|Home | About | Journals | Submit | Contact Us | Français|
Although the interactions of complement and viruses have been widely studied, the function of C5 and the membrane attack complex in the context of viral infection or antibody-mediated neutralization remains controversial. Using C5-depleted or deficient human or mouse sera we show that C5 does not contribute to the antibody-dependent or –independent neutralization of West Nile virus (WNV) in cell culture. Consistent with this, C5 neither contributed to protection against WNV pathogenesis nor augmented the neutralizing efficacy of complement-fixing anti-WNV neutralizing antibodies in mice. Although previous studies established that activation of the classical, lectin, and alternative complement pathways restricts WNV infection, our results show little effect of C5 and by inference the terminal lytic complement components. Overall, these results enhance our mechanistic understanding of how complement controls flavivirus infections.
The complement system is a group of ~30 serum and cell surface proteins that participate in recognition, immune system priming, immune complex formation, and clearance of altered-self ligands (Blue, Spiller, and Blackbourn, 2004). Complement activation can result in formation of classical/lectin and alternative pathway C5 convertases (Muller-Eberhard, 1986) that cleaves C5 and promotes assembly of the C5b-9 membrane attack complex (MAC) lytic pore (Tschopp, Masson, and Stanley, 1986). Assembly of the MAC and subsequent lysis of cellular or pathogen targets represents a direct effector function of the complement system. However, the role of MAC assembly in direct and/or antibody-mediated neutralization of viruses remains unclear.
Cell culture studies have shown that serum complement directly inactivates infectious HIV, West Nile virus (WNV), and influenza viruses (Jayasekera, Moseman, and Carroll, 2007; Mehlhop et al., 2005; Sullivan, Takefman, and Spear, 1998) and augments antibody-mediated neutralization of influenza (Beebe, Schreiber, and Cooper, 1983; Mozdzanowska et al., 2006), HIV (Aasa-Chapman et al., 2005; Posner et al., 1992; Spear et al., 1993; Verity et al., 2006), respiratory syncytial (Baughman et al., 1968; Yoder et al., 2004), varicella zoster (Beebe and Cooper, 1981; Grose, Edmond, and Brunell, 1979; Schmidt and Lennette, 1975), Epstein-Barr (Nemerow, Jensen, and Cooper, 1982; Sairenji, Sullivan, and Humphreys, 1984), herpes simplex (Lerner, Shippey, and Crane, 1974; Snyder, Myrup, and Dutta, 1981; Wallis and Melnick, 1971) and vaccinia (Benhnia et al., 2009) viruses. Despite the extensive in vitro characterization of the effects of complement on viral infection, the mechanism of complement-enhanced neutralization remains controversial. In some studies only the first four components of complement were necessary to increase the potency of antibody-mediated viral neutralization (Beebe, Schreiber, and Cooper, 1983; Daniels et al., 1969; Johnson, Capraro, and Parks, 2008; Leddy, Simons, and Douglas, 1977; Linscott and Levinson, 1969; Sullivan, Takefman, and Spear, 1998). Serumen-hanced antibody neutralization of influenza virus was found to depend solely on C1q, and correlate with IgG-subclasses that bound C1q avidly (Feng, Mozdzanowska, and Gerhard, 2002; Mozdzanowska et al., 2006). In contrast, antibody-dependent MAC-mediated virolysis of HIV and parainfluenza virus has been reported (Spear et al., 1990; Spear et al., 1993; Vasantha et al., 1988). Although it has been suggested to contribute to antibody potency, in vivo studies have not confirmed a role for C5 and the terminal complement components in augmenting the efficacy of antiviral antibody (Baldridge and Buchmeier, 1992; Mathews, Roehrig, and Trent, 1985; McKendall, 1985). Nonetheless, in the absence of antiviral antibody, C5 may restrict infection of influenza, Sindbis, and cowpox viruses (Hicks et al., 1978; Hirsch, Griffin, and Winkelstein, 1980; Miller et al., 1995) by undefined mechanisms.
WNV is a single stranded positive sense enveloped RNA arthropod-borne virus of the Flaviviridae family and is closely related to several other human pathogens including dengue, yellow fever, Japanese encephalitis, and tick-borne encephalitis viruses. Mature flaviviruses have small ~500 Å icosahedral virions, with little, if any exposed lipid as the surface is covered by a shell of tightly-packed envelope (E and prM/M) proteins (Mukhopadhyay, Kuhn, and Rossmann, 2005). Previous studies have reported that serum complement directly neutralizes infectious WNV in vitro (Mehlhop et al., 2005), and that classical, lectin, and alternative pathway complement components are required to restrict WNV pathogenesis in mice (Mehlhop and Diamond, 2006). However, the mechanism of protection against WNV remains incompletely understood although complement-dependent priming of adaptive immunity clearly contributes to control. Here, we used acquired and genetic deficiencies of C5 to evaluate the role of the terminal complement components in antibody-dependent and independent virus neutralization and restriction of pathogenesis. We find that complement protective mechanisms against WNV occur primarily through C5-independent mechanisms.
The lineage 1 New York WNV strain (WNV-NY) (3000.0259) was isolated in 2000 (Ebel et al., 2001) and passaged once in C6/36 Aedes albopictus cells to generate an experimental stock. BHK21-15 cells were used to measure viral titer of infected cells or tissues by plaque assay (Diamond et al., 2003). Raji cells stably expressing DC-SIGNR were maintained as described (Pierson et al., 2007). Infections were performed with WNV RVP produced using a previously described complementation strategy (Pierson et al., 2006).
Blood was collected by axillary venupuncture into serum separator tubes (Sarsted) from eight to twelve week-old male wild type and C3−/− C57BL/6 mice or wild type and C5−/− B10.D2 mice that were obtained commercially (Jackson Laboratories and Taconic, respectively) and from colleagues (C3−/−, H. Molina, St Louis, MO). Blood was clotted on ice and serum was pooled, aliquotted, and frozen at −80°C until use. Heat-inactivation of serum was achieved after incubation at 56°C for 30 minutes.
The neutralizing activity of serum complement was determined using a modified plaque reduction assay on BHK21-15 cells by mixing wild type or complement-deficient mouse sera (10% final serum concentration) with 6 × 101 PFU of WNV in gelatin veronal buffer containing Ca2+ and Mg2+ (GVB++; CompTech). Following incubation for one hour at 37°C, virus was added to BHK21-15 cell monolayers and incubated for one additional hour at 37°C. BHK21-15 cells were then washed with Dulbecco’s Modified Eagle Media, overlaid with 1% agarose in Minimal Essential Media, and cultured for three days at 37°C. Plaques were counted following formaldehyde fixation and staining of wells with 1% (w/v) crystal violet in a 20% ethanol solution. The effect of complement on antibody (mouse E16-IgG2b or humanized E16-IgG3) neutralization was evaluated using WNV RVP and Raji-SIGN-R cells in the presence or absence of 5% mouse or human serum using a high-throughput flow cytometry-based assay (Pierson et al., 2007). The human IgG subclass switch variant of mouse E16 (Oliphant et al., 2005) was generated as previously described (Mehlhop et al., 2007).
Sheep erythrocytes were coated with goat anti-sheep erythrocyte polyclonal antibody. Sensitized erythrocytes were exposed to C5-depleted human serum supplemented with excess normal mouse serum in the presence or absence of serial dilutions of BB5.1 mAb or murine IgG1 isotype control. C5-dependent hemolysis was assayed by measuring the optical density (OD) values at 415 nm of supernatants after 1 h at 37°C and the 50% value of classical pathway hemolytic complement activity (CH50) as described (Morgan, 2000).
All mice were housed in a pathogen-free facility at Washington University School of Medicine. Studies were performed in compliance and with approval of the Washington University School of Medicine Animal Safety Committee. Eight week-old wild type or congenic C5−/− B10.D2 mice were used for pathogenesis studies. Four or eight days after infection, spleens and brains were removed, weighed, homogenized using a bead beater apparatus (BioSpec Products, Inc), and titrated for virus by plaque assay on BHK21-15 cells (Diamond et al., 2003). For passive transfer studies with anti-WNV mAbs, two independent models were used: (a) five to six week-old wild type or congenic C5−/− B10.D2 mice; and (b) wild type C57BL/6 mice administered 50 mg/kg BB5.1 or murine IgG1 isotype control mAb by intraperitoneal injection on day −1. Both groups of mice were passively transferred increasing doses of E16 mAb by intraperitoneal injection at day −1 and then infected via footpad with 102 PFU of WNV on day 0.
For in vitro experiments, a paired T-test was used to determine significant differences. For viral burden analysis, the Mann-Whitney test was applied to evaluate differences in titers. Kaplan-Meyer analysis of survival data was performed using the log rank test. IC50 analysis was performed by non-linear regression analysis and statistical significance was determined using analysis of variance (ANOVA) and F-tests. All data were analyzed using Prism software (GraphPadPrism4).
A previous study suggested that a heat-labile component of serum could directly neutralize WNV infection in the absence of antibody (Mehlhop et al., 2005). To evaluate specifically the role of C5 and the terminal MAC in antibody-independent serum neutralization, WNV (strain New York, 2000) was incubated with naïve serum from C5−/− and congenic C5+/+ B10.D2 mice before addition to a BHK21-15 cell monolayer. WNV infectivity decreased by 90% (n = 3, P < 0.002) in the presence of naïve serum from wild type B10.D2 or C57BL/6 mice compared to heat-inactivated (56°C for 30 minutes) serum (Fig 1A). Notably, WNV infectivity was equivalently reduced with naïve serum from C5−/− B10.D2 mice whereas no reduction was measured when C3−/− mouse serum was substituted. Thus, C5 was dispensable for direct complement (C3)-dependent neutralization of WNV in vitro.
To assess the role of C5 in augmenting neutralizing activity of antibodies, WNV reporter virus particles (RVP) expressing GFP were incubated with serum complement and increasing concentrations of E16, a strongly neutralizing anti-WNV antibody (Oliphant et al., 2005) for 1 h at 37°C prior to infection of Raji-DC-SIGNR cells (Pierson et al., 2006). Parallel studies were performed using normal human and mouse sera and a humanized IgG3 isotype switch variant (hE16-IgG3) and the parent mAb (mouse IgG2b), both of which avidly fix complement (Mehlhop et al., 2007). Cells were harvested after 48 h and assayed for infection by flow cytometry. In the presence of normal human or mouse serum, the neutralization efficacy of either humanized or murine E16 increased 27-fold (P = 0.01) and 5-fold (P = 0.02), respectively (Fig 1B and C). However, incubation with C5-deficient human or mouse sera did not decrease this augmented neutralization potency of E16 (Fig 1B and C). Importantly, the C5-dependent lytic function of serum complement from B10.D2 mice was confirmed by a standard hemolysis assay (Fig 2D). Collectively, these cell culture studies suggest that C5 has no significant role in antibody-independent or –dependent complement-mediated WNV neutralization. Thus, and similar to experiments with influenza virus (Beebe, Schreiber, and Cooper, 1983; Feng, Mozdzanowska, and Gerhard, 2002), virolysis cannot explain complement-dependent WNV neutralization in vitro.
While C5 did not affect serum-dependent neutralization of influenza virus in vitro, in mice C5 was necessary for protection from disease, as prolonged infection and enhanced mortality were observed in C5−/− mice (Hicks et al., 1978). Additionally, prior experiments had shown essential roles of all three complement activation pathways in restricting WNV pathogenesis as increased mortality was observed in C1q−/−, C3−/−, C4−/−, mannose binding lectin A−/− × C−/−, and factor B−/− mice ((Mehlhop and Diamond, 2006), and A. Fuchs, M. Vogt, E. Mehlhop, and M. Diamond, unpublished results). To determine if C5 has a protective role against WNV in vivo, C5−/− and C5+/+ congenic B10.D2 mice were infected by a subcutaneous route with 102 PFU of WNV and followed for mortality. Notably, 44.4 ± 11.7% (N = 18) of eight week-old C5+/+, and 50.0 ± 10.2% (N= 24) of C5−/− B10.D2 mice survived WNV infection (Fig 2A, P > 0.5). This data suggested that C5 had little role in the control of WNV infection. Consistent with this, viral burden analysis of tissues at day 4 in the spleen and day 8 in the brain showed no significant difference of WNV replication in the C5−/− and C5+/+ mice (Fig 2B, P ≥ 0.7).
To confirm this using an independent model, C5 was depleted from C57BL/6 mice by passively transferring 50 mg/kg of BB5.1, a neutralizing anti-C5 mAb (Frei, Lambris, and Stockinger, 1987) one day prior to infection. This dose is approximately 1,000-fold greater than the EC50 value (4.1±1.4 ng/ml) in vitro as determined by a C5-dependent hemolysis assay (Fig 2D). Using this assay, we confirmed that a single administration of 1 mg of BB5.1 reduced C5-levels to less than 20% of endogenous levels even 17 days after treatment (Fig 2E). Analogous to the data in B10.D2 mice, 50.0 ± 9.4% (N = 28) of isotype control, and 61.9 ± 7.5% (N = 42) of anti-C5 mAb treated C57BL/6 mice survived WNV infection (Fig 2C, P > 0.2). Thus, a genetic or acquired deficiency of C5 does not increase susceptibility to lethal WNV infection. As a loss of C5 abolishes both MAC formation and generation of the C5a anaphylatoxin, our studies confirm and extend previous observations that C5a-mediated inflammatory and chemotactic responses are not required for the control of WNV infection (Mehlhop and Diamond, 2006).
C5 has been proposed to promote direct virolysis of HIV and parainfluenza virus in the context of antiviral antibody (Spear et al., 1990; Spear et al., 1993; Vasantha et al., 1988). To determine if C5 effector functions that occur downstream of classical pathway activation by a neutralizing anti-WNV antibody contribute to protection in vivo, C57BL/6 mice were depleted of C5 by BB5.1 antibody treatment and administered increasing doses (0.1 1, 10, 100, and 1,000 ng) of hE16-IgG3 one day prior to infection. We selected hE16-IgG3 because it binds C1q avidly and exhibits a marked shift to lower antibody requirements for neutralization in vitro in the presence of human or mouse serum (Fig 1B, and data not shown). Notably, the IC50 of isotype-control treated and C5-depleted mice were 20.4 ± 8.4 ng (N = 13/dose) and 16.3 ± 2.2 ng (N = 12/dose) of hE16-IgG3 (Fig 3A, P > 0.5). To confirm these results, C5−/− and C5+/+ B10.D2 mice were treated with increasing doses of hE16-IgG3 one day prior to subcutaneous infection with 102 PFU of WNV. Analogously, no statistical change in hE16-IgG3 potency was observed in C5−/− and C5+/+ mice (Fig 3B, C5+/+ B10.D2 IC50 = 4.1 ± 2.8, N = 22/dose; C5−/− B10.D2 IC50 = 9.0 ± 10.8, N = 9/dose; P > 0.05). Thus, C5 does not significantly affect the efficacy of a neutralizing, complement-fixing anti-WNV antibody in vivo. These results agree with prior studies with unrelated enveloped RNA and DNA viruses (Baldridge and Buchmeier, 1992; Mathews, Roehrig, and Trent, 1985; McKendall, 1985), and suggest that complement-dependent virolysis does not contribute to antibody neutralization or protection against WNV in vivo.
Our investigations suggest that C5 and MAC-mediated virolysis does not contribute to complement-mediated neutralization and protection from WNV disease, at least in mice. Given our analogous in vitro results with human serum (see Fig 1B), we hypothesize this will be sustained in humans and other animals. One speculation as to why the MAC does not promote antibody-dependent or –independent virolysis is because the icosahedral surface of the mature WNV virion is tightly-packed by a highly-ordered array E and prM/M proteins, with little, if any exposed lipid (Mukhopadhyay et al., 2003). However, the mature virion undergoes dynamic motion at 37°C, which could transiently expose hidden epitopes including surface lipids (Lok et al., 2008). Moreover, maturation of flavivirus virions is often incomplete (Nelson et al., 2008), and partially mature infectious particles also display surface lipid (Cherrier et al., 2009; Zhang et al., 2003). Thus, it remains uncertain as to why antibody-dependent or –independent virolysis fails to occur after complement deposition. Our experiments suggest that complement-mediated neutralization of WNV is likely an opsonic phenomenon and may alter binding interactions that are essential for virus attachment and/or fusion. This idea agrees with recent studies on both paramyxovirus (Johnson, Capraro, and Parks, 2008) and influenza virus (Jayasekera, Moseman, and Carroll, 2007) that described opsonic mechanisms of complement-mediated virus neutralization.
Finally, our data also suggests that C5-dependent lysis of WNV-infected cells does not contribute significantly to protection against disease either in the presence or absence of antiviral antibodies. These results are consistent with recent studies showing that mAbs recognizing cell surface forms of NS1 on infected cells protect in vivo through Fc-γ receptor-dependent yet C1q-independent mechanisms (Chung et al., 2006b; Chung et al., 2007). Although more studies are required, the viral NS1 protein on the surface of cells may limit MAC accumulation by antagonizing alternative pathway complement deposition through its binding to the negative regulator factor H (Chung et al., 2006a).
We thank T. Pierson for critical comments on the manuscript, J. Lambris for the BB5.1 hybridoma, J. Atkinson for experimental suggestions, S. Johnson for the hE16-IgG3, and K. O’Brien for technical assistance. This work was supported by the NIH (grants U01 AI061373 and the Midwest Regional Center of Excellence for Biodefense and Emerging Infectious Diseases Research (U54AI057160)).
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.