Local complement activation and modulation in mucosal immunity

Abstract

The complement system is an evolutionarily conserved arm of innate immunity, which forms one of the first lines of host response to pathogens, and assists in the clearance of debris. A deficiency in key activators/amplifiers of the cascade results in recurrent infection, whereas a deficiency in regulating the cascade predisposes to accelerated organ failure, as observed in colitis and transplant rejection. Given that there are over 60 proteins in this system, it has become an attractive target for immunotherapeutics, many of which are FDA approved or in multiple phase 2/3 clinical trials. Moreover, there have been key advances in the last few years in the understanding of how the complement system operates locally in tissues, independent of its activities in the circulation. In this review, we will put into perspective the abovementioned discoveries to optimally modulate the spatiotemporal nature of complement activation and regulation at mucosal surfaces.

Introduction

The complement system has long been acknowledged as a key component of the innate immune response, while also facilitating cross-talk with the adaptive immune system.¹ As a family of over 60 proteins, the system is activated and regulated through a coordinated series of steps to facilitate mucosal defense.² As a result, both children and adults with complement deficiencies develop infections.³ For example, deficiencies of key complement proteins (e.g., C3) predispose to severe, recurrent respiratory tract infections that begin soon after birth.^4,5 Severe infections are often dependent on the role of complement to opsonize pathogens. A notable feature of complement-mediated defense is the protection it offers against encapsulated pathogens such as Streptococcus pneumoniae, Hemophilus influenzae, Neisseria meningitidis, and to a lesser extent, Neisseria gonorrheae.^2,6 More recently, the importance of complement-mediated immune responses has also been demonstrated in infections driven by Klebsiella and Pseudomonas in the lung,^7,8 Salmonella and Shigella in the gut,^9,10 obligate intracellular bacteria such as chlamydiae^11–13 and even viruses such as herpes simplex^14,15 and coronaviruses.^16–18 On the other hand, genetic deficiencies in certain complement components (e.g., C1q, or regulatory proteins such as CD55) increase the susceptibility for certain autoimmune or autoinflammatory diseases, which highlights the homeostatic role for complement in the clearance of debris, a key process at mucosal tissues.^19–21 Thus, how the complement system can be leveraged to design therapeutics and vaccines to promote mucosal defense has gained considerable interest.

Canonical and non-canonical roles of complement.

Activation of the complement system generates anaphylatoxins (such as C3a and C5a) and C3b, which can both opsonize its target, and form additional enzymes (known as convertases) that ultimately facilitate the formation of the membrane attack complex (MAC, Figure 1). MAC deposition on a target results in membrane perturbation, which may culminate in lysis of the pathogen.²² These functions are known as ‘canonical’ roles of complement and are integral to host defense. However, the period from 2013 onwards has resulted in an appreciation of additional roles of complement proteins, including their ability to modulate cellular survival, effector immune responses, wound healing, and even neuroprotection.²³ These emerging roles are referred to as ‘non-canonical’ roles of complement. Newer interactions have been reported between complement proteins and other key components of cellular functioning (e.g., ATG16L1), as well as alternate isoforms that may explain how such a family of proteins may have both canonical and non-canonical roles.^24,25 In this context, the ‘complosome’ has been recognized as a group of intracellularly active complement proteins that orchestrate normal cell physiology. These proteins influence mitochondrial function, glycolysis, and oxidative phosphorylation in both immune and non-immune cells, which are integral to mucosal immune responses.²⁶ The abovementioned observations have led to speculations that the complement cascade—an evolutionarily conserved host response system—may first have developed intracellularly, primarily in unicellular organisms as a defense mechanism against damage, stress, or pathogens.²⁷ With the development of multicellular organisms, the system likely evolved into being secreted locally to promote host defense in microenvironments, and then systemically in mammals as a ‘guardian of the intravascular space’.²⁸ Moreover, the ‘complosome’ is also beginning to explain how mucosal sites may be predisposed to more chronic diseases such as fibrosis and cancer.^26,29

Figure 1. The complement cascade and its regulation.

The complement cascade can be initiated by either the classical pathway, primarily through antigen-antibody complexes; the lectin pathway when pattern recognition molecules such as ficolins (FCN) and mannose-binding lectin (MBL) bind to sugars, or through alternative pathway activation, which occurs via spontaneous tickover, but is amplified by properdin (P). These pathways converge on the formation of C3 convertases, which can cleave C3 to C3a and C3b. C3b facilitates formation of C5 convertases, which can cleave C5 into C5a and C5b. C5b forms complexes with C6, C7, C8 and C9 to form the membrane attack complex (MAC). At each step, the complement cascade is tightly regulated by fluid-phase regulators (AT: antithrombin; C1-INH: C1 inhibitor; C1QBP: C1q binding protein; C4BP: C4 binding protein; CLU: clusterin; CPN: carboxypeptidase N; CSMD1: CUB And Sushi Multiple Domains 1; FH: Factor H, FI: Factor I; VTN: vitronectin) and membrane regulators (CD46; CD55; CD59; CR1; VSIG4: VSIG4 V-set and immunoglobulin domain containing 4).

Upstream inducers of local complement activation

Many effector components of the complement cascade (e.g., C3, C4 and Factor B) are acute-phase proteins and are thus, are upregulated in the setting of stress. Transcriptional regulation of complement receptors and membrane regulators has been covered in other reviews.³⁰ Their levels are regulated at both the transcriptional and translational level. The promoter region of C3 has two CCAAT/enhancer binding protein (C/EBP) consensus sequences, basic DNA binding region and leucine zippers 1 and 2 (bZIP1 and bZIP2), which confers responsiveness to cytokines such as IL-1 and IL-6.³¹ More recently, even stores of intracellular C3—which has an alternate translational start site and lacks a signal peptide—were upregulated by IL-1β.²⁵ Additionally, the C3 promoter region also has an FXR-responsive IR-1 element,³² a PPAR response element (PPRE) that can bind to PPARα,³³ and an estrogen response element.³⁴ Both C3 and Factor B promoter regions have κB binding sites that confer responsiveness to TNF-α.^35,36 IFN-γ increases C4 levels via IRF-1-mediated upregulation of synthesis, and by also stabilizing C4 mRNA.^37,38 Previous work on how murine C3 and Factor B are STAT-1 dependent³⁹ has since been independently validated in human airway epithelial cells infected with SARS-CoV-2 infection.¹⁷ Specifically, both STAT1 and RELA (an NF-κB family transcription factor) bind to the promoter regions of C3, CFB, C1S and C1R, which are induced by SARS-CoV-2 infection.¹⁷ Ruxolitinib—a competitive inhibitor of the ATP-binding catalytic site on JAK1 and JAK2—normalizes the expression of the abovementioned genes in SARS-CoV-2 infected A549-ACE2+ cells, and also decreases the proportion of C3a⁺ cells, suggesting the JAK1/2 pathway amplifies local C3 activation.¹⁷ More recently, STAT1 (and STAT3) have been implicated in IL-6-mediated upregulation of C3 and CFB.⁴⁰ Similar to ruxolitinib, the observation that treatment with either anti-IL-6 (siltuximab) or anti-IL-6R (tocilizumab) can lower markers of complement activation—including alternative pathway activation—suggests JAK/STAT pathways are contributing to a feedforward loop of complement activation.⁴⁰

Downstream signaling events following complement activation

Both canonical as well as non-canonical signaling events contribute to innate and adaptive immune responses in the setting of complement activation. In the context of canonical signaling events, the anaphylatoxins C3a and C5a bind to their cognate receptors (C3aR, C5aR1, C5aR2) to promote smooth muscle contraction, vasodilation, increased vascular permeability, as well as primarily, the recruitment of innate immune cells such as neutrophils and monocytes in the setting of an acute injury; C5a being a much more potent anaphylatoxin than C3a.^41,42 These anaphylatoxins also facilitate an oxidative burst and promote the degranulation of neutrophils, eosinophils and mast cells.⁴¹ More recently, increased C3aR1 and C5aR1 activation on myeloid cells has been shown to induce NET formation.^17,18 At the same time, C5a induces the expression of adhesion molecules such as P-selectin and ICAM-1 on structural cells of the lung (such as endothelial and epithelial cells respectively), and facilitates IL-8 production through MAPK signaling pathways.^43,44 Additionally, although MAC primarily induces membrane perturbation and cellular lysis, sublytic MAC formation on epithelial cells results in increased intracellular Ca²⁺ concentration, triggering NLRP3 inflammasome activation and IL-1β release.⁴⁵ Moreover, MAC internalization by endothelial cells activates noncanonical NF-κB and augments IL-1β secretion.^46,47 These proinflammatory cytokines contribute to both autocrine and paracrine amplification loops that continue to perpetuate inflammation and eventually tissue damage.⁴⁸

Tissue damage can activate the clotting and coagulation pathways, which generate enzymes such as thrombin and kallikrein that can continue to perpetuate complement cascade activation, often independent of convertase assembly.^49,50 The net result of inflammation, tissue damage, blood flow stasis and hypercoagulability is increased thrombosis, which has notably been implicated in disorders of systemic complement activation involving endothelial injury, such as atypical hemolytic uremic syndrome, and severe SARS-CoV-2 infection.^16,51 In addition to the abovementioned examples of canonical signaling events, non-canonical, intracellular signaling events have also become increasingly acknowledged as contributing to effector cellular functions. Examples of these non-canonical signaling events include but are not restricted to C1q modulating the mitochondrial metabolism of CD8⁺ T cells to restrict autoimmunity,⁵² CD46-mediated T regulatory cell survival and function,⁵³ C5aR1 signaling on mitochondrial membranes facilitating ROS generation and anaerobic glycolysis to favor IL-1β production in macrophages,⁵⁴ and the binding of C3 to ligands such as Fyn-related kinase that promote survival.^24,25

Sourcing of complement proteins.

Prior human studies have demonstrated that the liver is the major source of most circulating complement proteins, such as C3, C4 and Factor B.⁵⁵ Transplantation studies from C3 F/S mismatched recipients suggested that potentially up to 5-10% of circulating C3 levels may be derived from either the bone marrow, or the kidney.⁵⁶ Positive amplifiers of the complement cascade such as properdin are brought in by myeloid cells, such as neutrophils.^57,58 Moreover, Factor D, which is a protease that cleaves Factor B and amplifies the alternative pathway, is produced by adipose tissue and cells of the macrophage/monocyte lineage.^59,60 However, the complement system and its components have primarily been considered as being present in the circulation (plasma). Recent work by us and others suggests various complement proteins are locally active at mucosal surfaces independent of circulating proteins.^8,61,62 Moreover, several single cell-RNA sequencing datasets also suggest that these genes are highly expressed by structural cells within mucosal surfaces, such as epithelial cells and fibroblasts.^17,63 Although these observations were conceptually suggested several decades ago based on data from culture systems,⁴² recent technological advances have facilitated the interrogation of what is produced locally, compared to what reaches the mucosal surfaces from the circulation post-infection. These advances have implications for therapeutics aimed at diseases affecting barrier surfaces such as pneumonia, lung transplant rejection and inflammatory bowel disease, among others.

Complement-mediated host defense at respiratory mucosal surfaces

The complement system plays a crucial role in defending respiratory mucosal surfaces during infections.^2,7 However, when complement activation becomes excessive, it can lead to lung damage, including acute lung injury (ALI) in animals, and acute respiratory distress syndrome (ARDS) in humans.⁶⁴ Several components of the complement system are detected in the sinonasal tissue, lungs and the bronchoalveolar lavage (BAL) of different mammalian species, including humans.^65–68 Both immune and non-immune cells in the respiratory tract produce complement proteins, which have both canonical functions in host defense, but also non-canonical functions such as affecting cellular survival, and metabolism.^25,69–72 Single cell transcriptomic analysis of human and mouse lung reveals epithelial cells, mesothelial cells and fibroblast as the main producers of C3 and C5, whereas interstitial macrophages strongly express C1q.⁶³ Additionally, structural cells in the lungs also express complement regulatory proteins.^73–75 In vivo experiments demonstrate lung epithelial cell-derived C3 being protective against Pseudomonas-induced acute lung injury.⁸ Several studies have reported anaphylatoxins C3a and C5a in the BAL.^68,76 These anaphylatoxins interact with their respective receptors, C3aR and C5aR1 and C5aR2, which are present on tissue-resident structural cells such as epithelial cells and fibroblasts, in addition to immune cells.^77–80 These interactions trigger the activation of pathways both associated with host defense, but also involved in tissue remodeling.^73,81,82

Caveats for complement-mediated host defense in the lung.

First, it is important to differentiate the sourcing of these proteins in the lung, given that individual cell types in the lungs produce these proteins, and in the setting of injury, circulating proteins also enter the lungs due to alveolar-capillary barrier disruption.^17,64,68 Second, there may be inherent differences in the time course of infection, as well as how the pathogen alters complement-mediated responses in the lung. For example, viruses such as SARS-CoV-2 upregulate both C3 and Factor B expression in the lung epithelial cells, and increase C5a/C5aR1 signaling increased locally in the lung, notably in myeloid cells, which then induces NET (neutrophil extracellular trap)-dependent lung damage.^17,18 Third, ongoing lung damage downregulates complement regulatory proteins such as decay accelerating factor (DAF), which promotes irreversible damage such as fibrosis.⁷⁵ Thus, a better understanding of the spatiotemporal induction and resolution of various complement components in the lung in response to infection, inflammation and injury is key for leveraging this system from both a diagnostic and therapeutic aspect in lung diseases.

Complement production by structural cells of the lung.

Airway epithelial cells have a unique ability in storing as well as secreting activators of the complement cascade, such as C3, Factor B and C5,^8,17,70,83 which may have evolved given their constant interaction with the environment. Airway epithelial C3 and Factor B facilitate cellular survival in the setting of infection (e.g., Pseudomonas) as well as sterile injury.^25,70 While secreted C3 facilitates opsonophagocytosis, intracellular C3 (potentially derived through a alternative translational start site, resulting in C3 lacking the signal peptide, which is therefore translated in the cytosol) can also opsonize invasive Staphylococcus aureus in epithelial cells,⁷² can bind to ATG16L1 facilitating autophagolysosomal formation,²⁴ and can interact with Fyn-related kinase (FRK), thereby facilitating survival via a PTEN-mediated process.²⁵ However, these intracellular stores may not always be protective. In the context of SARS-CoV-2 infection, primary human airway epithelial tissue cultures have shown increased intracellular C3 stores and elevated secretion of C3a-desArg, while also demonstrating enhanced MAC formation. This activation results in a highly proinflammatory microenvironment, causing significant tissue damage, which is decreased by antagonizing both C3aR and C5aR1.⁸³ Whether all the components that are required to form MAC are produced by epithelial cells in vivo, or some are derived from different sources in an orchestrated manner remains to be clarified. Yet, while initially observed in immune cells, it is conceivable that the “complement-metabolism-inflammasome” axis is also activated in the structural cells of the lung such as epithelial cells independent of the components required for MAC formation.⁸⁴ These non-canonical roles of complement are likely to supplement the existing functions of complement proteins secreted at the air-liquid interface, which serve as anaphylatoxins, promote opsonization and phagocytosis (Figure 2).

Figure 2. Complement-mediated host defense in the lung.

In the lung, structural and immune cells are local sources of complement, in addition to the circulating proteins that leak into the lungs during injury. Complement plays multiple roles in the host immune response and in maintaining homeostasis. Within lung cells, C3 is synthesized and cleaved into C3a and C3b. C3b plays a role in opsonizing pathogens to facilitate phagocytosis. It also initiates inflammation by binding to the receptor for C3a (C3aR), leading to the expression of cytokines such as TNF-α and IL-1β. Additionally, C3b participates in facilitating membrane attack complex (MAC) formation by serving as part of the C5 convertase, which cleaves C5 into C5a and C5b. C5b forms a part of the MAC, resulting in cellular lysis. Furthermore, C3a and C5a act as anaphylatoxins, recruiting leukocytes to the site of infection. Intracellular mitochondrial, lysosomal, and/or endosomal C3aR and C5aR1 execute their functions by initiating signaling cascades like those initiated extracellularly.

Complement production by immune cells in the lung.

Innate immune cells such as neutrophils and monocytes are the first responders to an infection in the lung. These infiltrating cells, in addition to tissue-resident immune cells such as alveolar macrophages, are being recognized as extrahepatic source of complement proteins in the lungs. Several studies have demonstrated that monocytes and macrophages produce complement proteins such as C1q, C1s, C2, C4, C3, Factor B, Properdin, Factor D, and C5.⁶⁹ In addition to secreting these proteins, these immune cells have been shown to possess intracellular stores of C3 and C5, which facilitate cellular survival, key metabolic pathways and effector function.^85–87 These C3 stores in myeloid cells are increased both via uptake of C3 as C3(H₂O)⁸⁸—which is increased at sites of inflammation⁸⁹—and also through transendothelial diapedesis, which is dependent on integrin lymphocyte-function-associated antigen 1 (LFA-1) signals,⁹⁰ thus facilitating enhanced mucosal responses. Moreover, the C5a-C5aR1 axis in monocytes modulates mitochondrial function and, both the C3a-C3aR and C5a-C5aR1 axes are involved in IL-1β production.^54,91 These autocrine and intracellular roles of complement are likely to have increasing implications for effective mucosal responses (Figure 2).

Myeloid cells not only synthesize complement but also secrete regulatory molecules like Factor H, Factor I, and C1INH, while expressing various complement receptors including C3aR, C5aR1, C5aR2, CR1, CR3, and CR4.⁶⁹ These complement receptors (notably the receptors for anaphylatoxins, such as C3aR and C5aR1) contribute to the migration of immune cells to sites of inflammation, with C5aR1 playing a significant role in regulating alveolar macrophage numbers and neutrophil infiltration into the alveolar space.⁹² The presence of these receptors makes them amenable to modulation in vivo, as demonstrated using C5aR1 antagonism in mouse models of SARS-CoV-2 infection.¹⁸ Moreover, the alternative pathway has been targeted via Factor D inhibition to reduce SARS-CoV-2-induced immunopathology in macaques.⁹³ Cell-autonomous Factor H dampens resolution of inflammation in atherosclerosis models by limiting efferocytosis;⁹⁴ whether similar effects are in play in the setting of models of airway or alveolar injury remains to be investigated. Overall, immune cell-derived complement emerges as a critical mediator in the rapid and effective response to lung infection and injury, orchestrating both innate and adaptive immune mechanisms to combat pathogens.

The abovementioned observations highlight the importance of local complement production in the lungs, its role in defending against infections, and the potential for detrimental effects if complement activation becomes excessive. Further research is needed to gain a deeper understanding of the mechanisms involved and explore potential therapeutic interventions associated with complement regulation in the lung.

Complement in the gut

Microbiota and infection.

The gastrointestinal tract is constantly balancing tolerogenic responses towards innocuous antigens such as gut microbes and dietary antigens, while being vigilant towards any infectious threats. Additionally, a single layer of epithelium divides the diverse array of microbial and dietary load from body’s largest collection of immune cells within the lamina propria. While a tolerogenic capacity is beneficial to the host to avoid inflammatory responses to innocuous antigens, the inability to dampen the tolerogenic capacity could be detrimental in setting of an enteric infection. Hence, innate immune responses—specifically complement proteins—play a vital role in host defense at the gut mucosal interface.

A number of studies have highlighted the involvement of intestinal C3 in restricting pathogen invasion and enhancing growth restriction through the autophagy pathway.^9,95 Complement-opsonized Listeria are trafficked intracellularly to autophagy pathways that restrict bacterial replication. Additionally, mice lacking intestinal expression of ATG7, a key autophagy mediator, have a greater mucosal burden of Listeria, similar to those seen in C3-deficient mice. Hence, coating of bacterial pathogens with C3 is critical for innate immune defense of epithelial cells (Figure 3). However, pathogens such as Shigella and Salmonella express proteases capable of degrading C3 and escaping complement-mediated autophagy.⁹

Figure 3. The role of gut-derived complement.

There is growing recognition that complement proteins play an important role in shaping the gut microbiome. During early life, newborn pups receive complement (for example, C3) from breast milk. The presence of complement proteins in the breast milk is critical to limit the bloom of gram-positive bacteria, such as Staphylococcus lentus B3, and is critical for providing protection from enteric pathogens such as Citrobacter rodentium. In fully grown adult mice during homeostasis, the gut microbiome influences the production of certain complement factors, such as C3 from stromal cells. This stromal cell-derived C3 is also critical to fight infection from Citrobacter rodentium. In animal models of chronic intestinal inflammation, gram-negative bacteria trigger TLR4 and C3 expression in colonic epithelial cells, which gets secreted in the lumen to affect the bacterial composition.

Complement proteins also play a critical role in influencing the gut microbial community. Prior studies have demonstrated presence of complement within the intestinal lumen. For example, patients with bacterial infection have elevated levels C3 and C5 in their small intestine.⁹⁶ However, we are just beginning to understand how luminal C3 levels are regulated by the composition of the microbiota. In fact, fecal C3 concentrations vary based on the vendors the mice are obtained from.⁶² Mice raised germ-free or treated with antibiotics to deplete gut microbiota have negligible amount of C3 in their feces, while fecal microbiota transplantation into germ-free mice results in restoration of C3 levels. In this study, a predominance of Prevotella sp contributed to increased fecal C3 levels. These findings indicate that the gut microbial composition can directly influence fecal C3 concentrations. Moreover, local complement secretion in the colon plays a critical role in protection from Citrobacter rodentium infection.⁶²

Upon interrogation of the local sources of C3 in the gut, stromal cells are found to be a major cell type producing C3 in the colon, with high density of C3-producing cells found in isolated lymphoid follicles (ILFs).⁶² Unlike Peyer’s patches that develop in utero and remain fixed in number throughout life, ILFs develop after birth arising from cryptopatches in response to stimuli and are dynamic: forming, regressing and changing in numbers throughout life. Mature ILFs have been observed to give rise to IgA-producing cells, and the absence of ILFs can result in dramatically altered microbial composition. Collectively, these stromal cells within ILFs produce C3—a process reliant on the microbiota—which protects the host against pathogenic infection. However, the studies till date have not interrogated the effects of deleting this tissue-specific sourcing on local C3 concentration in the gut. The utilization of available and emerging tools will help dissect to what extent these cell types contribute to the overall levels of C3 in the gut during homeostasis and with infection, versus to what extent, intracellular C3 levels in these cell types are modulating the host response to infection.

Events in early life, particularly those altering the gut microbiome and immune development, have an important role in protection from infection. Various studies have noted the presence of complement components in breast milk, while its pathophysiological relevance remains to be determined. A recent study described the importance of complement in breast milk in shaping the microbial community of an infant’s gut and thereby protecting newborns from pathogenic microbes.⁹⁷ Using an elegant system of cross-fostering newborn mouse pups with complement-deficient dams, the authors demonstrate the critical role of complement in breast milk on lysing specific gram-positive commensal bacteria via C1q-initiated, antibody-independent and membrane attack complex-dependent fashion. Yet, critical questions remain regarding the switch from breast milk-derived complement proteins to locally produced complement factors in adult gut tissue, and the role of complement proteins beyond protection from infection. Indeed, recent studies have highlighted a novel role of C1q produced by intestinal macrophages in peristaltic movement.⁹⁸ In this study, authors demonstrate that C1q-expressing macrophages are closely associated with enteric neurons in the intestinal muscularis. Interestingly it was observed the C1q produced by the macrophages does not alter response to enteric infection, intestinal injury or influence intestinal immune cell composition. However, loss of C1q production by intestinal macrophages results in altered gut motility. In the brain, C1q production by macrophages has been shown to regulate development and function of neurons.^99,100 However, the mechanism by which intestinal macrophage derived-C1q regulates enteric neuronal function remains to be identified.

Inflammatory bowel disease (IBD).

Given the central function of the complement system in detection, opsonization and elimination of bacteria as well as apoptotic cells, it would be logical for the complement system to promote homeostasis within the gut. However, uncontrolled and sustained complement activation evokes severe inflammatory processes and may result in tissue damage as seen in IBD. While its well established that IBD is associated with genetic variants of genes belonging to the innate immune system,¹⁰¹ its precise etiology remains unresolved. IBD represents a group of chronic relapsing autoimmune diseases that include two main phenotypes, Crohn’s diseases (CD) and ulcerative colitis (UC). CD is clinically heterogeneous, defined in part by transmural inflammation that many affect all layers of the intestine. However, UC is restricted to the colon and presents as severe mucosal inflammation and ulcerations. The question remains as to when the system is activated inappropriately, thereby turning complement from a defender to an aggressor that triggers inflammatory diseases.

Deposition and activation of complement proteins in IBD tissues have been noted since the 1970s. While the role of complement proteins in driving IBD pathogenesis has been studied using various animal models, the mechanism driving complement deposition and its contribution to disease progression has yet not been clearly understood. Interest in the role of complement in IBD has resurged now that there is a growing understanding of complement production by tissue-resident and infiltrating cells. Indeed, higher C3 and Factor B levels have been observed in the serum samples of CD patients compared to those with UC or healthy controls.^102,103 In a follow-up study, it was observed that CD patients had strong C3b deposition in the intestinal epithelium, but not IgG1, C1q or C4c, while all these proteins were observed in the tissue from UC patients.¹⁰⁴ Hence, the authors concluded that the alternative pathway of complement activation is more pronounced in CD, while antibody-mediated classical pathway activation on epithelial cells seems to be predominant in UC. Further, C4 mRNA transcripts were detected in normal and inflamed tissue in CD, whereas C3 mRNA was specifically detected only in inflamed mucosal areas.¹⁰⁵ However, these studies are largely correlative, and do not explain to what extent complement is an inciting factor in driving IBD, versus a bystander activated in the course of ongoing inflammation.

A recent immunochip-based meta-analysis on 4056 perianal CD and 11088 patients with CD from three independent cohorts demonstrated that a missense variant in CFB (G252S, rs4151651), which results in reduced binding of Factor B to C3b, is associated with perianal CD.¹⁰⁶ In these patients, activation of the alternative complement pathway is impaired, which leads to decreased phagocytosis and decreased cytokine secretion. The authors speculate this polymorphism results in increased local burden of Corynebacterium, which is abundant in patients perianal CD, presumably due to impaired complement-mediated phagocytosis. However, the source of Factor B, and how it can potentially be modulated in these patients, remains to be deciphered. Prior studies have reported Factor B production by intestinal epithelial cells, and correlation between TLR3 activation and CFB upregulation in inflamed regions of the gut.^107–109 Taken together, there is an increasing necessity to conduct large-scale studies examining connections between IBD-associated serologies, the gut microbiome and immune responses locally to develop novel drug interventions.

Complement in autoimmune or autoinflammatory diseases involving multiple mucosal surfaces

Although the deficiency of certain complement proteins results in recurrent infection, the complement system mediates a key role in certain autoimmune or autoinflammatory diseases which involve multiple mucosal surfaces. For example, anti-neutrophil cytoplasmic antibody (ANCA)-associated vasculitides (AAVs) involve the upper and lower respiratory tract, kidneys, skin, and in certain patients, the gastrointestinal tract.^110,111 AAVs are considered ‘pauci-immune’ due to scarce immunoglobulin and complement deposition. However, pre-clinical modeling, as well as recent studies in humans, have demonstrated complement activation in AAV through the alternative pathway, resulting in C3 and C5 activation.^112,113 Interestingly, the role of complement activation in AAV is primarily attributed to C5a generation and its downstream effects, rather than MAC formation.¹¹⁴ These observations have resulted in therapeutic targeting of C5aR1 in AAV.¹¹¹ Similarly, patients with a loss-of-function variant in CD55, a membrane regulator, are predisposed to complement hyperactivation, angiopathic thrombosis, and protein-losing enteropathy, termed as CHAPLE disease.²¹ Impaired complement regulation results in increased soluble C5a, as well as increased complement deposition on immune and non-immune cells. Moreover, costimulatory function and cytokine modulation by CD55 is defective in these patients. Interestingly, gut lymphatics are preferentially targeted in CHAPLE disease, resulting in the protein-losing enteropathy. Thrombosis in these patients is primarily attributed to a combination of increased proinflammatory cytokines (i.e., TNF-α), a reduction in thrombomodulin and increased tissue factor expression on endothelial cells.²¹ Gut microbiota are also altered in patients with CHAPLE disease, demonstrating a proinflammatory profile. Treatment with eculizumab (C5 inhibitor) shifted the composition away from this proinflammatory profile towards taxa associated with a healthy microbiome,¹¹⁵ indirectly suggesting that dysregulated complement activation may shape gut microbiome profiles, as has also been suggested elsewhere.⁶²

More recently, increased complement activation has also been implicated in severe SARS-CoV-2 infection,^16,40,93,116 multisystem inflammatory syndrome in children (MIS-C)^117–119 as well as long COVID.^120–122 All three conditions involve multiple organ systems. Alternative pathway activation is associated with increased endothelial injury and hypercoagulability in severe SARS-CoV-2 infection.¹⁶ However, the mechanisms by which certain patients develop increased complement activation, thus predisposing them to increased morbidity and mortality, remains to be clarified. Certain studies have identified putative complement-associated loci, including variants of critical regulators in patients with severe SARS-CoV-2 infection. For example, rs61821114 and rs61821041 map to expression quantitative trait loci (eQTLs) that result in decreased CD55 expression, and are associated with age-related macular degeneration.¹²³ Functional single nucleotide polymorphisms in Factor H that reduce the efficiency of C3b inactivation (rs1061170 and rs1065489) were enriched in a cohort of patients who developed MIS-C compared to those who had COVID-19 but were discharged home.¹¹⁸ In one cohort, patients with MIS-C had significantly higher levels of Factor H-associated autoantibodies when compared to patients with acute COVID-19.¹²⁴ Thus, whether we can prospectively identify a subgroup of patients developing MIS-C or long COVID who are predisposed towards increased complement activation, and target them using complement-modulating therapies needs to be assessed in larger studies.

Emerging roles of complement at other mucosal surfaces

Although a majority of work on the role of complement in mucosal responses has been focused on the respiratory and gastrointestinal tract, there is existing as well as emerging literature on its role at other mucosal surfaces, such as the eye as well as the genitourinary surfaces.

Eye.

The cornea can produce key components of the cascade such as C3.^15,125 Additionally, C3 and Factor B have been detected in tears of patients with vernal as well as giant papillary conjunctivitis.¹²⁶ Based on the absence of transferrin in the tears, most of C3 was deduced to be locally produced, as is the generation of anaphylatoxins (e.g. C3a, C3a-desArg, and C5a).¹²⁷ Locally produced C3 in the corneal/conjunctival epithelium is required not only for pathogen clearance and mitigating immunopathology in bacterial and viral infections, but is also required for effective humoral protection mediated by vaccines and monoclonal antibodies.^125,128 C3 deficiency results in impaired viral clearance of HSV-1 antigen from mouse corneas and is associated with an increase in inflammatory myeloid cells, and select pro-inflammatory (e.g., IL-1α, CXCL1, CCL2, MMP-2/3/8) and pro-angiogenic factors (e.g., VEGF-A) which coincides with increased neovascularization and corneal opacification, thus suggesting an important role for C3 in maintaining corneal homeostasis.¹⁵ Conversely, dysregulated C3 activation in HSV-1 keratitis contributes to corneal denervation, which can be reversed by localized depletion using daily ocular application of cobra venom factor in these mice.¹⁴ In this model, both tissue-resident non-hematopoietic cells (i.e., those epithelial cells forming the cornea) and resident/infiltrating CSF1R+ leukocytes contribute to local C3 expression in the cornea during acute HSV-1 infection. Regulators of complement activation, such as the decay accelerating factor (CD55), CD59, and CD46 are detected in the corneal epithelium at the limbus as well as in the central cornea, and are shed in tears, as is Factor H.^129,130 CD46 is detected at low levels, while the levels of CD55 and CD59 are comparable, which translate to their functions. For example, CD55 contributes to a majority of the C3 convertase regulatory activity in tear fluid, while a majority of C9 inhibitory activity in tear fluid is attributed to CD59. Understandably, increased levels of C4b and clusterin dominate an ‘immune response’ category in lacrimal fluid secretions of patients with dry eye disease,¹³¹ while in preclinical models, local CD55 antagonism results in intense conjunctival inflammation and iritis.¹³²

Genitourinary surface.

Complement proteins such as C3, C4 and MBL also facilitate host defense at the genitourinary mucosa. MBL serves as a pathogen recognition molecule in the genital microenvironment by demonstrating pH-dependent binding to fungi such as Candida.¹³³ Both C3 as well as complement activating-IgG are detected in cervicovaginal secretions as well as the lateral fornix of the vagina during pregnancy and labor, and in non-pregnant women.¹³⁴ C3 and C4 are detected in the uterine cervical plug.¹³⁵ Local levels of C3, C4 and Factor B and their activation correlate with the abundance of certain microbiota, as well as change with the onset of labor.^136,137 Specifically, during labor, Factor H and FHL1 are increased locally, which regulate complement activation. Soluble forms of DAF (CD55) and CD59 are also detected in cervical mucus,¹³⁸ while CD46, CD55 and CD59 are expressed in the epithelium lining the uterine cervix.¹³⁹ Certain statins can regulate complement activation in the cervix by upregulating CD55.¹⁴⁰ Of note, certain parasites such as Trichomonas vaginalis evade complement-mediated lysis through mechanisms such as acquiring CD59 from host cells¹⁴¹ and upregulating proteases that degrade C3,¹⁴² facilitating the persistence of the pathogen. Thus, although clinical reports suggest that elevated cervicovaginal levels of complement activation fragments such as C3a, C4a and C5a are associated with intra-amniotic infection and/or inflammation, and imminent spontaneous preterm birth in women with preterm premature rupture of membranes,¹⁴³ how these levels change, and what contributes to their activation, is likely to be important in understanding genitourinary mucosal immune responses in women, and potentially early detection and prevention of these disorders affecting both pregnant as well as non-pregnant women.

Opportunities for site-specific complement modulation at mucosal surfaces

Given the increasing understanding of local complement production and activation at barrier surfaces, there has been a renewed interest in targeting therapeutics to these mucosal sites. There are several approaches for modulating complement at barrier surfaces, including using monoclonal antibodies, receptor antagonism and augmenting complement regulators, including using epitope-specific targeting. The modes of action, indications, and side effects of FDA-approved complement inhibitors have already been summarized in other excellent reviews.^23,144 The traditional approach to modulating complement in disease, including in lung as well as gastrointestinal disorders, has been administering monoclonal antibodies, aptamers or antagonists intravenously. Certain complement therapeutics have demonstrated efficacy in multisystem disease. For example, avacopan, an oral C5aR1 inhibitor, is FDA-approved as an add-on treatment to standard therapy including glucocorticoids for adult patients with severe active anti-neutrophil cytoplasmic autoantibody (ANCA)-associated vasculitis. Similarly, the FDA has approved parenteral pozelimab, a monoclonal antibody to C5, for the treatment of adult and pediatric patients 1 year of age and older with CHAPLE disease. Unfortunately, most clinical trials employing these systemic approaches have resulted in disappointing outcomes in lung disease (Table). Among these agents, vilobelimab (an anti-C5a monoclonal antibody) has received emergency use authorization from the FDA based on the results of the PANAMO study that showed it improved survival of critically ill patients with COVID-19 on invasive mechanical ventilation,¹⁴⁵ although results from the FORCE study using avdoralimab (an anti-C5aR1 monoclonal antibody) did not show a mortality benefit.¹⁴⁶ Those with progressive chronic obstructive lung disease and active malignancy were excluded in PANAMO, although the prevalence of obesity and chronic kidney disease were similar in the two studies.¹⁴⁷ Yet, these treatments have not become mainstream in patients with severe SARS-CoV-2 infection, as first, the absolute mortality difference has been < 10% when adding these drugs to standard of care, and second, the overall mortality rate due to SARS-CoV-2 infection has reduced over time.

Table.

Published and pending clinical trials on complement inhibition in respiratory illness.

Clinical Trial	Drug	Mechanism	Route	Primary Outcome	Ref.
Published
NCT04333420 (Phase 3)	Vilobelimab (PANAMO)	Monoclonal antibody (mAb) to C5a	i.v.	In the predefined analysis without site-stratification, vilobelimab within 48 h post-intubation significantly reduced all-cause mortality at 28 days in COVID-19 patients requiring invasive mechanical ventilation	103
NCT04390464 (Phase 4)	Ravulizumab (TACTIC-R)	mAb to C5	i.v.	No significant difference in composite primary outcome in COVID-19 patients with a risk score indicating a 40% risk of admission to an ICU or death	104
NCT04369469 (Phase 3)	Ravulizumab	mAb to C5	i.v.	No significant difference in all-cause mortality at day 29 in COVID-19 patients requiring either invasive or non-invasive mechanical ventilation	105
NCT04382755 (Phase 2)	Zilucoplan	Peptide C5 inhibitor	s.c.	No significant improvement in oxygenation at day 6 and 15 in hospitalized COVID-19 patients	106
NCT04371367 (Phase 2)	Avdoralimab	mAb to C5aR1	i.v.	No significant improvement in clinical status at days 14 and 28 in hospitalized COVID-19 patients	107
NCT04414631 (Phase 2a)	Conestat	Recombinant C1 inhibitor	i.v.	No significant difference in day 7 disease severity in hospitalized adults at risk for progression to severe COVID-19	108
NCT04978051 (Phase 2)	Icatibant	Bradykinin B2 receptor antagonist	i.v.	No difference in clinical response on study day 10/discharge in patients COVID-19 pneumonia admitted in the early hypoxemic stage.	109
Pending Publication
Clinical Trial	Drug	Mechanism	Route	Primary Outcome
NCT05962606 (Phase 2)	AON-D21	C5a-neutralizing L-aptamer	i.v.	Safety and tolerability in severe community-acquired pneumonia
NCT04488081 (Phase 2)	Narsoplimab (I-SPY)	mAb to MASP-2	i.v.	Time to reach a durable COVID-19 level 4 or less on a WHO ordinal scale, or discharge at COVID-level 4 or lower (except for discharge to another hospital), and time to death (mortality).
NCT04988035 (Phase 2)	Danicopan (ACTIV-5)	Small molecule Factor D inhibitor	p.o.	Number of hospitalized COVID-19 patients meeting criteria for 8 ordinal scale categories on Day 8*
NCT04530136 (Phase 2)	Ruconest	Recombinant human C1 esterase inhibitor	i.v.	Day 7 disease severity (WHO Ordinal Scale) in patients hospitalized for stage II COVID-19*
NCT04402060 (Phase 1/2)	APL-9	Second generation C3 inhibitor (PEGylated peptide)	i.v.	Number of COVID-19 patients with respiratory failure who experience treatment-emergent adverse events (TEAEs) and serious TEAEs at Day 30*
NCT04574869 (Phase 1)	RLS-0071	Dual-targeting peptide inhibitor (C1, myeloperoxidase)	i.v.	Frequency and severity of AE, including SAE at Day 28 post-last dose*
NCT04395456 (Phase 2)	AMY-101	Cyclic C3-inhibitory peptide based on third-generation compstatin analog (Cp40)	i.v.	Composite mortality, oxygen requirement and clinical status (6 point ordinal scale) at Day 21 in COVID-19 patients with ARDS*
NCT03051698 (Phase 4)	Cinryze	Recombinant human C1 esterase inhibitor	i.v.	Terminated (Interim analysis showed no differences between groups in patients with intermittent-mild asthma in influx of inflammatory cells in the lung 7 hours after bronchial instillation of house dust mite and lipopolysaccharide
NCT00485576 (Phase 2)	Eculizumab	mAb to C5	i.v.	Completed but no results posted (primary outcome: allergen-induced late asthmatic response as measured by the AUC of FEV1 from 3 to 7 hours post-allergen challenge [time frame: 7 hours]) in patients with mild allergic asthma

^*study reported as being completed on www.clinicaltrials.gov with results posted on website.

Abbreviations: ARDS: acute respiratory distress syndrome; AUC: area under the curve; C5aR1: C5a receptor 1; FEV1: forced expiratory volume at the end of the first second of forced expiration; ICU: intensive care unit; i.v.: intravenous; MASP-2: mannose-binding protein-associated serine protease 2; p.o.: oral; s.c.: subcutaneous; WHO: World Health Organization

However, several preclinical studies suggest that innovative approaches such as inhaled drug administration may result in improved outcomes. Factor B inhibition using a monoclonal antibody administered via nebulization results in a significant decrease in airway hyperresponsiveness and an inhibition of airway inflammation and eosinophils in the airways and the lung tissue, as well as a reduction in the number of goblet cells.¹⁴⁸ An intranasally delivered C3aR antagonist reduced histologic evidence of sinonasal respiratory epithelial injury and inflammation as compared to untreated controls in a mouse model of Aspergillus fumigatus-induced chronic rhinosinusitis. C3aR antagonism decreased total inflammatory cells and eosinophils in the lamina propria, as well as goblet cell hyperplasia, and epithelial thickness compared to untreated controls.⁶⁶ A recombinant soluble form of the mouse membrane complement inhibitor complement receptor-related gene y (Crry) fused to the IgG1 hinge, CH2 and CH3 domains (Crry-Ig) has also been shown to prevent the development of airway hyperresponsiveness (determined by methacholine challenge testing), lower airway and lung eosinophilia, and decrease BAL interleukin (IL)-4, IL-5, and IL-13 levels.¹⁴⁹ Crry-Ig has decay-accelerating activity for both the classic and alternative pathways of complement, and also has cofactor activity for Factor I-mediated cleavage of C3b and C4b, suggesting that an approach of delivering complement regulators to mucosal sites may help reduce allergic airway inflammation, notably in the challenge phase. Although intrapulmonary delivery of anti-mouse C5 mAb achieves intrapulmonary C5 inhibition without affecting circulating C5 inhibition (abstract¹⁵⁰), none of the preclinical studies in airway complement modulation for asthma have yet been translated to a clinical setting.

Local C3aR antagonism in brain-dead donors reduces ischemia-reperfusion injury (IRI) in a preclinical mouse model of lung transplantation.¹⁵¹ Specifically, a single pretransplant nebulized dose of C3aR antagonist administered to the brain-dead (BD) donors significantly reduced IRI as compared to vehicle-treated BD donors, as well as reduced the severity of acute cellular rejection. The field of fusing complement regulators (e.g., Crry) to antibodies has advanced by utilizing single-chain antibodies derived from post-ischemic neoepitopes, such as modified annexin IV (B4) and a subset of phospholipids (C2). Since then, a novel complement inhibitor composed of single-chain antibody (scFv) derived from the C2 nAb linked to Crry (C2scFv-Crry) has been shown to inhibit IRI and delay acute rejection in a murine orthotopic lung transplant model.¹⁵²

In the context of gastrointestinal disease, a strategy of targeting complement inhibition to sites of complement activation by linking an iC3b/C3dg-binding fragment of mouse complement receptor 2 (CR2) to complement regulatory proteins, such as Crry or Factor H, has demonstrated promise in preclinical models. Intravenous administration of CR2-Crry 30 minutes after reperfusion reduces intestinal ischemia-reperfusion injury by influencing C3 deposition and neutrophil infiltration, without significantly affecting systemic complement inhibition.¹⁵³ Intraperitoneal administration of fusion protein CR2-Factor H to mice reduces complement activation (C5a levels in colon homogenate), inflammation and injury in the colon—measured by immune cell populations in the lamina propria such as CD19⁺ B cells, macrophages, and mature dendritic cells—and also reduces fibrosis (measured using picrosirius red staining).¹⁵⁴ Oral C5aR1 antagonism (PMX205) prevents DSS-induced colitis in mice, presumably through lowering proinflammatory cytokines and augmenting IL-4 and IL-10 levels.¹⁵⁵ However, clinical studies of complement modulation in the gut are just starting to emerge, and the field is open for carefully designed studies in the right patient population. ALXN1007, an anti-C5a monoclonal antibody, is well-tolerated by patients with newly diagnosed acute GVHD with lower gastrointestinal involvement (receiving concomitant steroids). However, it did not show efficacy in a phase 2 study and was associated with infection in nearly a quarter of the treated patients.¹⁵⁶ On the other hand, the use of eculizumab in patients with CHAPLE disease resulted in improvement in gastrointestinal pathology along with the re-establishment of a homeostatic gut microbiome.¹¹⁵ More recently, parenteral pozelimab (a fully human IgG4P (IgG4 with a proline substitution to promote light chain stabilization) antibody directed against C5) has been FDA-approved for this indication. Pozelimab administered to children who had markedly low weight-for-age and stature-for-age at baseline, as well as hypoalbuminemia due to protein-losing enteropathy, demonstrated serum albumin normalization and improvement with no worsening in clinical outcomes after having at least 52 weeks of treatment exposure.¹⁵⁷ However, these drugs are still administered systemically; as the field evolves, site-specific modulation will need to be tested to reduce systemic effects of complement inhibition.

Summary

Several independent preclinical and clinical studies have demonstrated that the complement system is locally active at mucosal sites, playing a role in infection, inflammation and repair. Moreover, systemic complement modulation has largely failed. Although the reasons for this failure need to be evaluated, as to whether it is due to inaccurate subphenotyping of patients, the inability of the drug to reach the site of inflammation or its cellular permeability, among other reasons. However, multiple preclinical studies have demonstrated that targeted modulation via site-specific delivery may be effective in modulating complement-mediated mucosal responses. Hence, research going forward should ideally take into consideration the kinetics of local complement-driven immune responses, the cell-type specific sourcing, the role of an intracellular complement system, and both canonical and non-canonical roles of this system at various mucosal sites, as we aim to promote both immune resistance and tissue resilience.¹⁵⁸