Regulation of antibody effector functions through IgG Fc N-glycosylation

Abstract

Immunoglobulin gamma (IgG) antibodies are key effector proteins of the immune system. They recognize antigens with high specificity and are indispensable for immunological memory following pathogen exposure or vaccination. The constant, crystallizable fragment (Fc) of IgG molecules mediates antibody effector functions such as complement-dependent cytotoxicity, antibody-mediated cellular cytotoxicity, and antibody-dependent cell-mediated phagocytosis. These functions are regulated by a single N-linked, biantennary glycan of the heavy chain, which resides just below the hinge region, and the presence of specific sugar moieties on the glycan has profound implications on IgG effector functions. Emerging knowledge of how Fc glycans contribute to IgG structure and functions has opened new avenues for the therapeutic exploitation of defined antibody glycoforms in the treatment of cancer and autoimmune diseases. Here, we review recent advances in understanding proinflammatory IgG effector functions and their regulation by Fc glycans.

Immunoglobulins

Immunoglobulins (Igs) are glycoproteins secreted by B cells and plasma cells and constitute one of the main effector mechanisms of the adaptive immune system. Igs evolved to specifically recognize target structures (antigens) and mediate appropriate actions by communicating with cellular and humoral components of the immune system. Antigen recognition is mediated by the “fragment antigen binding” (Fab) domains which contain the complementarity-determining regions (CDRs) located in the N-terminal region of heavy chains (HCs) and light chains (LCs) (Fig. 1, IgG). These areas are characterized by a high degree of variability in amino acid composition in between antibodies, which in turn leads to a broad spectrum of potential binding partners. The characteristic architecture of Igs, composed of disulfide bond-stabilized β-sheets as well as inter-chain disulfide bonds, guarantees the structural integrity required for the functionality of antibodies. The C-terminal regions of the two HCs constitute the “fragment crystallizable” (Fc), which contains the binding sites for immune effector molecules such as the complement system or Fc receptors. In some antibody types, including IgG, the Fab and Fc regions are separated by a less-structured stretch of amino acids called the “hinge region” which provides flexibility to the antibody and contains the disulfide bond(s) linking the two HCs [1, 2].

Fig. 1.

Structure of the immunoglobulin G (IgG) molecule. IgG is composed of two heavy and two light chains linked by disulfide bonds. The antigen-binding fragment (Fab) consists of two moieties with identical structure which define the antigen-specificity through their complementarity-determining regions (CDR). The crystallizable fragment (Fc) mediates antibody effector functions through binding to Fc receptors and interaction with the C1q component of the complement system. Each IgG molecule contains a single, highly conserved IgG-Fc N-glycan in each of the two CH2 domains (Fc glycan) and may carry additional glycans in the antigen-binding sites (Fab glycans). CH constant heavy, CL constant light, Fab antigen-binding fragment, Fc crystallizable fragment

Being amongst the most abundant serum proteins [3], immunoglobulins fulfill essential functions in protecting our body against invading pathogens. Depending on the type and stage of an infection as well as the anatomical site, different functional properties are required. This functional diversification is achieved by a process called antibody class switching. Thereby, tightly regulated DNA recombination events lead to the excision and replacement of the antibody’s constant region while keeping the CDRs and, therefore, the specificity of the antibody largely unchanged. The different constant regions are called isotypes and are differentiated in five classes (IgA, IgD, IgE, IgG and IgM) and six subclasses (IgG1–4 and IgA1–2). The decision which type of antibody is produced is dependent on the signals a B cell encounters during its maturation towards an antibody-secreting cell (ACS). A naïve mature B cell, previously selected in the bone marrow and spleen for functional integrity of the cell-surface-bound antibody (B-cell receptor, BCR) and low self-reactivity, expresses cell-surface IgD and/or IgM. Alternative splicing of the primary VH transcript results in the secreted antibody, which now lacks the cytoplasmic tail and the transmembrane region of the BCR. This allows B cells to simultaneously produce cell-surface-bound- (BCR) and secreted (antibodies) immunoglobulins with identical binding specificities. Upon cognate antigen encounter, B cells can enter several developmental programs eventually resulting in further diversification of the antibody repertoire by somatic hypermutation (SHM), a process which leads to mutation of the CDR sequences, and differentiation into long-lived memory B cells, terminally differentiated long-lived ASC (plasma cells) or short-lived ASC (plasmablasts). The sites of antigen-induced B-cell differentiation are secondary lymphoid organs such as the spleen, lymph nodes or gut-associated lymphoid tissues. Differentiation can take place independent of T cells, resulting in little SHM and limited class switch or in a T-cell dependent manner, which typically involves the generation of germinal centers where extensive SHM takes place and switching to all classes of antibodies can occur. The nature of the antigen, anatomic location, signaling via pattern recognition receptors and the cytokine milieu are crucial for the choice of the differentiation program and the decision which antibody isotypes develop [4]. Antibodies fulfill important functions within the immune system such as neutralization of toxins or microbes and assisting the killing of transformed cells or bacteria by opsonization, antibody-dependent cell-mediated cytotoxicity (ADCC), antibody-dependent cell-mediated phagocytosis (ADCP) or complement-dependent cytotoxicity (CDC) (Fig. 2). This review puts a focus on the generation and functions of the most abundant human antibody class in circulation, immunoglobulin G (IgG) and highlights the aspects of how these effector functions are regulated.

Fig. 2.

Effector functions of IgG antibodies. In addition to neutralization initiated by binding of the Fab domain to target molecules, the antibodies’ Fc fragment mediates IgG effector functions such as killing of transformed cells or bacteria by opsonization, antibody-dependent cell-mediated cytotoxicity (ADCC), antibody-dependent cell-mediated phagocytosis (ADCP) or complement-dependent cytotoxicity (CDC)

Immunoglobulin G

IgG is the prototypic antibody composed of two HCs and two LCs linked by disulfide bonds. Each IgG molecule contains a single, highly conserved IgG-Fc N-glycan in each of the two CH2 domains and may carry additional glycans in the antigen-binding sites [5, 6]. One reason for its predominance in serum is the exceptionally long half-life of 21 days due to recycling via the neonatal Fc receptor (FcRn) as loss of FcRn dramatically reduces the half-life of IgG [7–9]. FcRn binds to internalized IgG at low pH (6.5) and recycles to the cell surface where higher pH leads to the release of IgG back into circulation [8].

The four human IgG subclasses are named according to their frequency in serum IgG-1, -2, -3, and -4 and share more than 90 % amino acid (AA) sequence homology with some important differences: the cysteine- and proline-rich hinge region, which contains inter-HC disulfide bonds and determines Fab-arm flexibility, is the hotspot of diversity. IgG3 has the longest hinge constituted of up to 62 AA and 11 disulfide bonds (exact numbers vary in allelic variants called allotypes). The extended length allows the Fab domain of IgG3 to have a great amount of conformational flexibility relative to its Fc [2, 10]. In contrast, IgG2 has a rigid hinge composed of only 12 AA containing four disulfide bonds making it the IgG isotype with the most restricted Fab-arm flexibility [2, 10]. IgG1 (15 AA hinge, 2 disulfide bonds) and IgG4 (12 AA, 2 disulfide bonds) have intermediate properties [2, 10]. Further important AA differences are located in the binding region for complement proteins and Fc receptors, the CH2 domain of the HC. Additional AA sequence variations stem from allelic differences in the human population. These, if immunogenic, are called allotypes, some of which can influence functional properties, in particular for IgG3 [2, 11].

Fc-dependent effector functions of IgG

In addition to binding antigen via their Fab fragments, IgG antibodies regulate immune responses through their Fc domain (Fig. 2). First, IgG can initiate the activation of the complement pathway, resulting in the generation of the proinflammatory anaphylatoxins C3a and C5a and the membrane attack complex which may lead to lysis of the target cell by complement-dependent cytotoxicity (CDC). Second, IgG autoantibodies can cross-link cellular Fc receptors specific for IgG (FcγRs) that are present on most innate immune effector cells, including neutrophils, mast cells and macrophages. FcγRs mediate important IgG effector functions such as induction of antibody-dependent cellular cytotoxicity and opsonization and phagocytosis of antigens.

Receptors binding to the Fc domain of IgG are called Fc-gamma receptors (FcγRs) (Fig. 3). Humans express five or six classical FcγRs which are grouped into activating [FcγRI, FcγRIIa (FcγRIIc expressed by some individuals [12]) and FcγRIIIa], inhibitory (FcγRIIb) and glycosylphosphatidylinositol (GPI)-anchored (FcγRIIIb) Fc receptors (Fig. 3). With the exception of T cells, all major immune cells express FcγRs, which allows antibodies to explore the functions of many cells and exhibit a wide range of effector mechanisms. IgG isotypes bind to FcγRs with their CH2 domain [13–16]. Differences in the CH2 AA composition [17], Fc glycan structure [18, 19] and hinge region [15, 17] influence IgG binding to FcγRs. Activating FcγRs contain immunoreceptor tyrosine-based activation motifs (ITAM) in their intracellular domain or their adaptor proteins, whereas FcγRIIb contains an immunoreceptor tyrosine-based inhibitory motif (ITIM), which allows it to counteract activating signals. Based on the receptor`s affinity for IgG, FcγRI is referred to as “high affinity FcγR”, whereas all other FcγRs are considered “low affinity FcγRs”. The high affinity results in FcγRI being constantly associated with IgG, whereas firm binding to low-affinity FcγRs requires the formation of an immune complex [20]. Although due to the high serum-IgG concentration, even low-affinity FcγRs are almost saturated with IgG in blood [21], the faster off-rates allow cells expressing these receptors to rapidly sample IgGs in solution [22–25].

Fig. 3.

Human Fc receptors for IgG (FcγR). FcγRs differ in their cellular distribution, their function and their affinity for the IgG-Fc fragment

FcγRI (CD64) is expressed by myeloid cells (monocytes, DCs, macrophages) and granulocytes (neutrophils, mast cells) and binds with very high affinity to IgG1, 3 and 4 but does not bind IgG2. Signaling by FcγRI leads to the activation and differentiation of monocytes towards monocyte-derived DCs and may contribute to antigen presentation to T cells [26]. FcγRI is deregulated in several antibody-mediated autoimmune diseases implying a potential role in disease pathology [27, 28]. FcγRI-bound monomeric IgG is constantly internalized and recycled to the cell surface, whereas cross-linking with the cognate antigen leads to internalization and degradation [29]. Despite its high affinity for monomeric IgG, mice transgenic for human FcγRI suggest that the receptor retains its ability to bind IgG-ICs in vivo and contributes to IC-mediated cell activation [30].

FcγRIIa is expressed by granulocytes, monocytes, macrophages, DCs and platelets [31]. FcγRIIc is encoded by a gene that resulted from the crossover of fcgr2a and fcgr2b [32]. It is only found on NK cells, and allelic polymorphisms result in FcγRIIc being expressed by only approximately 45 % of individuals [12, 33] where it can contribute to cytotoxicity [33].

The FcγRIII (CD16) exists in two alternative forms encoded by two different genes, a transmembrane FcγRIIIa expressed on natural killer (NK) cells and macrophages, and a GPI-linked FcγRIIIb present on neutrophils [34, 35]. The activating low-affinity FcγRIIIa (CD16) mediates antibody-dependent cellular cytotoxicity (ADCC) and is highly expressed on the cytotoxic CD56^dim CD16⁺ NK-cell subset as well as on other hematopoietic cells (Fig 3). NK cells are thought to be the key mediators of ADCC, a mechanism harnessed in monoclonal antibody treatments of various cancers overexpressing unique antigens, such as neuroblastoma, breast cancer, B-cell lymphoma, and others.

FcγRIIb is the only FcγR expressed on B cells and plasma cells. Its co-engagement with the BCR delivers an inhibitory signal and can, therefore, be envisioned as a negative feedback from circulating antibodies, which may limit B-cell differentiation and antibody production [36]. Its importance for B-cell homeostasis is highlighted by the autoimmune susceptibility of mice lacking FcγRIIb [37] and reduced expression on B cells derived from patients with various antibody-mediated autoimmune diseases [27, 38, 39]. FcγRIIb expression on murine plasma cells regulates their persistence and apoptosis [40] and limits IgG autoantibody production [41].

On myeloid cells and granulocytes, such as monocytes, dendritic cells, neutrophils and basophils, FcγRIIb can be co-expressed with activating FcγRs [32]. Immune complex-mediated activation of monocytes was shown to be negatively regulated by FcγRIIb [42, 43], and consequently, absence of FcγRIIb on murine DC results in increased T-cell priming [36]. These studies suggest that loss of balanced FcγR signaling could result in uncontrolled responses that can lead to the damage of healthy tissues and the initiation of autoimmune disease.

Complement-mediated IgG effector functions

Complement activation is one of the most effective defense mechanisms of the immune system. It is initiated by the binding of so-called initiators of the complement system such as C1q, ficolins or mannose-binding lectin to target surfaces like infected cells or microbes. Their binding leads to a cascade of events ultimately culminating in the formation of the membrane attack complex and lyses of the target. IgG can assist the initiation of the complement system in three ways. (1) C1q binds with high affinity to antigen-bound, but not monomeric IgG [44, 45], (2) IgG carrying IgG-Fc N-glycan terminating in mannose may additionally be able to bind mannose-binding lectin [46, 47] and (3) C3b, a downstream component of the complement cascade, can directly bind to IgG [48]. IgG isotypes have different C1q binding affinities with IgG3 binding most potently followed by IgG1, very weak binding for IgG2 and none for IgG4 [49]. The main interaction points of C1q are located in IgG’s CH2 and are constituted by residues which are mostly conserved between isotypes [13, 50]. The structure of the hinge region, the IgG-Fc N-glycan and the relative orientation of the Fab domain influences the affinity of C1q and may explain binding differences between IgG isotypes [44, 51].

Fc glycan-modulated IgG effector functions

Human immunoglobulins are Fc-glycosylated and can, depending on the isotype and the sequence of the antigen-binding regions (complementarity-determining regions, CDR), carry additional glycans in the Fab domains. IgG is unique with respect to a single, highly conserved asparagine 297 (N297)-glycosylation site which points towards a hole in the Fc region formed by the CH2 and CH3 domains.

During protein translation, a pre-formed lipid-linked glycan is transferred and covalently attached to N297 in the lumen of the endoplasmic reticulum (ER). This initial glycan is composed of two N-acetylglucosamines (GlcNAc) followed by nine mannose (Man) and three glucose (Glc) residues [52, 53]. Its structure is highly conserved in eukaryotes and serves as an important mechanism for protein folding and quality control of proteins carrying N-glycans [52]. If folded properly, the IgG polypeptide is transferred from the ER to the Golgi, where glycosyl-hydrolases and -transferases can modify the glycan structure leading to such diverse and highly complex glycans as seen in the IgG Fc. In addition to the oligosaccharide core, more than 95 % of the biantennary complex-type structure of the final IgG glycan carries an N-acetylglucosamine on both arms [54, 55] and 85 % are fucosylated [56] (Fig. 4). In contrast, the presence of galactose is less homogenous with 40 % of glycans carrying one galactose (G1 glycan) and the frequency of non-galactosylated (G0) or bi-galactosylated glycans (G2) ranging between 20 and 40 % depending on age and gender [55, 57, 58]. The most distal sugar on the glycan is sialic acid (neuraminic acid, Neu5Ac). Around 5–10 % of glycans carry sialic acid on one arm, and approximately 1 % of serum IgG-Fc glycans are bi-sialylated [5, 58, 59]. In addition to the Fc domain, roughly 15–20 % of human serum IgGs are glycosylated in their Fab domain [60, 61]. The functional significance of Fab glycosylation is incompletely understood, but it has been suggested that it might impact binding affinities of antigen–antibody interactions [62, 63].

Fig. 4.

Structure of a fully processed IgG-Fc N-glycan. The Asparagine 297-linked complex-type glycan is located within the CH2 domains of the Fc fragment and consists of a complex, biantennary structure. In vivo, such a fully processed glycan will be found only in trace amounts as the majority of antibodies will carry either no, one or two galactose residues and a fraction of those carrying galactose will additionally possess sialic acid

Removal of the entire Fc N-glycan impairs antibody effector functions [13, 64], and the presence or absence of distinct IgG-Fc monosaccharides was shown to regulate IgG effector functions.

Fc fucose

The majority of circulating IgG antibodies are fucosylated [56] which, compared to afucosylated isotypes, reduces IgG´s binding affinitiy for the activating FcγRIII (CD16) and thereby its potential to induce antibody-dependent cellular cytotoxicity [19]. Fucosylation also appears to impair antibody-dependent cell-mediated phagocytosis [65, 66]. Consequently, clinical trials using afucosylated monoclonal antibodies were initiated and showed improved efficacy in target cell depletion [67–69].

Fc galactose

In adoptive transfer models of autoimmune diseases, non-pathogenic doses of autoantibodies become pathogenic when present as agalactosyl glycoforms [18, 70, 71]. Decreased levels of galactosylation are associated with several chronic inflammatory diseases such as rheumatoid arthritis, systemic lupus erythematosus, multiple sclerosis, and in patients with tuberculosis [55, 72–74]. In contrast, increased IgG-Fc galactosylation is detectable during pregnancy, and in rheumatoid arthritis patients who experience pregnancy-induced remission [75, 76] suggesting that Fc galactosylation might exert anti-inflammatory functions. In line with this hypothesis, a recent study showed that high galactosylation of IgG immune complexes in mice promotes the association of FcγRIIB and dectin-1, which blocks the proinflammatory effector functions of C5aR and CXCR2 [77]. The observation that antiviral activity and spontaneous control of HIV infection are associated with increased prevalence of total and antigen-specific agalactosylated antibodies additionally argues for a functional significance of antibody galactosylation in humans and is in line with the assumption that lack or loss of IgG Fc-linked galactose occurs during or promotes inflammation [6, 78]. Along these lines, Ho et al. recently reported that in patients with chronic hepatitis B low IgG-Fc galactosylation levels are associated with high-grade liver inflammation and fibrosis, suggesting that IgG-Fc galactosylation might be a potential noninvasive indicator of severe liver necroinflammation and fibrosis [79]. One important aspect to consider when investigating the impact of galactose is that it provides the basis for the addition of sialic acid, the most distal sugar moiety on the IgG-Fc glycan.

Fc sialic acid

Similar to Fc galactosylation, decreased levels of IgG sialylation are observed in chronic autoimmune diseases such as rheumatoid arthritis, juvenile idiopathic arthritis and Wegener’s granulomatosis [55, 80–82]. It has also been demonstrated that IgG sialylation increases during pregnancy and that this increase may be associated with the remission of rheumatoid arthritis during pregnancy [83]. While the aforementioned data clearly support the hypothesis that these terminal sugar residues are involved in modulating antibody activity, they suggest that it is not the lack of galactose residues itself but, rather, the concomitant absence of terminal sialic acid residues that may be responsible for the enhanced inflammatory activity exerted by aglycosylated glycoforms [18, 70]. The relevance of sialic acid residues for modulating immune responses is highlighted by the finding that intravenous immunoglobulins (IVIG) completely lose their immunosuppressive capacity upon removal of Fc-sialic acid residues by neuraminidase treatment in experimental autoimmune disease models [18]. Conversely, IVIG preparations as well as isolated Fc fragments enriched for terminal sialic acid residues appear to have a more than tenfold higher anti-inflammatory activity [70, 84]. Proposed mechanisms that mediate anti-inflammatory activities of Fc sialylation include the induction of an anti-inflammatory cytokine milieu following binding of sialylated IgGs to the murine C-type lectin receptor SIGNR1, which in turn induces interleukin (IL)-33 and IL-4 production and eventually leads to the upregulation of the inhibitory FcγRIIb on macrophages, thereby limiting antibody-mediated immunopathologies [85]. However, conflicting results exist concerning the ability of SIGNR1 and its human homolog DC-SIGN to recognize the sialylated IgG-Fc glycan [86, 87] and if these receptors are required for the anti-inflammatory properties of sialylated Fc [88]. In humans, Fc sialylation reduces proinflammatory IgG effector functions such as complement-dependent cytotoxicity (CDC) by inhibiting the binding of the antibody’s CH2 domain to C1q [89]. Thus, the mechanisms that mediate anti-inflammatory properties of Fc sialylation are not fully understood and might involve Fc receptor-dependent and -independent mechanisms. Therapeutic implications of the aforementioned findings lie in the possibility to modify Fc glycosylation to increase the anti-inflammatory efficacy of both IVIG and monoclonal antibody-mediated immunotherapies. It remains to be evaluated whether Fc sialylation can be harnessed to improve anti-inflammatory efficacy and the clinical response to IgG-mediated treatment strategies.

Therapeutic relevance and concluding remarks

The aforementioned studies clearly demonstrated that minor structural changes in IgG-Fc glycosylation profoundly affect antibody effector functions and opened up new opportunities for designing therapeutic antibodies with increased efficacies. Defucosylated antibodies which enhance ADCC are currently evaluated and increasingly utilized in cancer therapy. Obinutuzumab, a glycoengineered anti-CD20 antibody with reduced fucosylation and increased bisecting GlcNAc, has recently been approved as first-line treatment for patients with chronic lymphocytic leukemia (CLL) and follicular lymphoma (FL) who did not benefit from treatment with a fucosylated anti-CD20 antibody, i.e., rituximab [90]. Removal of the N-glycan impairs FcγR binding and complement activation, and this strategy has been increasingly recognized as a targeted treatment for autoimmune conditions. In vivo administration of the bacterial IgG glycan-hydrolysing enzyme EndoS, which cleaves the linkage between the two GlcNAc residues in the core of the N-linked glycan, was shown to ameliorate the development of various experimental models of autoimmune diseases [91–94], and this appears to be safe and well-tolerated in preclinical models [95] and represents a potential strategy to limit antibody-mediated autoimmune disease conditions [96].

The anti-inflammatory activity of sialylated IgGs was first demonstrated for intravenous immunoglubulins (IVIG): in contrast to fully sialylated IVIG preparations, desialylated IVIG failed to suppress autoimmune disease development in an antibody-mediated experimental arthritis model [18]. Subsequent studies confirmed the protective and crucial role of IVIG sialylation in various experimental autoimmune disease conditions [84, 87, 97, 98]. Fc-sialylated glycovariants were shown to mediate upregulation of the inhibitory FcgRIIB [18], to block B-cell proliferation independent of Fc receptors [99] in mice and to limit proinflammatory IgG effector functions through impairment of CDC in humans [89]. A scalable process to produce fully sialylated IVIG with consistent enhanced anti-inflammatory activity has recently been described, and the safety and efficacy of fully sialylated IgG will soon be evaluated in clinical trials [84]. Such trials, if well designed, are instrumental to evaluate the biological significance of IgG-Fc N-glycan modifications in human diseases and might generate strategies for tailoring IgG-based recombinant antibodies for the treatment of cancer and autoimmune diseases.

Compliance with ethical standards

Conflict of interest

The authors declare no conflict of interest.