A protective role for complement C3 protein during pandemic 2009 H1N1 and H5N1 influenza A virus infection.

Abstract

Highly pathogenic H5N1 influenza infections are associated with enhanced inflammatory and cytokine responses, severe lung damage, and an overall dysregulation of innate immunity. C3, a member of the complement system of serum proteins, is a major component of the innate immune and inflammatory responses. However, the role of this protein in the pathogenesis of H5N1 infection is unknown. Here we demonstrate that H5N1 influenza virus infected mice had increased levels of C5a and C3 activation byproducts as compared to mice infected with either seasonal or pandemic 2009 H1N1 influenza viruses. We hypothesized that the increased complement was associated with the enhanced disease associated with the H5N1 infection. However, studies in knockout mice demonstrated that C3 was required for protection from influenza infection, proper viral clearance, and associated with changes in cellular infiltration. These studies suggest that although the levels of complement activation may differ depending on the influenza virus subtype, complement is an important host defense mechanism.

Full text

Page 1

A Protective Role for Complement C3 Protein during
Pandemic 2009 H1N1 and H5N1 Influenza A Virus
Infection
Kevin B. O’Brien1,2, Thomas E. Morrison3, David Y. Dundore1, Mark T. Heise4,5,6, Stacey Schultz-Cherry2*
1Department of Medical Microbiology and Immunology, University of Wisconsin, Madison, Wisconsin, United States of America, 2Department of Infectious Diseases, St.
Jude Children’s Research Hospital, Memphis, Tennessee, United States of America, 3Department of Microbiology, University of Colorado–Denver, Aurora, Colorado,
United States of America, 4Department of Genetics, University of North Carolina, Chapel Hill, North Carolina, United States of America, 5Department of Microbiology and
Immunology, University of North Carolina, Chapel Hill, North Carolina, United States of America, 6Carolina Vaccine Institute, University of North Carolina, Chapel Hill,
North Carolina, United States of America
Abstract
Highly pathogenic H5N1 influenza infections are associated with enhanced inflammatory and cytokine responses, severe
lung damage, and an overall dysregulation of innate immunity. C3, a member of the complement system of serum proteins,
is a major component of the innate immune and inflammatory responses. However, the role of this protein in the
pathogenesis of H5N1 infection is unknown. Here we demonstrate that H5N1 influenza virus infected mice had increased
levels of C5a and C3 activation byproducts as compared to mice infected with either seasonal or pandemic 2009 H1N1
influenza viruses. We hypothesized that the increased complement was associated with the enhanced disease associated
with the H5N1 infection. However, studies in knockout mice demonstrated that C3 was required for protection from
influenza infection, proper viral clearance, and associated with changes in cellular infiltration. These studies suggest that
although the levels of complement activation may differ depending on the influenza virus subtype, complement is an
important host defense mechanism.
Citation: O’Brien KB, Morrison TE, Dundore DY, Heise MT, Schultz-Cherry S (2011) A Protective Role for Complement C3 Protein during Pandemic 2009 H1N1 and
H5N1 Influenza A Virus Infection. PLoS ONE 6(3): e17377. doi:10.1371/journal.pone.0017377
Editor: Jean-Pierre Vartanian, Institut Pasteur, France
Received August 27, 2010; Accepted February 1, 2011; Published March 9, 2011
Copyright: ? 2011 O’Brien 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 Howard Hughes Medical Institute Gilliam Fellowship, an Advanced Opportunity Fellowship (AOF) from the University of
Wisconsin-Madison, and a Sigma Xi Grants-in-Aid-of-Research grant to K. O’Brien, NIH grant AR047190 to M.T. Heise, and NIH AI059049, and in part by the NIH
NIAID contract number HHSN266200700005C and the American Lebanese Syrian Associated Charities (ALSAC) to S. Schultz-Cherry. 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: stacey.schultz-cherry@stjude.org
Introduction
Since emerging in 1997, highly pathogenic avian influenza
(HPAI) H5N1 viruses have been associated with over 500 human
infections with an unprecedented fatality rate exceeding 60%. The
significant virulence of these viruses, their continual evolution in
birds, and the co-circulation of the pandemic H1N1 virus lead to
concerns that the H5N1 viruses pose pandemic threats.
The severe disease associated with HPAI H5N1 infections in
humans and animals could result from several factors including
dissemination of the virus beyond the respiratory tract, higher and
prolonged viral replication leading to increased viral cytolytic
damage, differences in the host response induced by the H5N1
viruses, or a combination of all these factors. Although host
responses are clearly complex, the clinical data and animal models
suggest that the innate immune responses differ in H5N1 infected
individuals (reviewed in [1,2]). As compared to seasonal influenza
infection H5N1 infected patients have elevated serum levels of
several chemokines and cytokines [3,4,5]. Similar results were
observed in animal models where H5N1 infection is associated
with elevated cytokine and chemokine levels [2,6,7,8,9], enhanced
recruitment of macrophages and neutrophils into the lungs leading
to acute lung inflammation [10], and premature apoptosis of
dendritic cells [11]. This aberrant host response is reminiscent of
1918 influenza virus infected animals [10,12].
A major component of the innate immune response that has not
been evaluated during HPAI H5N1 infections is complement. The
complement system is comprised of more than 30 proteins
responsible for recognizing and eliminating pathogens while
stimulating early and late cellular functions (reviewed in [13]).
Three biochemical pathways activate the complement system: the
classical complement pathway, the alternative complement
pathway, and the mannose-binding lectin pathway [14]. The
hydrolysable C3 protein is the converging point for all three
complement activation pathways, making it the central player in
the complement cascade [15]. Upon activation, C3 is cleaved into
C3a and C3 convertase, which supports the further cleavage of C5
into C5a. C3a and C5a function similar to chemokines promoting
localized attraction and activation of immune cells including
neutrophils, which serve an important role in early and late
defense against pathogens including influenza virus [16,17,18,19].
A recent study by Boon et al demonstrated a protective role for
complement C5 in H5N1 influenza pathogenesis [20]. Thus, the
goal of these studies was to fill our gap in knowledge by
determining if mice infected with H5N1 influenza virus differed
in complement C3 activation as compared to a seasonal or the
PLoS ONE | www.plosone.org1March 2011 | Volume 6 | Issue 3 | e17377

Page 2

pandemic 2009 H1N1 influenza virus and if C3 was required for
protection from HPAI H5N1 influenza infection.
Results
H5N1 influenza virus increases C3 and C5a lung levels as
compared to seasonal or pandemic influenza strains
To quantitate C3 and C5a protein levels during influenza
infection, C57BL/6 mice were lightly anesthetized and intrana-
sally inoculated with PBS (uninfected control), seasonal H1N1 A/
Puerto Rico/8/1934 (PR/8), pandemic 2009 H1N1 A/Califor-
nia/7/2009 (CA/09), or HPAI H5N1 A/Vietnam/1194/04 (VN/
1194) and bronchioalveolar lavage (BAL) collected on days 1, 3,
and 6 post-infection (pi). Using a mouse-specific C3 sandwich
ELISA we found that within 3 days post-infection (dpi) C3 levels
were .3-fold higher in the H5N1 VN/1194 virus infected mice as
compared to controls and continued to increase to .4-fold by
6 dpi (Fig. 1A). This is in contrast to mice infected with seasonal
(PR/8) or pandemic (CA/09) viruses where C3 levels never
increased above control. A similar trend was observed for C5a.
H5N1 VN/1194 infected mice had increased C5a levels within
3 dpi (,2-fold) increasing to .4-fold by 6 dpi as compared to
uninfected controls (Fig. 1B). There was no significant increase in
PR/8 and CA/09 virus infected mice.
To monitor C3 processing in the BAL, immunoblotting was
performed under reducing conditions with a polyclonal anti-C3
antibody and protein loading determined by anti-actin antibody
([21]; Fig. 1C and Fig. S1). To quantitate differences, densitometry
was performed as compared to actin and results expressed as the
fold change over uninfected controls (Fig. 1D). The anti-C3
antibody detected 2 major proteins at ,65 and ,43 kDa, which
are likely bC3 and iC3b respectively. Mouse/rat C3 byproducts
run at slightly different sizes than human [22,23]. Similar to the
C3 ELISA data, H5N1 VN/1194 infected animals had increased
levels of both bC3 and iC3b in BAL at 3 and 6 dpi (Fig. 1C and
D). Mice infected with seasonal PR/8 virus had levels at or below
control uninfected throughout the course of infection. Comparable
results were observed with the pandemic CA/09 infected mice
except for a dramatic increase in iC3b levels at 3 dpi only. The
reason for this is unclear and no increase was seen in the ELISA.
Parallel trends were seen in BALB/c mice and mice infected with
other strains of HPAI H5N1 virus suggesting that these results are
not mouse or H5N1 VN/1194 strain specific (data not shown). In
summation, these results suggest that H5N1 virus elicits higher
levels of C5a and C3 in the lungs of infected mice.
Increased morbidity and exacerbated inflammation in
the lungs of C32/2infected mice
Increased complement has been associated with enhanced
inflammation and tissue destruction [24]. Thus, we hypothesized
that the elevated complement levels in the HPAI H5N1 infected
mice could be involved in the prominent inflammatory response
associated with these infections. To evaluate this, wild-type (WT)
and C3 knockout mice (C32/2) mice were intranasally inoculated
with PBS (control), CA/09, or H5N1 VN/1194 virus, lungs
collected at 3 (Fig. 2) and 6 dpi (Fig. 3) and histopathology
performed. To evaluate the infiltrating cell population, tissues
were stained for macrophages and neutrophils and quantitated by
ImageScope. By 3 dpi, both the CA/09 (Fig. 2H) and VN/1194
(Fig. 2N) infected C32/2mice had enhanced inflammation as
compared to WT (Fig. 2G and 2M) with the VN/1194 being more
severe. In the CA/09 infected C32/2mice, this was associated
with increased numbers of macrophages (Fig. 2J) as compared to
WT (Fig. 2I). Quantitative analysis showed an increase from
,17% positive cells in WT to ,34% in C32/2mice (Fig. 4A).
There was no change in neutrophil levels (Fig. 2K and L). Tissues
from uninfected control animals are shown in Fig. 2A–F.
Although inflammation was more apparent in the H5N1 VN/
1194 infected C32/2mice, this was not associated with significant
differences in the numbers of macrophage or neutrophil as
compared to WT (Fig. 2O–R). Both strains of mice had minimal
numbers of macrophages and ,10% staining for neutrophils
(Fig. 4A). These numbers were significantly lower than those seen
in CA/09 infected animals; regardless of the strain.
By 6 dpi, there was increased inflammation in the C32/2mice
as compared to WT; including uninfected C32/2mice (Fig. 3B
versus 3A). The CA/09 infected WT mice exhibited mild
bronchiolitis and perivasculitis (Fig. 3G) comprised of both
macrophages (Fig. 3I) and neutrophils (Fig. 3K). In contrast, the
CA/09 infected C32/2mice had moderate to moderately severe
bronchitis, bronchiolitis, and vasculitis, with perivasculitis, consol-
idation (Fig. 3H) and increased numbers of macrophages (Fig. 3J).
Quantitative analysis showed an increase from ,17% positive cells
in WT to ,42% in C32/2mice (Fig. 4A). What was most evident
was the decrease in neutrophil numbers in the C32/2mice from
,22% in WT to ,10% in C32/2mice (Fig. 4A).
Although the pathologic changes were similar, the most
prominent difference between VN/1194 infected WT and C32/2
mice was the extent of the damage. VN/1194 infected WT mice
had increased inflammation and damage in localized areas of the
lung, primarily at the edges of the bronchi (Fig. 3M). In contrast,
infected C32/2
mice displayed a more diffuse, non-focal
pathology with increased interstitial involvement (Fig. 3N). This
was associated with an increase in neutrophils (Fig. 3R versus 3Q).
Neutrophillevels increased from ,25% in WT to ,40% in C32/2
mice (Fig. 4A). To determine if the changes in inflammation played
a role in pathogenesis, infected mice were monitored for weight loss
(Fig. 4B). By 8 dpi, the influenza infected C32/2mice had lost
.25% of their day 0 weights (CA/09 p,0.009, VN/1194 p,0.05)
and had to be humanely euthanized. In contrast, the infected WT
mice never lost more than 8% of their day 0 weights (CA/09
p,0.005, VN/1194 p,0.0001).
Delayed viral clearance in C32/2infected mice
Finally to better understand the mechanism of C3-mediated
protection, viral titers were measured in the lungs on days 3, 6,
and 8 pi (Fig. 5). In CA/09-infected WT mice, lung titers were
highest 3 dpi (106.5TCID50/ml) and decreased to approximately
103.5TCID50/ml by 8 dpi (Fig. 5). In contrast, the C32/2infected
mice had significantly higher viral titers at days 6 (p,0.04) and
8 pi (p,0.03) suggesting a delay in viral clearance. Similar trends
were observed in the VN/1194-infected C32/2mice, although
these differences were not significant until 8 dpi when the viral
titers in WT mice were 103.5TCID50/ml as compared to 105.75
TCID50/ml (p,0.02) in C32/2infected mice. Overall, these
studies suggest that C3 is an important host response during
influenza infection; potentially by aiding in viral clearance and
regulating lung inflammation.
Discussion
Here, we present data that the complement component C3 is
required for protection from pandemic 2009 H1N1 and HPAI
H5N1 influenza virus infections by aiding in viral clearance and
regulating lung inflammation. Our work complements previous
studies demonstrating a protective role for C3 against mouse-
adapted strains of influenza virus [25,26,27]. These studies
demonstrated that complement C3 was important for T-cell
C3 Protects during Influenza Infection
PLoS ONE | www.plosone.org2 March 2011 | Volume 6 | Issue 3 | e17377

Page 3

Figure 1. Complement levels in BAL during influenza infection. On days 1, 3, and 6 post-infection, BAL was collected from mice inoculated
with PBS (control) or infected with PR/8, CA/09, or VN/1194 influenza virus and analyzed for complement C3 (A) or C5a levels (B) by sandwich ELISA.
Error bars represent standard deviation. (C) C3 processing in the BAL of uninfected and influenza-infected mice at different times post-infection were
analyzed by western blot analysis and results quantitated by densitometry (D). Results are representative of 2 separate experiments.
doi:10.1371/journal.pone.0017377.g001
C3 Protects during Influenza Infection
PLoS ONE | www.plosone.org3March 2011 | Volume 6 | Issue 3 | e17377

Page 4

priming and migration to the lung [27] and promoting the
expansion of CD8+ and CD4+ T cells during infection [28]. The
lack of T cell priming, expansion and migration in the C32/2
mice could explain the delayed viral clearance.
The complement system is a major component of innate
immunity and consists of both soluble factors and cell surface
receptors that interact to sense and respond to invading pathogens.
Three general activation pathways, referred to as the classical,
Figure 2. Increased inflammation and innate immune cells of influenza infected C32 2/ /2 2mice at 3 dpi. At 3 days post-infection, lung
tissue was collected, formalin-fixed, and paraffin-embedded. Sections (4 mm thick) were stained with hematoxylin and eosin (A, B, G, H, M, N) or for
macrophages (C, D, I, J, O, P), or neutrophils (E, F, K, L, Q, R). Representative pictures of each group are shown at 206magnification.
doi:10.1371/journal.pone.0017377.g002
C3 Protects during Influenza Infection
PLoS ONE | www.plosone.org4March 2011 | Volume 6 | Issue 3 | e17377

Page 5

alternative, and lectin pathways, converge on C3 (composed of an
alpha and beta chain), the central component of the complement
system. Formation of C3 convertases leads to cleavage of C3 to its
active fragments, the anaphylatoxin C3a and the opsonin C3b.
This cleavage event exposes a reactive thioester that allows
covalent attachment of C3b to target surfaces. C3b can be further
cleaved into the signaling fragments iC3b, C3dg, and C3d which
regulate phagocytosis and a variety of other immune cell effector
functions. In addition, C3b binds the C3 convertases resulting in a
substrate change from C3 to C5, which is cleaved to the
anaphylatoxin C5a and the initiator of the membrane attack
complex C5b.
Figure 3. Increased inflammation and innate immune cells of influenza infected C32 2/ /2 2mice at 6 dpi. At 6 days post-infection, lung
tissue was collected, formalin-fixed, and paraffin-embedded. Sections (4 mm thick) were stained with hematoxylin and eosin (A, B, G, H, M, N) or for
macrophages (C, D, I, J, O, P), or neutrophils (E, F, K, L, Q, R). Representative pictures of each group are shown at 206magnification.
doi:10.1371/journal.pone.0017377.g003
C3 Protects during Influenza Infection
PLoS ONE | www.plosone.org5 March 2011 | Volume 6 | Issue 3 | e17377