Mannose binding lectin is required for alphavirus-induced arthritis/myositis.

Abstract

Mosquito-borne alphaviruses such as chikungunya virus and Ross River virus (RRV) are emerging pathogens capable of causing large-scale epidemics of virus-induced arthritis and myositis. The pathology of RRV-induced disease in both humans and mice is associated with induction of the host inflammatory response within the muscle and joints, and prior studies have demonstrated that the host complement system contributes to development of disease. In this study, we have used a mouse model of RRV-induced disease to identify and characterize which complement activation pathways mediate disease progression after infection, and we have identified the mannose binding lectin (MBL) pathway, but not the classical or alternative complement activation pathways, as essential for development of RRV-induced disease. MBL deposition was enhanced in RRV infected muscle tissue from wild type mice and RRV infected MBL deficient mice exhibited reduced disease, tissue damage, and complement deposition compared to wild-type mice. In contrast, mice deficient for key components of the classical or alternative complement activation pathways still developed severe RRV-induced disease. Further characterization of MBL deficient mice demonstrated that similar to C3(-/-) mice, viral replication and inflammatory cell recruitment were equivalent to wild type animals, suggesting that RRV-mediated induction of complement dependent immune pathology is largely MBL dependent. Consistent with these findings, human patients diagnosed with RRV disease had elevated serum MBL levels compared to healthy controls, and MBL levels in the serum and synovial fluid correlated with severity of disease. These findings demonstrate a role for MBL in promoting RRV-induced disease in both mice and humans and suggest that the MBL pathway of complement activation may be an effective target for therapeutic intervention for humans suffering from RRV-induced arthritis and myositis.

Full text

Page 1

Mannose Binding Lectin Is Required for
Alphavirus-Induced Arthritis/Myositis
Bronwyn M. Gunn1,2., Thomas E. Morrison3., Alan C. Whitmore2, Lance K. Blevins2, Linda Hueston4,
Robert J. Fraser5, Lara J. Herrero6,7, Ruben Ramirez6, Paul N. Smith8, Suresh Mahalingam6,7*,
Mark T. Heise1,2,9*
1Department of Microbiology and Immunology, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, United States of America, 2Carolina Vaccine
Institute, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, United States of America, 3Department of Microbiology, University of Colorado School of
Medicine, Aurora, Colorado, United States of America, 4Arbovirus Emerging Disease Unit, CIDMLS-ICPMR, Westmead Hospital, Westmead, Australia, 5Department of
Medicine, Royal Melbourne Hospital, University of Melbourne, Melbourne, Australia, 6Virus and Inflammation Research Group, University of Canberra, Canberra, Australia,
7Emerging Viruses and Inflammation Research Group, Institute for Glycomics, Griffith University, Gold Coast, Australia, 8The Australian National University, Canberra,
Australia, 9Department of Genetics, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, United States of America
Abstract
Mosquito-borne alphaviruses such as chikungunya virus and Ross River virus (RRV) are emerging pathogens capable of
causing large-scale epidemics of virus-induced arthritis and myositis. The pathology of RRV-induced disease in both humans
and mice is associated with induction of the host inflammatory response within the muscle and joints, and prior studies
have demonstrated that the host complement system contributes to development of disease. In this study, we have used a
mouse model of RRV-induced disease to identify and characterize which complement activation pathways mediate disease
progression after infection, and we have identified the mannose binding lectin (MBL) pathway, but not the classical or
alternative complement activation pathways, as essential for development of RRV-induced disease. MBL deposition was
enhanced in RRV infected muscle tissue from wild type mice and RRV infected MBL deficient mice exhibited reduced
disease, tissue damage, and complement deposition compared to wild-type mice. In contrast, mice deficient for key
components of the classical or alternative complement activation pathways still developed severe RRV-induced disease.
Further characterization of MBL deficient mice demonstrated that similar to C32/2mice, viral replication and inflammatory
cell recruitment were equivalent to wild type animals, suggesting that RRV-mediated induction of complement dependent
immune pathology is largely MBL dependent. Consistent with these findings, human patients diagnosed with RRV disease
had elevated serum MBL levels compared to healthy controls, and MBL levels in the serum and synovial fluid correlated with
severity of disease. These findings demonstrate a role for MBL in promoting RRV-induced disease in both mice and humans
and suggest that the MBL pathway of complement activation may be an effective target for therapeutic intervention for
humans suffering from RRV-induced arthritis and myositis.
Citation: Gunn BM, Morrison TE, Whitmore AC, Blevins LK, Hueston L, et al. (2012) Mannose Binding Lectin Is Required for Alphavirus-Induced Arthritis/
Myositis. PLoS Pathog 8(3): e1002586. doi:10.1371/journal.ppat.1002586
Editor: Michael S. Diamond, Washington University School of Medicine, United States of America
Received October 7, 2011; Accepted January 29, 2012; Published March 22, 2012
Copyright: ? 2012 Gunn 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 NIH/NIAMS R01 AR 047190 awarded to MTH. 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: s.mahalingam@griffith.edu.au (SM); heisem@med.unc.edu (MTH)
. These authors contributed equally to this work.
Introduction
Arthritogenic alphaviruses, such as Ross River virus (RRV) and
chikungunya virus (CHIKV), are mosquito-borne viruses that
cause severe polyarthritis and myositis in humans. RRV causes
annual disease outbreaks in Australia and has caused sporadic
epidemics of debilitating polyarthritis, including one outbreak
involving over 60,000 people in Oceania [1]. RRV is transmitted
to humans primarily by the Aedes and Culex species of mosquitoes
that generally populate marsh areas, and CHIKV transmission has
been traditionally mediated by the urban Aedes aegypti, though the
virus has recently adapted for efficient transmission by the widely
distributed Aedes albopictus species [2], leading to an increased risk
for CHIKV spread into new areas, as illustrated by recent
outbreaks in Italy and southern France [3]. The expansion of
CHIKV into an additional mosquito vector and the subsequent
epidemic has highlighted the ability of the arthritic alphaviruses to
move into new geographic areas and cause large-scale outbreaks of
acute and persistent arthralgia and myalgia in humans.
RRV-induced arthritic disease presents predominantly as
painful stiffness, inflammation, and swelling in peripheral joints
that can last months after initial infection and the host
inflammatory response is thought to play a major role in disease
pathogenesis. Inflammatory monocytes constitute the bulk of
leukocytes isolated in synovial aspirates from RRV-infected
patients [4,5], and macrophage-cytotoxic drugs have been shown
to drastically reduce disease progression and severity in mice [6,7].
In addition, mice lacking C3, the central complement factor that is
essential for complement activation, exhibit reduced RRV-
induced disease and tissue destruction [8], implicating a role for
PLoS Pathogens | www.plospathogens.org1March 2012 | Volume 8 | Issue 3 | e1002586

Page 2

complement in development of the disease. Consistent with studies
in mice, synovial aspirates from patients with RRV-induced
arthritis have been shown to contain increased levels of the C3
cleavage product C3a [8]. Although macrophage recruitment to
infected tissues is markedly increased after RRV infection, the role
played by complement is independent of inflammatory cell
recruitment. Rather, tissue destruction and disease progression
requires complement receptor 3 (CR3), suggesting that comple-
ment interactions with CR3 on inflammatory cells promote tissue
destruction in RRV-infected tissues [9]. However, it is currently
unclear how the complement system is activated following RRV
infection.
There are three main activation pathways of the complement
cascade; the classical, alternative, and lectin dependent pathways,
that all converge on factor C3 and lead to activation of
complement effector functions (reviewed in [10]). The classical
pathway is initiated by C1q interactions with antigen-bound
complexes of IgG and IgM, and the proteases C1r and C1s cleave
C4 and C2 to generate the C3 convertase C4b2b. Binding of
factor B (fB) and spontaneously hydrolyzed C3 initiates the
alternative pathway, and fB binding to C3b leads to formation of
the alternative C3 convertase C3bBb that can amplify comple-
ment activation. In the lectin pathway, mannose binding lectin
(MBL) or the ficolins bind to carbohydrate moieties on foreign
bodies, such as viruses, or to host cells and apoptotic cells, and the
MBL-associated serine proteases (MASPs) cleave C4 and C2 to
form the C3 convertase C4b2b. Cleavage and processing of C3 by
the C3 convertases produce several C3-derived components that
are potent activators of the immune system. One such component
is iC3b, which is a ligand for several complement receptors, such
as CR3. Binding of iC3b to CR3 on cells such as monocytes/
macrophages, neutrophils, and NK cells, results in activation of
these cells, leading to enhanced phagocytosis and cytotoxic activity
against iC3b-opsonized cells [10].
MBL is a soluble C-type lectin that can initiate the complement
cascade through binding of the carbohydrate recognition domains
(CRD) to cell-surface sugars expressed on bacteria and viruses and
some endogenous host ligands (reviewed in [11]). The complement
system is notorious for having both a protective and pathologic
role and frequently leads to additional tissue injury and damage
once activated. Similarly, MBL appears to be able to have the
ability to protect as well as harm the host cells. In the context of
sterile inflammatory diseases such as myocardial and gastrointes-
tinal ischemic reperfusion injury, MBL and the lectin pathway
mediate development of disease through complement-mediated
regulation of pro-inflammatory cytokines and inflammation,
leading to exacerbated pathology and tissue injury [12–15]. In
contrast to the pathologic role of MBL in sterile inflammatory
diseases, MBL is thought to primarily play a protective role in
response to infectious pathogens. MBL has been shown to be
essential for host protection from many different viral and
bacterial infections either through direct binding to pathogens or
by limiting spread through complement effector functions. MBL
has been shown to bind directly to many different viruses,
including human immunodeficiency virus (HIV), Ebola virus, and
arboviruses such as dengue virus and West Nile virus (WNV), and
MBL can either directly neutralize these viruses through activation
of complement or interfere with their binding to host cells
(reviewed in [16]). Furthermore, studies using mice deficient in
both MBL genes (MBL-A2/2and MBL-C2/2; MBL-DKO) have
revealed that MBL can have a protective role during WNV and
herpes simplex virus infections [17–19]. Though MBL neutralizes
flaviviruses, such as WNV and dengue virus [18], and plays a
protective role during WNV infection [19], these same studies
found no detectable interactions between MBL and alphaviruses,
suggesting that MBL does not play a protective role during
alphavirus infection. However, the role of MBL role in the
pathogenesis of alphavirus-induced arthritis/myositis has not been
evaluated.
The goal of this study was to further assess the role of the host
complement system in the pathogenesis of alphavirus-induced
inflammatory disease and to determine which complement
activation pathways are required for virus-induced disease. In a
mouse model of RRV-induced arthritis and myositis, mice
deficient in either the classical or alternative pathways developed
severe disease, while mice deficient in both genes of MBL (MBL-
DKO) were resistant to disease, suggesting that MBL plays a
major role in RRV-induced disease. Similar to previous findings
with C32/2mice [8], RRV-infected MBL-DKO mice had similar
levels of viral burden and inflammation compared to wild-type
(WT) but exhibited significantly less complement deposition, tissue
damage, and disease. Further analysis found that MBL levels are
enhanced in RRV infected tissues and that MBL binds to RRV
infected cells, suggesting that RRV infection leads to MBL
deposition and subsequent complement activation. Importantly,
studies in human patients suffering from RRV-induced disease
found that levels of MBL were elevated in the serum of RRV-
infected patients compared to healthy controls. In addition, serum
and synovial fluid MBL levels correlated with the severity of RRV
disease, while no differences were observed in classical or
alternative pathway activation, suggesting that MBL contributes
to RRV-induced disease in human populations.
Results
The MBL pathway is essential for RRV-induced disease
and inflammatory tissue destruction
Complement activation products are elevated in the synovial
fluid of persons suffering from RRV-induced arthritis and
complement activation is required for virus-induced arthritis/
myositis in a mouse model [8,9,20]. Although other alphaviruses,
such as the neurovirulent Sindbis virus, have been shown to
activate complement via both the classical and alternative
Author Summary
Arthritogenic alphaviruses such as Ross River virus (RRV)
and chikungunya virus are transmitted to humans by
mosquitoes and cause epidemics of debilitating infectious
arthritis and myositis in various areas around the world.
Studies in humans and mice indicate that the host
inflammatory response is critical for development of RRV-
induced arthritis and myositis, and the host complement
system, a component of the host inflammatory response,
plays an essential role in the development of RRV-induced
disease through activation of complement receptor 3
(CR3)-bearing inflammatory cells. Of the three main
complement activation pathways, only the lectin pathway
activated by mannose binding lectin (MBL) was essential
for RRV-induced complement activation, tissue destruc-
tion, and disease. Furthermore, we found that levels of
MBL were elevated in human patients suffering from RRV-
induced polyarthritis and MBL levels correlated with
disease severity. Taken together, our data implicates a
role for MBL in mediating RRV-induced disease in both
humans and mice, and suggests that therapeutic targeting
of MBL may be an effective strategy for disease treatment
in humans.
MBL Is Required for RRV-Induced Arthritis/Myositis
PLoS Pathogens | www.plospathogens.org2March 2012 | Volume 8 | Issue 3 | e1002586

Page 3

pathways [21], the pathway(s) leading to complement activation by
arthritic alphaviruses is currently unknown. Therefore, mice
deficient in key components of the classical (C1q2/2), alternative
(factor B, fB2/2), or lectin (MBL-A/C2/2, MBL-DKO) pathways
were assessed for their susceptibility to RRV-induced disease.
C1q2/2, fB2/2, MBL-DKO, or WT C57BL/6 mice were
inoculated with RRV in the footpad and assessed for weight loss
and scored for hind limb function as previously described [20].
RRV-infected WT mice showed signs of hind-limb weakness by 5
days post infection (dpi) and developed severe hind-limb weakness
by 7 dpi through 10 dpi (Figure 1A). RRV causes disease
independently of B cells and antibody [20], suggesting that the
classical pathway does not contribute to disease during RRV
infection. Consistent with this, C1q2/2mice exhibited severe
disease signs and hind-limb weakness similar to WT animals
(Figure 1A). Likewise, fB2/2mice developed severe RRV-induced
disease (Figure 1A), demonstrating that the alternative pathway of
complement activation is not required for RRV-induced disease,
though it is important to note that RRV-infected fB2/2mice
tended to develop more severe disease compared to WT mice,
suggesting that the alternative pathway is activated and may
actually play a protective role during RRV infection. In contrast to
C1q2/2and fB2/2mice, MBL-DKO mice infected with RRV
developed mild hind-limb weakness and exhibited reduced weight
loss compared to WT mice (Figure 1, A and B). Mock-infected
WT and MBL-DKO mice did not differ in weight gain (Figure S1)
and showed no signs of disease throughout the course of infection
(data not shown). While we cannot rule out a minor contribution
of the classical and alternative activation pathways or activation of
the lectin pathway through ficolins in development of RRV
disease, this data demonstrates that the lectin pathway initiated by
MBL plays an essential role in driving RRV-induced disease.
MBL contributes to damage within quadriceps muscle
Following infection of WT C57BL/6 mice, RRV replicates to
high levels within both the joints and skeletal muscle and elicits an
inflammatory infiltrate into these tissues [20]. Following the onset
of inflammatory cell infiltration, wild type mice develop severe
destructive myositis, which is a major aspect of virus-induced
disease in this mouse model, and we have previously shown that
muscle cell killing and disease is dependent upon both C3
activation and CR3 [8,9]. Therefore, to confirm the role of MBL
in driving RRV-induced disease, inflammatory pathology was
assessed within the quadriceps muscles of RRV-infected fB2/2,
C1q2/2, MBL-DKO, or WT mice by H&E staining of paraffin
embedded sections. At 10 dpi, a time point of peak RRV disease,
we observed similar inflammation and tissue pathology in fB2/2,
C1q2/2, and WT mice (Figure 1C). Inflammatory cells are
present in the quadriceps muscles of infected mice from all strains,
as indicated by the solid arrowheads. We observed tissue damage
in the quadriceps muscles in fB2/2, C1q2/2, and WT mice as
evidenced by the degeneration of the fibrous architecture of the
skeletal muscle (marked by open arrowheads). In contrast, RRV
infected MBL-DKO mice maintained the architecture of the
skeletal muscle with very little tissue damage, despite the presence
of inflammatory cells (Figure 1C), which was strikingly similar to
previous results demonstrating an essential role for C3 in RRV-
induced disease [8]. To confirm that MBL-DKO mice have
decreased tissue damage following RRV infection compared to
WT mice, we used Evans Blue dye (EBD) uptake to detect areas of
damage. Consistent with the clinical scores and histological
analyses at 10 dpi, RRV-infected WT mice had abundant EBD
positive muscle fibers within the quadriceps muscle whereas EBD
positive cells were rare in RRV-infected MBL-DKO mice
(Figure 1D), further demonstrating that MBL is required for the
induction of tissue damage during RRV-induced disease.
RRV infection induces MBL deposition onto tissues and
cells
To determine if MBL is deposited onto tissues following RRV
infection, we evaluated tissue homogenates of quadriceps muscle
from RRV-infected WT mice at 7 dpi for levels of MBL by
immunoblot analysis. We observed an increase in the amount of
MBL in the quadriceps muscle of infected mice compared to that
of mock-infected mice (Figure 2A), indicating that RRV infection
results in elevated amounts of MBL within target tissues. Given the
enhanced MBL within RRV infected muscle tissue, we next
evaluated whether MBL would directly bind to RRV or RRV
infected cells. Studies with two other alphaviruses, CHIKV and
Sindbis virus, found no evidence for interactions with MBL, while
mosquito derived West Nile virus was efficiently bound and
neutralized by MBL [18]. Consistent with these findings, we were
unable to detect direct binding between MBL and either
mammalian cell or mosquito cell derived RRV virions by ELISA
(data not shown) and we also found no evidence for MBL-
mediated neutralization of RRV (Figure S2). Given the lack of
detectable interactions between MBL and the RRV virion, we
assessed whether MBL bound to RRV-infected cells. Differenti-
ated C2C12 murine skeletal muscle cells were infected with RRV
for 24 hours and then incubated with medium containing either
C57BL/6 wild-type serum or MBL-DKO serum for 30 minutes.
Cell lysates were harvested and analyzed for presence of MBL-C
by immunoblot analysis. As shown in Figure 2B, the amount of
MBL-C deposition was enhanced in cells infected with RRV
compared to mock infected cells, indicating that RRV infection
results in increased MBL binding to cells. No deposition of MBL
was detected onto cells incubated with MBL-DKO serum,
indicating the specificity of detection. Therefore, though MBL
does not appear to interact with the RRV virion, RRV infection
does lead to MBL deposition on infected cells.
Complement deposition and activation is reduced in
RRV-infected MBL-DKO compared to WT mice
MBL, but not the alternatively or classical complement
activation pathways, was required for RRV-induced disease and
tissue pathology (Figure 1), and the phenotype in MBL-DKO mice
is strikingly similar to C32/2mice, suggesting that MBL plays a
major role in driving complement activation during RRV
infection. Therefore, we directly assessed whether MBL was
required for RRV-dependent complement deposition. Western
blot analysis of skeletal muscle from wild type or MBL-DKO mice
indicated that C3 levels, including the a and b chains of C3 were
present at reduced levels in the skeletal muscle of RRV infected
MBL-DKO mice compared to wild type mice (Figure S3A–B).
However, since inflammatory macrophages produce C3 [22],
western blot analysis was not able to clearly differentiate between
complement deposition within the tissue and de novo production
of complement by the infiltrating inflammatory cells in both wild
type and MBL-DKO animals. Therefore, we directly assessed the
impact of MBL deficiency on complement deposition within the
RRV infected muscle by performing immunohistochemistry on
quadriceps muscle from WT and MBL-DKO mice using an anti-
mouse C3 antibody. Abundant C3 staining localized to damaged
skeletal muscle at 7 dpi in WT mice while C3 staining was
substantially reduced in muscle from RRV-infected MBL-DKO
mice (Figure 2C). Importantly, we observed comparable C3
staining between RRV-infected C1q2/2, fB2/2, and WT mice
MBL Is Required for RRV-Induced Arthritis/Myositis
PLoS Pathogens | www.plospathogens.org3 March 2012 | Volume 8 | Issue 3 | e1002586

Page 4

(Figure S3C), suggesting that neither C1q nor fB are required for
C3 deposition on muscle tissue following RRV infection.
Therefore, these results suggest that MBL is the major mediator
of complement activation and deposition within RRV infected
muscle tissue.
Viral replication is unaffected in the muscle tissue of
MBL-DKO mice
Prior studies with C32/2and CR32/2mice demonstrated that
complement activation and CR3-dependent signaling is essential
for RRV-induced disease and tissue destruction, but complement
deficiency had no effect on viral burden or tropism. To determine
whether this was also the case in MBL-DKO mice, we evaluated
WT and MBL-DKO mice for viral load within the quadriceps
muscles, ankle joints, and serum. As shown in Figure 3A, MBL-
DKO mice exhibited no significant difference in the amount of
infectious virus in the quadriceps muscle through days 7 and
10 dpi, which represent the times when RRV-induced muscle
destruction peaks. Furthermore, analysis of the viral distribution
within wild type and MBL-DKO animals by in situ hybridization
found no differences in the localization of RRV specific signal
between the two mouse strains. (Figure 3D). Therefore, the
Figure 1. MBL is required for development of severe RRV-induced disease and tissue damage. (A–B) Twenty-four day old WT C57BL/6
(solid square, n=6), C1q2/2(open square, n=6), fB2/2(solid circle, n=5), or MBL-DKO (open circle, n=6) mice were infected with RRV, scored for
hind limb function based on a scale described in the Materials and Methods, and assessed for weight loss. Each data point represents the arithmetic
mean 6SD and is representative of at least three independent experiments. We performed a Mann-Whitney analysis with multiple comparison
corrections (p,0.01 was considered significant) on clinical scores at 10 dpi and a one-way ANOVA analysis with Bonferroni’s correction on percent of
starting weight at 10 dpi (p,0.05 was considered significant) to determine significance between the various knock-out lines compared to WT
**p,0.01; *** p,0.001; n.s. not significant. (C) Tissue pathology and inflammation was examined at 10 dpi by H&E staining of paraffin embedded
sections of quadriceps muscle. A representative section from each knockout strain is shown. A section from RRV-infected C32/2is shown for
comparison. Solid arrowheads point to areas of inflammation; open arrowheads indicate tissue damage (D) To assess damage within the muscle,
mock or RRV-infected WT or MBL-DKO mice were injected with EBD at 10 dpi, and frozen sections were generated. EBD positive muscle fibers were
identified by fluorescence microscopy. Representative sections of mock- and RRV-infected mice are shown and are representative of two
independent experiments.
doi:10.1371/journal.ppat.1002586.g001
MBL Is Required for RRV-Induced Arthritis/Myositis
PLoS Pathogens | www.plospathogens.org4 March 2012 | Volume 8 | Issue 3 | e1002586

Page 5

Figure 2. RRV infection induces MBL deposition onto cells, resulting in MBL-dependent C3 deposition onto infected tissues. (A) To
determine if MBL levels are elevated within the quadriceps muscles of RRV-infected wild-type animals, homogenized quadriceps muscles from either
MBL Is Required for RRV-Induced Arthritis/Myositis
PLoS Pathogens | www.plospathogens.org5March 2012 | Volume 8 | Issue 3 | e1002586