IgG4 can induce an M2-like phenotype in human monocyte-derived macrophages through FcγRI

Abstract

Antibodies evoke cellular responses through the binding of their Fc region to Fc receptors, most of which contain immunoreceptor tyrosine-based activation motif domains and are thus considered “activating.” However, there is a growing appreciation of these receptors for their ability to deliver an inhibitory signal as well. We previously described one such phenomenon whereby interferon (IFN)γ signaling is inhibited by immune complex signaling through FcγRI. To understand the implications of this in the context of therapeutic antibodies, we assessed individual IgG subclasses to determine their ability to deliver this anti-inflammatory signal in monocyte-derived macrophages. Like IgG1, we found that IgG4 is fully capable of inhibiting IFNγ-mediated events. In addition, F(ab’)2 fragments that interfere with FcγRI signaling reversed this effect. For mAbs developed with either an IgG1 or an IgG4 constant region for indications where inflammation is undesirable, further examination of a potential Fc-dependent contribution to their mechanism of action is warranted.

Abbreviations

IC

immune complexes

ITAM

immunoreceptor tyrosine-based activation motif

ITIM

immunoreceptor tyrosine-based inhibitory motif

TLR

Toll Like Receptor

IP-10

Interferon Inducible Protein-10

Introduction

Antibodies elicit a variety of cellular responses through the binding of their Fc region to cognate receptors. Members of the Fcγ receptor family (FcγRs) are found on most immune cell subtypes, as well as a growing number of non-immune cells,^1-3 enabling a variety of cellular responses depending on the antibody and on the context. There are currently hundreds of mAbs under clinical development, including both intact mAbs as well as Fc fusion proteins, and the majority possess a human IgG Fc region capable of binding to one or more of the known FcγRs. Many of these molecules have been designed to harness activating effector functions such as antibody-dependent cell-mediated cytotoxicity (ADCC) or complement-dependent cytotoxicity (CDC) to aid or enhance their mechanism of action.⁴ Other antibody-related therapeutics are designed to work primarily through neutralization of their target by Fab-region binding; the Fc region is assumed to be “silent,” other than its contribution to FcRn binding that grants IgG its long half-life. Quite frequently, these so-called neutralizing antibodies are directed toward inflammatory cytokines, their receptors, or other molecules involved in the inflammation that underlies many chronic pathologies.⁵ However, anecdotal evidence of Fc-region involvement in the mechanism of action of some members of this latter class has begun to emerge.

Human antibodies of the IgG class, either in monovalent form or in multivalent form complexed with antigen, bind various members of the family of FcγRs. It is thought that most often, these interactions result in activation of the immune system by triggering the immunoreceptor tyrosine-based activation motif (ITAM)-containing Fcγ receptors or accessory molecules.⁶ Only one member of this family, FcγRIIb, signals instead through an immunoreceptor tyrosine-based inhibitory motif (ITIM) domain and has been shown in mice to play a role in inhibition of inflammatory responses and in the establishment of tolerance.^7,8 However, there is growing evidence that some ITAM-containing activating receptors have an intrinsic ability to deliver an alternate inhibitory signal as well.⁹ Our lab previously described one such phenomenon whereby IFNγ signaling events are inhibited by immune complex (IC)-mediated signaling through FcγRI.¹⁰ These lines of inquiry have contributed to a growing appreciation of the complexity of Fc-encoded functions present in therapeutic monoclonal antibodies. In addition, each IgG subclass has a different specificity toward each of the different FcγRI that is also dependent on the type and size of the immune complex in which it is presented.¹¹

To further our understanding of these inhibitory pathways, we extended our earlier studies to human macrophages because they display more complete expression of all of the human FcγRI than monocytes and are found in nearly every tissue.¹² We looked not only at the ability of IC to inhibit IFNγ receptor signal transduction, but also whether other inflammatory receptor systems were affected by examining the effect of IC on Toll Like Receptor (TLR)4-mediated events in the macrophage. Finally, as therapeutic monoclonal antibodies create ICs of a single IgG subtype by binding their antigen or a cell bearing that antigen on its surface, we analyzed the effect of individual IgG subtypes as well as whole human IgG in our IC model. We found that the IC model using human IgG evoked equally strong inhibition of IFNγ signaling in macrophages as was previously seen in monocytes, and that although interleukin (IL)4 receptor signaling remained unaffected, TLR4-dependent events were dramatically inhibited. Most surprisingly, however, we found that complexed IgG4 was fully capable of inhibiting IFNγ-mediated events in primary monocyte-derived macrophages. Although it has been known that IgG4 is able to bind the high affinity FcγRI, this is the first report that IgG4 is capable of transducing any type of signal through an Fcγ receptor.

Results

IC-mediated inhibition of IFNγ signaling has previously been studied using freshly elutriated human monocytes.¹⁰ However, in vivo, macrophages may represent the more relevant and functionally mature population responsible for establishing inflammation in response to invasion or injury.^13,14 More importantly, they are responsible for the establishment and maintenance of the chronic inflammation underlying numerous pathologies targeted by therapeutic intervention strategies.¹⁵ We therefore extended our studies to assess IC-mediated inhibition of IFNγ signaling in mature macrophages and whether it affected other macrophage inflammatory receptor systems, such as TLR4. Although previous studies demonstrated a lack of effect on signaling downstream of more immunomodulatory cytokines such as GM-CSF and IL4,¹⁶ it was of interest to determine whether additional inflammatory pathways, such as the Toll-like receptors (TLRs), responsible for microbial pattern recognition, are affected by the presence of IC in this model. Interferon Inducible Protein 10 (IP10) is rapidly and vigorously produced in response to IFNγ.¹⁷ However, in IC treated monocytes, the amount of IP10 produced is dramatically inhibited.¹⁰ In macrophages, the inhibition of IP10 mRNA accumulation and protein secretion was likewise inhibited to a similar degree in the presence of IC (Fig. 1A). IP10 mRNA and protein are also induced following TLR4 triggering by lipopolysaccharide (LPS); this IP10 production was decreased to an even greater degree by immune complex pretreatment than was seen downstream of IFNγ (Fig. 1B). Indeed, the levels of other inflammatory chemokines and cytokines downstream of LPS treatment were similarly reduced, including RANTES and IL8 (Fig. 1C); this was also true for the macrophage activation marker CD80 as well as the scavenger receptor CD163 (Fig. 1D). These data demonstrate that Fc-dependent inhibitory mechanisms are intact and even augmented in differentiated human monocyte-derived macrophages, reducing the production of IP10 and other cytokines, chemokines, and cell surface molecules that activate the immune system in response to host or microbial inflammatory activation signals such as IFNγ or LPS.

Figure 1.

IC-dependent changes in macrophage cytokine production and surface expression in response to inflammatory stimuli. One-day old monocytes or 7-day-old macrophages were plated on IC-treated or untreated plates for 1 hour prior to treatment with 10 ng/mL IFNγ or LPS, as indicated. (A and B) RNA was isolated after 1.5 hours at 37°C, reverse transcribed, and levels of the IP10 were analyzed by real-time polymerase chain reaction. Alternatively, supernatants were removed and clarified after 6 hours at 37°C and analyzed for IP10. (C) Supernatants collected after 20 hours were assessed for IL-10, RANTES, and IL-8 by multiplex analysis. Results are the mean ± SEM of 5 experiments with separate donors, performed in triplicate. For A and B, the data for each donor are normalized to the amount of IP10 RNA or protein induced by IFNγ or LPS treatment alone (100%). For C, the data for each donor are normalized to the amount of secreted protein induced by LPS treatment (100%). (D) Macrophages exposed to IFNγ or LPS treatment for 20 hours were stained with Alexa 647 mouse anti-human CD163, PE-Cy5 conjugated mouse anti-human CD80, or APC conjugated mouse anti-human CD105, as indicated. Results are the mean ± SEM of 3 experiments with separate donors.

In contrast, IL-10 production was induced by IC independent of the second stimulus (Fig. 1C), in agreement with other studies reporting IC-dependent production of this anti-inflammatory cytokine.^18-20 IL-10 is known to induce an alternatively activated, regulatory or “M2” macrophage phenotype that, in turn, leads other cells toward a more Th2-type immune response.¹⁸ In keeping with these observations, IC also induced the expression of CD105, a transforming growth factor β receptor family member, on the macrophage surface (Fig. 1D).

Many studies performed to date have used pooled human IgG to mimic normal immune complexes, as would happen in the normal course of infection. However, treatment with therapeutic antibodies presumably results in the formation of immune complexes- involving the therapeutic and its target- that are composed of antibodies of a single IgG subclass. To investigate the relevance of these immunomodulatory mechanisms to therapeutic monoclonal antibodies, we tested the ability of myeloma-derived IgG1, IgG2, and IgG4 to mediate IC-dependent inhibition of IFNγ and LPS signaling. These immunoglobulins were chosen as they represent those currently found among licensed therapeutic monoclonal antibodies (http://www.antibodysociety.org/news/approved_mabs.php). IgG1 has been shown to display superior binding to all FcγRs, whereas IgG2 is thought to only bind to the H131 allotype of the FcγR2a.²¹ In contrast, IgG4 is commonly believed to be devoid of appreciable FcγR binding.^22,23 Surprisingly, our results demonstrate significant concentration-dependent immunomodulatory behavior of each IgG subclass, including IgG4 and, to a lesser degree, IgG2 (Fig. 2). IgG4 antibodies have never been previously shown capable of transducing a signal through any Fcγ receptor, including FcγRI.

Figure 2.

IgG subclass specificity of IC inhibition of IFNγ-dependent signaling. Seven-day-old macrophages were trypsinized and plated on IC-treated or untreated plates for 1 hour prior to treatment with 10 ng/mL IFNγ as indicated. Five concentrations of each indicated IgG subtype were used to coat the wells: 50.0, 16.6, 5.5, 1.9, 0.7 μg/mL, from left to right. Supernatants were removed and clarified after 6 hours at 37°C and analyzed for IP10 by multiplex. Results are the mean ± SEM of 3 experiments with separate donors, performed in triplicate. The data for each donor are normalized to the amount of IP10 protein induced by IFNγ treatment alone (100%).

It is frequently stated in the literature that glycosylation of the IgG Fc region is required for Fcγ receptor family binding.^{21, 24} To determine whether IC comprising either IgG1 or IgG4 bore this dependence, antibody of each subclass was subjected to complete deglycosylation with PNGase F. Upon testing deglycosylated IgG1 and IgG4 against their native (mock-treated) counterparts as well as whole human IgG, we found that the ability of IgG1 to inhibit IFNγ signaling in this model was completely independent of glycosylation, whereas the ability of IgG4 to affect this outcome was only minimally affected by removal of Fc-associated glycans (Fig. 3). As glycosylation dependence of FcγR binding is widely accepted, our data suggest that either a different receptor complex is responsible for this ability to inhibit inflammatory signaling in macrophages, or that glycosylation was not absolutely required for FcγRI binding by these IgG subtypes in this IC model.

Figure 3.

IC-dependent inhibition of IFN-dependent IP10 production does not require Fc glycosylation. Myeloma-derived IgG1 and IgG4 were either treated with PNGase for 48 hrs or mock-treated in buffer under the same conditions but without enzyme. Seven-day-old macrophages were trypsinized and plated on plates coated with the indicated antibody preparation for 1h prior to treatment with 10 ng/mL IFNγ as indicated. Supernatants were removed and clarified after 6 hours at 37°C and analyzed for IP10 by multiplex. Results are the mean ± SEM of 3 experiments performed in triplicate, normalized to IP10 protein induced by IFNγ treatment alone (100%).

Whereas ICs derived from pooled IgG inhibit IFNγ signaling in monocytes specifically through FcγRI,¹⁰ it was not clear that this would be the case in macrophages. Not only do macrophages display the broadest FcγR expression of any cell type and higher expression of the FcγRs that are common to monocytes,²¹ but published reports suggest that all 3 classes of ITAM-containing Fcγ receptors can evoke inhibitory events.^10,25,26 IgG4 antibodies (either monomeric or as immune complexes), however, have never previously been demonstrated to transduce a signal through any Fcγ receptor. To determine the dependence of IC-mediated inhibition of IFNγ signaling on FcγR, blocking F(ab’)2 fragments specific to each FcγR class were used alone and in combination to assess the relative contribution of each (Fig. 4A). Consistent with previous data from human monocytes, the reduction of IFNγ signaling by polyclonal IgG or IgG1 IC showed greater dependence on FcγRI than the other Fcγ receptors. Examination of the raw data suggests that FcγRII or FcγRIII may be responsible for significant contributions to IP10 reduction in individual donors when IC are composed of IgG1 (data not shown). However, when data are combined across multiple donors, only the F(ab’)2 that interferes with FcγRI demonstrates significant reversal of IC-dependent signal inhibition. In contrast, IgG4-mediated inhibition of IFN-induced signaling was reduced only when FcγRI was blocked, with no similar variable dependence on the other FcγRs, indicating the differences in the binding behavior of IC from these 2 subclasses for the different FcγRs.

Figure 4.

IgG4 IC delivers an inhibitory signal through FcγRI. (A) Seven-day-old macrophages were trypsinized and treated with the indicated F(ab’)2 fragments (at 10 μg/mL each) for 30′ on ice prior to plating on indicated IC for 1 hour prior to treatment with 10 ng/mL IFNγ as indicated. Supernatants were removed and clarified after 6 hours at 37°C and analyzed for IP10 by multiplex. Results are the mean ± SEM of 5 experiments with separate donors performed in triplicate, normalized for each individual donor to the amount of IP10 protein induced by IFNγ treatment alone (100%). *p ≤ 0.05, **p ≤ 0.005. (B) Macrophages were treated with the Syk inhibitor Bay 61–3606 (10 μM) for 1 hour prior to plating on IC-treated or untreated plates for one hour. Cells were then treated with 10 ng/mL IFNγ or LPS for 1.5 hours followed by RNA isolation, reverse transcription, and analysis of IP10 mRNA levels by real-time polymerase chain reaction. (C) Supernatants of these macrophages collected after 20 hours were analyzed for RANTES and IL-12 p40 by multiplex. Results are the mean ± SEM of 3 experiments with separate donors, performed in triplicate. The data for each donor are normalized to the amount of protein induced by LPS treatment alone (100%). (D) Macrophages incubated with LPS for 20 hours were stained with Alexa 647 mouse anti-human CD163 or PE-Cy5 conjugated mouse anti-human CD80. Results are the mean ± SEM of 3 experiments with separate donors.

To examine whether the ability of IgG4 to inhibit IFN-induced IP10 transcription indeed depends on signaling components necessary for the activating pathways of FcγRs, the effect of Syk inhibitors on this phenomenon were investigated. The inhibition of IP10 mRNA accumulation that was invoked by IgG4 IC was indeed reversed by treatment of macrophages with the highly specific Syk inhibitor Bay 61–3606 prior to IgG4 IC exposure (Fig. 4B). IgG4 IC treatment was also able to reduce inflammatory cytokine secretion (Fig. 4C) and surface marker expression (Fig. 4D) in response to LPS, providing further evidence that IgG4 works through the same general pathway as whole IgG from human serum.

Discussion

Decades of research into how intravenous immunoglobulin (IVIG) exerts its therapeutic effect in numerous pathologies have uncovered multiple mechanisms whereby immunoglobulins can suppress inflammation.^27,28 One of these involves antibody and immune complex binding to the ITIM-containing inhibitory receptor FcγRIIb, which acts by recruiting phosphatases to counter nearby signaling events.⁸ Sialylated IgGs can in turn bind glycosylation-dependent receptors such as DC-SIGN in humans or SIGN-R1 in mice, which causes increased expression of this inhibitory receptor, presumably skewing the overall effects of antibodies and immune complexes away from immune activation and toward tolerance.^{7, 29} Although the importance of this inhibitory circuit has been convincingly demonstrated in mouse models of inflammation,³⁰ its relative contribution in humans is not yet as clear. The neonatal Fc receptor, FcRn, is responsible for the endosomal recycling of antibodies that imparts their long half-life.^31,32 Another proposed mechanism of action of IVIG is competition with problematic auto-antibodies for FcRn binding, resulting in their accelerated clearance. Related to this is the notion that IVIG may compete with autoantibody immune complexes for activating FcγR binding.²⁸ The fact that IVIG needs to be used at much higher doses to treat autoimmune pathologies compared with those used to complement immune deficiencies is in agreement with these mechanisms of general competition that involve the antibody Fc region. In addition, Fc-mediated inhibitory signaling has been described downstream of multiple ITAM-containing, or “activating” Fcγ receptors.

Work from our lab and others have made clear that each of the human FcγRs has the capacity to affect this balance in either direction.^{9, 10, 33} Previous results from our lab demonstrated that IC made of human IgG dampen the response of human monocytes to IFNγ through FcγRI.¹⁰ The data reported in this paper extend this finding to confirm that the response to IFNγ is similarly inhibited in primary human monocyte-derived macrophages (Fig. 1A). We further confirm the veracity of our model system by demonstrating that IC treatment diminishes inflammatory cytokine production (Fig. 1B and 1C) and activation marker expression (Fig. 1D) following LPS treatment. Conversely, IC treatment increases secretion of anti-inflammatory IL-10 and the surface expression of TGF-β receptor family member CD105 (Fig. 1B and 1C). These changes are consistent with the proposed immune complex-dependent “M2b” macrophage phenotype that has been previously described.³⁴ Interestingly, although increased CD163 surface expression is a marker for some types of M2 macrophages, its expression has been shown to be decreased by immune complexes,³⁵ in keeping with the idea that the M2 phenotype induced by immune complexes represents a hybrid between classical and alternative macrophage activation.

It would appear as though the low affinity activating receptors FcγRIIa (through its own ITAM) and FcγRIIIa (through the ITAM-containing accessory molecule, the common gamma chain) achieve inhibitory signaling through a different mechanism than FcγRI; it requires low signal strength due to monovalent or low avidity/affinity ligand.⁹ In contrast, occupancy of FcγRI on macrophages by monovalent IgG produces a tonic signal required for full IFNγ receptor signal strength.³⁶ In agreement with this paradigm, monovalent IgG has no significant effect on IFNγ signaling in our system (unpublished data) whereas there is a direct positive correlation between the density of the IC and the degree of inhibition of IP10 production (Fig. 3). Independent evidence from other work suggests that this may be a hallmark of inhibitory immune complex signaling through FcγRI,³⁷ especially in monocyte-derived or tissue resident macrophages.

FcγRI is often ignored when general statements are made about antibody-dependent cellular effector function. One reason for this may be the outdated perception that since the high-affinity FcγRI can bind monovalent IgG, it will always be occupied due to the high circulating concentration of monovalent IgG and thus unavailable for immune complex binding. In truth, the off-rate of monomeric IgG1, the subclass that binds FcγRI with the highest affinity, is such that it is constantly leaving the receptor available for higher-avidity interactions with immune complex.³⁸ In this regard, it has been shown that immune complexes are able to effectively compete for FcγRI binding and signaling in the presence of physiologically-relevant concentrations of IgG.³⁹ It has also been repeatedly demonstrated that IC binding to macrophages induces IL-10 secretion and can induce a regulatory macrophage phenotype,^18,40 with the data suggesting that these effects occur downstream of FcγRI.³⁷ A recent investigation of TNF antagonists, 2 human antibodies, an Fc-receptor fusion, and a PEGylated Fab’, demonstrated that the therapeutics containing a fully functional human Fc region induced the formation of regulatory macrophages and inhibited T cell proliferation in a mixed-lymphocyte reaction to a greater degree.⁴¹ All of these observed responses would depend on the interaction of the Fc region of the therapeutic mAb with the FcγR on the human macrophage.

Phagocytosis of immune complexes as a result of FcγRI engagement on macrophages is also known to be a trigger toward adoption of an “alternatively activated” or M2 phenotype⁴² that results in reduction of inflammatory cytokine and chemokine levels and induction of immunomodulatory IL-10. This might be important not only for antibodies designed simply to neutralize inflammatory signaling pathways. It has been shown that M2 macrophages, especially those induced by IL-10 treatment, are superior to classically activated M1 macrophages for the ingestion of Rituximab-coated cancer cells.⁴³ There is also a growing appreciation for the role of macrophages in the phagocytosis of therapeutic antibody-coated tumor cells,^44,45 and if the M2 fate is superior for this task, Fc binding to FcγRI may encourage its development. Taken together, these data suggest that the phagocytic effector function (or “frustrated phagocytosis,” in the case of our IC model) is the normal consequence of CD64-dependent FcRγ-chain ITAM stimulation, which produces an alternatively activated- or M2- fate in macrophages, the only cell that naturally expresses FcγRI.

Human IgG4 antibodies (either soluble or as immune complexes) have never been previously demonstrated to transduce a signal through any Fcγ receptor, despite reports of their relatively high affinity for FcγRI.⁴⁶ Our data demonstrates that immobilized (or “complexed”) IgG4 can effectively inhibit IFNγ-stimulated IP10 production (Fig. 2) or inflammatory events downstream of LPS treatment (Fig. 4), and that these phenomena are dependent on the activation of FcγRI through Syk (Fig. 4). Although the incomplete ability of the FcγRI blocking F(ab’)2 fragments (clone 10.1) to prevent the action of IgG4 leaves room for the possibility of involvement of another receptor, a more likely explanation would be that these results are due to an artifact of this particular antibody.⁴⁷ The similarly surprising finding that glycosylation is not required for induction of this inhibitory pathway, either by IgG1 or IgG4, is potentially explained by recent findings following careful investigation of the size- and glycosylation-dependence of FcγRI binding by IC: it was demonstrated that for IgG1 IC binding to FcγRI, glycosylation is relatively unimportant, whereas for IgG4, the larger the IC, the less important Fc glycosylation was for FcγRI binding.¹¹ Consistent with our general findings regarding IgG4, a recent report showed that a human IgG4 form of a therapeutic anti-CD20 antibody is capable of B cell depletion in a humanized mouse model, whereas F(ab’)2 fragments of the same antibody are not.⁴⁸ Data from this work also suggest that macrophages are the critical effector cell for B cell depletion in the bone marrow. IgG4 is often selected as the backbone of choice for those therapeutic antibodies designed to neutralize an inflammatory mediator or its receptor without unwanted activation of the immune system.⁴ Our data suggest that therapeutic IgG4s could in fact actively contribute to the inhibition of inflammation through Fc-dependent signaling. The presumed inability of IgG4 to invoke any FcγR-dependent signal has likely been based both on an incomplete understanding of its affinity for FcγRI and the functionally monovalent nature of endogenous IgG4.^49,50 However, updates describing its high affinity to FcγRI are accompanied by new crystal structures of the IgG4 Fc that suggest the possibility of a unique additional interaction site with this receptor. ^51,52 Additionally, although natural IgG4s undergo Fab-arm exchange that renders them hetero-bivalent in vivo, therapeutic IgG4s are now frequently stabilized by mutations that assure their integrity during manufacture and subsequent administration. These new findings highlight how therapeutic IgG4s could presumably form immune complexes with an increased ability for functional multivalent FcγRI engagement, perhaps nearly as capable as IgG1s. Growing evidence suggests that a major outcome of FcγRI-dependent immune complex signaling in macrophages is anti-inflammatory,⁵³ and this new evidence that complexed IgG4 may be capable of inducing an M2 macrophage phenotype in vivo should be more thoroughly investigated. Such studies would improve understanding of all Fc-mediated contributions to therapeutic antibody efficacy.

Materials and Methods

Isolation and treatment of human monocytes

Human peripheral blood monocytes were obtained from healthy volunteers by leukophoresis and further purified from mononuclear cells by Ficoll-Hypaque sedimentation, followed by countercurrent centrifugal elutriation, as described previously.¹⁶ The use of human volunteers for this purpose was reviewed and approved by both the National Institutes of Health and Food and Drug Administration internal review boards. Monocytes used for experiments were allowed to adhere for 1 hr in serum-free Dulbecco modified Eagle medium (DMEM; Lonza, #12–733F) onto untreated plates or plates coated with serum-derived human IgG or human myeloma-derived IgG1 or IgG4 (Athens Research and Technology, #16-16-090707 (IgG), #16-16-090707-1M (IgG1) or # 16-16-090707-4M (IgG4)) at 50 μg/ml. Alternatively, human monocyte-derived macrophages were cultured by plating freshly elutriated monocytes in DMEM supplemented with 10 μg/mL gentamicin (Invitrogen, #15750-060) at 6 × 10⁶/2 mL/well in 6-well plates. After 4 to 6 hours, heat-inactivated fetal bovine serum (Valley Biomedical, #BS3032HI) was added to 10%, and monocytes were cultured for 7 d to enable differentiation into mature macrophages. Macrophages were then removed using Accutase (Innovative Cell Technologies, #AT104) and allowed to recover in DMEM with 10% FBS for 3h at 37°C. Blocking F(ab’)2 (Ancell, #216–520 (α-FcγRI, clone 10.1), #181–520 (α-FcγRII, clone 7.3), # 165–520 (α-FcγRIII, clone 3G8) were applied to macrophages at 10 μg/mL each for 45′ on ice. Where indicated, cells were treated with Syk inhibitor Bay 61–3606 (Santa Cruz, #4152655) at 10 μM for 1 hour in suspension. Cells were then plated on IC-coated plates and or uncoated plates for 1 hour and subsequently treated with human recombinant IFNγ (kindly provided by Genentech, Inc., South San Francisco, CA) or ultrapurified LPS from E. coli K12 (Invivogen, #tlrl-peklps) at a concentration of 10 ng/ml.

Antibody deglycosylation

Enzymatic deglycosylation of serum-derived human IgG and myeloma-derived IgG1 and IgG4 was performed as adapted from the literature.⁵⁴ Briefly, 5 ml (2500 units) PNGase F (New England Biolabs #P07045) was added to 100 μl IgG at 20 mg/ml in 50 mM Tris, 150 mM NaCl pH 8.5 and incubated at 37°C for 42 h. Untreated control antibody was incubated in this mixture in the absence of PNGase under the same conditions. The samples were then added to 200 μl 0.01 M HEPES pH 7.4, 0.15 M NaCl, 0.005% v/v Surfactant P20 (HBS-P buffer) and 200 μl pre-washed Protein A/G beads (Santa Cruz, SC-2003). Samples were rotated for 30 min at RT, spun down at 2500 rpm for 2 min, and washed 2× with HBS-P. IgG was then sequentially eluted with 10 mM glycine pH 3.0, 2.5, and 2.0, each for 5 min at RT, and 1M Tris pH 8 was added to eluates to bring them to pH 7.4.

RNA isolation and analysis

Total RNA was isolated from cells using the RNeasy kit (Qiagen, #74106) according to the manufacturer's instructions. Reverse transcription with random hexamers was conducted using the SuperScript™ II Reverse Transcriptase (Invitrogen, #18064-014) as per the manufacturer's established protocols. Two micrograms of RNA per sample were used and a control parallel reaction was included where no reverse transcriptase was added. One-fortieth of the reverse transcriptase reaction material was analyzed by Quantitative RT-PCR using SYBR Green and detection with the real-time PCR System 7900 (Applied Biosystems). Quantitative RT-PCR primers were as follows: for IFN-inducible protein 10 (IP-10), 5′-GGAACCTCCAGTCTCAGCACC-3′ and 5′-CAGCCTCTGTGTGGTCCATCC-3′, and for β-actin, 5′- TCGTCGACAACGGCTCCGGCATGTGC-3′ and 5′-TTCTGCAGGGAGGAGCTGGAAGCAGC-3′. PCR conditions (45 cycles) included 15 s at 95°C, 30 s at 60°C, and 30 s at 72°C. The threshold cycle number for hIP-10 was normalized to that of the β-actin transcripts and then set to a percentage of the maximum transcription detected in samples treated with 10 ng/ml IFNγ. Relative fold increases were determined using the 2-ΔΔCT method as described elsewhere.⁵⁵ Two-tailed Student's T tests were performed where statistically-significant differences are noted.

Multiplex analysis

After indicated treatments, supernatants were clarified and diluted 10 to 20 times such that values obtained fit in the linear range of the standard curve for each cytokine reported. Assays were performed in duplicate using multiplex bead-based kits for the indicated human cytokines and chemokines (Panomics, #PC1008) per manufacturers' instructions. Fluorescence of beads was measured using a Bioplex 200 analyzer (Bio-Rad). Cytokine data analysis was performed using the BioPlex Manager software (Bio-Rad). Concentrations were determined with a 5-parametric logistic nonlinear regression curve-fitting algorithm, and results were entered into GraphPad Prism 5 software (GraphPad Software) for plotting and statistical analysis.

Flow cytometry

For surface marker staining, cells were detached using Accutase (Innovative Cell Technologies, #3155493), washed, and stained with Alexa 647 mouse anti-human CD163 (BD; # 562669), PE-Cy5 conjugated mouse anti-human CD80 (# 559370), or APC conjugated mouse anti-human CD105 (#562408) for 30 minutes at 4°C. Samples were resuspended in phosphate-buffered saline containing 1.0% paraformaldehyde, and analyzed on a FACSCalibur flow cytometer (BD Biosciences), using a gate set for single, intact macrophages.

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.

Acknowledgments

This project was supported in part by an appointment of DAH and AGM to the Research Participation Program at the Center for Biologics Evaluation and Research administered by the Oak Ridge Institute for Science and Education through an interagency agreement between the U.S. Department of Energy and the U.S. Food and Drug Administration.

We thank Silvia Bacot and members of Dr. Kathleen Clouse's lab for elutriation of human monocytes and other technical support, as well as Drs. Peter Adams and Tao Xie for critical review of the manuscript.

The findings and conclusions in this paper have not been formally disseminated by the U.S. Food and Drug Administration and should not be construed to represent any Agency determination or policy.

Authors’ Contributions

JFAS designed the experiments, performed the statistical analyses, helped analyze and interpret the data, and helped write the manuscript. DAH and AGM performed most of the experiments, and helped analyze the data. GHB helped analyze and interpret the data. GMF helped design the experiments, analyzed and interpreted the data, and wrote the manuscript.