THE FULL PAPER
Proteins of the XMRV retrovirus implicated in chronic fatigue syndrome and prostate
cancer are homologous to human proteins relevant to both conditions.
C.J.Carter
The XMRV retrovirus has been implicated in chronic fatigue syndrome and
prostate cancer. A homology search comparing retroviral with human proteins
revealed short contiguous amino acid strings (typically 5-8 aa) matching human
proteins whose dysfunction might be expected to cause fatigue, including
mitochondrial proteins related to oxidative phosphorylation, glutamate receptors and
their synaptic scaffolds, muscular acetylcholine receptor scaffolds and structural
proteins, components of the immune system, and phosphatidylinositol signalling inter
alia. Viral proteins are also homologous to members of the oestrogen, peroxisome
proliferator, and CREB activated receptor networks, all of which are implicated in
prostate cancer, and to a protein, SRCAP, that controls the expression of the prostatespecific
antigen. These short matches are often predicted to be antigenic, and
antibodies to XMRV proteins may target their human homologues. This is supported
by the presence of autoantibodies to muscarinic receptors , vimentin and LAMINB1
(all XMRV homologues) in chronic fatigue syndrome sufferers. Homologous XMRV
proteins might also interfere with the protein interactomes of their human
homologues. Viral mimicry of human proteins is extensive and often relevant to
disease. For example Epstein-Barr viral proteins aligns with multiple sclerosis
autoantigens, while HIV-1 proteins align with several components of the immune
system. Mutant proteins in Huntington’s disease and cystic fibrosis also align with
proteins from common phages or viruses. This suggests a common theme of viral
derived autoimmunity/network interference in many human disorders, which could
radically change the shape of future therapy. Such viral mimicry likely relates to the
idea that life evolved from viruses, leaving behind a legacy of viral derived human
proteins whose homology to the current virome may be responsible for many human
diseases and syndromes. Vaccination programmes or immunosuppression may be
beneficial in many of these conditions.
Introduction.
Chronic fatigue syndrome is an unexplained medical condition characterised
by extreme mental and physical fatigue 25. and by cognitive impairment, depression
and muscular pain and excessive sensitivity to light, sound and smell 41. The condition
has a very high prevalence, for example a figure of 30% of the general population was
recently reported in a study from the Netherlands 46 and there are as many as four
million sufferers in the USA. 7 Perhaps most of the population will experience this
condition at some time in their lives. It may be triggered by infection or trauma or by
vaccination against hepatitis b 33 or against multiple infectious agents (suggested as
the cause of Gulf War syndrome31,49) . The syndrome appears to have an autoimmune
component and antibodies to silicone, squalene , muscarinic receptors and nuclear
envelope antigens have all been reported
The syndrome is treated by low dose antidepressants which are palliative at
best 48, has been considered as a psychiatric phenomenon 47 and can indeed benefit
from cognitive therapy 49. The lack of effective therapy has encouraged the use of
alternative medicine whose effectiveness remains to be verified by the scientific
community 49. Many studies have linked viral infection to the syndrome and
implicated the cytomegalovirus, the Epstein-Barr virus, human herpes virus 6 and 7 ,
hepatitis C, an intestinal enterovirus , the Nipah virus and parvovirus B19
5,10,11,15,32,42,49. Microbial infections are also common in these patients 52,
Perhaps the most intensively studied and hotly debated 21 viral cause of chronic
fatigue syndrome is the XMRV retrovirus (Xenotropic murine leukemia virus-related
virus) initially reported by Lombardi and colleagues 24. This virus has also been
implicated in prostate cancer 4. As shown below, this virus expresses proteins which
are homologous to several human proteins which are relevant to all of the symptoms
encountered in chronic fatigue syndrome and to the proteins implicated in prostate
cancer.
Methods
A homology search between the XMRV genome (NC_007815.1) or the
XMRV proteins and human proteins (blastx; nucleotide search vs proteins or Blastp;
protein vs protein) was undertaken using the NCBI server 2. Only proteins with
matches containing pentapeptide strings were included, except in the case of longer
strings with few gaps or where several contiguous strings were identified. The results
in Tables 1-10 record the position of the match within the viral genome or viral
protein and the position within the matching human proteins, (whose DNA will also
evidently match that of the viral genome) Accession numbers and a brief definition
of function, as recorded in Entrez gene are provided. The B-cell immunogenecity of
different amino acids was calculated using the B-epitope prediction server (Bepipred)
23
http://www.cbs.dtu.dk/services/BepiPred/
and high scoring amino acids are tagged (*) in the various tables
Results
XMRV proteins are homologous to a variety of human proteins as shown in
Table 1-10. Perhaps the most relevant in relation to fatigue are the mitochondrial
proteins involved in respiration and oxidative phosphorylation (eg ATP801, COX11)
(Table 1, Fig 1). Glutamate is the primary excitatory brain neurotransmitter and
XMRV homology to the AMPA receptor GRIA4 and to members of the presynaptic
(Bassoon, piccolo) and postsynaptic machinery (DLGAP3)might also contribute to
cognitive defects (Table 2, Fig 1) as might members of the phosphatidylinositol,
phosphodiesterase and Rho signalling networks (Table 3, Fig 1).. The muscular pain
experienced by chronic fatigue sufferers might well be related to homology with
proteins involved in acetylcholine receptor scaffolding (Table 2, Fig 1) and to
structural muscle elements (Table 4, Fig 1, and the sensitivity to smell to homology
with several olfactory receptors (Table 2 , Fig 1).
XMRV proteins are also homologous to proteins of the growth factor signalling
networks (e.g. tyrosine kinases FLT3 and TYRO3) (Table 5) which are relevant to
cancer-related growth (Fig 2)
They are also homologous to the TAP1 and TAP2 antigen transporter and to a number
of immunoglobulins and cytokine-related proteins (Table 6, Fig 2).
Other classes involved include a number of proteins related to nuclear receptors,
several of which (oestrogen and PPAR receptors and CREB) are directly implicated in
prostate cancer 6,30,50, and to a protein (SRCAP) that controls the expression of the
prostate-specific antigen the 39 ,marker for prostate cancer (Table 7, Fig 2) . Other
homologous classes include transcription factors (Table 8, Fig 2),adhesion molecules
(Table 9, Fig 1 and 2) , proteases and protein processors (Table 11) and a number of
miscellaneous or unknown proteins (Table 13).
Some of the viral translated genome matches to human proteins do not appear
to relate to any known XMRV protein (certain olfactory receptors the phosphatase
PPAPDC2 and the phosphodiesterase ENPP6). This may be related to viral open
reading frames that have not yet been characterised , for example a new ORF has just
been described for the cytomagalovirus 29 , or that may shift with mutation. However
in these cases viral DNA remains homologous to that of the human target.
Different amino acids have different antigenicity depending on their charge
and hydrophobicity characteristics and the B-epitope antigenicity index for each
amino acid is shown in Table 14 23. It should be noted that such indices can change
markedly depending on the number of amino acids in the contiguous string or on the
identity of the neighbouring amino acids. Marked synergy exists when antigenic
amino acids form contiguous stretches. This individual index serves as a rough gauge
of the antigenic potential of the peptide. The top 5 antigenic amino acids are marked
with an asterisk in the various tables, and contiguous antigenic amino acids, which are
the most likely epitope and cross-reactive candidates, are highlighted in grey. It can
be seen that a number of matching proteins are predicted to be highly antigenic.
These include the mitochondrial proteins CHCHD10 ,acetyl CoA carboxylase and the
sulfite oxidase SUOX, (Table 1), the glutamate receptor GRIA4 (Table 2) the TBCC
tumour suppressor (Table 7) numerous growth regulators (FLT3, TYRO3,WNT10B
and EIF4B ) (Table 5) the antigen transporter TAP1 (Table 6) the PPAR and
oestrogen receptor and PSA regulators, PELP1, PPRC1 and SRCAP (Table 7), the
transcription factor FOXO6 (Table 8) , pleckstrin 3 (Table 9) the metalloprotease
ADAMTS9 and ubiquilin 3 (Table 10).
Discussion.
The XMRV virus expresses predicted proteins with homology to human
proteins that are clearly highly relevant to the symptoms encountered in this disease,
including fatigue (mitochondrial respiration), cognitive deficits 49 (glutamate, PI
signalling) problems related to olfaction 41(olfactory receptors) muscular pain 41
(acetylcholine receptor and muscle related structural proteins) and the association
with many active viral and microbial infections 19 (immune related proteins).
.XMRV viral proteins are also highly homologous to components of peroxisome
proliferator-activated receptor gamma (PPARG) or oestrogen receptor signalling
networks both of which have been implicated in prostate cancer 6,30 A viral protein is
also homologous to SRCAP which controls the expression of the prostate specific
antigen, the cardinal marker of prostate cancer39.
Although these matching strings are short, they are often contiguous and the
viral homologues may interfere with their human counterparts in a number of ways.
Firstly, many of these stretches are predicted to be antigenic and antibodies to the
virus may also target the human homologues. Indeed autoantibodies to muscarinic
receptors, cellular cytoskeletal components, including vimentin and other Lamina
related proteins and to nuclear envelope proteins including Lamin A and LAMIN B1
have been reported in chronic fatigue syndrome 22,44. These were all identified as
XMRV homologues in this survey (specifically muscarinic receptor, CHRM2, Lamin
B1 and Vimentin and generally the cytoskeletal components in Table 4).
Viral DNA will also match that of the human target and possibly sequester host
transcription factors or microRNA of influence splicing of the homologous human
gene.
These homologies, targeted at highly relevant proteins support the implication
of the virus in both chronic fatigue syndrome and prostate cancer.
Several studies have failed to detect the virus in chronic fatigue syndrome and
the issue is hotly debated 14,21,28. Because the viral proteins are homologous to human
proteins, it is likely that any anybodies generated in response to the virus would target
their human homologues. This is supported by the high predicted antigenicity of a
number of viral matching proteins and by the presence of autoantibodies to viral
matching proteins in chronic fatigue sufferers. As the human proteins are persistently
encountered by the antibody, such an autoimmune response would become selfsustaining,
thus no longer requiring the presence of the virus. Indeed, the more
successful the immune response against the virus, the greater the risk of
autoimmunity, a disastrous pyrrhic victory. This scenario could thus explain the
failure to detect the virus in several studies.
It has recently been shown that Raltegravir potently inhibits XMRV
replication and clinical trials in chronic fatigue syndrome and prostate cancer have
already been proposed 38 a suggestion supported by the results of this survey.
However, it is possible that the effects of the virus, in the form of autoimmunity
persist after viral elimination and different strategies including immunosuppression
might be considered in such cases.
Viral mimicry of human proteins appears to be a near universal phenomenon.
For example several viruses (almost 100), including those implicated in late-onset
Alzheimer’s disease, express proteins that are homologous to beta-amyloid .The
autoantigens in multiple sclerosis, myasthenia gravis, systemic lupus erythromatosus,
Pemphigus Vulgaris and Sjogrens syndrome align with proteins from their respective
reported viral risk factors (and novel suspects), and HIV-1 proteins align with
important components from all compartments of the human immune system. This
even applies to human genetic disorders, as the APP mutations in Alzheimer’s disease
convert the surrounding peptide to sequences that match proteins from commensal
bacteria and from the Norovirus and common cold virus. The mutant proteins in
Huntington’s disease and cystic fibrosis also align with proteins expressed by very
common viruses or phages 8,9. This suggests that a large number of diseases have an
autoimmune component triggered by viral antigens with homology to important
human proteins.
Viruses or phages were long ago proposed as the origin of life 13,17. While
responsible for our existence, they appear to have left behind a legacy of viral derived
human proteins containing short but contiguous and often immunogenic amino acid
stretches homologous to current viral antigens. The autoimmune defence so triggered
may be responsible for a large number of human illnesses. The therapeutic
implications of such mimicry are clearly extensive as it suggests an autoimmune
component in a variety of disorders which might thus benefit from vaccination against
the appropriate pathogen, immunosuppression and pathogen elimination.
Table 1 to 10
Human proteins that match proteins from the translated XMRV genome. The
alignment is shown and accession numbers are provided together with a brief
description of the function of the protein. Unless specifically referenced, these are
copied from the gene descriptions in ENTREZ gene. The matching peptide sequence
in each case is highlighted in bold and contiguous amino acids of 5 or more or longer
stretches of contiguity with few gaps are boxed. The asterisks represent the 5 highest
scoring antigenic amino acids (Table 14) and the grey highlights illustrate contiguous
antigenic stretches that are the most likely B-cell epitopes, and the most likely
candidates for cross-reactivity.