IgG4 drives M2a macrophages to a regulatory M2b‐like phenotype: potential implication in immune tolerance

Abstract Background Macrophages can be converted in vitro into immunoregulatory M2b macrophages in the presence of immune complexes (ICs), but the role of the specific subclasses IgG1 or IgG4 in this phenotypic and functional change is not known. Objective We aimed to refine the original method by applying precisely defined ICs of the subclasses IgG4 or IgG1 constructed by two independent methods. Methods Monocyte‐derived macrophages (MDMs) were treated with M‐CSF, followed by IL‐4/IL‐13 to induce the M2a allergic phenotype. To mimic unspecific or allergen‐specific ICs, plates were coated with myeloma IgG1 or IgG4, or with grass pollen allergen Phl p 5 followed by recombinant human Phl p 5‐specific IgG1 or IgG4. M2a polarized macrophages were then added, cultured, and examined for cellular markers and cytokines by flow cytometry, ELISA, and rtPCR. Alternatively, immune complexes with IgG1 or IgG4 were formed using protein L. Results IgG4 ICs down regulated CD163 and CD206 on M2a cells, and significantly increased IL‐10, IL‐6, TNFα, and CCL1 secretion, indicating a shift to an M2b‐like phenotype. Treatment with IgG4 ICs resulted in expression of FcγRII and down modulation of FcγRII compared with IgG1 treated cells (P = 0.0335) or untreated cells (P < 0.00001). Conclusion Immune complexes with subclasses IgG1 and IgG4 can in vitro be generated by plate absorption, and in fluid form by protein L. Cross‐linking of FcγRIIb by the IgG4 subclass redirects pro‐allergic M2a macrophages to an M2b‐like immunosuppressive phenotype. This suggests an interplay of macrophages with IgG4 in immune tolerance, likely relevant in allergen immunotherapy.


| INTRODUCTION
In immediate-type allergy, a particular Th2 cytokine microenvironment is established by inflammatory cytokines such as IL-4, IL-5, and IL-13.
These cytokines induce allergen-specific IgE, eosinophilia, mucus production, and the recruitment of inflammatory cells to inflamed tissues. [1][2][3] They polarize monocytes and macrophages into alternatively activated M2a macrophages typically characterized by downregulated expression of the hemoglobin-haptoglobin scavenger receptor (CD163) 4,5 and upregulated expression of mannose receptor 1 (CD206) as well as of the B7-2 costimulatory protein CD86. The CD86 upregulation is supported by basophil-and mast cell-derived IL-4 as shown for allergic pulmonary diseases. 6 Moreover, these M2a macrophages secrete cytokines such as IL-1 receptor antagonist (IL-1ra) 7 and chemokines such as the pulmonary and activation-regulated chemokine (PARC; CCL18), the macrophage-derived chemokine (MDC; CCL22), and the thymus and activation-regulated chemokine (TARC; CCL17). 8 Whereas the symptoms of allergic reactions can be combated by different pharmacological treatments, allergen immunotherapy (AIT) represents the only curative approach in type I allergy. 9 AIT results in longterm clinical benefits, and numerous publications highlight the induction of cellular responses within regulatory T cells (Treg), especially inducible IL-10-and TGF-β-producing type 1 Treg (Tr1), and regulatory B cells (Bregs) as mechanisms of inducible tolerance. 9,10 Bregs are not only a source of IL-10, but they also sense IL-10 as a switching factor for IgG4 production. 11 This is important as IgG4 is a hallmark of AIT, although its tolerogenic function is still insufficiently understood. 10 While it is accepted that Tregs and Bregs with IL-10 have a critical role in dampening the allergic inflammatory response in AIT, the role of macrophages in this process is not entirely defined. In particular, the M2b macrophages could take part in tolerance induction. M2b macrophages are specifically characterized by the secretion of CCL1 chemokine, 12 a ligand of the cognate chemokine receptor CCR8. 13,14 Notably, CCR8 is not only essential for maintaining the M2b characteristics, it is also expressed by CCR8 + FOXP3 + Treg cells. 15 These master drivers of immune regulation could, therefore, be ignited by macrophages via CCL1 and CCR8.
Decisive for the present study were the observations that immune complexes (IC), without further specification, were compulsory to differentiate the M2b subtype from monocyte-derived macrophages in vitro. 12,16 When recently subclass-specific effects were addressed on M1 macrophages, IgG4 like IgG1 inhibited IFNγ signaling via FcγRI, favoring an M2-like phenotype. 17 When treating allergy, more or less the immunological opposite is desired, which is conversion from pro-allergic M2a to M2b macrophages.
Our hypothesis combined all this background information considering the possibility to convert the pro-allergic M2a subtype, highly present in the Th2 environment, into an immunoregulatory M2b-like subtype. We (a) refined the original method, in which undefined immune complexes were applied for activating M2a macrophages, by using precisely defined immune complexes of subclasses IgG4 or IgG1; and (b) we applied two independent methods for the formation of such immune complexes, either plate bound or in solution.
Subsequently, we screened for cellular markers and cytokines correlating with the induction of a tolerogenic microenvironment.

| Reagents and antibodies
All reagents and antibodies used in this study are reported in Tables S1-S2.

G R A P H I C A L A B S T R A C T
This study highlights the central role of IgG4 in immune tolerance. The treatment of M2a with ICs consisting of IgG4 shows a considerably greater effect in promoting an M2b-like phenotype than with IgG1-containing ICs. The secretion of CCL1 and IL-10 by M2b-like macrophages is important to establish a tolerogenic microenvironment. Human peripheral blood was obtained either by venipuncture using vacuum tubes coated with lithium heparin (Greiner Bio-One, Kremsmuenster, Austria) or by leukocyte reduction chambers (LRS) cones generated by the TrimaAccel automated blood collection system (Terumo BCT, Lakewood, CA, USA) during the process of single platelet apheresis. Peripheral blood mononuclear cells (PBMCs) were isolated from each donor using the cushion of Ficoll-Paque (Data S1). Total PBMCs were seeded, and after 2 hours, the non adherent cells were washed away twice. The adherent cells (monocytes) were maintained in RPMI 1640 with 10% heat-inactivated FBS, and 1% of P/S (cRPMI), supplemented with 20 ng/mL rh-M-CSF for 7-9 days.

| Isolation and treatment of human monocytes
Half of the medium was refreshed every 2-3 days ( Figure S1A-B).
The purity of monocytes was checked after 3 days, by flow cytometric analysis as CD3 − CD11b + CD86 + adherent cells and determined as 80% ±5 of total live cells ( Figure S2).  USA) was coated with these antibodies at 50 μg/mL in HBSS and washed twice with cRPMI (Data S1).

| Creation of soluble IgG1 and IgG4 immune complexes by protein L
To mimic soluble immune complexes (ICs) consisting of human myeloma IgG1 (mIgG1) or IgG4 (mIgG4), 7-9 days differentiated monocytes were treated at room temperature for 30 minutes with mIgG1 or mIgG4 antibodies and then washed twice with cRPMI. The rProtein-L (Thermo Fisher Scientific, Waltham, MA, USA) at a ratio of 1 : 4 (rProtein-L : antibodies) was used to cross-link the bound antibodies. 18 (Data S1).

| Creation of allergen-specific immune complexes
To mimic immune complexes with human anti-Phl p 5 IgG1

| Staining and flow cytometric analysis
After 72 hours, cells incubated on IgG1-or IgG4-IC-coated plates were detached using ice-cold PBS/EDTA and washed twice with HBSS plus 3% FBS as staining buffer, for surface marker phenotypization. Then, the cells were incubated with a multicolor staining mix of monoclonal antibodies against CD14, CD86, CD11b, CD163, and CD206 or their isotype controls (BioLegend) diluted 1:100 in staining buffer for 30 minutes at 4°C followed by 2× washing with staining buffer.
For the FcγR staining, the cells were detached as described above, washed with an ice-cold staining buffer (PBS −Ca−Mg plus 1% BSA, 0.02% NaN 3 ), and then blocked with 2.4 mg/mL human IgG (Beriglobin P; CSL Behring, King of Prussia, PA, USA) in staining buffer for 30 minutes on ice. The cells were then incubated with a multicolor staining mix of primary monoclonal antibodies against CD64, CD32, and CD16 or isotype controls (as specified in Table S2; M2a + IgG1 (Data S1).

| Real-time polymerase chain reaction
The primers (Table S4) were designed using Primer-Blast tool 19 and evaluated with Beacon Designer Free Edition (Premier Biosoft, Palo Alto, CA, USA). Real-time PCR (rtPCR) was performed using the Solis Biodyne Supermix (Solis Biodyne, Tartu, Estonia) in accordance with manufacturer's recommendations. rtPCR was performed on a QuantStudio 12K Flex system (Thermo Fisher Scientific) (Data S1).
The 2 −ΔΔCt analysis was performed using the QuantStudio 12k Flex Software v1.2.3 (Thermo Fisher Scientific) to obtain the cycle of threshold (Ct) for each sample investigated. Actin beta (ACTB) was used as reference gene. The mean fold change expression was calculated for M2a, M2a + IgG1, and M2a + IgG4.

| Graphs and statistical analysis
The graphs and the statistical analyses were performed with GraphPad Prism version 6.00 for Macintosh (GraphPad Software, La Jolla, CA, USA).

M2b macrophages
The flow cytometric analyses of nine donors in four independent experiments were performed to distinguish between the different alternatively activated macrophages polarized in vitro. The analyses evaluated the surface marker expression of CD206, CD163, and the co stimulatory molecule B7-2 (CD86), considered as specific markers of M2 alternatively activated population. 8 The results showed that M2a macrophages had a higher expression of CD206 compared with M2o polarized macrophages, the CD163 expression was relatively higher in M2a macrophages than in M2b, but lower than in M2c, and the expression of CD86 was higher in M2a macrophages than in M2b or M2c ( Figure 1A). Additionally, we analyzed the surface expression of high-affinity Fc gamma receptors FcγRI-CD64, low-affi-  Table S5A-B).

| Contrasting cytokine expression between M2a
and M2c, and M2b The method of choice to assess the differences between M2a proallergic and M2b immunoregulatory macrophage subpopulation is their cytokine and chemokine patterns. 12 M2a and M2c macrophages slightly expressed IL-10, but almost no IL-6 and TNFα, and no CCL1 (    Table S6. An almost exclusive secretion of CCL1 was observed when M2a macrophages were stimulated with ICs containing IgG4 than IgG1 compared with M2a alone (Table S6 and Figure S3C).
Therefore, ICs formed with mIgG4 also functionally drive the polarization status of pro-allergic M2a macrophages toward an M2b-like regulatory subtype. production. 1,10 We addressed here whether IgG4 might even have a more prominent role in immune tolerance by directly influencing IL-10 production. Corresponding to the Th1-Th2 paradigm, also macrophages can functionally differentiate into M1, and M2 cells, with the M2a, b, c subtypes, which can be mimicked by in vitro stimulation protocols. 16,27 It caught our attention that especially the tolerogenic M2b subtype is activated in the presence of IgG immune complexes, which so far were not further classified.
The study hypothesis was, therefore, that pro-allergic M2a macrophages could by immune complexes containing IgG4, as induced by AIT, be functionally converted into an immunoregulatory M2b subtype. M2b-derived secretion of CCL1 and IL-10 may contribute to the development of a tolerogenic microenvironment.
We first determined the phenotype and the secretion pattern of all in vitro polarized macrophages because of some overlap in their characteristics. For instance, the expression of CD206 that is considered high in the M2a subtype is also expressed by M2o or M2c (Figure 1A) and even by M1 cells; 5 CD163, which is medium-low expressed by M2a and in M2o subtypes, is upregulated by glucocorticoid treatment in the M2c subtype. 4,5,16 In accordance with previous studies, only the M2b macrophage subtype produced CCL1 chemokine and higher IL-10 levels than any other subtypes and thus had the most pronounced signature. [12][13][14] It is known that macrophages have a degree of plasticity, but the phenotypic and functional change in pro-allergic M2a toward an M2b-like phenotype only in the presence of IgG4, but not IgG1, was striking. This subclass-specific phenomenon was observed with complexes containing myeloma IgG4, as well as when using recombinant anti-Phl p 5 specific IgG4 when complexed to its specific allergen.
Therefore, IgG4 complexes are potent stimulators of macrophages to CCL-1 and IL-10 production and capable to turn them from a proallergic to a regulatory phenotype.
To investigate the molecular mechanism, we examined the surface expression of the FcγR family on M2a macrophages. Only M2o and M2c macrophages expressed high-affinity FcγRI and low-affinity FcγRIII ( Figure 1B)  The results from three independent experiments were combined for statistical analysis. Repeatedmeasures one-way ANOVA statistical analysis and Tukey multiple comparison post-hoc test were performed, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 IgG subclasses. 12 Our data show that IgG4-mediated FcγRII stimulation is decisive for the phenotypic and functional conversion of M2a High doses of allergens are needed in AIT to achieve immune tolerance, 28 and both IgG1 and IgG4 are formed in different ratios. Only when enough IgG4 compared to IgG1 is produced, the FcγRIIb on M2a macrophages will be engaged and lead to a M2b conversion, resulting in secretion of IL-10 and CCL1. While IL-10 further supports the class switch of B cells in IgG4 producing cells, CCL1 will recall CCR8 + FOXP3 + Tregs from the periphery.
Our results are in line with studies on IgG4-related disease (IgG4-RD), a fibro-inflammatory condition characterized by abundant IgG4 + plasma cells in affected tissues, and by specific involvement of Th2, Treg cells, and M2 macrophages. [33][34][35] A Japanese genomewide association study of IgG4-RD reported three susceptibility loci consistent with an antigen-driven disease. Currently, IgG4 is rather regarded as a protective antibody in IgG4-RD, dampening the more harmful effects of IgG1. 36 In accordance, FCGR2B is considered to play a critical role in the control of IgG4-RD. 37 A recent study correlates the high expression of IgG4 from IgG4-RD sclerosing cholangitis patients' B cells with the high expression of CCL1 in the tissues and, in turn, with the recruitment of CCR8 + FOXP3 + Treg cells. 38 The high affinity of CCL1 for CCR8 as well as the emerging opinion that CCL1 is indispensable for Treg function, further emphasizes the link of CCL1 and IL-10 whit the creation of an immunoregulatory microenvironment. 15,39 Moreover, the importance of CCL1, produced by M2b macrophages, in this process is also documented by the fact that treatment with antisense CCL1-oligonucleotide inhibited M2b macrophages to maintain their characteristics such as IL-10 production and inflammation inhibition. 13,14 Thus, our data argue for a central role of IgG4 in immune tolerance as induced by AIT, prompting cross talk between macrophages and regulatory lymphocytes.

ACKNOWLEDGMENTS
We are obliged to Prof. Fred Finkelman and to Prof. Sophia Karagiannis for their essential scientific input and discussion.

CONF LICTS OF INTEREST
The authors declare that they have no conflicts of interest.

AUTHOR CONTRIBU TI ONS
EJJ and RB designed the study; RB, EJJ, and FRW conceived and designed the experiments; RB and AOR performed the experiments; MBF, KH, AOR, and SF contributed to reagents and materials; RB, AOR, and EJJ analyzed the data; RB, EJJ, and HS contributed to the interpretation of the data; RB and EJJ drafted the paper. All authors contributed to texts and approved the manuscript.
O R C I D