Medical Hypotheses 110 (2018) 60–63

Contents lists available at ScienceDirect

Medical Hypotheses
journal homepage: www.elsevier.com/locate/mehy

Hypothesis for the cause and therapy of neurodegenerative diseases MARK

⁎
Philip Serwer
Department of Biochemistry and Structural Biology, The University of Texas Health Science Center, San Antonio, TX, United States

A R T I C L E I N F O A B S T R A C T

Keywords: The cause and therapy of neurodegenerative diseases remain unsolved puzzles. These diseases are correlated
Amyloid with presence of beta sheet-rich amyloid assemblies. Here, I derive and assemble puzzle pieces to obtain a loose
Bacteriophages end-tying hypothesis for cause with direct implications for therapy. I use the following extrapolations to find
Directed evolution connectable puzzle pieces: (a) the traditional extrapolation that amyloid/amyloid precursors cause disease, (b) a
Innate immune systems
recent extrapolation that amyloid-forming proteins, some of which are virus protein homologs, are components
Protein structure
Viral assembly intermediates, obscure
of an empirically obscure innate immune system that counters insults, including those by both viruses and
bacteria, (c) a new extrapolation that various insults produce assemblies with structural features in common and
that amyloid-forming, innate immune system proteins recognize these features and, then, counter insults by co-
assembly, (d, 1) a second new extrapolation that beta sheet is a common structural feature and is extended
during insult-neutralizing co-assembly and (d, 2) an appendix, derived from studies of phages T3 and T4, that
most insult-produced assemblies are obscure to current biochemical analysis. The hypothesis is the following.
One function of amyloid-forming proteins is non-classical innate immunity to biological insults. This immunity
works via beta sheet-extending co-assembly of amyloid-forming proteins with beta sheet-containing insult
products. For example, co-assembly with beta sheet-containing viral assembly intermediates inhibits virus
production. Amyloid-forming proteins cause neurodegenerative disease when errant, typically overproduced.
Other innate immunity systems sometimes exacerbate symptoms. This hypothesis suggests the following
therapy, based on manipulating Nature’s chemistry. First, conduct directed evolution to obtain low-pathogeni-
city, chronic symptom-producing viruses with assembly intermediates that co-assemble with and destabilize
both amyloid and amyloid sub-assemblies. Then, infect patients with these viruses.

Introduction
Neurodegenerative disease mechanism and therapy are puzzles that
have evaded solutions. Furthermore, solutions do not appear to be on
the horizon with current strategies. A new perspective appears to be
needed. I will propose a perspective and hypothesis that lead to the
novel conclusion that therapy should be based on manipulating
Nature’s chemistry via directed evolution, rather than using human-
design chemistry.
Neurodegenerative diseases are correlated with presence of cell-
damaging amyloid assemblies. However, the normal, non-disease-
causing function(s) of these assemblies, and sometimes of unassembled
amyloid-forming proteins, is not known [1–5]. This statement includes
Alzheimer disease [1], Parkinson disease [2], ALS [3], Huntington
disease [4] and prion-associated diseases [5]. Here, I derive a hypoth-
esis by assembling connectable puzzle pieces via extrapolations from
data. I use four puzzle pieces, two newly derived here.
Taking ALS as an example, one might look for puzzle pieces by
asking why this disease sequentially spreads from one motor nerve to
another. Possible answers include [1] virus infection and [2] self-acti-
vation of the synthesis of a toxic molecule.
Indeed, ALS resembles an ultra-slow polio, a disease caused by virus
infection [6]. But, self-activation of the synthesis of a toxic molecule is
also possible. In early work on bacterial viruses (also called bacter-
iophages or phages), most investigators thought, although incorrectly,
that phages were self-activating bacterial proteins [7].
Origin and foundation of the hypothesis
First and second puzzle pieces
Self-activation of the synthesis of a toxic molecule(s) is, however,
the most reasonable explanation, based on (a) correlation of amyloid/
amyloid sub-assembly presence with neurodegenerative disease and (b)
low-to-zero inter-organism transmission [1–5]. My first extrapolation/
puzzle piece (summarized in Fig. 1) is the conventional one, although

⁎
Address: Department of Biochemistry and Structural Biology, The University of Texas Health Science Center, 7703 Floyd Curl Drive, San Antonio, TX 78229-3900, United States.
E-mail address: serwer@uthscsa.edu.

https://doi.org/10.1016/j.mehy.2017.11.001
Received 17 July 2017; Accepted 5 November 2017
0306-9877/ © 2017 Published by Elsevier Ltd.
P. Serwer
Fig. 1. The puzzle pieces are summarized. The four extrapolations involved are separately
re-stated.
not universally accepted [1], that this correlation is not accidental and
that either amyloid or, more likely, an amyloid sub-assembly causes
disease. But, this extrapolation leads one to ask what are the normal
functions, if any, of amyloid-forming proteins. These proteins undergo a
change in secondary/tertiary structure (to be called a change in state)
when they assemble to form amyloid [1–5]. The data suggest that at
least Alzheimer disease amyloid-forming protein (Aβ, a cleavage pro-
duct) and prions are descended from viral proteins. This raises the
possibility that these proteins inhibit viral propagation [review [8]].
Historically, the assumption is that disease-causing, amyloid-forming
proteins do not have functions when they are in the amyloid-forming
state; protein misfolding is the major theme [1–5].
However, this assumption is not always correct because bacteria
61
Medical Hypotheses 110 (2018) 60–63
produce amyloids that have functions. These functions include forma-
tion of external fibers [curli, for example, in the case of Escherichia coli
[9]] needed for formation of biofilms. Mechanisms exist to inhibit in-
tracellular (toxic) formation of bacterial amyloid and the typically more
toxic, smaller amyloid sub-assemblies [10].
Obtaining the next puzzle pieces begins with the following previous
[8] extrapolation. In addition to pre-amyloid state functions, amyloid-
forming proteins have amyloid state function. This latter function is
within a (thus far empirically obscure) system of innate immunity. After
beneficial activation within this system, amyloid-forming proteins act
to counter external insults by both microbes and physical traumas.
Neurodegenerative diseases occur after beneficial activation, when non-
beneficial hyper-synthesis and, sometimes, misfolding of these innate
immunity proteins occur. Erroneous self-countering (a) sometimes is
the source of this malfunction [8] and (b) produces self-activation of
synthesis. Thus, this previous hypothesis provides a simple explanation
for the progression of ALS. The initially beneficial activation provides a
simple explanation for the previously observed [11], broad-based cor-
relation of Alzheimer disease with infections by viruses, bacteria and
fungi.
Therefore, my second puzzle piece is the accepting of this previous,
relatively recent extrapolation (Fig. 1). This extrapolation is in-
dependent of further conformational change that can (a) be construed
as misfolding and (b) exacerbate the disease via its propagation (tem-
plate toxicity), as shown for an amyloid of ALS [3], as well as Aβ [1]
and prions [5].
Third and fourth puzzle pieces (new)
However, the number of innate immune system proteins of this type
cannot possibly be as high as the number of different insults. In Ref. [8],
this problem is anticipated by proposing that the amyloids adapt, via
conformational changes, to various insults. But, the observed amyloid-
producing conformational changes have limited variability and are
biased toward beta sheet structure [1–5].
Fig. 2. An example is given of the proposed mechanism for
responding to an insult caused by a virus. (a) A virus
subassembly converts to a virus. HIV is chosen as the ex-
ample [19,20]. (b) An innate immunity protein (red)
counters this conversion by co-assembly with a beta sheet-
rich intermediate of the capsid protein (light blue). This
intermediate is currently hypothetical for most viruses. (c)
The innate immunity protein forms amyloid precursors and
amyloid when it co-assembles with itself, typically in a
mistaken attempt to counter itself. The result in (c) is for-
mation of toxic assemblies that cause neurodegenerative
diseases. The proposed therapy is based on a virus that
converts condition (c) back to a condition approximating
(b), except for formation of a less stable (disposable) ver-
sion of the amyloid and its sub-assemblies. (For inter-
pretation of the references to colour in this figure legend,
the reader is referred to the web version of this article.)
P. Serwer
Therefore, I introduce the following new extrapolation and third
puzzle piece. The various insults produce protein structures with lim-
ited variability (common structural features) and limited targets for
insult countering co-assembly of amyloid-forming proteins (Fig. 1).
During the replication of viruses (e.g., Fig. 2a), this co-assembly inhibits
the function of viral replication intermediates and, therefore, inhibits
the replication of the virus. One example is the co-assembly with in-
termediates in viral capsid assembly (Fig. 2b).
A fourth puzzle piece-producing extrapolation is that beta sheet is a
common structural feature targeted by this innate immune system. This
extrapolation is based on the known high beta sheet component in
eukaryotic amyloids [1–5] and bacterial curli [9]. Parenthetically, the
RNA-binding function of ALS amyloid-forming proteins (including TDP-
43) suggests that in-common viral RNA structure is sometimes also
targeted [3]. The data of the above references are suggestive, but not
complete, for this extrapolation. One reason for the incompleteness is
that, in the case of virus-derived insults, the proposed replication in-
termediates have not been isolated. However, additional supporting
preliminary data exist, as discussed below.
The proposed, but not isolated, replication intermediates generate
the following appendix to the fourth puzzle piece-producing extra-
polation. [1] First, historically (beginning in 1967), some DNA mole-
cules in phage T4-infected cells are found packaged by altered state
capsid shells [12,13]. These shells do not survive lysis of the cells, but
are observed in electron micrographs of thin sections of T4-infected
cells permeabilized by low concentration osmium tetroxide [12,13]. [2]
Second, recently (2016), possibly related, altered state phage T3 cap-
sids are isolated, but only after introduction of new procedures. The
latter capsids are isolated from lysates of cells infected in atypical la-
boratory conditions with a directed evolution-modified T3 [14,15]. The
T3 altered state capsids have characteristics that suggest predominant
beta sheet structure of the major capsid protein [14]. This protein is not
beta sheet-rich in mature T3 phage [16]. Although discovered 49 years
earlier, the T4 altered state shells are still not isolated, i.e., they are
biochemically obscure. Phage systems are more biochemically tractable
than eukaryotic systems. Thus, my appendix to the fourth extrapolation
is that the intermediates discussed above for puzzle pieces 3 and 4 are
present, but are similarly obscure, possibly because of either instability
or short lifetime.
Hypothesis
Based on the above puzzle pieces, my hypothesis for mechanism is
the following. (a) One function of amyloid-forming proteins is the
production of non-classical innate immunity to various insults. (b) This
immunity typically works via beta sheet-extending co-assembly of
amyloid-forming proteins with beta sheet-containing insult products.
(c) The amyloid-forming proteins cause neurodegenerative disease
when they become errant, typically overproduced. A source of this
malfunction is activation for self-countering.
To provide additional perspective, I note that extensive evidence
exists for a contribution to neurodegenerative disease from innate im-
mune and support systems other than the system proposed here. The
focus is typically on microglial clearance and astrocyte support func-
tions. Amyloid formation activates these systems, but does not ne-
cessarily contribute directly to their toxic effects [1,17,18], although it
does in the case of ALS and microglia [3]. I do not question these latter
aspects. I consider them, however, to be secondary contributors to the
primary disease-generating process described in the above hypothesis.
This statement also applies to analysis of several mutations that in-
crease susceptibility to ALS [3], including those that add hexanucleo-
tide repeats to non-coding regions of genes. In addition, other anti-viral
innate immune systems have been demonstrated [19,20].
62
Medical Hypotheses 110 (2018) 60–63
Evolutionary rationale: Acquisition of innate immunity proteins
The above hypothesis raises the question of how the proposed in-
nate immunity proteins are originally acquired. In the case of anti-viral
and related innate immunity proteins, I propose that the genes for these
proteins (a) are originally acquired by horizontal transfer from infecting
virus pathogens and (b) are retained via selection for pathogen re-
sistance. At the level of non-coding DNA, this concept is already verified
for the CRISPR-Cas adaptive bacterial immune systems [21]. With these
systems, bacteria retain short phage DNA segments from past phage
infections. The bacteria use this retained DNA to generate specificity
needed to cleave the DNA genomes of phages that subsequently infect,
while not cleaving the bacterial DNA.
But, protein-based immunity requires over an order of magnitude
more genomic information per insult. This is one reason for the as-
sumption that innate immunity proteins target a common structural
feature for co-assembly (assembly homology), when acting as part of an
innate immune system.
Empirical testing for predicted intermediates
Testing for the predicted beta sheet-rich viral (and other) inter-
mediates should be possible via isolation and characterization from
cells. But, this strategy has the following apparent limitations. We do
not have the relevant data to design isolation procedures, especially
given potential short lifetime and instability. For viruses, I expect this
limitation to be increased by use for virus propagation of rapid-growth,
typical laboratory conditions, instead of slower-growth, more en-
vironmental conditions. But, vexingly, we must perform the isolation to
get the relevant data. To escape this trap, one strategy is to isolate viral
assembly intermediates from cells that are infected in conditions not
typically used, preferably closer to conditions in the environment.
This is the strategy used above to obtain previously obscure capsid-
containing particles of phage T3 [14,15]. These particles are either not
detected [14] or more difficult to detect [15] when using conventional,
rapid-growth, laboratory conditions. If, analogously, eukaryotic virus
intermediates of this type exist, they are possibly targets of amyloid-
forming proteins (Fig. 2b). They would also be logical targets for de-
velopment of anti-viral drugs.
Therapy suggested by the hypothesis
The above hypothesis suggests the following, thus far untried
therapy. Infect a patient with a therapeutic virus that (a) produces re-
latively mild, but chronic symptoms, and (b) assembles via inter-
mediates that cause disassembly of amyloid and amyloid sub-assem-
blies by co-assembling with them via assembly homology. In other
words, convert the Fig. 2c scenario to the Fig. 2b scenario, but with
destabilized amyloid in the latter. This therapeutic virus might be
generated by multiple passages through primates, i.e., by directed
evolution. The selection would be for both mild, chronic symptoms and
reduced amyloid amount in hosts with neurodegenerative disease. This
strategy (a) is apparently not suggested by any other current concept of
neurodegenerative disease and (b) bypasses therapy-delivery limita-
tions of drug-based strategies [examples [4]].
Indeed, some success has been reported [22,23] in managing
murine Alzheimer and Parkinson disease with filamentous phage M13.
This phage also disassembles several amyloids in vitro, as does, though
less effectively, an artificial multimer of a protein, gp3, at one end of
the filamentous viral capsid [24]. This latter work is the result of an
accidental observation and is not driven by the ideas introduced here.
One possibility is that gp3, together with the major M13 capsid protein,
acts to assist removal of errant innate immune protein complexes by
extending beta sheet structure while making this structure less stable.
But, phages will not replicate in patients, which creates problems in
delivery of sufficient protein [23]. Projected success is greater with a
P. Serwer
virus that does replicate in patients.
Acknowledgments
I thank Dr. Martin Adamo for comments on a draft of this manu-
script. The author received support from the Welch Foundation (AQ-
764). The Welch Foundation did not participate in the design, writing
and submission of this manuscript.
Author contributions
Philip Serwer conceived and developed the hypothesis and wrote
the manuscript.
Conflicts of interest
None.
References
[1] Van Dam D, Vermeiren Y, Dekker AD, Naudé PJW, De Deyn PP. Neuropsychiatric
disturbances in Alzheimer’s disease: What have we learned from neuropathological
studies? Curr Alzheimer Res 2016;13:1145–64.
[2] Kalia LV, Lang AE. Parkinson disease in 2015: Evolving basic, pathological and
clinical concepts in PD. Nat Rev Neurol 2016;12:65–6.
[3] Taylor PJ, Brown Jr. RH, Cleveland DW. Decoding ALS: from genes to mechanism.
Nature 2016;539:197–206.
[4] Bates GP, Dorsey R, Gusella JF, et al. Huntington disease. Nature Rev Disease
Primers 2015;1:15005.
[5] Collinge J. Mammalian prions and their wider relevance in neurodegenerative
diseases. Nature 2016;539:217–26.
[6] Van den Pol AN. Polio, still lurking in the shadows. J Neuro Sci 2013;33:855–62.
[7] Summers WC. The strange history of phage therapy. Bacteriophage 2012;2:130–3.
[8] Bandea CI. Aβ, tau, α-synuclein, huntingtin, TDP-43, PrP and AA are members of
63
Medical Hypotheses 110 (2018) 60–63
the innate immune system: a unifying hypothesis on the etiology of AD, PD, HD,
ALS, CJD and RSA as innate immunity disorders. bioRχiv, http://biorxiv.org/
content/early/2013/11/18/000604, doi: https://doi.org/10.1101/000604.
[9] Schwartz K, Boles BR. Microbial amyloids–functions and interactions within the
host. Curr Opin Microbiol 2013;16:93–9.
[10] Evans ML, Chorell E, Taylor JD, et al. The bacterial curli system possesses a potent
and selective inhibitor of amyloid formation. Mol Cell 2015;57:445–55.
[11] Itzhaki RF, Lathe R, Balin BJ, et al. Microbes and Alzheimer’s disease. J Alzheimer’s
Dis 2016;51:979–84.
[12] Kellenberger E, Eiserling FA, Boy de la Tour E. Studies on the morphopoiesis of the
head of phage T-even: III, The cores of head-related structures. J Ultrastruct Res
1967;21:335–60.
[13] Schaerli C, Kellenberger E. Head maturation pathway of bacteriophages T4 and T2:
V. Maturable ε-particle accumulating in acridine-treated bacteriophage T4-infected
cells. J Virol 1980;33:830–44.
[14] Serwer P, Wright ET. Testing a proposed paradigm shift in analysis of phage DNA
packaging. Bacteriophage 2016;6:e1268664.
[15] Serwer P, Wright ET. ATP-driven contraction of phage T3 capsids with DNA in-
completely packaged in vivo. Viruses 2017;9:119.
[16] Guo F, Liu Z, Fang PA, et al. Capsid expansion mechanism of bacteriophage T7
revealed by multistate atomic models derived from cryo-EM reconstructions. Proc
Natl Acad Sci USA 2014;111:E4606–14.
[17] Heneka MT, Golenbock DT, Latz E. Innate immunity in Alzheimer’s disease. Nat
Immunol 2015;16:229–36.
[18] Minter MR, Taylor JM, Crack PJ. The contribution of neuroinflammation to amyloid
toxicity in Alzheimer's disease. J Neurochem 2016;136:457–74.
[19] Altfeld M, Gale Jr. M. Innate immunity against HIV-1 infection. Nat Immunol
2015;16:554–62.
[20] Kovalskyy DB, Ivanov DN. Recognition of the HIV capsid by the TRIM5α restriction
factor is mediated by a subset of re-existing conformations of the TRIM5α SPRY
Domain. Biochemistry 2014;53:1466–76.
[21] Doudna JA, Charpentier E. The new frontier of genome engineering with CRISPR-
Cas9. Science 2014;346:1258096.
[22] Frenkel D, Katz O, Solomon B. Immunization against Alzheimer's beta-amyloid
plaques via EFRH phage administration. Proc Natl Acad Sci USA 2000;97:11455–9.
[23] Messing J. Phage M13 for the treatment of Alzheimer and Parkinson disease. Gene
2016;583:85–9.
[24] Krishnan R, Tsubery H, Proschitsky MY, et al. A bacteriophage capsid protein
provides a general amyloid interaction motif (GAIM) that binds and remodels
misfolded protein assemblies. J Mol Biol 2014;426:2500–19.