Antibody-mediated complement activation in pathology and protection

Abstract

Antibody-dependent complement activity is associated with autoimmune morbidity, but also anti-tumor efficacy. In infectious disease, both recombinant monoclonal antibodies and polyclonal antibodies generated in natural adaptive responses can mediate complement activity to protective, therapeutic, or disease-enhancing effect. Recent advances have contributed to the structural resolution of molecular complexes involved in antibody-mediated complement activation, defining the avid nature of participating interactions, and pointing to how antibody isotype, subclass, hinge flexibility, glycosylation state, amino acid sequence, and the contextual nature of the cognate antigen/epitope are all factors that can determine complement activity through impact on antibody multimerization and subsequent recruitment of C1q. Beyond the efficiency of activation, complement activation products interact with various cell types that mediate immune adherence, trafficking, immune education, and innate functions. Similarly, depending on the anatomical location and extent of activation, complement can support homeostatic restoration or be leveraged by pathogens or neoplasms to enhance infection or promote tumorigenic microenvironments, respectively. Advances in means to suppress complement activation by intravenous immunoglobulin (IVIG), IVIG mimetics, and complement-intervening antibodies represent proven and promising exploratory therapeutic strategies, while antibody engineering has likewise offered frameworks to enhance, eliminate, or isolate complement activation to interrogate in vivo mechanisms of action. Such strategies promise to support the optimization of antibody-based drugs that are able to tackle emerging and difficult-to-treat diseases by improving our understanding of the synergistic and antagonistic relationships between antibody mechanisms mediated by Fc receptors, direct binding, and the products of complement activation.

Graphical Abstract

Determinants & Outcomes of Antibody-mediated Complement Activation. This review discusses the factors specific to antibodies, antigenic targets, pathogen defenses, and tissue microenvironments that determine the consequences of complement activation. Furthermore, it explores some approaches to capture these complexities in in vitro assays and screens of antibody-mediated complement functions in host-neoplasm, host-graft/auto-host, and host-pathogen interactions. Image reproduced with permission⁴

Introduction

The complement system is an evolutionarily ancient defense network that is distributed nearly ubiquitously throughout host extracellular tissues and intracellularly. This integral innate immune system is comprised of more than thirty plasma proteins and cell surface receptors that collectively enhance or complement the ability of antibodies and innate immune cells to clear microbes and host cellular debris, promote inflammation, and directly disrupt the membrane of microbial pathogens and enveloped viruses¹ (Figure 1). Following initiation by one of three canonical mechanisms - the alternative, lectin, or classical pathways - a proteolytic cascade involving zymogens and amplifying feedback loops results in the intermediate outcome of coating or opsonizing the target in complement protein fragments that can interface with both innate and adaptive immune elements, and drive a variety of possible terminal outcomes including target lysis¹ (Figure 1a).

Figure 1: Overview of Complement Initiation Pathways & Possible Outcomes.

Figure reproduced with permission¹.

(a) The complement cascade be initiated by three defined pathways, which commonly amplify through the C3 convertase, and terminate in target opsonization and/or membrane disruption. The classical pathway denotes activation of the complement cascade by the multi-functional C1q molecule, where C1q-mediated activation was classically defined by an initial recognition event by C1q of target-bound antibody. The Lectin pathway is initiated when specific polysaccharides are directly recognized by mannose-binding lectin (MBL), collectin-11 (CL-11), or by members of the ficolin family (Fcn). The alternative pathway is generally characterized by low levels of continual activation that is typically tempered by host regulatory proteins to protect from self-inflicted tissue damage. (b) Gradually, the products of this low-level activation build up on surfaces lacking host complement regulatory proteins and reach a crescendo, resulting in (c) exponential amplification and cascade progression leading to (d, e) soluble inflammatory molecule generation, target opsonization, lytic insertion of the membrane attack complex (MAC), destruction, adherence and trafficking, adaptive immune education, and host clearance. Ag: antigen; Ab: antibody; C1q: complement component 1q; C1r/s: complement component 1q-associated serine proteases r and s; MBL: mannose binding lectin; MASP: MBL-associated serine proteases; Fcn: ficolins; CL-11: collectin-11; PAMP: pathogen-associated molecular pattern; FP: factor P (properdin); C3(H₂O): hydrolyzed C3; FB: factor B, FD: factor D; RCA: regulators of complement activation; C4: complement component 4; C2: complement component 2; FI: factor I; CR1: complement receptor 1 (CD35); C3dg: complement fragment 3dg; C3aR: C3a receptor; C5aR1 and 2: C5a receptor 1 and 2 (CD88 and C5L2, respectively); CRIg: CR of the immunoglobulin family; MAC: membrane attack complex (C5b-9); TCR: T-cell receptor; BCR: B-cell receptor.

The classical complement cascade represents a powerful link between adaptive antibody-mediated immunity and effectors of innate immunity. In fact, target-coated complement fragments can regulate adaptive effector B and T lymphocyte development via immune complex trafficking and antigen presentation, thus augmenting adaptive responses to achieve sustained and robust immune memory¹ (Figure 1d). Despite being the subject of decades of investigation, understanding of the complement system is by no means complete, and novel activities and relationships continue to be discovered and defined for diverse complement molecules and activation byproducts. In particular, as monoclonal antibodies now represent a growing and effective class of clinical intervention, understanding the diverse interactions and outcomes of their interactions with the complement system is of high translational relevance.

Antibody-Mediated Complement Activation (AMCA)

Immune complexes (ICs) formed by antigen-bound IgA, IgG, and/or IgM are able to activate complement². While IgA and IgM isotypes are inherently high avidity immunoglobulins that can recruit pathway initiation molecules and activate complement upon stable surface binding, we will focus on IgG. In its monomeric state, IgG can exhibit bivalent antigen recognition but can only weakly interact with a single globular head of the C1q hexamer³.

Classical and Lectin pathway initiation molecules C1q and mannose-binding lectin (MBL) or ficolins, respectively, share the structural feature of multivalent recognition limbs. Avid recognition functions to limit non-specific engagement by requiring a high local concentration of danger signals, and to increase the stability of the complex by decreasing the off-rate proportional to the number of target-engaged limbs^{4, 5}. In fact, monovalent binding affinity between these pattern recognition molecules and their targets is negligibly low^{3, 6}, requiring antigen clustering for avid engagement of MBL/ficolins, and antigen-driven antibody clustering for avid C1q engagement. Therefore, the structure of these cascade initiation molecules facilitates sustained cascade initiation upon specific target recognition. While MBL and ficolins recognize IgM⁷, which is in some sense pre-oligomerized, and carbohydrate components of bacterial and fungal cell walls that are usually present in regularly repeating patterns, the same cannot be said of IgG nor all of its targets, respectively. This puts the onus on antibody clustering on the target surface to facilitate avid C1q binding and efficient activation, suggesting that important factors in antigen selection would include expression level, spatial distribution, and membrane mobility. Important antibody factors (Figure 2a) would include antigen-binding and Fc valency as defined by the intrinsic affinity between Fc and C1q, and between antibody and antigen, as well as spatial aspects affected by epitope placement, angle of binding, and hinge length and flexibility between antigen-binding and C1q-binding (Fc) domains^{5, 8–10} (Figure 2b).

Figure 2: Structural Determinants of Antibody-mediated Complement Activation (AMCA).

(a) IgG1 antibody structure denoting antigen-binding (Fab) and Fc Receptor-binding (Fc) domains and antibody-intrinsic factors known to affect the efficiency of AMCA. Boxed regions indicate the site of interactions with C1q globular heads (orange) and between Fc domains (purple). Amino acid residues known to impact Fc-C1q (orange) or Fc-Fc (purple) interactions in the context of IgG1 numbering are indicated. (b) Structure of C1q globular head (orange) and IgG1 Fc (purple) complex. (c). Structure of Fc-Fc dimer complex. (d). (Left) Top down view of Fc hexamer (purple) interacting with six C1q globular heads (orange). (Right) Side view of C1 complex C1r and C1s (blue). IgG1 Fc, PDB 1HZH. IgG1 Fc and gC1q co-complex, PDB 6FCZ. Cryo-EM reconstruction of C1-IgG1 complex, EMD-4232, reproduced with permission¹³⁰.

Appreciation of such molecular determinants of classical complement activation has been dramatically refined with insights provided by the structure of the C1-(IgG-Fc)₆ hexamer complex formed at the binding surface⁴, which directed attention to Fc-Fc interactions of adjacently-bound antibodies (Figure 2c). While the formation of IgG hexamers was suggested by the crystal structure of an IgG¹¹, Diebolder et al. illustrated that IgG-RGY, IgG containing three point-mutations – E345R, E430G, and S440Y – in the CH2-CH3 interface, enhanced non-covalent Fc-Fc interactions to the point where IgG spontaneously oligomerized in solution in the absence of antigen, recruiting C1 and resulting in classical complement activation. The fact that IgG oligomerization is fundamental to the efficiency of complement activation explains, in part, the dependency of epitope specificity, and the ability of antibody to cluster antigen into lipid rafts on antibody-mediated activation^{12, 13}. A notable observation regarding epitope specificity’s contribution to differential complement activity is the impact of antibody distance from the cell membrane, where membrane-proximal epitopes appear to be favored⁹.

Beyond such sophisticated spatial factors, a number of antibody-intrinsic factors are known to impact the ability of a given immunoglobulin to exhibit Antibody-Mediated Complement Activation (AMCA). The four subclasses of IgG (e.g. IgG1, IgG2, etc.), display distinct complement activation profiles. In general, IgG1 and IgG3 are considered to be the most efficient activators, while IgG2 can be weakly activating, and IgG4 is often considered to be somewhat incapable of driving complement activation². Additionally, IgG Fc domains contain a conserved N-linked glycan in the vicinity of the C1q binding interface, which is required for efficient C1 activation². Interestingly, the effect of deglycosylation on reduced C1q binding has been linked to the glycans’ role in stabilizing the structure of Fc, and mediate interactions impacting IgG oligomerization rather than a direct impact of glycan on C1q-Fc binding affinity^{2, 5}. Beyond this requirement for glycan, controlled studies of monoclonal antibodies with modified glycosylation profiles have enabled mapping of the impact of specific glycoform compositions on complement activation. These studies have revealed that, whereas Fc-galactosylation enhances the C1q-binding, C4b-fixation, and downstream effects, fucosylation and bisection have no effect, and conflicting results have been reported as to the effect of Fc sialylation^14–16. Interestingly, glycoforms lacking terminal sialic acid and galactose residues have been shown to bind with higher affinity for lectin pathway initiation molecules¹⁷.

While Fc features such as these offer useful rules of thumb for anticipating AMCA, it has been long clear that they are not adequate to accurately predict the function of a given antibody of interest. Overall, the diversity of antibody-intrinsic factors that regulate classical complement activation, from structural and Fc glycosylation differences dictating adjacent Fc-to-Fc association to inherent affinity for antigen, antibody valency, and epitope specificity point to difficulty in faithful assessment of AMCA in vitro, even before subsequent impacts are considered.

Downstream Signal Amplification and Outcomes

The magnitude of activation depends on the balance between pathway activation complexes and complement inhibitory factors. The net magnitude and surface distribution of activation will determine the effector outcome(s). A product of classical initiation, and a commonly used tissue biomarker in autoantibody-mediated pathology diagnoses, C4b is covalently attached to functional groups in the near vicinity of C4 cleavage on the target surface via a reactive thioester exposed by C1q-associated serine proteases C1r and C1s. Bound C4b forms the basis of the enzymatically active C4bC2b complex, known as the classical C3 convertase, which itself acts as a serine protease cleaving plasma C3 into the small pro-inflammatory C3a fragment, and the large reactive C3b fragment that, like C4b, also binds to the target surface (Figure 1a). From here, the alternative C3 convertase (C3bBb) can be formed, representing an important step in signal amplification. In addition to amplification, cascade progression occurs through both C3 convertases, which can serve as a foundation for C5 convertase formation (Figure 1a). Thus, cascade amplification, mainly through C3 convertases but also through C5 convertases, plays the dual functional role of generating potent signaling soluble peptides, anaphylatoxins C3a and C5a, and cascade-progressing target surface-reactive and effector function-mediating complement fragments C3b, C4b, and C5b (Figure 1a).

The soluble C3a and C5a anaphylatoxins interact with cognate G-protein-coupled receptors, C5aR and C3aR, which are primarily expressed by myeloid-derived cells. Importantly, anaphylatoxin C5a acts as a chemoattractant for the recruitment of inflammatory cells (Figure 1a, 1e), and acts on immune effector cells to upregulate activating and downregulate inactivating Fc-receptors, resulting in monocyte infiltration and differentiation into macrophages, and recruited effectors of antibody-dependent cellular cytotoxicity (ADCC) and antibody-dependent cellular phagocytosis (ADCP)¹⁸ (Figure 1d). In this way, therapeutic AMCA can orchestrate an amplification loop by attracting additional innate elements to the site of antibody target recognition. The important functions of anaphylatoxins in host defense are carefully balanced against the consequences of misdirected, excessive, or sustained activation resulting in host tissue damage and exacerbation of undesired inflammatory conditions.

Complement fragments C3b, C4b, and C5b covalently coat target IC or particle surfaces and can mediate a variety of functions (Figure 1d). Several distinct species of cell-bound C3 and C4 can exist (i.e. iC3b, C3dg) depending on the levels of regulatory activity encountered during the cascade and in the post-activation microenvironment. These fragments maintain IC solubility and can mediate the recognition, trafficking, processing, and elimination of ICs through engagement with cognate complement receptors (CRs), including CR1 (CD35), CR2 (CD21), CR3 (CD11b/CD18), CR4 (CD11c/CD18), and complement receptor of immunoglobulin family (CRIg), expressed by a variety of leukocytes (Figure 1a). Given this wide expression profile, and the co-expression of antibody receptors on these cells, it is perhaps not surprising that opsonization can affect pro-inflammatory processes of phagocytosis, antigen presentation, cytokine stimulation, neutrophil extracellular trap expulsion, NK activation, and IC adherence and trafficking (Figure 1d). Studies of the impact of various opsonins on other functions make it clear that a variety of phenotypes can be observed, ranging from competitive¹⁹ to synergistic²⁰. Without requiring additional proteolytic steps, fragment C5b can readily recruit downstream complement components, C6 through C8 and multiple copies of C9 yielding lytic – including the membrane attack complex (MAC) (Figure 1e) – as well as sub-lytic terminal complement complexes (TCCs). Sub-lytic TCCs can induce intracellular signaling.

Classical Complement Pathologies and Interventions

Given lytic outcomes, the dysregulation of complement or the generation of autoantibodies that activate complement can lead to serious damage to host tissues and organs. In fact, classical complement C4d fragment is a clinical correlate for the diagnosis of, among other autoimmune disorders, systemic lupus erythematosus (SLE)²¹, lupus nephritis²², IgA nephropathy²³, anti-phospholipid antibody (APLA) syndrome²⁴, and C3-glomerulopathies^{25, 26}, as well as in organ transplant rejection²⁷.

Damage generated by autoantibodies can be treated by B cell depletion (e.g. SLE and RA)²⁸. Alternatively, inhibiting the progression to the lytic end of the complement cascade with the anti-C5 mAb eculizumab has been shown to be a clinically effective strategy in preventing complement-mediated hemolysis in paroxysmal nocturnal hemoglobinuria (PNH) patients^{29, 30}, and in preventing hemolysis leading to thrombotic angiopathy and kidney disease in patients with atypical hemolytic-uremic syndrome (aHUS)³¹. More recently, eculizumab has been shown to reduce the rate of relapse in patients with the autoantibody anti-aquaporin-4-mediated neuromyelitis optica (NMO) spectrum disorder³². Inhibition by eculizumab can be overridden by strong complement activation in vitro, and lytic complement mechanisms are unlikely the sole contributor to autoantibody-dependent complement-mediated disease states which may explain the differential, incomplete, or lack of therapeutic efficacy seen in other diseases^33–35. Intuitively, blocking terminal complement may leave patients susceptible to infection³⁶. In fact, late complement component deficiency is strongly associated with increased risk of developing meningococcal disease³⁷ where bactericidal complement antibody is a correlate of protection³⁸.

Nonetheless, the clinical successes of eculizumab portends both the scientific utility of such inhibitors in understanding the precise etiology of diseases in which complement is implicated, and the potential value of exploring other suppression strategies. To this end, additional strategies to suppress complement activity include a small-molecule C5 inhibitor³⁹, a single-chain antibody linked to a complement inhibitor⁴⁰, siRNA against C5⁴¹, inhibitory anti-C1s mAb⁴², C3 inhibition⁴³, a variety of IVIG-mimetic IgG1-Fc multimers^44–47, and disruption of ICs by relatively complement-inert subclasses such as human IgG4⁴⁸. Beyond canonical complement regulatory factors, a role for a histological biomarker in Alzheimer’s disease and atherosclerosis, apolipoprotein-E (ApoE), in attenuation of unresolvable inflammatory diseases was recently identified. ApoE was shown to selectively complex with activated C1q at picomolar affinity and inhibit its activity⁴¹. Lastly, intravenous immunoglobulin (IVIG) has been extensively used in the clinic to suppress inflammation across a range of disorders. In autoimmune neuropathies involving anti-ganglioside antibodies, both polyclonal IgM and IgA⁴⁹, and IVIG⁵⁰ were shown to inhibit complement activation in a glycoform-dependent manner^{51, 52}. The major critique of strategies to suppress classical complement activity is that the complement system is critical to the systemic defense, clearance of apoptotic debris, and general maintenance of host homeostasis, such as in tumor surveillance. However, while natural complement deficiencies in humans are associated with increased risk of meningococcal infection that can be reduced by vaccination⁵³, prolonged pharmacologic C3 inhibition does not appear to increase risk of infection in otherwise healthy adult non-human primates⁴³.

Role in tumor promotion & elimination

The complement system can be directly activated by neoplastic host cells, as well as indirectly via antibody opsonization. While demonstrated to be antagonistic of some Fc receptor-mediated functions⁵⁴, complement dependent cytotoxicity is thought to be a main mechanism of action for some monoclonal therapeutics⁵⁵. In fact, tumor cell vulnerability to complement-mediated lysis is implicated as a predictor of mAb-based therapeutic efficacy⁵⁶. However, aberrant complement activation and anaphylatoxin generation has been linked to augmented tumorigenesis⁵⁷ as demonstrated by tumor inhibition by pharmacological blockade of the C5a receptor⁵⁸. In chronic inflammation, anaphylatoxins - particularly C5a - can promote a pro-tumor microenvironment through a variety of mechanisms, including chemotactic recruitment of tissue-infiltrating myeloid cells that are stimulated to secrete immunosuppressive molecules thereby sustaining a pro-tumor microenvironment⁵⁸. Additionally, sub-lytic TCCs can have oncogenic effects, including cell cycle modulation and resistance to apoptosis and complement⁵⁹. Mechanisms of tumor cell complement evasion include antigenic modulation⁶⁰ and overexpression of complement regulatory proteins (CRPs), of which CD59 (MAC-inhibitory protein, or protectin) inhibits C9 polymerization and can induce internalization of affected areas of membrane in order to escape MAC lysis⁶¹. What is known regarding the interplay between complement and various cancers has been reviewed recently⁶².

Cancer cell populations are inherently heterogeneous in their susceptibility to classical complement fixation, and can be shielded by tumor microenvironmental factors (i.e. hypoxia, pH, etc.) that can impact the complement cascade⁶³. Nonetheless, targets such as CD20 have established the importance of AMCA as a mechanism of action⁶⁴, as well as point to epitope-dependent factors impacting efficient membrane clustering and Fc-Fc oligomerization^{12, 65}. In fact, the degree to which anti-CD20 antibodies induce classical activation has raised concerns that exhaustion of plasma levels of complement components might be rate- and efficacy-limiting⁶⁶. Targeting multiple antigens or epitopes simultaneously can also influence the efficiency of activation⁶⁷, and in some cases, overcome tumor heterogeneity and plasticity⁶⁸. These cocktails contribute to cross-linking-dependent Fc-effector function enhancement⁶⁹, but also increase the effective local concentration of IgG Fcs and their orientation heterogeneity on the target surface, likely influencing the probability of hexameric Fc organization⁷⁰.

Role in Infectious Disease: from elimination to evasion to enhancement

Natural host immunity is critically reliant upon the complement system to protect and eliminate infections from a number of microorganisms⁷¹ and viruses; the classical complement initiation pathway has been demonstrated to contribute to anti-viral activity and protection against West Nile Virus (WNV)⁷², Zika⁷³, Influenza⁷⁴, vaccinia virus⁷⁵, CMV⁷⁶, RSV⁷⁷, and HIV^{78, 79}. Additionally, classical complement has been implicated in host defense against the pathogenic bacteria Salmonella⁷¹, Gonorrhea⁸⁰, as well as against parasites^81–83. However, genetic variation affecting antigen identity and character can dramatically influence the susceptibility of a virus or pathogen to classical complement neutralization or lysis, even in the instance of closely related viruses⁸⁴. When confronted with an invasive pathogen, it is commonly understood that antibodies can directly interfere with mechanisms of infection (e.g. host cell entry by virions, replication, bacterial quorum-sensing, etc.), usually by steric interference. Similarly, complement opsoninization can theoretically potentiate the ability of an antibody to neutralize pathogens by steric mechanisms, as has been experimentally observed^{76, 85}.

Unfortunately, evolutionary competition with host immunity has equipped infectious agents with the means to evade, and in some cases, even repurpose the complement system to gain a fitness advantage through a variety of mechanisms. Herpes simplex virus type 1 expresses a glycoprotein, gE, that mimics human Fcγ receptor, and has been demonstrated to block human IgG Fc-mediated complement activation⁸⁶. Streptococcus pyogenes binds IgG from serum, which increases C4b-binding protein (C4BP) recruitment, and enhances bacterial virulence⁸⁷. Other complement evasion mechanisms include the incorporation of host membrane CRPs by budding enveloped virions⁸⁸, molecular mimicry⁸⁹ or recruitment of such host regulatory factors^{87, 90}, and production of other complement-disrupting factors such as complement-cleaving enzymes⁹¹, and interference with C1q-associated serine proteases⁹². Another mechanism of complement evasion, and evasion of antibody-mediated effector function more generally, is antibody-induced antigen internalization on virally-infected host cells^{93, 94}. These pathogen evasion mechanisms, and others, have been extensively reviewed recently⁹⁵. Understanding such mechanisms of evasion and enhancement of infection offers the prospect of aiding the development and optimization of next-generation vaccines and antibody drugs.

Despite controversy around the existence or degree of infection enhancement influenced by antibody-mediated complement activation, evidence indeed exists for such a phenomenon across a range of infectious agents, including HIV⁹⁶. Mechanistic explanations of infection enhancement include instances in which host complement regulators directly facilitate entry⁹⁷, but also sub-lytic opsonization can result in CR-mediated entry directly⁹⁸ or via phagocytic uptake⁹⁹. In this way, sub-lytic AMCA can enhance infectivity by increasing the efficiency of intracellular invasion, for example, by transporting infectious ICs to susceptible target cells¹⁰⁰. Several bacterial pathogens, such as F. tularensis, exploit CR3-mediated macrophage internalization and downstream anti-inflammatory signaling to achieve intracellular persistence¹⁰¹. Other mechanisms of complement enhanced infection have been reported, but these activities often occur alongside protective mechanisms. In fact, the difference between WNV neutralization and infection-enhancement depends on factors that dictate the degree of complement activation¹⁰². This balance between protective and pathogenic outcomes of AMCA represents yet another challenge to in vitro models and in vivo understanding.

Ab engineering

Modulating AMCA, whether beneficial or detrimental to antibody drug efficacy, is a logical strategy to improve antibody drugs. Downstream pathway interventions include inhibition of regulatory factor function, such as of Factor I, a strategy which has been demonstrated to increase the lytic efficiency of rituximab¹⁰³. Complement regulatory factor-neutralizing and tumor-targeting antibodies have been delivered in mAb cocktails¹⁰⁴, as well as combined in bi-specific mAbs¹⁰⁵. A novel bi-specific, targeting both CD20 antigen and the C1q globular head, was recently reported that provides a direct means of C1q recruitment that may additionally provide enhanced flexibility and relieve geometric constraints tied to antigen and epitope specificity¹⁰⁶. Direct modification of the antibody molecule in bispecific strategies such as this, or more commonly by Fc engineering (Table 1) not only contributes to drug optimization, but has also served as a key tool for determining antibody mechanism of action.

Table 1:

Engineered antibodies and their complement phenotypes

		ID	Mutations / Alterations	Complement Phenotype		Sources and Important Citations	Comments
				C1q-binding	CDC potency
Complement Phenotype	Reduced	LALA	L234A, L235A	n.d.	- - -	125 Xu et al. Cell Immunol. (2000)	Significant reduction in ADCC; has been used in vivo in nonhuman primates, 126 Hessell Nature (2007)
						127 Hezareh M., et al. J. Virol. (2001)
		LALA-PG	L234A, L235A, P329G	- - -	n.d.	128 Schlothauer et al. PEDS (2016)	Residual FcgRI binding by SPR. This, and related variants are being used in human clinical trials for complete effector knockout.
		KA	K322A	n.d.	- - -	CDC potency: 4 Diebolder, C. A., et al. Science (2014); ADCC: 127 Hezereh M. et al. J. Virol. (2001)	Significant reduction in ADCC
		Glycoform	N297Q, aglycosylation	- - -	- - -	129 Sazinsky S. et al. PNAS (2008); CDC potency: 130 Ugurlar D. et al. Science (2018)	n.b. to low affinity FcgRs
		Glycoform	Sialylation	- -	- -	15 Quast I. et al., J Clin Invest. (2015);
	Enhanced	Glycoform	Sialylation	+	+	16 Dekkers G. et al. Front Immunol. (2017)
		Glycoform	Galactosylation	+	+	14 Peschke B., et al. Front Immunol. (2017)
		IgG Fc-μzTP	L309C (IgG1-Fc + IgM tailpiece)			44 Spirig, R. et al., J Immunol. (2018)	IVIG mimetic - depletes classical complement pathway components to ameliorate inflammation & autoimmune disorders. Evaluated in mice in immune thrombocytopenia and inflammatory arthritis.
		RGY	E345R, E430G, S440Y	n.d.	n.d.	4 Diebolder, C. A., et al. Science (2014)	Triple-mutant that forms solution-phase hexamers, recruits and activates C1. Seminal paper in elucidating structure of classical complement activation complex
		EG	E430G	no change	++	108 de Jong, R. N. et al., PLoS Biology (2016)	Point mutation that facilitates IgG1 oligomerization at the target surface, but not in solution at relevant concentration levels.
		EFT	S267E, H268F, S324T	+++	++	131 Moore L. et al. mAbs (2010)	Ablated ADCC
		EFT+AE	S267E, H268F, S324T + G236A, I332E	n.d.	+++		Broad enhancement to low affinity FcgRs
			K326A, E333A	++	+	107 Idusogie et al. J Immunol. (2001)	Unaltered ADCC
		113F	IgG1 CH1 & Hinge; IgG3 Fc; part of C-term. CH3 substituted with IgG1 residues; See ref.	+	+	Source - 132 Natsume et al. Cancer Res (2008); CDC potency - 131 Moore L. et al. mAbs (2010)	Selection of mutations from paper demonstrating importance of K326 & E333 residues in C1q-IgG interaction. Enhanced B-cell depletion in cynomolgus monkeys.
		IgGA	IgG1-CH2, IgA1-CH3; See ref.	++	+	133 Kelton W., Georgiou G., et al. Cell Chem. Bio. (2014)	Chosen from panel of IgG1/IgG3 chimeras for being the most potent complement activator that retained protein A binding.
	Selective/Enhanced	IgG G801	K320E, Q386R Glycosylated	++	n.d.	134 Lee C.H., Georgiou G., et al. Nature Immuno. (2017)	Residual binding to FcgRI (G801); No detectable FcgR binding (A801). Allows for differentiation between FcR-mediated and CR-mediated cellular cytotoxicity and phagocytosis. Evaluated in mice.
		IgG A801	K320E, Q386R Aglycosylated	+++	+

Not determined (n.d.)

Fc domains lacking complement binding, and/or other effector functions entirely have been used extensively in in vivo studies in animal models to parse the effects of complement and FcγR functions versus Fab-mediated activities alone, and some modifications are even advancing clinically.

Conversely, other mutations have been explored to increase the efficiency of AMCA activation—both through avidity modulation based on engineering of the Fc-Fc interface and propensity to hexamerize, and through affinity for the C1q globular head (Figure 2b)¹⁰⁷. Such potentiating strategies can have the effect of overcoming the action of complement regulatory factors, effectively driving the cascade towards terminal complex formation with direct lytic and signaling impacts. The triple mutant RGY used in the structural studies enhanced Fc-to-Fc affinity to the point where antibody hexamerization occurred in solution, complexing with C1q in the absence of antigen. A point mutation E430G, representing the ‘G’ in the RGY triple mutant, demonstrated enhanced hexamerization on the target surface and not in solution (Table 1). Additionally, this variant was shown to exhibit reduced immunogenicity and possess other traits desirable for biopharmaceutical development¹⁰⁸. E430G was shown to convert an anti-EGFR IgG1, 2F8/Zalutumumab, from complement-inert to -activating¹⁰⁹. This mutation has been used to demonstrate the contribution of AMCA effector mechanisms to the antitumor and antimicrobial efficacy of specific antibodies^{80, 109–111}, and as a means to demonstrate inherent susceptibility of a given target to terminal classical complement functions¹⁰⁹. Efficiency-enhancing mutations can result in lysis of targets even in the context of late-complement-component-depleted and C9-deficient serum, perhaps indicating that swift and sustained complement activation is important and sufficient for efficient lysis^{112, 113}. An early stage clinical trial (NCT03576131) of dual non-competing anti-DR5 combination therapy utilizing this avidity-modulating mutation is ongoing¹¹¹.

Modeling the Contribution of Complement

Collectively, the library of ADCA-modifying Fc mutations is further supported by a suite of in vitro assays with diverse readouts, and animal models useful to evaluation of the contribution of complement activation to pathology or protection driven by antibodies. The detection of complement fixation products has provided clinically-useful biomarkers for numerous autoimmune diseases^{24, 114} and cancer progression¹¹⁵, as well as in the prognosis of tissue/organ transplant tolerability and post-transplant monitoring^{27, 116, 117}. While detection of complement activation end-products on cells and tissues can be correlated with certain autoimmune disease states, pre-clinical characterization of drug candidates AMCA requires recapitulating the context of natural systems in which interventional AMCA would take place.

Given the physiological complexity and highly interconnected nature of the complement system, it would appear to be difficult to rely on in vitro experimentation as a model of in vivo outcomes. In order to best approximate their biological relevance, the design of in vitro experiments must take into consideration a number of factors. For example, there are individual- and tissue-specific differences in the concentration of complement components. To this end, reconstitution of the classical and alternative pathway by a defined mixture of the individual components might be useful in the standardization of complement assays, and may also reduce artefacts originating from other serum components that vary between serum sources¹¹⁸.

Recapitulating native contexts is likely to produce more relevant in vitro results, but such target presentations and assay readouts are often less amenable to high-throughput screening. A common means to observe complement activation is through the use of C3 fragment-specific monoclonal detection reagents¹¹⁹, which is adaptable to bead-based antigen-multiplexed antibody screening¹²⁰. The throughput of these methods come at the expense of capturing the functional consequences of complement activation, whether opsonophagocytosis, immune adherence and trafficking, lysis, or others, which can be characterized through lower-throughput specialized assays¹²¹. Even specialized assessment of complement functions in vitro may fail to fully capture the clinical context, as they are often performed on lab time-scales in the absence of effector cells, and using susceptible or sensitized targets. In vivo models can substantially increase the clinical relevance when characterizing the varied outcomes of AMCA, or lack thereof.

Animal models have been extensively used to parse the mechanisms of action of pathogenic or protective antibodies with a diversity of genetic knockouts^{122, 123}, human cascade component knock-ins¹⁰⁶, and complement-modifying interventions including C5 and/or C3 depletion using cobra venom factors¹²⁴. While animal models have demonstrated an indispensable role here, species-specific and human allelic differences in each of the many complement factors can also influence outcomes. Once a relevant difference is identified, models can be generated to more closely reflect native interactions. For example, a humanized C1q mouse, in which mouse C1qA, C1qB, and C1qC genes were replaced with chimeric versions containing human globular head domains, has been generated¹⁰⁶.

One certainty remains, determining when, where, and how complement matters in disease and antibody-based interventions is complicated. If carefully designed, experiments modulating AMCA in vitro and in vivo can provide unique insights that lead to the identification of druggable targets or assist in clinical translation of antibody drugs.

Conclusions

The complement system is an amazingly complex, potent, and intricately regulated aspect of immunity. In part encouraged by the high-profile therapeutic success of anti-CD20 therapy, the factors controlling activation efficiency are being more deeply explored in the context of infectious diseases and autoimmunity. Continued development of assays, models, novel interventions, and engineered antibodies promise to elucidate how complement-dependent antibody mechanisms contribute to disease neutralization, eradication, and in some cases, enhancement as the diversity of interventions and models that have been engineered to explore, exploit, or suppress antibody-mediated complement activity are deployed across diverse disease states.