Clinical promise of next-generation complement therapeutics

Ischaemia–reperfusion (I/R) organ injury

Inappropriate complement activation on host surfaces exposed to I/R has been well documented in several preclinical models, and therapeutic intervention has shown promising results ameliorating key pathological markers³⁴. Complement activation is thought to predominantly involve the lectin and alternative pathways by virtue of the recognition of certain carbohydrate signatures (for example, terminal mannose or fucose)³⁵ on ischemic endothelial tissue by mannose-binding lectin or collectin 11 (REF.³⁶) and the capacity of the AP to amplify complement deposition on infarcted tissue³⁷ (FIG. 2b). The tight linkage of the LP and AP is made apparent by the ability of the LP-associated protease, mannose-binding lectin-associated serine protease 3 (MASP3), to act as the enzyme that supplies mature FD to the circulation through the cleavage of pro-FD³⁸ (FIG. 1), further supporting their overlapping pathogenic involvement in I/R injury. The CP has also been implicated in the pathogenesis of I/R injury through the binding of natural IgM to certain neoantigens exposed on ischemic tissue³⁹. The role of the terminal pathway has been debated, with conflicting results, in preclinical models of ischemic stroke³⁷.

I/R injury is the major cause of delayed graft function (DGF) in transplanted kidneys, an early event significantly affecting the graft’s long-term function and survival⁴⁰. Use of C1-INH, a broad serine protease inhibitor with activity against CP-associated and LP-associated proteases, attenuated inflammatory markers and prevented complement deposition in a porcine model of kidney I/R injury⁴⁰. Following up on these promising results, a phase II trial has evaluated the effect of C1-INH (Berinert) in reducing I/R damage in recipients of deceased donor kidney transplants, who are at high risk for DGF⁴¹ (TABLE 1).

Earlier efforts to target complement in the ischemic heart were translated into a phase II clinical trial in patients undergoing cardiopulmonary bypass that evaluated the efficacy of TP10/CDX1135 (Celldex), a recombinant form of soluble complement receptor 1 inhibiting the CP and AP C3/C5 convertases^42,43 (TABLE 1). TP10 treatment resulted in a reduction in myocardial infarction in male patients undergoing cardiopulmonary bypass, suggesting that central inhibition at the level of C3 could likely offer broader protection to ischemic organs by shutting down early C3 opsonization of ischemic tissues⁴⁴. However, limited efficacy and the observed gender specificity of TP10’s activity restricted the true translational value of this trial⁴⁵, and the development of this drug candidate for cardiovascular and other indications was later discontinued by Celldex.

Mirococept (APT070), a cytotopic complement inhibitor encompassing the first three short consensus repeat domains of CR1 fused to a membrane-tethering peptide and a membrane-inserting myristoyl group⁴⁶, has also entered clinical development as a treatment option for I/R injury during transplantation. A multicentre phase II trial is assessing the efficacy of APT070 in preventing kidney I/R injury and reducing the incidence of DGF in cadaveric renal allografts⁴⁷ (TABLE 1).

Proof of concept for the therapeutic efficacy of targeting C3 in I/R injury has largely been provided by studies of surface-directed AP inhibitors in models of intestinal and cerebral ischaemia–reperfusion and post-ischemic stroke. Indeed, surface-directed inhibition of C3 convertase activity via chimeric recombinant constructs that combine regulatory and C3 opsonin-binding moieties has ameliorated key pathological indices in preclinical models of stroke^48,49. C3 deficiency and site-targeted complement inhibition with either CR2-Crry (inhibiting all complement pathways) or CR2-FH (inhibiting the AP) have been found to significantly reduce infarct size and improve neurological recovery in the acute phase after stroke in a model of transient middle cerebral artery occlusion⁵⁰. A recent preclinical study has provided further leverage for the translational potential of C3 inhibition in ischemic stroke⁵¹ by employing a ‘fusion’ complement inhibitor (B4Crry) that targets all three complement pathways at the level of the C3 convertase⁵². Its inhibitory moiety, Crry, is a functional analogue of the human C3 regulators CD46 and CD55. By virtue of its single-chain variable fragment moiety (B4) that specifically recognizes a stroke-associated neoepitope in the ischemic brain, this inhibitor homes into the ischemic region, blocking C3 opsonization, and improving long-term motor and cognitive recovery after systemic delivery⁵². Overall, a series of preclinical I/R studies in various organs have provided a robust conceptual basis for developing C3-based therapeutics as new treatment options for ameliorating the early neurodegenerative consequences of ischemic and haemorrhagic stroke.

Adding further diversity to the toolbox of complement therapeutics tested in cerebral I/R injury, antibody blockade of the key LP enzyme, MASP2, has markedly improved neurological and histopathological outcomes after focal cerebral ischaemia, suggesting that LP targeting may be therapeutically beneficial in patients with ischemic stroke⁵³. Of note, the inhibitory MASP2 antibody (HG4) used in these studies is a derivative of the human MASP2-targeting mAb OMS721, which has been clinically developed by Omeros for several complement-mediated diseases⁵⁴.

Organ transplantation

It is increasingly appreciated that solid organ transplantation triggers several pathogenic pathways that are tightly intertwined with various effectors of complement activation, both in the vasculature and on the allograft surface^55,56. In addition to its cardinal role in triggering donor organ inflammatory damage via CP/LP-mediated neoepitope recognition during I/R, complement activation is also considered a major pathogenic driver in acute antibody-mediated rejection (ABMR) following allogeneic organ transplantation^32,57,58 (FIG. 2b). Moreover, its multifaceted role in driving adaptive immune stimulation and humoral responses is growingly appreciated as a factor further affecting long-term graft function³². Sensitization of transplant recipients by donor-specific human leukocyte antigen (HLA)-directed or ABO-directed alloantibodies marks a key initial trigger for both acute transplant rejection and for fuelling chronic cell-mediated organ injury and rejection (that is, across HLA and ABO transplantation barriers)^56,58. Given the pervasive nature of complement’s involvement in the transplant process, it is hardly surprising that pharmaceutical companies are devoting attention to targeted therapeutics for organ transplantation. Defining the optimal point for complement modulation in transplantation is still highly debated; however, effective complement modulation could likely be achieved in a narrow time window that could be vital for promoting early graft accommodation and thus prolonging the transplant’s survival^32,55.

Prompted by the clinical success of anti-C5 therapy in haematological and renal indications, recent studies have evaluated the safety and efficacy of complement intervention in mitigating allograft injury in conditions ranging from ABMR and recurrent post-transplant aHUS, to post-transplant recurrent C3 glomerulopathies and antiphospholipid antibody syndrome⁵⁹. Therapeutic modulation at the level of C5 has shown promise in alleviating complications of early ABMR and reducing incidence rates of kidney graft rejection in transplant patients with circulating anti-HLA donor-specific antibodies^60,61 (TABLE 1). However, a fair proportion of graft recipients fail to respond to anti-C5 treatment, implying that a broader and more comprehensive intervention may be warranted. Two pilot studies have shown that C1-INH treatment, in combination with intravenous (IV) immunoglobulin, is well tolerated in the treatment of ABMR and that this regimen is associated with improved renal allograft function^62,63. Moreover, Cinryze, a plasma-derived preparation of C1-INH, has advanced to a multicentre phase III trial assessing its efficacy for the treatment of ABMR in sensitized kidney transplant patients⁶⁴ (TABLE 1). It should be noted, however, that the beneficial impact of C1-INH treatment on kidney graft survival may well extend beyond its inhibitory activity in the CP or LP, summoning also its ability to intercept proteases of the coagulation and fibrinolytic/kinin pathways.

While C1-INH could potentially reduce transplant-associated CP and LP activation (for example, during I/R injury), given its broad inhibitory activity on the serine proteases C1s/r and MASPs, respectively, the amplifying capacity of the AP on the opsonized surfaces of the donor organ would remain unhindered, thus imposing an inflammatory burden on graft function and survival³². In this regard, C3 convertase modulation or direct C3 interception could feature as plausible strategies for improving graft accommodation in transplant recipients, given the multifaceted impact of C3 activation on cell-mediated and humoral inflammatory pathways operating locally on the graft’s surface and in the vascular endothelium⁶⁵. Endorsing the potential of C3 intervention in clinical transplantation across ABO barriers, Amyndas has already announced plans to develop its C3-targeted inhibitor AMY-101 as a treatment option for increasing graft accommodation in patients undergoing ABO-incompatible kidney transplantation⁶⁶ (TABLE 1).

Sepsis-associated inflammation

Polymicrobial sepsis is a systemic inflammatory condition associated with high morbidity and mortality⁶⁷. It evokes excessive intravascular activation of the complement system that, coupled to overwhelming cytokine signalling, contributes to systemic innate immune dysfunction and thromboinflammation that are disseminated to distal organs, culminating in multiorgan failure⁶⁸. The pathogenic role of distinct complement effectors in sepsis-induced inflammatory damage has been demonstrated in both rodent and primate models of Escherichia coli-induced or polymicrobial (caecal ligation puncture) sepsis^69,70. From a translational perspective, small-sized C3 inhibitors of the compstatin family and C5 inhibitors (for example, RA101295, Ra Pharma) have both shown therapeutic efficacy in preclinical models of sepsis, particularly in combination with CD14 blockade, an approach that abrogates proinflammatory lipopolysaccharide signalling^71,72. Of note, complement inhibition at the level of C3 or C5 has shown promising results in primate models of sepsis by maintaining vital organ integrity and attenuating thromboinflammatory damage when delivered as an early rescue regimen^70,73.

While all these approaches have yielded promising results at a preclinical stage, they are still far from being translated into clinically meaningful therapies. Nonetheless, clinical efforts have been spearheaded by a humanized mAb targeting C5a (CaCP29/IFX-1, InflaRx), a key inflammatory effector in sepsis (FIG. 2a). This antibody therapeutic has undergone phase II trials in patients with severe sepsis or septic shock with evidence of organ dysfunction and pulmonary or abdominal infection⁷⁴, but no further plans have been announced for these indications thus far⁷⁴ (TABLE 1). Of note, clinical development of IFX-1 for another inflammatory indication, the chronic skin disorder hidradenitis suppurativa, recently led to a major setback, as this anti-C5a agent failed to meet its primary end point (that is, clinical response score) in a large multicentre phase II study involving patients with moderate to severe hidradenitis suppurativa⁷⁵ (TABLE 1). C1-INH has also been evaluated as a potential modulatory strategy for sepsis in a pilot study of human endotoxaemia closely resembling lipopolysaccharide-induced septic shock models. Although this study advanced to phase III, further plans for the clinical development of C1-INH as a treatment for septicaemia have not been announced⁷⁶ (TABLE 1). C3 interception by the peptidic C3 inhibitor Cp40 afforded early organ protection, reducing systemic thromboinflammation in non-human primates (NHPs) subjected to trauma-induced haemorrhagic shock, thereby indicating that C3 may be a tractable target for developing ‘rescue’ therapies for trauma-associated multiorgan failure⁶⁹. However, the complex nature of the inflammatory circuits involved in septicaemia and sepsis or trauma-induced multiorgan dysfunction poses a significant challenge in developing targeted anti-inflammatories based exclusively on complement modulation. This challenging situation is reflected by the limited commitment of complement drug makers to pursuing clinical programmes in this disease area. From a different perspective, complement activation can be exploited as a beneficial effector mechanism for containing hospital-acquired bacterial infections and dampening systemic inflammation. In this regard, Aridis has been developing mAbs against Pseudomonas aeruginosa (that is, AR-101/Aerumab) with enhanced complement-fixing properties that facilitate opsonophagocytic clearance of the pathogen. AR-101 has completed phase IIa testing and shows promise as a new adjunctive anti-infective treatment for patients with severe P. aeruginosa pneumonia⁷⁷.

Haemodialysis-induced inflammation

Haemodialysis (HD) represents the mainstay of clinical management for alleviating the complications of a failing kidney in patients with end-stage renal disease (ESRD)^78,79. The clinical challenge in treating patients with ESRD is further accentuated by the staggering shortage of donor organs worldwide⁸⁰. Interestingly, even modern ‘biocompatible’ HD filters evoke appreciable levels of complement activation and fuel thromboinflammatory responses that can propagate inflammation to distal organs¹⁰. The parallel activation of multiple innate immune systems, including the contact and coagulation cascades, imposes a growing burden on patients with ESRD that causes deterioration in their quality of life over time, increasing their risk for anaemia, atherosclerosis and cardiovascular disease¹⁰.

Several preclinical studies have supported the pathogenic involvement of all three complement pathways in HD-induced inflammation, while indicating a critical role for the AP in amplifying the complement-driven proinflammatory response elicited during HD^33,78. Recent translational efforts in primate and human models of HD-induced inflammation have provided proof of concept for the therapeutic targeting of complement. Administration of a single dose of C1-INH in an ex vivo perfusion model of HD resulted in attenuated C3 activation and decreased levels of key proinflammatory and prothrombotic markers implicated in cardiovascular disease predisposition⁸¹. These findings are in line with earlier observations in a clinically relevant NHP model of HD-induced inflammation in which C3 inhibition has yielded promising results⁸². A single intravenous bolus injection of the compstatin analogue Cp40 (that is, the active component of AMY-101) completely abrogated complement activation in cynomolgus monkeys subjected to HD and led to an increase of the anti-inflammatory cytokine IL-10 during the HD session⁸². Taken together, these studies have identified a window of opportunity for developing anti-inflammatory therapies in HD that could target C3 activation on the biomaterial surface³³ (TABLE 1). It should be clarified, however, that the beneficial activity of C1-INH in HD may extend well beyond its capacity to modulate the CP and LP on the biomaterial surface because of its promiscuous inhibitory activity on other proteases of the fibrinolytic, clotting and kinin pathways¹⁰. Of note, the transient nature of complement inhibition during an HD session significantly lowers the risk of infectious complications, thus obviating the need for prophylactic vaccination or other antimicrobial treatments. With the production costs of peptides being lower than those of classical biologics (for example, antibodies)^83,84, and with the prospect of injecting C3-inhibitory peptides directly into the dialysis circuit for a short duration of treatment, this immunomodulatory strategy could likely offer an effective and comparatively affordable treatment for patients undergoing HD.

Periodontal disease

A plethora of preclinical and translational studies are reshaping our understanding of the fine immunoregulatory role of tissue-resident microbiota in health and disease⁸⁵. Periodontitis is a prevalent oral inflammatory disease that afflicts millions of people in the Western world⁸⁶. It is instigated by dysbiotic microbiota that thrive on local inflammation in the gingival cavity, promoting the gradual destruction of the tooth-supporting tissues and eventually leads to bone loss⁸⁷. The inflammophilic microbiome causes periodontal disease by subverting the innate immune response at the level of complement activation and Toll-like receptor signalling⁸⁸. Preclinical genetic and pharmacological studies of complement inhibition in rodent and NHP models of naturally occurring and ligature-induced periodontitis have unveiled the key pathogenic role of the complement–Toll-like receptor crosstalk in driving microbial dysbiosis, periodontal inflammation and osteoclastogenic bone loss¹⁴. More importantly, these studies have identified potential points of therapeutic intervention such as C3, the central hub of the system, and the downstream proinflammatory C5a–C5aR1 signalling axis¹⁴. Local injection of the C3-targeted inhibitory peptide Cp40 in the interdental papilla of primates afflicted with periodontitis has led to clinically meaningful reduction of key inflammatory indices, attenuating the osteoclastogenic bone loss that occurs in periodontal disease⁸⁹. Following up on preclinical safety and tolerability studies⁹⁰, the Cp40-based therapeutic AMY-101 recently received Investigational New Drug approval by the FDA for the conduct of the first clinical study to evaluate its efficacy in adults with gingivitis (ClinicalTrials.gov identifier: NCT03694444) (TABLE 1).

Taken together, these translational studies attest to the safety and efficacy of transiently applied local C3 intervention in oral inflammatory diseases (for example, periodontitis) and pave the way for a novel adjunctive, host-modulatory therapy that could benefit patients with periodontal diseases by targeting complement in the oral cavity.

Haematological disorders

The field of complement-mediated haematological diseases is justifiably dominated by PNH and the clinical success of anti-C5 therapy (eculizumab)^20,100. PNH is driven by the spontaneous intravascular haemolysis of erythrocytes that lack the glycosylphosphatidylinositol-anchored complement regulators CD55 and CD59 and therefore succumb to autologous complement attack and MAC-mediated intravascular haemolysis⁹⁵ (FIG. 2d). The approval of eculizumab drastically changed the landscape of PNH by introducing the first aetiological therapy for these patients^20,101. Despite this progress, unmet clinical needs have emerged, including the genetically driven refractory phenotype to anti-C5 in certain patients, the residual haemolysis of C3-opsonized PNH cells in extravascular compartments, and the pharmacokinetic/pharmacodynamic (PK/PD) breakthrough haemolysis observed in certain cases that evoke strong complement activation (that is, acute infections), irrespective of drug dosing levels^93,102,103. Moreover, the excessive costs of current anti-C5 therapy (typically, some US$500,000 per year and patient, depending on the market)¹⁰⁴ cannot be overlooked, since they impose a sizeable economic burden on health-care systems and limit drug access in developing countries¹⁰⁵ (an overview of regulatory and economic aspects pertaining to complement drug development is provided in BOX 2). Altogether, these clinico-economic challenges have rekindled the interest of drug companies in developing biosimilars of eculizumab, alternative anti-C5 therapeutics, or drug candidates targeting upstream components of the AP or the central protein of the complement cascade, C3 (REF.¹⁸).

Addressing the challenges of patient compliance and drug plasma residence, Alexion has recently launched Soliris’ successor, Ultomiris (ravulizumab), a long-acting version of eculizumab that differs only in four amino acids but exhibits longer plasma residence thanks to neonatal Fc receptor-dependent antibody recycling¹⁰⁶. This antibody engineering allows for an extended IV dosing interval of 8 weeks, considerably longer than that of eculizumab^29,30,107. After two multicentre phase III studies established the non-inferiority of ravulizumab to eculizumab in both patients with clinically stable PNH during previous eculizumab therapy and in complement inhibitor-naive patients^29,30, Ultomiris gained FDA approval for PNH in December 2018. An approval in other markets and for other indications, including aHUS and gMG, is expected to follow soon.

Since details about the expected treatment costs are scarce, it remains to be seen whether ravulizumab will alleviate the burden on the health-care system, in addition to the burden for patients regarding treatment frequency.

The arsenal of clinically developed anti-C5 agents for PNH features several mAbs that target a different epitope of C5 than does eculizumab, thus projecting clinical benefit even for those patients bearing the therapy-refractory p.R885H C5 mutation⁹³. SKY59/RO7112689, a mAb developed by Roche and Chugai, displays a favourable PK profile for chronic application by exploiting a pH-dependent mechanism of neonatal Fc receptor-mediated antibody recycling¹⁰⁸. Currently, SKY59 is being evaluated in phase I/II trials as a treatment option for PNH, with initial reports of high subcutaneous bioavailability and prolonged plasma residence in healthy volunteers predicting a promising clinical development path (NCT03157635). Other C5-targeting antibodies in clinical development for PNH include tesidolumab (LFG316, Novartis)¹⁰⁹ and pozelimab (REGN3918, Regeneron), which are registered in phase II and I studies, respectively¹¹⁰ (TABLE 1).

Besides larger biologics, the diversified toolkit of C5-targeting agents comprises smaller proteins and peptides with the capacity to block C5 cleavage. Specifically, coversin (Akari Therapeutics), a recombinant C5/leukotriene B4-targeted inhibitor derived from the soft tick Ornithodoros moubata, is currently being developed for PNH as a subcutaneous treatment option¹¹¹. After achieving its primary end point — lactate dehydrogenase reduction — in a phase Ib trial in patients with PNH, coversin is now advancing to phase II trials, offering another treatment option with increased patient compliance (due to subcutaneous (SQ) dosing) that can also circumvent the genetic resistance to eculizumab¹¹² (TABLE 1). Constrained peptides are also gaining momentum and entering late-stage clinical development¹¹³. Zilucoplan (RA101495, Ra Pharma) is a macrocyclic peptide targeting C5 in a region distinct from that targeted by eculizumab¹¹⁴. This C5 inhibitor has entered clinical development for PNH as an SQ formulation and shown promising results in two phase Ib/II trials in treatment-naive and eculizumab-treated patients^115,116; it has now advanced to an extension phase II study in patients with PNH (TABLE 1). From a different mechanistic perspective, Alnylam is developing an RNAi therapeutic (cemdisiran/ALN-CC5) that targets the hepatic expression of C5 at the mRNA level¹¹⁷. Whereas monotherapy with cemdisiran may hold promise in other indications (that is, aHUS), this strategy has not yielded optimal results in PNH, and Alnylam has redirected its programme to combine cemdisiran dosing with concomitant eculizumab treatment, albeit at a lower dosage (TABLE 1).

Clearly, all the above-mentioned approaches are capitalizing on the success of anti-C5 therapy, without significant deviation from a mechanistic standpoint. Their rationale resides in the extension of dosing intervals and in increasing patient compliance through self-administration of the drug.

Therapeutic strategies that target the AP amplification loop or the C3 protein itself offer a robust mechanistic basis for inhibiting complement in PNH, with potentially broader therapeutic coverage by blocking both intravascular and extravascular haemolysis⁹². Considerable therapeutic insight in this direction has been gained from a family of cyclic peptides targeting C3, the compstatins (for updated reviews on the design and preclinical development of compstatins, please see REFS^118,119). Compstatins bind to a shallow pocket formed by the macroglobulin ring of the β chain of human and NHP C3 and selectively block convertase-mediated cleavage of C3 independent of the initiation pathway¹²⁰. In fact, C3-targeted inhibition by these small peptidic inhibitors has shown clinical promise as a broader therapeutic approach by preventing intravascular MAC-mediated haemolysis and also abrogating the residual haemolysis that can result from extravascular phagocytosis of C3-opsonized PNH cells exposed to anti-C5 agents¹²¹.

These mechanistic attributes have propelled the clinical development of two compstatin-based drug leads, that is, the PEGylated therapeutic APL-2 (Apellis) and the non-PEGylated AMY-101/Cp40 (REF.¹¹⁸) (Amyndas). APL-2 has been derivatized from POT-4, corresponding to the second-generation compstatin analogue 4(1MeW)¹¹⁸, through conjugation of two peptide moieties to an extended PEG linker¹²². Both the 4(1MeW) analogue and AMY-101/Cp40 have been developed by purely academic efforts at the University of Pennsylvania and subsequently licensed to Apellis and Amyndas, respectively. APL-2 is currently in phase II studies as an add-on therapy to eculizumab¹²³ and as a monotherapy in treatment-naive patients with PNH¹²⁴ (TABLE 1). After reporting successful weaning of patients from eculizumab and switching to APL-2 monotherapy with favourable haematological results, Apellis has announced the enrolment of patients in a phase III study in which the C3 inhibitor’s efficacy will be benchmarked against eculizumab¹²⁵. It has also registered a multicentre extension study to evaluate the long-term safety and efficacy of APL-2 in PNH¹²⁶. Amyndas has completed phase I studies in human volunteers, with good safety and tolerability profiles for their non-PEGylated third-generation compstatin analogue AMY-101 and has announced plans for phase II studies in PNH and other complement-mediated indications¹²⁷ (TABLE 1). AMY-101 is based on the Cp40 compstatin analogue, which has shown excellent PK/PD profiles in NHP, sustained inhibitory potency and broad therapeutic efficacy in blocking both intravascular and extravascular haemolysis of PNH erythrocytes¹²¹. Of note, the design of Cp40 analogues with targeted mini-PEGylation in the N-terminus or the addition of charged amino acids at the C-terminus has led to derivatives with higher solubility and improved PK profiles, thus broadening the spectrum of administration routes and likely reducing the dosing frequency of these peptidic drugs in chronic regimens¹²⁸.

Efforts to develop complement therapeutics amenable to chronic administration have culminated in two orally available, small-sized inhibitors that target the AP amplification loop at different steps¹⁸. To this end, both Novartis and Achillion are advancing two small-molecule inhibitors to late-stage clinical development that block the serine proteases Factor B (FB) and FD, respectively¹²⁹. The FB inhibitor LNP023 (Novartis) blocks the active site of FB and the Bb fragment and has entered phase II trials for PNH^130,131 (TABLE 1). Meanwhile, Achillion is dosing patients with PNH with its orally available FD inhibitor, ACH-4471, which blocks the catalytic site of FD, preventing AP C3 convertase formation¹³². ACH-4471 is being evaluated in two phase II trials, both as a monotherapy in treatment-naive patients with PNH and as an add-on to eculizumab in patients with suboptimal response to anti-C5 therapy¹³³ (TABLE 1). Achillion has also announced plans to advance its next-generation FD inhibitors ACH-5228 and ACH-5548 to phase I trials, aiming for higher AP inhibition along with a reduced dosing frequency¹³⁴. Of note, a series of compstatin derivatives with improved PK properties and appreciable levels of oral bioavailability in NHP are now in preclinical development¹³⁵. The further development of oral formulations for enhancing the bioavailability of these peptidic C3 inhibitors is expected to largely improve complement therapeutic intervention, especially in those indications that rely on chronic treatment (for example, PNH).

Whereas PNH represents an archetypal disease for testing AP therapeutics, other haemolytic conditions with prominent complement involvement require tailored approaches that better align with their distinct pathophysiology. For example, autoimmune haemolytic anaemias (AIHAs) such as cold agglutinin disease (CAD) or warm antibody AIHA are fuelled by autoantibodies that trigger CP activation on erythrocytes (FIG. 2c). This attribute has propelled the clinical development of CP-targeted inhibitors for these indications. After receiving orphan drug designation for CAD in 2017, sutimlimab (BIV009/TNT009, Sanofi/Bioverativ), a humanized mAb targeting the C1s protease of the CP, entered late-stage clinical development, being evaluated as an IV infusion in two separate phase III trials in patients with CAD^136,137 (TABLE 1). Further expanding the spectrum of therapeutics for AIHAs, Apellis has also instigated a phase II trial (PLAUDIT) to assess the biological efficacy of its C3 inhibitor, APL-2, in patients with CAD or warm antibody AIHA¹³⁸ (TABLE 1). While the advent of clinical-stage drug candidates targeting various enzymes of the complement cascade (that is, C1s, FB and FD) marks an important milestone in complement drug discovery, the potential lack of therapeutic coverage due to residual enzymatic activity in cases of strong complement activation or insufficient dosing (for example, a missed dose in a daily oral administration schedule) should not be ruled out and may warrant further investigations.

The spectrum of haemolytic conditions fuelled by complement is continuously expanding, with new indications entering the clinical development spotlight⁹⁵. Sickle cell disease (SCD), a genetically driven haemolytic disorder, has long been associated with an overactive AP but, until recently, has been neglected from a translational standpoint¹³⁹. AP activation is likely triggered by increased exposure of certain membrane phospho lipids (for example, phosphatidylserine) on SCD erythrocytes and further fuelled by the release of haem from lysed erythrocytes^139,140. Moreover, the contribution of the AP to vaso-occlusive episodes (that is, SCD crises) and delayed haemolytic transfusion reactions in patients with SCD has been proposed¹⁴¹. In this regard, the translational value of AP modulation in SCD has recently been demonstrated in a study showing that both Factor H (FH) and smaller FH constructs can block the adhesion of sickled red blood cells to the inflamed vascular endothelium of patients with SCD¹⁴². Future investigations will have to determine the clinical potential of such modulatory strategies in genetic disorders such as SCD. Of note, haem-induced AP activation is now recognized as an exacerbating factor that leads to bystander haemolysis in both SCD and malaria-induced anaemia¹⁴³. In this regard, C3 inhibition by the compstatin analogue Cp40 can abrogate haem-induced complement activation and C3 fragment deposition on erythrocytes, pointing to the therapeutic potential of C3-based inhibitors for ameliorating malaria-induced anaemia in patients infected with Plasmodium¹⁴⁴. While direct C3 inhibition or C3 convertase modulation might be viable approaches to protect the endothelium from complement damage during intravascular haemolysis, targeting surface ‘modifiers’, such a P-selectin, that are upregulated in response to haem exposure, might point to an alternative therapeutic approach for quenching complement-mediated endothelial injury. In this respect, P-selectin was shown to act as a scaffold for C3 deposition on endothelial surfaces, and antibody blockade of its interaction with C3 resulted in attenuated complement deposition and reduced endothelial damage in haemolytic mice¹⁴⁵. These data strongly suggest that blocking tissue-specific docking sites for C3 fragment deposition might have additive or stand-alone therapeutic potential in diseases such as SCD.

Chronic renal disorders

Given its unique anatomical and functional blueprint, the kidney represents one of the organs most susceptible to complement-mediated inflammatory damage^1,146. Disorders that predominantly affect the glomerular function, leading to acute or chronic kidney injury, have therefore been in the crosshairs of pharmaceutical companies developing complement therapeutics¹⁸. Under steady-state conditions, the glomerular endothelium is protected from autologous complement activation by the glycocalyx, an overlying layer of proteoglycans and glycolipids that sequester fluid-phase regulators (for example, FH), thus preventing C3 opsonization and AP amplification¹⁴⁶. This structural adaptation, combined with the endothelial expression of surface-bound regulators of the C3/C5 convertases, maintains renal homeostasis. In renal disorders involving high complement turnover, this regulatory shield is compromised by genetic alterations and acquired defects (that is, rare or common gene variants in complement activators or regulators, or autoantibodies)^147,148 that collectively render the glomeruli prone to complement damage.

Ocular inflammatory disorders

In many ways, the ocular disease space has galvanized complement drug discovery in the post-eculizumab era¹⁸. The eye is an immune-privileged organ with anatomical barriers that pose certain challenges in terms of local drug administration, bioavailability and sustained activity¹⁸⁶. Complement has long been known to contribute to ocular tissue homeostasis in conjunction with innate immune sentinels such as resident microglia and other phagocytes¹⁵. However, more than a decade ago, it was thrust into the spotlight of pharmaceutical research when a common gene variation in the complement regulator FH, a p.Y402H polymorphism, became associated with increased risk for age-related macular degeneration (AMD)^187–189. AMD is a chronic inflammatory disease of the retina that represents a leading cause of irreversible vision loss in the industrialized world¹⁵. In its early stages, it manifests as dense protein-rich and lipid-rich deposits, termed drusen, that accumulate in Bruch’s membrane, a specialized extracellular matrix-rich layer that separates the retinal epithelium from the underlying choroid¹⁵. AMD progresses from its early form, marked by drusen deposition and sub-retinal inflammation, into two advanced forms of AMD. In the advanced ‘dry’ form (geographic atrophy, GA) choroidal vessels deteriorate, causing widespread retinal pigmented epithelial cell death and photoreceptor loss. The second advanced form, termed ‘wet’ (neovascular) AMD, features predominantly leaky blood vessels protruding into the sub-retinal space, vascular endothelial growth factor (VEGF)-driven neovascularization and further inflammatory damage, leading to retinal atrophy and vision loss^15,186. Eyes with GA can also develop wet AMD and vice versa. Dry AMD/GA still lacks effective treatment, while the mainstay of treatment for wet AMD is the intravitreal injection of anti-VEGF agents¹⁵.

Whereas the role of complement in AMD pathogenesis is not fully elucidated, genome-wide association studies have identified a series of common or rare risk variants in complement genes that enhance the predisposition to AMD, likely through an overactive AP or loss of AP regulatory control in the sub-retinal space¹⁹⁰. Encouraged by these observations, Genentech/Roche coordinated a series of clinical trials in GA with lampalizumab, a humanized IgG Fab fragment directed against FD, the rate-limiting protease of the AP. Lampalizumab binds to an FD exosite, blocking substrate access and activation of the AP C3 proconvertase (C3bB)¹⁹¹. Although lampalizumab reduced GA lesion size in the MAHALO phase II trial, with greater drug benefit shown in a subgroup of CFI risk-allele carriers¹⁹², it failed to achieve its primary end point (reduction in GA lesions) in two large multicentre phase III trials in patients with GA²⁷ (TABLE 1). Whereas the lampalizumab case has constituted a setback in complement drug discovery for AMD, it has also raised awareness about comprehensive patient stratification in trials, drug bioavailability and delivery into ocular tissues and the potential contribution of FD bypass mechanisms that may become operative in such pathologies, thus skewing drug efficacy¹⁹³. Furthermore, targeting serine proteases that amplify complement responses (that is, FD and FB) may entail a risk of reducing the enzyme’s activity to an extent that a residual, minimal amount of enzyme will still suffice to fuel AP activation, thereby causing pharmacodynamic breakthrough and exacerbation of the disease. These potential issues surrounding lampalizumab’s trials in AMD remain relevant to other indications and need further investigation to resolve any ambiguities.

While FD targeting in ocular indications remains within the scope of certain companies¹²⁹, the drug industry is devoting more attention to C3 inhibition as a promising new platform for treating AMD. Earlier translational efforts with the compstatin analogue POT-4 (Potentia/Alcon)¹⁹⁴ had shown initial efficacy in NHP models of early AMD¹⁹⁵, but failed to recapitulate this response in a phase I study in wet AMD, likely because of insufficient dosing¹⁹⁴. To compensate for the lower activity of POT-4, APL-2 has incorporated two moieties of this peptide via a bridging PEG linker¹²². A safety and efficacy phase II trial of APL-2 in patients with GA (FILLY) has recently been completed, showing reductions in GA lesion size independently of genetic variants that could modulate disease progression¹⁹⁶. This PEGylated C3 therapeutic has now been advanced to two multicentre phase III studies in patients with GA on a monthly and every 2 months intravitreal dosing scheme¹⁹⁷ (TABLE 1). Whereas PEGylation may have afforded a longer half-life to APL-2 compared with that of its predecessor APL-1/POT-4, this approach potentially carries a risk of fuelling adverse events related to PEG accumulation (for example, tissue vacuolation)¹⁹⁸ or induction of choroidal neovascularization (conversion to wet AMD) after repeated subretinal injections¹⁹⁹. While other PEGylated antiangiogenic agents have been effectively used in patients with wet AMD (for example, Macugen, pegaptanib sodium), the relative PEG ‘burden’ of a drug, dictated by differences in dosing amounts, should be taken into account in each case. Furthermore, the possibility that C3 inhibition in the eye might favour a local M1 to M2 macrophage conversion, thereby promoting neovascularization as part of the tissue repair process, cannot be ruled out^200,201. Although the presence of a small percentage of wet AMD conversions in the FILLY trial could be attributed to a subclinical course of disease undetectable at the baseline, concerns about neovascular AMD conversion in APL-2-treated subjects warrant further investigation.

Non-PEGylated third-generation and fourth-generation compstatins^128,202 (developed by Amyndas) with improved target affinity and PK profiles for chronic administration may confer benefits regarding dosing frequency and PEG-related effects⁶⁶. The prolonged residence of a novel C3 inhibitor of undisclosed structure, termed AMY-106 (Amyndas) in ocular tissues at C3-saturating levels, extending over 3 months after a single intravitreal injection in NHP, indicates its clinical potential for the treatment of ocular indications associated with C3 dysregulation (for example, AMD)⁶⁶. Of note, prolonged exposure of patients to C3 inhibitors (for example, APL-2 dosing in patients with PNH or GA) has not evoked any safety concerns thus far, further endorsing the clinical potential of this targeting strategy both systemically and locally.

AMD remains a fertile backdrop for testing new complement therapeutics spanning the entire spectrum of targets¹²⁹. IONIS-FB-L_Rx (Ionis/Roche), an antisense oligonucleotide targeting hepatic expression of FB, has completed phase I studies and is scheduled for a phase II study in patients with GA²⁰³ (TABLE 1). Systemic deli very of IONIS-FB-L_Rx has achieved FB reduction in both blood and ocular tissues of NHP, indicating its potential for modulating AMD progression²⁰⁴. However, in view of the local production of AP components in ocular tissues (that is, FB)²⁰⁵ and the controversy surrounding the relative contribution of systemic versus local complement stores to retinal degeneration²⁰⁶, it remains to be seen whether this elegant AP modulatory strategy will show efficacy in human AMD trials. Despite the failure of eculizumab to show clinical benefit in a phase II AMD trial²⁰⁷, the ocular drug space is considering alternative anti-C5 agents such as the antibody LFG316 (Novartis)²⁰⁸ and the C5-targeting aptamer (single-stranded RNA) Zimura/avacincaptad pegol (Ophthotech)²⁰⁹. Combinations of these anti-C5 agents with other therapeutics targeting the AP (anti-properdin, GLC561) or VEGF (Lucentis, anti-VEGF) are also in clinical evaluation in patients with AMD, with trial results eagerly awaited (further details on these trials can be found in REF.¹²⁹) (TABLE 1).

Safety

Given the key role of complement in pathogen clearance, safety has long been debated as a potential bottleneck for advancing complement therapeutics to the clinic. However, the clinical record of eculizumab has largely dissolved these concerns, showing that prophylactic vaccination against certain encapsulated bacteria (for example, Neisseria spp., Haemophilus spp. or Streptococcus spp.) can effectively protect patients undergoing chronic treatment¹⁰¹. The advent of C3-based therapeutics in chronic indications such as PNH has reignited this discussion, with hitherto largely hypothetical arguments⁹¹. Notably, primary C3 deficiencies in humans have only been associated with increased susceptibility to opportunistic infections during childhood, while this concern is largely mitigated during adulthood because of a fully matured adaptive immune compartment²¹⁰. More importantly, insights from ongoing clinical trials applying C3-based inhibitors with prophylactic vaccination regimens have reconciled these concerns with actual clinical data. The safety and tolerability of the C3-inhibitor APL-2 in ongoing phase II/III trials in patients with PNH dosed for an extended time window and the recently reported safety of AMY-101/Cp40 in NHP under prolonged treatment have significantly negated the risk originally feared with this therapeutic approach, opening avenues for exploring such inhibitors, particularly as local treatments (for example, in AMD) or when applied to acute-phase indications (for example, haemodialysis treatment)^125,211. The increased opsonophagocytic killing of meningococci in vaccinated individuals treated with AP inhibitors, when compared to anti-C5 agents, further supports the low infectious risk of upstream (AP) inhibition and adds more options for tailoring anti-complement therapies²¹².

Towards personalized complement therapies

New structural insights

Structural advances in understanding complement mechanisms at the atomic level can point the way to new targets and treatment modalities. For instance, the recent elucidation of the structural basis for the formation of the initiating complexes of the CP, such as C1-IgG1 or C1-IgM^222–225, and the hexameric arrangement of IgG-antigen complexes on surfaces²²⁶, have not only provided important insights into the mode of interaction between the main recognition unit of C1, C1q and the C1s/r serine proteases, but also revealed how intact C1 recognizes various CP activators (for example, IgG or IgM-containing immune complexes). This knowledge could leverage the design of tailored CP inhibitors for diseases driven by IgM neoantigen recognition (for example, ischaemias). Altogether, these structural studies have not only indicated a conserved mechanism of complement activation through the CP and LP²²⁵ but have also enabled the design of more effective antibody therapeutics, termed ‘HexaBodies’, which elicit potent complement-dependent cytotoxic responses against certain haematological tumours (for example, anti-CD20, anti-CD38)²²⁷. Extending the spectrum of targeted indications, GenMab has recently registered its HexaBody GEN1029 in first-in-human phase I/II trials in patients with solid cancers²²⁸. For many years, the structure of C5aR1/CD88, a key signalling effector involved in a range of inflammatory or autoimmune pathologies and already targeted by clinical-stage therapeutics such as Avacopan/CCX118, remained elusive²²⁹. The recent report of the crystal structures of human C5aR1 in complex with peptidic and non-peptidic antagonists has revealed unique structural features underlying C5aR1’s signalling activity²³⁰. These structural insights may enable the design of a new class of orthosteric or allosteric C5aR1 inhibitors that could be amenable to clinical development, providing promising drug leads that can modulate the proinflammatory C5a–C5aR1 axis.