Mouse Fcγ R expression and function
mFcγR exhibit differential expression on immune cell populations, which can be positively and negatively regulated by cytokines. mFcγRI is expressed in a restricted fashion on monocytes, some
monocyte-derived cells and macrophages; inhibitory mFcγRIIB is highly expressed on B cells, monocyte/macrophages, mast cells, basophils, dendritic cells and eosinophils, and low on neutrophils, plus considerably expressed on endothelial cells in the liver; mFcγRIII is highly expressed on monocyte/macrophages, neutrophils, mast cells, basophils, dendritic cells and eosinophils, and low on NK cells; mFcγRIV is highly expressed on Ly6Clow monocytes, macrophages and neutrophils, and absent on other cells.
The function of these receptors in vivo has been extensively studied using different receptor knock-out animals, as well as the administration of specific blocking antibodies, in the context of models of infection and immune challenge (reviewed in ). mFcγRIII is the major activating IgG receptor, and in its absence mice are more susceptible to infection, resistant to the anti-inflammatory effects of intravenous immunoglobulin therapy, and have impaired tumour clearance with therapeutic antibody treatment. Furthermore, mFcγRIII contributes to several models of antibody-dependent pathologies including the Arthus reaction, passive cutaneous anaphylaxis and passive systemic anaphylaxis, autoimmune anaemia, thrombocytopenia and arthritis. Similarly, the high affinity receptor mFcγRIV can contribute to nephritis, autoimmune thrombocytopenia and arthritis, lung inflammation, anti-melanoma therapy, and systemic anaphylaxis. These ranges of functions for mFcγRIII and mFcγRIV in defence and pathology align with the broad expression profile of these receptors: on neutrophils and macrophages, important inflammatory and phagocytic cells. mFcγRI, too, has been reported to contribute to the severity of arthritis, anaphylaxis and the Arthus reaction, in addition to the elimination of tumour metastases; yet this is at odds with the restricted expression pattern of this receptor on monocytes and some monocyte-derived resident cells. mFcγRI-/- mice also experience impaired clearance of pathogenic bacteria and helminth infection. Clarification of the in vivo function of mFcγRI in antibody-dependent pathologies, using mFcγRIonly mice, is the subject of a supplemental project to this main thesis work (Annex 7.2).
The importance of the inhibitory receptor mFcγRIIB is highlighted in knockout mice, which experience increased resistance to infection, greater bacterial clearance, and enhanced antibodymediated tumour therapy, yet also a much greater susceptibility to various autoimmune manifestations [148-150]. Indeed, on the C57Bl/6 background, mFcγRIIB-/- mice spontaneously develop glomerulonephritis . Attempts at blocking this receptor by the in vivo administration of blocking antibodies have been deterred by the rapid internalisation of such antibodies after binding to the target, mFcγRIIB, which is highly expressed on the liver sinusoidal epithelium, and probably ‘soaks up’ these blocking antibodies before they can reach immune cell targets .
Interestingly, mFcγRIIB, mFcγRIII and mFcγRIV exhibit dual specificity for IgE and IgG. Although these receptors have only low affinity for IgE, and their affinity for IgG is considerably higher, this factor is important to keep in mind, particularly in the context of anaphylaxis models, as the IgE interactions can have biological consequences. Human IgG receptors are more ‘strict’ in their binding to this subclass alone. Finally, mFcRn has a primary function in recycling IgG and thereby protecting it from catabolism, and as an antenatal IgG transporter, but also transports IgG into the tissues of adults, thereby facilitating both autoimmunity and anti-tumour therapy .
Human Fcγ R expression and function
Not only do mouse and human FcγR exhibit significant structural diversity (Figure 1.2), these receptors have considerably different patterns of expression on immune cell populations, thereby determining potentially disparate contributions to immune pathologies. hFcγRI (CD64) is restricted to monocytes/macrophages and DCs, but is highly inducible, compared to its murine counterpart, and can also be expressed on neutrophils. The uniquely human hFcγRIIA is expressed on all myeloid cells but not on lymphocytes, and is crucially also expressed on platelets. Inhibitory hFcγRIIB is highly expressed only on circulating B cells and basophils, expressed at low levels on monocytes and neutrophils, and expressed on tissue macrophages and DCs but not mast cells. hFcγRIIC is expressed on NK cells, monocytes, and neutrophils in only 20-25% of individuals those who carry the polymorphism Q13 or ORF; in the remaining persons, a SNP at position 13 generates a stop codon (stop13), in which case FCGR2C represents a pseudogene . hFcγRIIIA is expressed on NK cells and monocytes/ macrophages; and hFcγRIIIB is highly expressed on neutrophils and exhibits low expression on some basophils. Finally, hFcRn is expressed on antigen-presenting cells, monocytes/ macrophages, neutrophils, vascular endothelial cells, intestinal epithelial cells, and on syncytiotrophoblasts, the latter which permits antenatal transfer of IgG from mother to fetus.
The multiplicity of human FcγRs is increased by a series of genetic polymorphisms. Several FCGR polymorphisms modify the affinity between FcγRs and human IgG, and different polymorphisms may predispose to the development of disease, or determine responsiveness to therapy; these associations in humans can provide some insight of the role of these receptors in vivo (Reviewed in , included as Annex 7.3). Transgenic mouse studies have likewise greatly enhanced our understanding of the in vivo function of hFcγRs. In particular, these studies have highlighted the respective contributions of FcγR, to antibody-mediated inflammatory and allergic diseases, albeit with some major caveats according to reproducibility of transgene expression in the mice compared to the human, ligation by non-physiological ligands (mouse IgG), and confounding effects of expression in conjunction with mouse FcγR (Reviewed in ). hFcRI in transgenic mice retains its high affinity properties, can mediate phagocytosis in vitro and antibody-mediated cell destruction (ADCC) and clearance of opsonized red blood cells or platelets in vivo [142, 155’Heijnen, 1996 #277, 156]. hFcRI seems to have important roles in antigen presentation, amplifying antibody production and T cell responses via DC cross priming.[157, 158]. hFc RIIA is a promiscuously expressed and dominant activating human IgG receptor. Expressed alone, it can restore susceptibility to autoimmune thrombocytopenia, arthritis, airway inflammation, and local and systemic IgG-dependent allergic inflammation [159-161]. Dramatically, hFcRIIA expression on neutrophils alone, driven by a transgene associated with the restrictive hMRP8 promoter, or by adoptive transfer of human neutrophils, was sufficient to mediate IgGdependent pathologies arthritis, glomerulonephritis, and reverse passive arthus reaction, as well as restoring susceptibility to systemic IgG-dependent anaphylaxis [161-163]. In vitro, hFcRIIA can mediate uptake of immune complexes, and can trigger the formation of NETs in vivo . The only hFcR expressed on platelets, hFcRIIA endows human platelets with the capacity to interact with soluble immune complexes, and to be activated to promote thrombus formation or thrombocytopenia [159, 165, 166]. .hFcRIIIA has been investigated using transgenic mice predominantly for its involvement in anti-tumour therapy: hFcRIIIA can promote tumour cell destruction via ADCC, and antibodies [167, 168] with enhanced binding to hFcRIIIA demonstrated improved efficacy in vivo. hFc RIIIA and hFcRIIIB share a homologous extracellular domain, but whereas hFcRIIIA contains an ITAM motif in its intracellular tail, hFc RIIIB is a GPI-linked membrane receptor whose function is less well elucidated. Originally viewed as a decoy receptor, it is clear that hFcRIIIB can affect neutrophil migration and interaction with immune complexes [162, 164, 169], perhaps by association with surface integrins .
The differential roles of hFcRIIA and hFcRIIIB on neutrophils were studied using hMRP8cre promoter to restrict transgene expression only to neutrophils and some monocytes [162, 164]. hFcRIIIB-dependent neutrophil accumulation in the kidney or skin, during nephritis or passive Arthus models respectively, occurred in the absence of prominent edema, macrophage recruitment or tissue injury . On the other hand, hFcRIIA and hFcRIIIB co-expression cooperatively promoted tissue injury. In particular, neutrophil slow rolling and adhesion responses to immune complexes deposited within the vasculature were mediated by hFcRIIIB, followed by ‘quiescent’ immune complex uptake and clearance, in direct contrast to proinflammatory NET formation that was triggered by hFcRIIA engagement of soluble immune complexes . These studies serve to highlight particularly the cooperative effects of hFcR during inflammation in vivo, albeit in an artificial context with a false transgene promoter.
In an attempt to recapitulate the full range of classical hFcR diversity in a mouse model, individual transgenic mouse strains were bred to create hFcRItgIIAtgIIBtgIIIAtgIIIBtg mice on a background deficient in endogenous mouse FcR (FcRnull) . This approach unfortunately preserves the aberrant expression of individual transgenes, including exceptionally high hFcγRIIB on monocytes, erroneous expression of hFcγRIIB and hFcγRIIIA on eosinophils, hFcγRIIIA and hFcγRIIIB on some DCs, hFcγRIIIB on monocytes, and constitutive expression of hFcγRI on neutrophils [171-173]. Still, this study demonstrated that hFcγRItgIIAtgIIBtgIIIAtgIIIBtg mFcγRnull mice had normal spleen architecture, and could generate specific IgM and IgG antibody responses upon immunisation with a hapten-protein conjugate. In addition, these mice could mount cytotoxic effector functions via human IgG targeting B cells, T cells or platelets for depletion, or to eliminate tumour cells. Crucially, the injection of immune complexes formed by aggregated human IgG could trigger severe hypothermia and systemic shock in hFcγRItgIIAtgIIBtgIIIAtgIIIBtg mFcγRnull mice: that is, IgGdependent anaphylaxis  (see section Systemic inflammation: Anaphylaxis). This previous study did not, however, address the receptors, cells or mediators responsible for the models used therein.
The pathogenesis of endotoxic shock and the role of neutrophils
TLR4 expression on monocyte/macrophages and neutrophils drives the inflammatory cytokine response during mouse sepsis models: specific deletion of TLR on these cells demonstrated its requirement for effective bacterial clearance, but also ameliorated morbidity when combined with microbe clearance by antibiotics . Both monocyte/macrophages and neutrophils can produce the numerous inflammatory cytokines involved in the systemic response to endotoxin exposure: IL-1, IL-8, IL-6, TNF and IFN. These cells are also important sources of lipid mediators (PAF and eicosanoids) as well as nitric oxide, which can contribute to increased vascular permeability and a loss of vasomotor tone. The relative cell-specific source of cytokines and inflammatory mediators is difficult to delineate. However, considering that monocyte/macrophages express high levels of CD14, exhibit great capacity for pro inflammatory cytokine production, and are the dominant cell population in the peritoneal cavity, where systemic endotoxemia models are initiated, it is likely that these cells are the initial major source. Neutrophils may contribute to the inflammatory milieu after their recruitment following chemokine release by macrophages, cytokine stimulation, and accordingly increased CD14, whether expressed on neutrophils themselves or soluble CD14 in the environment.
Macrophages, as well as endothelial cells, smooth muscle cells, hepatocytes, and myocardiocytes, can synthesise NO from iNOS following LPS or inflammatory cytokine stimulation. iNOS-derived NO has a critical contribution to the vascular dysfunction associated with endotoxic shock. iNOS deficient mice were found to be resistant to endotoxin lethality , although this was refuted by others . Altogether the beneficial effects of NO inhibition during endotoxemia are highly debated, likely reflecting a functional multiplicity for this vasoactive compound during shock. Of all the inflammatory cytokines released during systemic inflammation, TNF is critical to the pathology of shock after endotoxin exposure [205, 206]. Despite this, TNF inhibitors are not under widespread usage to treat clinical shock, due to lack of efficacy in clinical trials. HMGB1 is a DAMP released in the later stages of shock and was observed to amplify LPS-induced inflammation , although recently it was shown that, contrary to antibody-mediated blockade, conditional HMGB1 ablation does not affect the shock phenotype following LPS challenge . Still, HMGB1 remained critically implicated in the amplification of neutrophil-mediated tissue injury during sterile inflammation. These findings exemplify many that highlight the complexity of unravelling inflammatory pathways during systemic endotoxemia.
The classical pathway of IgE-dependent mast cell and basophil activation: from the local reaction to systemic shock
High affinity IgE receptors FcεRI, expressed on mast cells and basophils, can bind monomeric IgE. In allergic individuals, it is understood that sensitisation elicits the production of allergen-specific IgE. Upon allergen re-exposure, the recognition of bivalent or multivalent antigen by FcεRI-bound IgE causes aggregation of FcεRI and triggers the activation of mast cells and basophils, resulting in the rapid and sustained release of diverse vasoactive mediators, including mast cell tryptase and histamine, and cytokine release . Certainly IgE has a prominent role in the development and maintenance of allergic inflammation, and activation of mast cells and eosinophils via specific IgE is a central event in many acute allergic reactions: highlighted by the diverse anti-inflammatory effects of anti-IgE treatment in patients with allergic asthma or allergic rhinitis, both of which are linked to reactions at mucosal or barrier surfaces.
The widely accepted paradigm of a systemic anaphylactic reaction considers that the pathophysiology is driven by IgE-dependent mast cell and basophil activation and histamine release.
In animal models, the transfer or passive sensitisation with specific IgE prior to challenge with the corresponding antigen can recapitulate the systemic signs of shock (passive systemic anaphylaxis, PSA): oedema, hypotension, loss of mobility, and severe hypothermia. IgE-induced PSA observed in wild type (wt) mice was abrogated in mice deficient for FcεRI  and in mast cell deficient W/Wv mice . It was also abrogated by pharmaceutical or genetic histamine inhibition, while intravenous injection of histamine alone can induce anaphylactic shock in mice. Anaphylaxis can proceed also through the human IgE receptor, as has been demonstrated using FcεRItg mice , and IgEdependent mast cell activation can contribute to severe passive systemic reactions in humanised mouse models .
The passive transfer of hypersensitivity can be achieved in humans, in 1921 Prausnitz and Künster demonstrated that intradermal injection of a serum regent (later identified as IgE) could transfer sensitivity to a reaction elicited by injection of a corresponding allergen (reviewed in ; Annex 7.4). As a test for patient allergy, a modified form of the cutaneous reaction is still used today: cutaneous anaphylaxis thereby results from a specific sensitisation protocol and route of allergen exposure. Systemic anaphylaxis in patients results from exposure not only at cutaneous surfaces (insect stings) but also at mucosal linings (gut epithelium) and systemically (injectable drugs). Indeed, systemic allergen absorption is necessary for anaphylactic shock to ingested allergens .
Considering that food and drug exposure accounts for the majority of anaphylactic reactions, and studies have indicated that drug and medication-induced anaphylaxis is the most common cause of
anaphylaxis fatalities [237, 238] it is necessary to understand the immunological mechanisms underlying these systemic reactions. IgE is only a very minor proportion of the total systemic immunoglobulin, present in the serum at less than 0.002 mg/mL, and with a very short half-life of the highest synthetic rate and longest biological half-life: IgG1 concentrations can range from 5- 12mg/mL.
Human IgE-independent anaphylaxis:
It remains under debate whether IgG can contribute to anaphylaxis in humans, and the relative importance of this immunopathological mechanism. Importantly, in the mouse anaphylaxis models described above, challenge is by administration of a relatively large amount of antigen via the intravenous route. The most likely case in which IgG-dependent anaphylaxis pathways may be invoked in humans, accordingly, is via intravenous exposure to injectable drugs and therapeutic agents.
Indeed, patients with antigen-specific IgG antibodies in the absence of IgE have been reported to experience anaphylactic reactions to mAb therapeutics [252, 253], aprotinin , dextran  or even total serum transfer  (reviewed in ). Anaphylaxis to the fluid resuscitation agent dextran, in particular, seems a prototypical IgG-dependent reaction, since dextran-specific IgE are rarely detected, and serum levels of IgG subclasses correlate highly with anaphylaxis severity .
Anaphylaxis following serum transfer resulted from IgG antibodies directed against IgA antibodies, in IgA-deficient individuals . Moreover, anaphylaxis associated with anti-IgA IgG has been documented in patients with common variable immunodeficiency (CVID) that received intravenous immunoglobulin therapy (IVIG) . Critically, a gain-of-function allele in FcγRIIA was associated with increased risk of anaphylaxis in CVID patients receiving IVIG therapy .
Platelet activating factor (PAF) is the dominant mediator responsible for mouse models of IgG- dependent anaphylaxis [250, 251] and is particularly associated with fatal outcomes [261, 262]. PAF has been proposed also as a central mediator in human anaphylaxis pathogenesis (reviewed in ). Patient studies indicate that PAF levels strongly correlate with anaphylaxis severity, and indeed that PAF provides a more specific and sensitive diagnostic marker than either mast cell tryptase or histamine . Furthermore, activity of its inactivating enzyme PAF acetylhydrolase was significantly lower in patients with severe and fatal reactions [230, 265]. A broad range of cells can release PAF, including leukocytes, lymphocytes and endothelial cells, and it can be secreted by mast cells and basophils following antibody-dependent activation. Neutrophils, however, are a major source of PAF, but also express PAF-R on the surface, thus PAF can have an autocrine effect on these cells to enhance release of other lipid mediators. PAF is therefore critically implicated in putative pathways of neutrophil-dependent anaphylaxis.
Patient anaphylaxis to neuromuscular blocking agents (NMBAs)
Particularly owing to the variability in presentation and eliciting agents, as well as the emergency nature of the reaction, anaphylaxis is very difficult to study in human patients. In cases of anaphylaxis arising in a clinical setting, however, more homogenous groups of patients may facilitate clinical studies, ie by controlled allergen exposure with defined route, dose and timing. Immediate hypersensitivity reactions during the perioperative period have been reported with increasing frequency, and may be attributable to anaesthetic drugs, antibiotics, latex, antiseptics, radio-contrast agents, colloids for intravascular volume expansion, blood products or disinfectants. The mostcommon causes are neuromuscular blocking agents (NMBAs, 60-70%), followed by latex (12-18%) and antibiotics (8-15%) .
To investigate the potential contribution of IgG-mediated pathways to anaphylaxis in a patient cohort, one may therefore consider anaphylaxis to drugs, e.g. curare-based NMBA. The most common cause of anaphylaxis during surgery, the incidence of NMBA-dependent reactions lies between 1 in 1,250 and 1 in 18,000 surgeries, with substantial geographical variability , and moreover, these reactions are fatal in up to 10% of cases [268-270]. Specific IgE, and also specific IgG, have been detected in the sera of patients who developed shock to NMBAs. In addition, allergen-specific IgE may be absent in 10-15% of patients [271, 272].
Table of contents :
1.1 Myeloid cells of the innate immune system
1.1.2 Monocytes, Macrophages and Mononuclear Phagocytes
1.1.3 Mast cells, Basophils and Eosinophils
1.2 Neutrophil death & inflammation resolution: a focus on lipid mediators
1.3 Antibodies & their receptors: conferring innate immune cells with adaptive specificity
1.3.1 B cells and the BCR
1.3.2 Antibodies and their classes
1.3.3 IgG antibody receptors (FcgR)
1.4 Systemic Inflammation
1.4.1 Inflammation-associated circulatory shock
2 Summary and objectives
3 A novel model of inducible neutropenia reveals a protective role for neutrophils during systemic inflammation
3.1 PAPER I
4 Neutrophils contribute to IgG-dependent anaphylaxis in FcγR-humanised mice
4.1 PAPER II
4.2 Considering high affinity FcγRI: the Audrey mouse
4.2.1 Audrey mice exhibit hFcγRI expression patterns comparable to that of humans, and retain hFcγRIIA/IIB/III expression of VG1543 mice
4.2.2 hFcγRI does not contribute to IVIG-PSA in Audrey mice, which proceeds via a dominant pathway involving hFcγRIIA, Neutrophils and PAF
4.2.3 Supplemental Methodology and Data
5.1 Part I: A protective role for neutrophils in LPS endotoxemia
5.2 Part II: The inflammatory effects of neutrophils: novel mouse models to study neutrophil function in vivo
5.2.1 A novel model of inducible neutropenia: PMNDTR mice
5.2.2 PMNDTR mice to study neutrophils in antibody-dependent pathologies: deciphering the contribution of neutrophils to systemic anaphylaxis
5.2.3 Audrey & humanised mouse models to study FcγR: limitations and potential
5.3 Part III: Neutrophils as protective or pathological agents of systemic inflammation
5.4 Part IV: Towards the clinic: systemic anaphylaxis to neuromuscular blocking drugs
5.4.1 Developing a mouse model of systemic anaphylaxis to Rocuronium Bromide
5.4.2 Evidence from the clinical study NASA: Neutrophil activation in systemic anaphylaxis
5.5 Final Considerations and Perspectives
7.1 IgG subclasses determine pathways of anaphylaxis in mice
7.2 In vivo effector functions of high-affinity mouse IgG receptor FcgRI in disease and therapy models
7.3 Review – Contribution of human FcγRs to disease with evidence from human polymorphisms and transgenic animal studies
7.4 Book chapter – Anaphylaxis (Immediate hypersensitivity): from old to new mechanisms