Antivirals reduce the formation of key Alzheimer's disease molecules in cell cultures acutely infected with herpes simplex virus type 1.

Abstract

Alzheimer's disease (AD) afflicts around 20 million people worldwide and so there is an urgent need for effective treatment. Our research showing that herpes simplex virus type 1 (HSV1) is a risk factor for AD for the brains of people who possess a specific genetic factor and that the virus causes accumulation of key AD proteins (β-amyloid (Aβ) and abnormally phosphorylated tau (P-tau)), suggests that anti-HSV1 antiviral agents might slow AD progression. However, currently available antiviral agents target HSV1 DNA replication and so might be successful in AD only if Aβ and P-tau accumulation depend on viral DNA replication. Therefore, we investigated firstly the stage(s) of the virus replication cycle required for Aβ and P-tau accumulation, and secondly whether antiviral agents prevent these changes using recombinant strains of HSV1 that progress only partly through the replication cycle and antiviral agents that inhibit HSV1 DNA replication. By quantitative immunocytochemistry we demonstrated that entry, fusion and uncoating of HSV1, are insufficient to induce Aβ and P-tau production. We showed also that none of the "immediate early" viral proteins is directly responsible, and that Aβ and P-tau are produced at a subsequent stage of the HSV1 replication cycle. Importantly, the anti-HSV1 antiviral agents acyclovir, penciclovir and foscarnet reduced Aβ and P-tau accumulation, as well as HSV1, with foscarnet being less effective in each case. P-tau accumulation was found to depend on HSV1 DNA replication, whereas Aβ accumulation was not. The antiviral-induced decrease in Aβ is attributable to the reduced number of new viruses, and hence the reduction in viral spread. Since antiviral agents reduce greatly Aβ and P-tau accumulation in HSV1-infected cells, they would be suitable for treating AD with great advantage unlike current AD therapies, only the virus, not the host cell, would be targeted.

Full text

Page 1

Antivirals Reduce the Formation of Key Alzheimer’s
Disease Molecules in Cell Cultures Acutely Infected with
Herpes Simplex Virus Type 1
Matthew A. Wozniak1, Alison L. Frost1, Chris M. Preston2, Ruth F. Itzhaki1*
1Faculty of Life Sciences, The University of Manchester, Manchester, United Kingdom, 2Centre for Virus Research, Medical Research Council-University of Glasgow,
Glasgow, United Kingdom
Abstract
Alzheimer’s disease (AD) afflicts around 20 million people worldwide and so there is an urgent need for effective treatment.
Our research showing that herpes simplex virus type 1 (HSV1) is a risk factor for AD for the brains of people who possess a
specific genetic factor and that the virus causes accumulation of key AD proteins (b-amyloid (Ab) and abnormally
phosphorylated tau (P-tau)), suggests that anti-HSV1 antiviral agents might slow AD progression. However, currently
available antiviral agents target HSV1 DNA replication and so might be successful in AD only if Ab and P-tau accumulation
depend on viral DNA replication. Therefore, we investigated firstly the stage(s) of the virus replication cycle required for Ab
and P-tau accumulation, and secondly whether antiviral agents prevent these changes using recombinant strains of HSV1
that progress only partly through the replication cycle and antiviral agents that inhibit HSV1 DNA replication. By
quantitative immunocytochemistry we demonstrated that entry, fusion and uncoating of HSV1, are insufficient to induce Ab
and P-tau production. We showed also that none of the ‘‘immediate early’’ viral proteins is directly responsible, and that Ab
and P-tau are produced at a subsequent stage of the HSV1 replication cycle. Importantly, the anti-HSV1 antiviral agents
acyclovir, penciclovir and foscarnet reduced Ab and P-tau accumulation, as well as HSV1, with foscarnet being less effective
in each case. P-tau accumulation was found to depend on HSV1 DNA replication, whereas Ab accumulation was not. The
antiviral-induced decrease in Ab is attributable to the reduced number of new viruses, and hence the reduction in viral
spread. Since antiviral agents reduce greatly Ab and P-tau accumulation in HSV1-infected cells, they would be suitable for
treating AD with great advantage unlike current AD therapies, only the virus, not the host cell, would be targeted.
Citation: Wozniak MA, Frost AL, Preston CM, Itzhaki RF (2011) Antivirals Reduce the Formation of Key Alzheimer’s Disease Molecules in Cell Cultures Acutely
Infected with Herpes Simplex Virus Type 1. PLoS ONE 6(10): e25152. doi:10.1371/journal.pone.0025152
Editor: Stacey Efstathiou, University of Cambridge, United Kingdom
Received April 21, 2011; Accepted August 26, 2011; Published October 7, 2011
Copyright: ? 2011 Wozniak et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits
unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This work was supported by a grant from the Henry Smith Charity, an award from the Olympus Foundation, donations from the estate of Mrs. Louise
de Launay and from the Visual Sciences Fund. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the
manuscript.
Competing Interests: The authors have declared that no competing interests exist.
* E-mail: ruth.itzhaki@manchester.ac.uk
Introduction
Alzheimer’s disease (AD) is a prevalent neuropsychiatric
disorder that primarily affects the elderly. It is characterised by
memory loss and cognitive dysfunction, and examination of
sufferers’ brains reveals an abundance of two neuropathological
features – senile plaques and neurofibrillary tangles. Plaques and
tangles, and their components – b-amyloid (Ab) and abnormally
phosphorylated tau (P-tau), respectively – are thought to be central
to disease pathogenesis but their cause and thus the underlying
cause of AD is unknown.
Current treatments for AD are merely palliative and thus there
is an urgent need for medications that delay disease progression.
One possible treatment strategy is the use of antiviral agents, in
particular agents that are effective against herpes simplex virus
type 1 (HSV1). This suggestion is based on the increasing body of
evidence implicating HSV1 in the aetiology of AD.
HSV1 is a neurotropic virus that infects most humans. It is
responsible for a number of diseases including herpes labialis (cold
sores), herpes simplex encephalitis (HSE) and some cases of genital
herpes. The first suggestion that HSV1 might have a role in AD
was based on the observation that in HSE the brain regions
damaged are the same ones as those that are affected in AD [1].
Subsequent research demonstrated that the virus is present in the
brain of most elderly subjects, including AD sufferers, and that it is
present within the brain regions affected by AD [2]. Moreover, the
virus was found to be a risk factor for the disease when in brain of
possessors of a specific genetic factor – the type 4 allele of the
apolipoprotein E gene [3–5]. Additional studies have linked HSV1
directly to the neuropathological features of AD. Specifically,
HSV1 infection of cells in culture causes Ab and P-tau
accumulation [6–12] and the brains of HSV1-infected mice
exhibit Ab deposits [7]. Consistently, the enzymes involved in the
generation of Ab and of (P-tau are increased in HSV1-infected
cells [6,7,13]. Further, the distribution of HSV1 DNA in human
brains in relation to senile plaques has been investigated using a
combination of in situ polymerase chain reaction (which amplifies
and thereby detects low levels of target DNA sequences in tissues)
with immunohistochemistry for Ab, or thioflavin S staining for
plaques. This revealed that HSV1 DNA is located very specifically
within senile plaques in AD frontal and temporal cortex [14]. The
in situ polymerase chain reaction findings do not prove HSV1-AD
PLoS ONE | www.plosone.org1 October 2011 | Volume 6 | Issue 10 | e25152

Page 2

causality, but taken with the increased Ab caused by infection,
they suggest that HSV1 is involved in formation of plaques (and
toxic Ab oligomers), and hence support a causal role for in AD.
As mentioned above, the role of HSV1 in AD points to the use
of antiviral agents to slow the progression of the disease. However,
currently available antiviral agents target viral DNA replication
and so might only be effective in AD if the accumulation of Ab and
P-tau depends on viral DNA replication. If, on the other hand, Ab
and P-tau are produced independently of viral DNA replication
then current antiviral agents might not completely inhibit the
production of these proteins, and thus might have a limited effect
on disease progression. The aim of the present study was to
determine the stage of the virus replication cycle necessary for the
accumulation of Ab and P-tau and to determine whether antiviral
agents do reduce the accumulation of Ab and P-tau. Using HSV1
recombinants, we show that Ab and P-tau accumulation requires
the initiation of early and/or late protein synthesis. We found also
that antiviral agents indeed reduce the accumulation of Ab and P-
tau in HSV1-infected cell cultures – supporting the usage of
antiviral agents to treat AD.
Results
Ab and P-tau accumulation requires early and/or late
HSV1 protein synthesis
The infectious cycle of HSV1 is a complex process, which
begins with binding of the virus to heparan sulfate proteoglycan
molecules on the cell surface. Subsequently the virus binds to more
specific receptors which facilitate virus entry into the cell by fusion
[15]. Once inside the cell the virus uncoats, the viral DNA moves
to the nucleus and circularises, and virus gene expression occurs.
Gene expression is initiated by a protein complex consisting of a
viral protein (virion protein (VP16)) and two cellular proteins (host
cell factor 1 and octomer-binding protein 1 (oct1)) [16]. There are
three main classes of genes which are expressed at different times
during infection, and are termed immediate early (IE), early and
late (or alpha, beta and gamma) [17]. IE proteins include infected
cell polypeptide (ICP) 0 and ICP4 which are transcriptional
activators [18,19]. Early proteins include those proteins required
for viral DNA replication such as the protein expressed from the
unique long 42 (UL42) gene [20]. Late proteins include
components of the virus particle such as glycoprotein C (gC)
and proteins required for virus assembly and egress. Expression of
gC is strictly dependent upon viral DNA synthesis [21]. Once
sufficient levels of virion proteins accumulate, virus particles
assemble and the virus exits the cell. Ab and P-tau accumulation
might require any of the stages or components of the infectious
cycle of HSV1.
In order to find which stage of the infection cycle is involved in
the production of Ab and P-tau, we used HSV-1 recombinants
blocked at early stages in the infection cycle. in1374 is a multiple
recombinant that enters cells and is uncoated but does not
synthesise any HSV1 proteins to detectable levels in most infected
cells [22]. Infection of Vero cells with this recombinant, at the
restrictive temperature, did not result in accumulation of Ab or P-
tau (Figure 1A and 2A), whereas infection with the wild-type (WT)
virus in1863 did cause Ab and P-tau accumulation (Figure 1B and
2B). Recombinant tsK/lacZ, a virus with a temperature-sensitive (ts)
mutation in ICP4, at the restrictive temperature expresses IE
proteins at high levels but does not synthesize subsequent early or
late proteins [23]. This recombinant did not show Ab or P-tau
accumulation (Figure 1C and 2C). These results suggest that
neither binding and entry of HSV1 nor the production of IE
proteins is sufficient to cause accumulation of Ab and P-tau. Their
accumulation therefore requires the production of early and/or
late HSV-1-specified proteins.
A viable HSV1 recombinant that specifies a truncated
glycoprotein E (gE) (in1404), and fails to assemble a functional
Fc receptor [24], did produce Ab and P-tau accumulation
(Figure 1D and 2D), suggesting that the Ab and P-tau staining
we observe is not an artefact due to the binding of the primary or
secondary antibodies to the gE-glycoprotein I Fc receptor encoded
by HSV1.
The efficiency of infection with in1863, tsK/lacZ and in1374 was
investigated by staining monolayers for the presence of b-gal. In
cultures infected with in1863 or tsK/lacZ, most cells were positive
for the enzyme (Figure 1 G and H). This demonstrates efficient
primary infection at MOI 5, since tsK/lacZ does not replicate at
the nonpermissive temperature of 38.5uC. Few cells expressed
detectable levels of b-gal in in1374-infected cultures (Figure 1 I).
However, many more were positive for b-gal after co-infection
with WT HSV1 lacking a lacZ insertion (Figure 1 J), because co-
infection provided functional VP16, ICP0 and ICP4.
Inhibition of HSV1 DNA replication reduces the
accumulation of Ab and P-tau in HSV1-infected cells
We investigatedwhether
necessary for the accumulation of Ab and P-tau by using the
antiviral agent acyclovir (ACV), which inhibits HSV1 DNA
replication. ACV targets specifically cells that contain replicating
HSV1 in that its action requires phosphorylation by the viral
thymidine kinase (TK), a more effective phosphorylating agent
than cellular purine or pyrimidine kinases. The monophosphate
form of ACV is then further phosphorylated by cellular kinases
to the active triphosphate form, which has a greater affinity for
viral than for cellular DNA polymerase. Acyclo-GTP is
incorporated into viral DNA, causing chain termination; thus
virus replication and spread are aborted and consequent tissue
damage is halted.
We infected Vero cells with HSV1 at a multiplicity of infection
(MOI) of 1 and examined the effects of varying doses of ACV on
HSV1, Ab and P-tau by immunocytochemistry (ICC). Figures 3,
4 and 5 clearly show a reduction in amounts of HSV1 proteins,
Ab and P-tau in the presence of ACV. Also, at 200 mM ACV the
cells exhibit far less cytopathology. Quantification of these ICC
results using Image J reveals that there is a significant reduction
in HSV1 proteins and abnormal tau phosphorylation at all
concentrations used (p,0.0001 in both cases) and in Ab at
100 mM and 200 mM (p,0.0001) (Figure 6). The staining for P-
tau dropped to almost zero at 50 mM, but that of Ab decreased
less: at 200 mM ACV its level was 28% of the infected cell value in
the absence of antiviral, significantly different (p,0.05) from the
value for mock-infected cells, which corresponded to about 10%
of the infected cell value.
Similarly, ACV treatment did not cause HSV1 staining to
disappear completely. This might reflect the possibility that the
HSV1 strain we used was ACV-resistant, or that the staining was
caused by proteins produced independently of viral DNA
replication. Therefore the reduction of Ab to only 28% might
be related also to the use of an ACV-resistant strain or to the
possibility that it is produced independently of viral DNA
replication. We checked that the strain of virus we were using
was not resistant to some extent to the antiviral agent using plaque
reduction assay (PRA) to determine the dose of ACV which
reduces infectivity to 50% (IC50). In this method, 24-well plates
were seeded with Vero cells, HSV1 was added and the spread of
the virus was restricted to two dimensions by the use of a dense
medium which results in the formation of plaques (holes within the
HSV1 DNA replication was
Antivirals Reduce Key Alzheimer’s Disease Proteins
PLoS ONE | www.plosone.org2October 2011 | Volume 6 | Issue 10 | e25152

Page 3

monolayer of cells). The dose of HSV1 used was chosen in order
to produce a number of plaques that is easy to count (usually
around 80 plaques per well). Addition of an antiviral agent
reduces the number of plaques per well and, if several doses of
antiviral are used, the IC50 can be determined. Using PRA we
found that the IC50 value of ACV was 0.83 mM, which is
consistent with published values [25], and suggests that the strain
of HSV1 we used was not resistant to ACV. An alternative
explanation for why HSV1 staining did not disappear completely
could be because the antibody we used detected multiple HSV1
proteins. Thus the staining detected might represent proteins that
are expressed independently of viral DNA replication. Therefore
we looked at the effects of ACV on specific HSV1 proteins using
monoclonal antibodies specific for ICP0, UL42 and gC, which
represent proteins from the immediate early, early and late stages,
respectively, of the virus infectious cycle. We found that staining
for gC (a late protein that is dependent on DNA replication) was
reduced to less than 1% of the value in the absence of antiviral,
whereas staining for ICP0 (an IE protein) and UL42 (an early
protein), which are independent of viral DNA replication [17],
was reduced but not to the same extent (to ,20% in each case –
data not shown). This shows that the HSV1 staining occurring
even at high ACV doses presumably reflects these and other
proteins that are synthesised independently of viral replication,
Figure 1. Initiation of Ab accumulation requires the initiation of early and/or late protein synthesis. Vero cells were infected for 16
hours with HSV1 strain 17 or with HSV1 recombinants, each at an MOI of 5, and tested for Ab by immunocytochemistry. Infection with in1374
did not lead to Ab accumulation (A) whereas infection with wild type HSV-1 (in1863) did show Ab accumulation (B). Infection with recombinant
tsK/lacZ showed no Ab accumulation at the restrictive temperature (C). Infection with recombinant in1404 did show Ab staining (D). Mock-
infected cells at 37uC (E) or at 38.5uC (F) showed no staining. Scale bar: 50 mm. To confirm infection of cells, monolayers were stained for the
expression of b-gal. Infection with in1863 (G) or tsK/lacZ (H) resulted in b-gal expression in most cells. In cultures infected with in1374 only a
small proportion of cells expressed b-gal (I), whereas many more were positive after co-infection with in1374 and WT HSV1 (MOI 5 for each
virus), demonstrating that in1374 had entered a high proportion of cells (J). Cells show typical HSV1-cytopathic effects, i.e., they become
rounded and contracted.
doi:10.1371/journal.pone.0025152.g001
Antivirals Reduce Key Alzheimer’s Disease Proteins
PLoS ONE | www.plosone.org3 October 2011 | Volume 6 | Issue 10 | e25152

Page 4

and suggests that Ab accumulation too is independent of viral
replication.
Other antiviral agents produce effects similar to those of
ACV
To discover whether other antiviral agents affect Ab and P-tau
accumulation, we investigated penciclovir (PCV) and foscarnet
(FOS). PCV, another guanine analogue, like ACV depends on the
presence of the HSV1 TK, and its products block chain elongation
of the viral DNA, but the two agents differ in their affinity for TK
and the viral DNA polymerase: that of ACV is much greater for
the TK, and that of PCV (as triphosphate) is much greater for the
polymerase [26]. FOS acts by inhibiting the pyrophosphate
binding site on viral DNA polymerases, at levels that do not affect
human DNA polymerases [27].
Both PCV and FOS reduced accumulation of Ab (Figure 7A).
The decreases in Ab level were statistically significant at all
concentrations of PCV used (p,0.0001) whereas FOS reduced Ab
levels significantly only at 200 mM (p,0.0001). Comparison of the
different antiviral agents reveals that for Ab levels, at a
concentration of 50 mM the reduction caused by PCV was
significantly different from those of the other antiviral agents
(p=0.025); at 100 mM, the reductions caused by PCV and ACV
were significantly different from that of FOS (p=0.003), but at
200 mM, there were no significant differences amongst the
antiviral agents. Consistently, the concentration of antiviral agent
required to reduce the ICC signal of Ab to 50% was lowest for
PCV (40.8 mM), followed by ACV (84.4 mM) and then FOS
(135.3 mM).
PCV and FOS reduced P-tau level too. The decreases in P-tau
level were statistically significant at all concentrations of PCV used
(p,0.0001) whereas FOS reduced P-tau levels significantly at
100 mM and 200 mM (p=0.012). However, there were no
significant differences in P-tau staining amongst the antiviral
agents at each concentration. Consistently, the concentrations of
antiviral agents required to reduce the ICC signal of P-tau to 50%
were fairly similar for ACV (26.1 mM), PCV (37.6 mM) and FOS
(48.9 mM).
HSV1 levels were statistically significantly reduced at 100 mM
and 200 mM PCV (p,0.0001) but not at any concentration of
FOS (p=0.182). The difference in HSV1 protein staining amongst
the antiviral agents at 50 mM was not significant (p=0.139). At
100 mM and 200 mM the differences were statistically significant
(p=0.001 and p=0.005, respectively), with PCV and ACV
causing a greater reduction than FOS. Consistently, the
concentrations of ACV and PCV required to reduce the ICC
signal of HSV1 proteins to 50% were similar (86.1 mM and
84.5 mM, respectively) and FOS was lower (.200 mM).
These results show that FOS is less effective than ACV or PCV
not only at reducing HSV1 replication but also at reducing Ab and
P-tau production.
As with ACV, PCV reduced P-tau staining to levels similar to
those in mock-infected cells (Figure 7B) but Ab level reached a
plateau of ,20% (Figures 7A), which was statistically significantly
Figure 2. Initiation of abnormal tau phosphorylation requires the initiation of early and/or late protein synthesis. Vero cells were
infected for 16 hours with HSV1 strain 17 or with HSV1 recombinants, each at an MOI of 5, and tested for abnormal tau phosphorylation (pS214) by
immunocytochemistry. Infection with in1374 did not lead to abnormal tau phosphorylation (A) whereas infection with in1863 did show
phosphorylation (B). Infection with recombinant tsK/lacZ showed no abnormal tau phosphorylation at the restrictive temperature (C). Infection with
recombinant in1404 did show staining (D). Mock-infected cells at 37uC (E) or at 38.5uC (F) showed no staining. Scale bar: 50 mm.
doi:10.1371/journal.pone.0025152.g002
Antivirals Reduce Key Alzheimer’s Disease Proteins
PLoS ONE | www.plosone.org 4 October 2011 | Volume 6 | Issue 10 | e25152

Page 5

higher than the values obtained for mock-infected cells. As with
ACV, we used PRA to check that the strain of HSV1 was not
resistant to PCV. We found that the IC50 for PCV was 7.2 mM,
indicating that the strain was not PCV-resistant. We investigated
also the effects of PCV on HSV1 ICP0, UL42 and gC and found
similar results (data not shown) to those of ACV, suggesting again
that Ab production is produced independently of HSV1 DNA
replication.
In the case of FOS treatment, P-tau and Ab were reduced to
,15% and ,20%, respectively (Figures 8A and 8B). We
examined the effects of FOS on HSV1 gC, UL42 and ICP0.
FOS reduced the levels of all these proteins but in all cases the
levels did not decrease to the levels found in mock-infected cells
(data not shown). This might reflect the use of a FOS-resistant
strain of HSV1 (the IC50 in PRA was .100 mM but this is
consistent with published values), the poor antiviral ability of FOS
and/or DNA replication-independent accumulation.
Comparison of quantitative immunocytochemistry with
HSV1 ELISA
As quantitative ICC (qICC) is not a standard method, we
compared the results of the qICC for HSV1 proteins with HSV1
ELISA (PRA could not be used as conditions comparable to qICC
(i.e., an MOI of 1, and 16 hours infection would be unsuitable as
at this dose, plaque numbers would be vastly too high). HSV1
ELISA revealed that the IC50 values for ACV, PCV and FOS
were 81.6 mM, 183 mM and .200 mM, respectively. These
ELISA values are in agreement with those obtained by qICC,
confirming the validity of the latter.
Figure 3. Acyclovir inhibits HSV1 replication. Vero cells were infected with HSV1 SC16 at an MOI of 1 for 16 hours. Cells were treated with 0 mM,
50 mM, 100 mM or 200 mM acyclovir (ACV), which was present throughout infection. After fixation the slides were tested for HSV1 proteins using
immunocytochemistry. Reactivity with HSV1 proteins was significantly reduced in the presence of ACV. Scale bar: 50 mm.
doi:10.1371/journal.pone.0025152.g003
Antivirals Reduce Key Alzheimer’s Disease Proteins
PLoS ONE | www.plosone.org5October 2011 | Volume 6 | Issue 10 | e25152