Vendruscolo M, Iwatsubo T, Masters CL, Cho M, Lannfelt L, Cummings JL, Vergallo A (2021) The amyloid-beta pathway in Alzheimer's disease. Mol Psychiatry 26, 5481-5503.
[3] Jagust W, Jack CR Jr, Bennett DA, Blennow K, Haeberlein SB, Holtzman DM, Jessen F, Karlawish J, Liu E, Molinuevo JL, Montine T, Phelps C, Rankin KP, Rowe CC, Scheltens P, Siemers E, Sperling R (2019) "Alzheimer's disease" is neither "Alzheimer's clinical syndrome" nor "dementia". Alzheimers Dement 15, 153-157.
[4] The Oskar Fischer Prize. https://oskarfischerprize.com/ (accessed on December 1, 2021).
[5] Castellani RJ, Plascencia-Villa G, Perry G (2019) The amyloid cascade and Alzheimer's disease therapeutics: theory versus observation. Lab Invest 99, 958-970.
[6] Joseph J, Shukitt-Hale B, Denisova NA, Martin A, Perry G, Smith MA (2001) Copernicus revisited: amyloid beta in Alzheimer's disease. Neurobiol Aging 22, 131-146.
[7] Smith MA, Joseph JA, Perry G (2006) Arson: tracking the culprit in Alzheimer's disease. Ann N Y Acad Sci 924, 35-38.
[8] Lee HG, Castellani RJ, Zhu X, Perry G, Smith MA (2004) Amyloid-beta in Alzheimer's disease: the horse or the cart? Pathogenic or protective? Int J Exp Pathol 86, 33-138.
[9] Morris GP, Clark IA, Vissel B (2014) Inconsistencies and controversies surrounding the amyloid hypothesis of Alzheimer's disease. Acta Neuropathol Commun 18, 135.
[10] Tse KH, Herrup K (2017) Re-imagining Alzheimer's disease - the diminishing importance of amyloid and a glimpse of what lies ahead. J Neurochem 143, 432-444.
[11] Ayton S, Bush AI (2021) β-amyloid: The known unknowns. Ageing Res Rev 65, 101212.
[12] National Plan to Address Alzheimer's Disease: 2021 Update (2021) ASPE, https://aspe.hhs.gov/reports/national-plan-2021-update
[13] Whitehouse P (2021) Why We Must Lance the Alzheimer’s Boil, https://www.j-alz.com/editors-blog/posts/why-we-must-lance-alzheimers-boil
[14] de la Torre J (2021) The FDA Approves Aducanumab for Alzheimer’s Disease, Raising Important Scientific Questions, https://www.j-alz.com/editors-blog/posts/fda-approves-aducanumab-alzheimers-disease
[15] Dobson CM (2017) The amyloid phenomenon and its links with human disease. Cold Spring Harb Perspect Biol 9, a023648.
[16] Prusiner SB (1998) Prions. Proc Natl Acad Sci U S A 95, 13363-13383.
[17] Marshall LR, Korendovych IV (2021) Catalytic amyloids: Is misfolding folding? Curr Opin Chem Biol 64, 145-153.
[18] Lee EY, Srinivasan Y, de Anda J, Nicastro LK, Tükel Ç, Wong GCL (2020) Functional reciprocity of amyloids and antimicrobial peptides: rethinking the role of supramolecular assembly in host defense, immune activation, and inflammation. Front Immunol 11, 1629.
[19] Malmberg M, Malm T, Gustafsson O, Sturchio A, Graff C, Espay AJ, Wright AP, El Andaloussi S, Lindén A, Ezzat K (2020) Disentangling the amyloid pathways: a mechanistic approach to etiology. Front Neurosci 14, 256.
[20] Billant O, Friocourt G, Roux P, Voisset C (2021) p53, a victim of the prion fashion. Cancers (Basel) 13, 269.
[21] Bandea CI (1986) From prions to prionic viruses. Med Hypotheses 20, 139-142.
[22] Bandea CI (2009) Endogenous viral etiology of prion diseases. Nat Prec, https://www.nature.com/articles/npre.2009.3887.1
[23] Bandea CI (2013) Aβ, tau, α-synuclein, huntingtin, TDP-43, PrP and AA are members of the innate immune system: a unifying hypothesis on the etiology of AD, PD, HD, ALS, CJD and RSA as innate immunity disorders. BioRxiv, https://www.biorxiv.org/content/10.1101/000604v1
[24] Jucker M, Walker LC (2013) Self-propagation of pathogenic protein aggregates in neurodegenerative diseases. Nature 501, 45-51.
[25] Aguzzi A, Heikenwalder M (2003) Prion diseases: Cannibals and garbage piles. Nature 423, 127-129.
[26] Hardy J (2009) The amyloid hypothesis for Alzheimer's disease: a critical reappraisal. J Neurochem 110, 1129-1134.
[27] Espay AJ, Sturchio A, Schneider LS, Ezzat K (2021) Soluble amyloid-β consumption in Alzheimer's disease. J Alzheimers Dis 82, 1403-1415.
[28] Benestad, SL, Austbø L, Tranulis MA, Espenes A, Olsaker I (2012) Healthy goats naturally devoid of prion protein. Vet Res 43, 87-91.
[29] Huang YM, Hong XZ, Shen J, Geng LJ, Pan YH, Ling W, Zhao HL (2020) Amyloids in site-specific autoimmune reactions and inflammatory responses. Front Immunol 10, 2980.
[30] Fulop T, Tripathi S, Rodrigues S, Desroches M, Bunt T, Eiser A, Bernier F, Beauregard PB, Barron AE, Khalil A, Plotka A, Hirokawa K, Larbi A, Bocti C, Laurent B, Frost EH, Witkowski JM (2021) Targeting impaired antimicrobial immunity in the brain for the treatment of Alzheimer's disease. Neuropsychiatr Dis Treat 17, 1311-1339.
[31] Backman JD, Li AH, Marcketta A, Sun D, Mbatchou J, Kessler MD, Benner C, Liu D, Locke AE, Balasubramanian S, Yadav A, Banerjee N, Gillies CE, Damask A, Liu S, Bai X, Hawes A, Maxwell E, Gurski L, Watanabe K, Kosmicki JA, Rajagopal V, Mighty J; Regeneron Genetics Center; DiscovEHR, Jones M, Mitnaul L, Stahl E, Coppola G, Jorgenson E, Habegger L, Salerno WJ, Shuldiner AR, Lotta LA, Overton JD, Cantor MN, Reid JG, Yancopoulos G, Kang HM, Marchini J, Baras A, Abecasis GR, Ferreira MAR (2021) Exome sequencing and analysis of 454,787 UK Biobank participants. Nature 599, 628-634.
[32] Alkuraya F (2019) Leading edge conversations: national efforts with global implications. Cell 177, 16-19.
[33] Wonkam A (2021) Sequence three million genomes across Africa. Nature 590, 209-211.
[34] Collins FS, Schwetz TA, Tabak LA, Lander ES (2021) ARPA-H: Accelerating biomedical breakthroughs. Science 373, 165-167.

Last comment on 18 June 2023 by Markku Kurkinen, PhD



Thanks Dr. Bandea for the great article and for bringing the loss-of-function framework up for discussion. I fully agree with your assessment of the inherent problems of the amyloid cascade hypothesis and that the genetic data does not have the power to differentiate between GOF and LOF mechanisms. I would just like to comment on this paragraph:

If the GOF model is indeed problematic, then the alternative paradigm, the LOF (see above), must be correct. Correct? Not so.. In 2012, the first natural genetic variation, a nonsense mutation in the PrP gene, which leads to a PrP -/- phenotype, was discovered among healthy homozygous dairy goats in Norway [28]. These goats did not develop any pathology/disease analogous to TSE, nor did the numerous transgenic PrP knockouts generated in mice, sheep, goats, cows, and other animal models. This also appears to be the case with many Aβ, tau, α-synuclein, and TDP-43 transgenic knockouts, but the findings need to be corroborated by natural knockouts. 

A more recent study of the Norwegian goats which lack PrP found that they suffer from the same phenotype as the knockout mice: demyelinating pathology, vacuolated fibers, shrunken axons, and onion bulbs (https://faseb.onlinelibrary.wiley.com/doi/10.1096/fj.201902588R). Although this is not fully analogous to the pathology observed in TSE, it still hints at important neuronal functions of PrP. In other models, adult knock-downs reveal a better disease analogy, probably because it resembles adult deficiency due to aggregation. For example, knocking out alpha-synuclein in adult rats and primates replicates Parkinson’s disease phenotype to a good degree (https://pubmed.ncbi.nlm.nih.gov/20551914/ and https://www.frontiersin.org/articles/10.3389/fnins.2016.00012/full ), and targeting Abeta42 in adult mice induces memory problems analogous to AD (https://onlinelibrary.wiley.com/doi/10.1002/ana.22313 ). Probably the best unfortunate experiment in AD is the worsening of cognition in healthy participants in the beta and gamma-secretase inhibitor trials (https://www.nejm.org/doi/full/10.1056/NEJMoa1812840 ), which shows that pharmacological adult knockdown of Abeta42 in humans leads to cognitive problems. 



In a series of inspiring articles, Dr. Ezzat and his colleagues present new data, observations, and arguments supporting the paradigm that the loss of (biological) function (LOF) of the primary proteins implicated in neurodegenerative diseases, including Aβ peptides in AD, α-synuclein in PD, and PrP in CJD, is the proximate cause for neurodegeneration (1-3). This is a major departure from the Amyloid hypothesis (AHyp) and the other dominant hypotheses in these fields, which invoke a gain of (toxic) function (GOF) as the proximate cause for neurodegeneration (4-5).

According to the LOF hypothesis, the soluble Aβ peptides, α-synuclein, and PrP misfold into insoluble aggregates (e.g., amyloids), thereby losing their biological function. Presumably, the LOF induces cellular/tissue dysfunctions, which eventually lead to neurodegeneration by processes that are yet to be clearly defined. In this scenario, the insoluble aggregates (e.g., amyloids) play a secondary role, and their removal would not stop the progression of disease. Instead, the hypothesis prescribes an increase in the production of soluble proteins, or adding soluble proteins (i.e., protein replacement), as therapeutic approaches.

According to the GOF hypothesis, the soluble Aβ peptides, α-synuclein, and PrP misfold into aggregates (e.g., amyloids), which have toxic effects at the cellular/tissue level and, eventually, lead to neurodegeneration by processes that are not clearly defined. In this scenario, the LOF plays a secondary role, if any, and the therapy focuses on reducing the production of the soluble proteins, or removal of proteins and their aggregates.

Clearly, when it comes to therapy, the two hypotheses are irreconcilable. However, although the immediate, or proximate, cause for neurodegeneration in the GOF and LOF hypotheses are opposite (i.e., gain of function versus loss of function, respectively), they share the primary causation, which is the misfolding and aggregation of Aβ peptides, tau, α-synuclein, and PrP. Therefore, in terms of prevention, both paradigms support the elimination or reduction of the factors/processes that induce protein misfolding and aggregation. Fortunately, this common goal is in line with the aims of the other investigators in the field, who consider many of the same factors/processes (e.g., toxins, infectious agents, and oxidative stress) as the direct cause neurodegeneration, while dismissing the participation of these proteins and their aggregates in neurodegeneration.

Therefore, as emphasized in the essay above, focusing on the multifactorial etiology of AD and other neurodegenerative diseases offers common ground for developing effective prevention approaches, which would also benefit other health conditions and diseases.

Another way to unify the field, and in particular the supporters of the GOF and LOF paradigms, is to find a common approach to address the fundamental question of whether or not these proteins are implicated in neurodegeneration, and, if they are, what is the proximate pathogenic mechanism: GOF, LOF, or a more complex intermediary process?

As pleaded in the post above, given decades of futile attempts, continuing to use clinical trials to address these critical questions makes little sense, particularly when they can be effectively addressed using a science-based approach that explores the natural genetic variation in humans. This project would be particularly effective in differentiating between the GOF and LOF hypotheses, since the natural heterozygous knockouts in the DNA sequences coding for Aβ, tau, α-synuclein, PrP, and other proteins implicated in neurodegeneration are expected to decrease the steady-state level of these proteins, which will either reduce (as would be expected in the context of GOF) or enhance (in the context of LOF) the pathogenic process.

This brings us to Dr. Ezzat’s specific comment on the pathophysiological changes observed in the PrP-free dairy goats, which exhibit subclinical demyelinating polyneuropathy (6). Additionally, these otherwise healthy goats show subtle biological and immunological changes or imbalances, such as induction of type I interferon-responsive genes (7), increased susceptibility to inflammatory stimuli (8), and modulation of innate immunity signaling pathways (9, 10). However, these changes are not analogous to pathophysiological changes seen in scrapie, and, as discussed below, unlike PrP producing goats, these relatively healthy breeding goats are not susceptible to scrapie. Also, given the fact that, similarly to Aβ, tau, α-synuclein and TDP-43, PrP is evolutionarily highly conserved, these findings are not unexpected, since these proteins play important biological functions under natural conditions (see below). The key question here is what kind of biological functions best explain these results, along with all the other related results in the field?

This question is particularly critical in the context of the LOF hypothesis because it postulates specifically that it is the loss of the proteins’ normal function that leads to neurodegeneration, so understanding the function of these proteins is essential; on the contrary, although the AHyp and the other GOF hypotheses disregard the role of the natural function of these proteins in the disease process, in their context, this function is secondary. Also, given that the LOF advocates regard the biological function of PrP, Aβ peptides, and α-synuclein to be at the center of their hypothesis, it is rather surprising that they do not consider this function to be central in understanding their alleged misfolding and aggregation.

As mentioned in the post, and as previously discussed (11, 12), I proposed that: (i) Aβ, tau, α-synuclein, TDP-43, and PrP are members of the innate immune system; (ii) the isomeric conformational changes of these proteins and their assembly into various oligomers and amyloid aggregates are not protein misfolding events as defined for decades, and their aggregation does not represent prion-replication activities, but they are part of these proteins’ normal, evolutionarily selected innate immune repertoire; and, (iii) the reactions and activities associated with the function of these proteins in innate immunity leads to pathological and neurodegenerative events that define AD, PD, ALS and CJD biologically as immune system disorders.

Nevertheless, in the spirit of finding common ground, I submit that as members of the immune system, and in context of their complex immune activities that are often exercised within the narrow gap between protection and injury, PrP, Aβ, tau, α-synuclein, and TDP-4 perform reactions that could be narrowly defined as GOF or LOF events. However, as is the case with most immune system activities, including autoimmune reactions, these events are part and parcel of their evolutionarily selected biological functions.

To address Dr. Ezzat’s comment further, the phenotypic changes observed in the PrP-free goats are not analogous to scrapie pathology and, as has recently been shown, these goats are completely resistant to scrapie (13); also, a reduction of PrP production in heterozygous PrP knockout goats extends the scrapie incubation period. These results clearly indicate that the scrapie pathology is dependent on the presence of PrP, and that a reduction in the soluble PrP by 50% does not exacerbate the disease. Together, these findings are difficult to explain through the LOF. Nevertheless, it remains to be seen whether similar reductions in the production of PrP, as well as of Aβ, tau, α-synuclein, and TDP-43, which can be assessed by studying natural genetic variation, produce analogous results in humans.


[1] Malmberg M, Malm T, Gustafsson O, Sturchio A, Graff C, Espay AJ, Wright AP, El Andaloussi S, Lindén A, Ezzat K (2020) Disentangling the amyloid pathways: a mechanistic approach to etiology. Front Neurosci 14, 256.

[2] Espay AJ, Sturchio A, Schneider LS, Ezzat K (2021) Soluble amyloid-β consumption in Alzheimer's disease. J Alzheimers Dis 82, 1403-1415.

[3] Sturchio A, Dwivedi AK, Young CB, Malm T, Marsili L, Sharma JS, Mahajan A, Hill EJ, Andaloussi SE, Poston KL, Manfredsson FP, Schneider LS, Ezzat K, Espay AJ (2021) High cerebrospinal amyloid-β 42 is associated with normal cognition in individuals with brain amyloidosis. EClinicalMedicine 38, DOI: 10.1016/j.eclinm.2021.100988

[4] Hampel H, Hardy J, Blennow K, Chen C, Perry G, Kim SH, Villemagne VL, Aisen P, Vendruscolo M, Iwatsubo T, Masters CL, Cho M, Lannfelt L, Cummings JL, Vergallo A (2021) The amyloid-beta pathway in Alzheimer's disease. Mol Psychiatry 26, 5481-5503.

[5] Jucker M, Walker LC (2018) Propagation and spread of pathogenic protein assemblies in neurodegenerative diseases. Nat Neurosci. 21, 1341-1349.

[6] Skedsmo FS, Malachin G, Våge DI, Hammervold MM, Salvesen Ø, Ersdal C, Ranheim B, Stafsnes MH, Bartosova Z, Bruheim P, Jäderlund KH, Matiasek K, Espenes A, Tranulis MA (2020) Demyelinating polyneuropathy in goats lacking prion protein. FASEB J. 34, 2359-2375.

[7] Malachin G, Reiten MR, Salvesen Ø, Aanes H, Kamstra JH, Skovgaard K, Heegaard PMH, Ersdal C, Espenes A, Tranulis MA, Bakkebø MK (2017) Loss of prion protein induces a primed state of type I interferon-responsive genes. PLoS One. 12, DOI: 10.1371/journal.pone.0179881

[8] Salvesen Ø, Reiten MR, Espenes A, Bakkebø MK, Tranulis MA, Ersdal C (2017) LPS-induced systemic inflammation reveals an immunomodulatory role for the prion protein at the blood-brain interface. J Neuroinflammation. 22, 106.

[9] Salvesen Ø, Reiten MR, Kamstra JH, Bakkebø MK, Espenes A, Tranulis MA, Ersdal C (2017) Goats without Prion Protein Display Enhanced Proinflammatory Pulmonary Signaling and Extracellular Matrix Remodeling upon Systemic Lipopolysaccharide Challenge. Front Immunol.8, 1722.

[10] Salvesen Ø, Tatzelt J, Tranulis MA (2019) The prion protein in neuroimmune crosstalk. Neurochem Int. 130, 104335.

[11] Bandea CI (2009) Endogenous viral etiology of prion diseases. Nat Pre. https://www.nature.com/articles/npre.2009.3887.1

[12] Bandea CI (2013) Aβ, tau, α-synuclein, huntingtin, TDP-43, PrP and AA are members of the innate immune system: a unifying hypothesis on the etiology of AD, PD, HD, ALS, CJD and RSA as innate immunity disorders. BioRxivhttps://www.biorxiv.org/content/10.1101/000604v1

[13] Salvesen Ø, Espenes A, Reiten MR, Vuong TT, Malachin G, Tran L, Andréoletti O, Olsaker I, Benestad SL, Tranulis MA, Ersdal C (2020) Goats naturally devoid of PrPC are resistant to scrapie. Vet Res. 51, doi: 10.1186/s13567-019-0731-2

Alzheimer never suggested plaques and tangles were the cause of dementia. Indeed, this is what he wrote in 1911: “So scheint wirklich kein stichhaltiger Grund vorhanden, diese Fälle als durch einen besonderen Krankheitzprozeβ verursacht zu betrachten” [1, p. 378]. “There is then no tenable reason to consider these cases as caused by a specific disease process” [2, p. 93]. 


Eleanor Drummond and colleagues [3] have catalogued the proteins in amyloid plaques and neurofibrillary tangles, which were localized in formalin-fixed paraffin-embedded (FFPE) brain samples by immunostaining, microdissected by laser capture, solubilized with formic acid, deparaffinized, reduced, alkylated, proteinase digested, and analyzed by quantitative LC-MS. This unbiased and simultaneous quantification detected ~900 proteins in plaques and ~500 proteins in tangles. 


Inherited dominant mutations in the APP, PS1, or PS2 gene cause Alzheimer dementia (AD) at the patient age of 19–55, at about the same age as their mother or father, and their mother or father developed dementia, the exact timing of onset being dictated by the gene and the particular mutation.These observations provide the best evidence for the amyloid hypothesis of AD, the ‘die-hard’ hypothesis that has almost singularly misguided AD research and drug development for 30 years.


According to the amyloid hypothesis, AD begins in the brain with Aβ peptides accumulation, aggregation, and amyloid formation. Yet, in clinical trial studies, reducing Aβ  peptides production and brain amyloid did not slow cognitive decline or improve daily living of AD patients. Similarly, preventive trials in cognitively unimpaired people at high risk, or genetically destined, of developing AD have failed to slow cognitive decline. The results of these studies are against the amyloid hypothesis. The amyloid hypothesis is too good to be true [4], but is it “too big to fail” [5], as Rudy Castellani and Mark A. Smith said in 2011, or should we change our thinking about AD etiology, its origin and disease mechanisms [6].


Of course, a hypothesis can never been proven right, but it can be proven wrong. When Sun et al. [7] studied 138 pathogenic PS1 mutations on the in vitro production of the Aβ42 and Aβ40 peptides by γ-secretase, they could not find any correlation between the amount of Aβ peptides produced or the Aβ42/40 ratio and the age of onset of AD. Remarkably, one third of the PS1 mutations produced no Aβ peptides, and yet they can cause AD at different ages of onset. This observation agrees with a study in which inactivation of one PS1 gene in the adult mouse brain caused neurodegeneration and progressive memory loss [8]. 


Osaka mutation (E693Δ) deletes glutamate from position 693 in APP, position 22 in Aβ peptide, which enhances oligomerization while it reduces fibril formation. This unique property  of the mutant (E22Δ) Aβ peptide is explained by its higher hydrophobicity resulting in faster oligomerization.  This agrees with the observation that in familial AD caused by Osaka mutation, patients have very low levels of Aβ amyloid in the brain [9]. Not suprisingly, these cases of AD without amyloid are taken as an evidence for Aβ oligomers causing AD [10].


 Naturally, of course, APP is more than the precursor protein for Aβ peptides, and Aβ peptides do much more than form amyloid [11], In addition, besides the γ-secretase, PS1 has other proteins to interact with, such as the glutamate transporter EAAT2 [12], and there are 149 substrates for PS1 [13].



  1. Alzheimer A (1911) Über eigenartige Krankheitsfalle des spateren Alters. Z Gesamte Neurol Psychiat 4:356–385 
  2. Förstl H, Levy R (1991) On certain peculiar diseases of old age. Hist Psychiatry 2:74–99

3. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5811767/pdf/nihms940664.pdf

4. Kurkinen M (2017). The amyloid hypothesis is too good to be true. Alzheimer’s Dement Cogn Neurol 1, 1–9. 


5. Castellani RJ, Smith MA (2011) Compounding artefacts with uncertainty, and an amyloid cascade hypothesis that is ‘too big to fail’. J Pathol224, 147-152. 

6. Kurkinen M et al. (2023) The amyloid cascade hypothesis in Alzheimer’s disease: Should we change our thinking? Biomolecule 13:453

7. Sun et al. (2017). Analysis of 138 pathogenic mutations in presenilin-1 on the in vitro production of Aβ42 and Aβ40 peptides by γ-secretase. Proc Natl Acad Sci USA 114, E476–E485.

8. Shen J, Kelleher RJ 3rd (2007) The presenilin hypothesis of Alzheimer’s disease: Evidence for a loss-of-function pathogenic mechanism. Proc Natl Acad Sci USA 104, 403–409. 

9. Tomiyama T, Shimada H (2020) APP Osaka mutation in familial Alzheimer;s disease - its discovery, phenotypes, and mechanisms of recessive inheritance. Int J Mol Sci 21:1413

10. Ghosh S, Ali R, Verma S (2023) Aβ-oligomers: A potential therapeutic target for Alzheimer's disease. Int J Biol Macromol 239:124231

11. Kurkinen M (2022). Astrocyte glutamate transporter EAAT2 in Alzheimer dementia. Glutamate and Neuropsychiatric Disorders—Current and Emerging Treatments; Pavlovic, Z.M., Ed.; Springer Nature: Berlin, Germany, pp. 229–260

12. Zoltowska et al. (2018) Novel interaction between Alzheimer’s disease-related protein presenilin 1 and glutamate transporter 1. Sci Rep 8, 8718. 

13. Güner G, Lichtenthaler SF (2020) The substrate repertoire of γ-secretase/presenilin. Semi Cell Dev Biol 105, 27–42.