The role of Fc–FcγR interactions in IgG-mediated microbial neutralization

IgG antibodies engaged in immune complexes or antibodies coating the surface of opsonized cells or microbes mediate downstream effector functions by binding to either type I or type II FcγRs (Pincetic et al., 2014). Type I FcγRs include the canonical FcγRs for IgG and are members of the immunoglobulin superfamily. Type II FcγRs include CD209 (DC-SIGN in humans and SIGN-R1 in mice) and CD23 and fall into the C-type lectin receptor family. The conformational state of the IgG Fc domain, physiologically regulated by the precise composition of the complex biantennary N-linked glycan attached to Asn297, determines whether type I or type II FcγRs are engaged by an IgG Fc. IgG Fc domains with a terminal sialic acid attached to the core heptasaccharide conjugated to the amino acid backbone of the IgG Fc domain at Asn297 adopt a flexible, “closed” conformational state, thereby permitting binding to type II FcγRs while engagement of type I FcγRs is inhibited. In contrast, nonsialylated IgG Fcs adopt an “open” conformation that suppresses engagement of type II FcγRs and promotes interactions with type I FcγR members (Sondermann et al., 2013; Ahmed et al., 2014; Pincetic et al., 2014).

Several previous studies clearly demonstrated that type II FcγR engagement by sialylated IgG Fcs results in active suppression of antibody-mediated and T cell–mediated inflammation (Anthony and Ravetch, 2010; Anthony et al., 2012; Schwab and Nimmerjahn, 2013; Fiebiger et al., 2015). In contrast, engagement of type I FcγRs results in an array of pleiotropic proinflammatory and immunomodulatory consequences. FcγR cross-linking triggers antibody-dependent cellular cytotoxicity and/or phagocytosis (ADCC/ADCP), in which IgG bridges target cells or microbes and FcγR-expressing effector cells to mediate cytotoxicity or phagocytosis (Nimmerjahn and Ravetch, 2008). Other FcγR-mediated effector functions include maturation and activation of APCs, including DCs, enhanced antigen uptake and presentation by APCs, cellular activation and release of cytokines and chemokines by innate effector cells, regulation of affinity maturation of B cells in the germinal center by setting thresholds for B cell activation, and plasma cell survival and regulation of antibody production (see Fig. 2; Pincetic et al., 2014).

The type I FcγR family is a group of structurally and functionally related receptors that belong to the immunoglobulin superfamily and share highly conserved intracellular signaling components (Fig. 1 A). FcγRs are broadly classified into either activating or inhibitory FcγRs based on the presence of intracellular immunoreceptor tyrosine-based activation motifs (ITAMs) or immunoreceptor tyrosine-based inhibitory motifs (ITIMs), respectively, that have the capacity to transduce immunostimulatory or immunosuppressive signals after receptor cross-linking by IgG immune complexes (Nimmerjahn and Ravetch, 2008). For humans, activating FcγRs include FcγRI, FcγRIIa, FcγRIIc, and FcγRIIIa, whereas FcγRIIb is the sole inhibitory FcγR. FcγRIIIb, a GPI-linked receptor exclusively expressed at high levels on neutrophils, functions in concert with the activation receptor FcγRIIa to mediate cellular responses. Both activating and inhibitory FcγRs are expressed by most cells of the innate immune system. Notable exceptions include NK cells, which express only activating FcγRs (FcγRIIIa in humans), and B cells, which express only the inhibitory FcγRIIb. With the exception of FcγRI, the FcγRs are of low affinity for IgG and thus do not engage monomeric IgG at physiological conditions. These receptors must engage multimeric immune complexes or IgG-coated targets to trigger receptor cross-linking and subsequent cellular responses. The cellular outcome of IgG interactions with FcγRs is governed by the affinity of the Fc for the specific FcγR and the expression pattern of those receptors on the effector cells. Therefore, the outcome of IgG-mediated inflammation and immunity is largely determined by balancing activating or inhibitory signals transduced by activating or inhibitory FcγRs, respectively (Nimmerjahn and Ravetch, 2006).

Because most effector leukocytes express both activating and inhibitory FcγRs, the relative ratio of the binding affinities of these receptors for a specific IgG Fc determines the outcome of the IgG–FcγR interaction. These binding affinities are regulated by the amino acid sequences of the different IgG Fc subclasses and the IgG Fc’s N-linked glycan structure. Thus, the IgG Fc composition can dramatically influence the in vivo outcome of FcγR engagement by immune complexes or opsonized cells by directing the effector cell into either a pro- or antiinflammatory state. For example, the mouse (m)IgG2a subclass engages the activating mFcγRIV (orthologue of human FcγRIIIa) with 100-fold greater affinity compared with the inhibitory mFcγRIIb receptor, whereas mIgG1 preferentially engages the inhibitory mFcγRIIb receptor (Nimmerjahn and Ravetch, 2005). Thus, mIgG2a antibodies have an activating/inhibitory (A/I) ratio (ratio of the Fc’s affinity for the relevant activating FcγR versus the affinity for the inhibitory FcγR) of ~70. In contrast, mIgG1 antibodies have an A/I ratio of 0.1. The in vivo activities of mIgG2a and mIgG1 antibodies correspond to their A/I ratios, and cytotoxic antibodies of the mIgG2a subclass potently kill target cells in vivo, whereas mIgG1 antibodies have minimal activity (Nimmerjahn and Ravetch, 2005). Similarly in humans, human (h)IgG2 and hIgG4 poorly interact with human FcγRs, whereas hIgG1 and hIgG3 interact more strongly (Nimmerjahn and Ravetch, 2007).

Apart from the IgG subclass and glycan composition, additional determinants for the regulation of IgG–FcγR interactions exist, including IgG immune complex size and the localization of FcγRs within membrane microdomains that favor interactions with intracellular signaling proteins (Floto et al., 2005; Bournazos et al., 2009a; Lux et al., 2013). Furthermore, allelic and copy number variants of the human FcγR genes affect binding affinities to human IgG, as well as receptor activity and expression (Bournazos et al., 2009b). These variants modulate in vivo antibody effector function, as cancer patients carrying the high-affinity allelic variants of FcγRIIIa and FcγRIIa (V158 and H131, respectively) demonstrate greater responses to antitumor mAb therapies (Cartron et al., 2002; Weng and Levy, 2003; Musolino et al., 2008). Indeed, selective engagement of certain FcγR classes by IgG has been shown to determine the outcome of passive antibody treatment in vivo, suggesting that Fc-mediated pathways play a key role in modulating the effector activities of an antibody. For example, enhanced engagement of activating FcγRs by antitumor mAbs greatly improves their cytotoxic activity (Clynes et al., 2000; Smith et al., 2012; Goede et al., 2014), whereas selective binding to the inhibitory FcγRIIb is necessary for optimal activity of agonistic anti-CD40 mAbs (Li and Ravetch, 2011). In the context of infectious diseases, Fc–FcγR interactions readily amplify the in vivo protective activity of neutralizing antibodies through opsonization and clearance of microbes and their toxins, as well as by cytotoxic killing of infected cells (Fig. 2; Bournazos et al., 2014a,c; DiLillo et al., 2014). Additionally, FcγR engagement by immune complexes may induce pleiotropic proinflammatory and immunomodulatory functions with the potential to initiate sustained antimicrobial immune responses (Fig. 2).

With the exception of the high-affinity FcγRI, all members of the FcγR family are mapped at a common FcγR locus (located at 1q23 for Homo sapiens) that is highly conserved among species. Indeed, the ancestral FcγR locus that encompasses the genes coding for the low-affinity FcγRs (FcγRIIa/b/c and FcγRIIIa/b) shares a common genomic organization that can be traced back early in evolutionary history and presents substantial similarities among different mammalian species (Fig. 1 B). Furthermore, the capacity of other mammalian FcγRs to interact with human IgG molecules is indicative of the high degree of FcγR gene similarity (Fig. 1 C).

Despite such similarities in the FcγR structure and function, there are several fundamental differences among mammalian species stemming mainly from the unique characteristics of human FcγRs. For example, the human FcγR locus exhibits a unique organization pattern that is the result of a nonhomologous recombination event of the ancestral FcγR locus that gave rise to human FCGR2C and FCGR3B genes encoding for FcγRIIc and FcγRIIIb, respectively (Qiu et al., 1990). Likewise, FcγR expression patterns on different leukocyte types differ substantially between different species. For example, nonhuman primate neutrophils express FcγRI, FcγRIIa, and FcγRIIb, whereas in humans, neutrophils express FcγRIIa, FcγRIIb, and FcγRIIIb (unpublished data). Likewise, mouse monocyte-derived DCs express FcγRI, FcγRIIb, FcγRIII, and FcγRIV, contrary to their human counterparts that express only FcγRIIa and FcγRIIb. Similar interspecies differences have been observed in other cell types, including NK cells, macrophages, and other effector leukocytes (Smith et al., 2012).

Such differences in the FcγR structure and expression pattern among different mammalian species commonly used as infectious disease models do not allow for the precise evaluation of the in vivo Fc effector activities of neutralizing antibodies. Careful characterization of the FcγR expression pattern and their capacity to interact with human IgG molecules is therefore required before the assessment of the in vivo activity of antibodies in other experimental species. Additionally, recently developed mouse strains humanized for FcγR genes could be used to accurately assess the contribution of Fc effector pathways in the in vivo protective activity of neutralizing antibodies (Smith et al., 2012; Bournazos et al., 2014b). Such strains encompass specific deletions of all mouse FcγR genes, with human FcγRs introduced as BAC transgenes with expression under the control of their endogenous promoters and regulatory elements (Smith et al., 2012). This approach ensures proper human FcγR expression patterns and cell type specificity. Indeed, FcγR-humanized mice faithfully recapitulate the unique characteristics of human FcγR functional and structural diversity, with their FcγR expression profiles among the various leukocyte types mirroring precisely those observed in humans (Smith et al., 2012). FcγR-humanized mice have been successfully used to assess the contribution of Fc effector function of passively administered antibodies to their in vivo protective activity in models of influenza, HIV-1, and anthrax infection (Bournazos et al., 2014a,c; DiLillo et al., 2014). Additionally, the FcγR-humanized mouse model has become a useful preclinical evaluation platform for comparing the in vivo activity of Fc domain–engineered variants of antibodies with enhanced capacity to engage particular human FcγR classes (Smith et al., 2012; Bournazos et al., 2014a; DiLillo and Ravetch, 2015).

The traditional notion that antibody-mediated neutralization of a microbe or toxin is the result of Fab–antigen interactions has relied on in vitro assays, which fail to reproduce the diversity of receptors, effector cells, and microenvironments that contribute to protection during in vivo conditions. Such in vitro assays provide limited information on the precise Fc effector mechanisms that participate in vivo to mediate host protection against microbial or toxin challenge. By focusing on in vivo studies in the following sections, using examples from influenza, HIV-1, and other infectious diseases, we will review recent data demonstrating that Fc–FcγR interactions are crucial for mediating optimal in vivo protection by neutralizing mAbs. In view of this evidence, it is now clear that ideal antibody-based therapeutics need to combine potent and broad neutralization capabilities with optimal Fc effector function to mediate maximal in vivo protective effects.

Antibodies that neutralize influenza virus traditionally have been thought to function in one of three ways, based on in vitro neutralization experiments. First, anti-hemagglutinin (HA) head antibodies disrupt virus attachment to sialic acids on the surface of target cells to block viral entry (Wiley et al., 1981; Knossow et al., 2002). Second, some anti-HA head and anti-neuraminidase (NA) antibodies may function to prevent viral budding and the release of progeny. Finally, anti-HA stalk antibodies can block viral fusion with the target cell (Ekiert et al., 2009). However, recent studies have more carefully clarified the in vivo mechanisms of viral neutralization and revealed that Fc–FcγR interactions are required for effective anti-influenza mAb-mediated protection from lethal influenza virus infection in vivo. For example, vaccines targeting the ectodomain of matrix protein 2 (M2e) of influenza A virus induce anti-M2e antibodies. M2e is not expressed on the viral particle, and anti-M2e antibodies do not neutralize virus in vitro. However, anti-M2e antibodies are protective against influenza infection in mice; this protection is mediated by alveolar macrophages expressing activating FcγRs (El Bakkouri et al., 2011; Lee et al., 2014), which likely mediate cytotoxicity or phagocytosis of antibody-coated infected cells. Furthermore, anti-M2e antibodies of the mIgG2a subclass, which preferentially engage mouse-activating FcγRs, mediate in vivo protection, with no activity observed for the mIgG1 subclass, which engages the inhibitory FcγRIIb (Schmitz et al., 2012). Furthermore, FcγR expression by macrophages was required for protection mediated by passive serum transfers from mice immunized with H1N1 virus (Huber et al., 2001). Collectively, these results suggest that infection of cells with influenza virus leads to cell surface expression of viral proteins, such as M2e, that serve as targets for antibodies that trigger FcγR-mediated clearance of infected cells.

HA is the most abundant glycoprotein expressed on the influenza viral envelope, as well as on influenza-infected cells. Influenza HA is composed of two major domains: the antigenically variable globular head and the relatively conserved stalk. Because of its immunodominance and high level of antigenic variation, the majority of the immune response is directed at the HA head domain, and most antibodies reactive with the HA head domain only neutralize a single influenza strain. In contrast, antibodies targeting the conserved HA stalk domain have been found to be broadly neutralizing, reactive with a wide range of influenza subtypes. Recent studies have mechanistically assessed the requirements for Fc–FcγR interactions during in vivo protection mediated by either anti-HA stalk broadly neutralizing antibodies or strain-specific anti-HA head antibodies (Corti et al., 2011; DiLillo et al., 2014). Although unnecessary for in vitro neutralization, interactions with activating FcγRs were required for in vivo protection mediated by a panel of broadly neutralizing anti-HA stalk antibodies after lethal influenza infection (DiLillo et al., 2014). This FcγR-mediated protection required interactions with the IgG Fc domain after viral entry into target cells, suggesting that HA-expressing, antibody-opsonized infected cells were targeted for clearance by ADCC/ADCP mechanisms, consistent with an earlier study (Corti et al., 2011). Interestingly, strain-specific anti-HA head mAbs protected mice from lethal infection independently of Fc–FcγR interactions, a property of these mAbs that correlates with their inability to interact with FcγRs and trigger ADCC during in vitro assays (DiLillo et al., 2014). In contrast, anti-HA stalk broadly neutralizing mAbs potently interact with FcγRs and trigger ADCC in vitro. Thus, at least two classes of neutralizing antibodies are co-selected in response to viral challenge to provide multiple pathways for antiviral protection: those that require the FcγR effector system and induce clearance of infected cells and those that do not. Recent results have shown that all antibodies with a wide breadth of reactivity (broadly neutralizing antibodies) analyzed thus far demonstrate a dependence on Fc–FcγR interactions to mediate protection in vivo (unpublished data), suggesting that FcγR engagement is a common mechanism of action for all anti-HA broadly neutralizing antibodies.

Because of their dependence on Fc–FcγR interactions to mediate protection in vivo, a recent study has addressed whether a broadly neutralizing hIgG1 antibody could be Fc engineered for augmented protection in vivo in the context of the human FcγR system in FcγR-humanized mice (DiLillo et al., 2014). By introducing point mutations in the hIgG1 Fc domain of an anti-HA stalk broadly neutralizing antibody to selectively enhance interactions with activating human FcγRs, survival was enhanced at least twofold, and weight loss was decreased after lethal influenza challenge compared with anti-HA stalk mAb with a wild-type human IgG1 Fc domain. Thus, optimizing interactions between neutralizing antibodies and the FcγR system is an attractive approach to enhancing the efficacy of passively administered anti-influenza antibody therapeutics.

For years, the development and use of potent neutralizing antibodies against HIV have been a challenge for efforts to control HIV-1 infection. Indeed, HIV-1 presents several unique structural and functional determinants that conventional antibody strategies must overcome to block viral entry to target cells. For example, the envelope glycoprotein of HIV-1 (env) is present at remarkably low density on the virus surface, thereby precluding high-avidity concurrent interactions of both IgG Fab arms (Klein and Bjorkman, 2010). In addition, its highly glycosylated structure forms a glycan shield that masks sites of potential vulnerability, further contributing to the limited immunogenicity of the HIV-1 env (Burton et al., 2012; Moore et al., 2012). HIV-1 env also exhibits substantial diversity (up to 35% amino acid sequence divergence), attributed to the high virus mutation rate and its capacity to remain latent for several years, even in chronically treated patients, as well as the chronicity of infection, during which antibody responses exert selection pressure on the virus (Korber et al., 2001; Gaschen et al., 2002). Although these immune evasion mechanisms greatly compromise the host’s capacity to mount potent broadly neutralizing antibody responses, early clinical studies have identified a small fraction of infected individuals that develop affinity-matured antibodies with broad activity against diverse, cross-clade virus isolates (Simek et al., 2009). These individuals, commonly referred to as “elite neutralizers,” have inspired the development and use of broadly neutralizing antibodies as a therapeutic modality to prevent and control HIV-1 infection. Thanks to recent advancements in B cell cloning techniques, the systematic isolation and characterization of broadly neutralizing antibodies from elite neutralizers has become a reality, leading to the development over the past 5 yr of several dozen anti–HIV-1 env mAbs with potent and broad activity (reviewed in Klein et al. [2013]). Passive administration of these broadly neutralizing mAbs has been shown to confer sterilizing immunity against SHIV challenge in macaques and HIV-1 infection in humanized mouse models (Mascola et al., 2000; Hessell et al., 2007; Balazs et al., 2012). More importantly, effective control of virus replication in HIV-1–infected humanized mice and in SHIV-infected nonhuman primates by these broadly neutralizing mAbs clearly suggests their potential clinical use to control HIV-1 infection in humans (Klein et al., 2012; Barouch et al., 2013; Horwitz et al., 2013; Shingai et al., 2013). Indeed, administration of the broadly neutralizing anti-CD4bs mAb, 3BNC117, in chronically infected HIV-1 patients successfully suppressed viremia for several days after mAb infusion (Caskey et al., 2015).

Previous studies have suggested a role for FcγRs during HIV-1 infection, as indicated by the association of genetic variants of FcγRIIa with the clinical progression of AIDS (Forthal et al., 2007). Given broadly neutralizing anti-HIV env mAbs were isolated and characterized only recently, the precise Fc effector mechanisms that contribute to their in vivo activity have also only recently been elucidated. Despite the lack of an in vivo HIV-1 infection model that faithfully recapitulates the complete virus infection cycle, human effector cells, and FcγR diversity, the role of Fc–FcγR interactions during in vivo mAb activity has been assessed using complementary in vivo models (Klein et al., 2012; Pietzsch et al., 2012; Smith et al., 2012). In particular, using a mouse model for HIV-1 entry, comparison of the in vivo protective activity of a panel of broadly neutralizing mAbs targeting different HIV-1 env epitopes revealed increased activity for the mIgG2a subclass compared with mIgG1 (Bournazos et al., 2014c). This finding is consistent with the capacity of the IgG2a subclass to interact preferentially with activating mFcγRs, thereby inducing clearance of mAb-opsonized viral particles through activating FcγR engagement (Nimmerjahn and Ravetch, 2005). Similar effects have been observed in studies using HIV-1–infected humanized mice in models of mAb-mediated post-exposure prophylaxis and therapy (Bournazos et al., 2014c; Halper-Stromberg et al., 2014). Assessment of the in vivo protective activity of Fc variants of anti–HIV-1 mAbs with differential FcγR binding capacities revealed that administration of mAb variants optimized for enhanced affinity to activating FcγRs resulted in substantially improved and durable suppression of viremia in humanized mice with established HIV-1 infection (Bournazos et al., 2014c). Likewise, in a model of post-exposure prophylaxis using HIV-1–infected humanized mice, mAb variants with diminished FcγR binding capacity failed to suppress viremia, contrary to wild-type hIgG1 mAbs (Halper-Stromberg et al., 2014). Similar effects have been previously observed in a nonhuman primate preexposure prophylaxis model, in which Fc domain variants with minimal capacity for FcγR engagement provided no protection against SHIV challenge (Hessell et al., 2007). These findings clearly suggest a key role for FcγR pathways during the in vivo protective activity of anti–HIV-1 mAbs and indicate that selective engagement of activating FcγRs greatly enhances their in vivo activity, guiding the development of a new generation of broadly neutralizing mAbs optimized for Fc effector function.

Similar to the influenza and HIV-1 infection systems, the contribution of FcγR-mediated interactions to the in vivo protective activity of mAbs extends to other viral, fungal, and bacterial diseases. For example, the generation and evaluation of a nonfucosylated glycovariant of the FDA-approved anti-RSV IgG, palivizumab, resulted in significantly improved in vivo protective activity compared with the parental mAb (Hiatt et al., 2014). Furthermore, mAbs against the pneumococcal capsular polysaccharide have been shown to require Fc–FcγR interactions to protect mice from pneumococcal pneumonia after challenge with Streptococcus pneumonia (Weber et al., 2012). Common to divergent infectious disease models, several studies have noted differential in vivo activity of mouse IgG subclass variants of protective mAbs, thereby providing evidence for a crucial role for FcγR effector pathways. Despite having comparable activities in vitro, the in vivo protective activity of these mAbs correlated precisely with their capacity to engage activating FcγRs; mIgG2a subclass variants always exhibited improved efficacy in vivo compared with mIgG1. This phenomenon has been observed for mAbs protective against bacterial (Staphylococcus aureus) and fungal (Cryptococcus neoformans) infections, as well as the anthrax toxin (Sanford et al., 1990; Schlageter and Kozel, 1990; Abboud et al., 2010; Varshney et al., 2014). In all cases, an increase in the capacity of a mAb to interact with activating FcγRs (increased A/I ratio) was accompanied by improved in vivo activity without any impact on their Fab-mediated functions. Based on these findings, efforts to generate mAbs with enhanced in vivo activity led to the development of Fc domain–engineered variants with selective binding to activating FcγRs (Bournazos et al., 2014a). For example, Fc optimization of an anti-anthrax lethal toxin mAb to selectively engage activating human FcγRs substantially improved its protective activity when tested in FcγR-humanized mice challenged with Bacillus anthracis (Bournazos et al., 2014a). Collectively, these findings provide substantial evidence for the contribution of Fc effector function to the in vivo activities of protective mAbs in several infectious disease model systems.

Antibodies have become major therapeutic tools with potential treatment applications in a variety of disease settings, especially infectious diseases. Ideal therapeutic antibodies to treat infection should be selected based not only on their antigen-binding characteristics (i.e., specificity, affinity, and broad binding capability) and neutralization ability (based on in vitro assays), but also for optimal engagement of the appropriate FcγR family members and recruitment of effector cells. Such Fc optimization can be achieved by altering the Fc glycan or introducing Fc point mutations that augment the Fc domain’s affinity for the appropriate FcγR, as is currently being tested in cancer therapeutics (Natsume et al., 2009). Thus, combining superior Fc-independent neutralization and Fc engineering for optimal Fc-dependent effector function will produce the most effective therapeutic antimicrobial antibodies to treat infectious disease in patients.

We would like to thank and acknowledge the Bill and Melinda Gates Foundation for its partial support in our research endeavors, specifically under award OPP1123905, “Characterization of guinea pig FcγR affinity and specificity for human IgG1 Fc glycovariants.”

The authors declare no competing financial interests.