Novel strategies to target the humoral alloimmune response

Antibody‐mediated rejection (ABMR) represents a major cause of late allograft loss in solid organ transplantation worldwide. This process is driven by donor‐specific antibodies (DSA), which develop either de‐novo or, in sensitized patients, are preformed at the time of transplantation. Effective targeting of ABMR has been hampered by a lack of robust randomized controlled trials (RCT), required for the regulatory approval of new therapeutics. In this review, we discuss the evidence behind the present “standard” of care and recent progress in the development of novel strategies targeting different aspects of the alloimmune humoral response, including naïve and memory B‐cell activation, the germinal centre reaction, plasma cell survival and antibody effector functions. In particular, we focus on co‐stimulation blockade and its combination with next‐generation proteasome inhibitors, new depleting monoclonal antibodies (anti‐CD19, anti‐BCMA, anti‐CD38, anti‐CD138), interleukin‐6 blockade, complement inhibition and DSA degradation. These treatment modalities, when used in the appropriate clinical context and combination, have the potential to finally improve long‐term allograft survival.


| INTRODUCTION
In recent decades, short-term kidney transplant outcomes have improved, because of technical advances and more effective prevention and treatment of acute cellular rejection (TCMR), with 1-year graft survival exceeding 95%. In contrast, long-term kidney allograft survival has only marginally improved. 1,2 Recent advances in both histological and molecular diagnostics, including singleantigen bead assays and tissue transcriptomics, have implicated antibody-mediated rejection (ABMR) as a leading cause of late allograft loss. 3,4 However, a proven, effective treatment for ABMR is lacking, because of a dearth of rigorous well-controlled, sufficiently powered randomized clinical trials (RCT).
Humoral alloimmune responses are characterized by the presence of donor-specific antibodies (DSA), which interact with graft endothelium, activate complement and trigger local inflammation, leading to transplant dysfunction. 5

This
This review was invited and edited by the Reviews Editor Katharina Fleischhauer. process can be hyperacute (minutes post-transplant) in the context of pre-formed DSAs (e.g., in AB0 blood group incompatibility), acute (the first 6 months post-transplant) or chronic (months to years post-transplant). Both acute and chronic ABMR may be mediated by memory responses in sensitized recipients, or by de-novo DSA generation in nonsensitized patients, for example, because of non-adherence and/or early T-cell mediated rejection. 6

| B CELL BIOLOGY IN ALLOIMMUNE RESPONSES
De-novo complement-binding DSAs arise following activation of naïve B cells and their subsequent maturation into alloantigen-specific memory B cells and antibodysecreting cells (ASCs), including both short-lived plasmablasts and long-lived plasma cells. 7 This process begins when alloantigen is delivered into secondary lymphoid organs (Figure 1). Donor-specific naïve B cells recognize antigen with their B-cell receptor (BCR), generating an activating signal (signal 1) that is modulated by BCR co-receptors such as activating CR2/CD21 or inhibitory CD22, CD72 and FcγRIIB. Activated B cells migrate from the follicle to the border of the T-cell zone, where they present alloantigen peptides in the context of MHCII to helper T cells, and receive an activating signal through CD40/CD40L in return (signal 2). Alloreactive helper T cells are generated when engaging cognate antigen-MHCII complex on the surface of activated F I G U R E 1 Alloimmune humoral response and previous alloantigenic experience. In non-sensitized patients, alloantigens released from damaged allograft tissue (e.g., during complement-mediated ischaemia-reperfusion injury, IRI) activate antigen-presenting cells (APC) and alloreactive naïve B cells (via BCR). Both cells present alloantigen epitopes on their surface MHCII to cognate T cells. APCs activate alloantigen-specific naïve CD4 + T cells, which then as helper T cells (Th) provide co-stimulatory (second) signal to activated B cells. These B cells can have two possible fates-either extrafollicular development into a short-lived plasmablast secreting early DSAs, or a GC reaction (de-novo sensitized patients). Aided by follicular helper T cell (Tfh) co-stimulation (via CD40/CD40L) and cytokines (IL-21), GC B cells undergo somatic hypermutation (+/− class-switch recombination) leading to higher BCR affinity, and ultimately differentiate into either memory B cells or plasma cells (sensitized patients). Plasma cells migrate into their niches (e.g., bone marrow) where they receive essential survival signals from several other cell types, including stromal cells and granulocytes. DSAs generated by plasma cells bind to graft endothelium and trigger tissue-damaging complement activation and FcR-mediated cytotoxicity. Upon alloantigen encounter, memory B cells can either refine their BCR affinity trough another GC reaction, or differentiate directly into plasma cells antigen-presenting cells with a co-stimulatory signal (CD28).
Within 5 days, activated alloreactive T and B cells clonally expand, with B cells differentiating into shortlived (extrafollicular) plasmablasts, or returning to the lymphoid follicle to establish a new germinal centre (GC). 8 Although extrafollicular plasmablasts provide the first wave of antibodies, these are less mutated and with lower affinity. 9 The GC is compartmentalized into a dark (DZ) and light (LZ) zone. The DZ contains intensely proliferating activated B cells (centroblasts), which undergo somatic hypermutation (SHM), changing the antigenbinding region of antibodies to improve affinity, a process known as affinity maturation. When SHM is finished, centroblasts become centrocytes and transit to the LZ. 10 The LZ contains follicular helper and follicular regulatory T cells that orchestrate the positive selection of centrocytes via CD40-CD40L interaction and cytokines such as IL-4 and IL-21. 11 Only centrocytes that find cognate T-cell help survive and either commit to another round of proliferation and SHM within the DZ, or exit the GC giving rise to long-lived memory B cells or plasma cells. Re-arrangement of heavy chain loci also enables mature B cells to express antibodies of non-IgM isotypes that differ in their effector functions.
Long-lived plasma cells migrate from their original GCs into specific niches enabling their survival and sustained antibody production ( Figure 1). Bone marrow plasma cell niches consist of CXCL12 + VCAM1 + stromal cells and a variety of hematopoietic cells that secrete survival factors such as APRIL, BAFF, and IL-6. 12 More recently, plasma cell niches have also been observed in both inflamed and non-inflamed non-lymphoid organs (NLOs). 13,14 Memory B cells recirculate in a similar manner to their naïve counterparts, but can become temporarily NLO-resident after encountering cognate antigen. 15,16 In contrast to memory B cells, plasma cells lack CD20, with implications for their depletion. 17,18 A recall alloantibody response in sensitized patients is initiated when alloreactive memory B cells are re-exposed to their cognate antigen and either directly or indirectly (via new GC reaction) differentiate into ASCs. 19 Generally, memory B cells with high-affinity BCRs differentiate directly into ASCs while others enter GCs for further affinity maturation. 20,21 B cells may also participate in alloimmune responses via antibody-independent functions such as antigen presentation, 22,23 production of pro-inflammatory cytokines (including TNFα, IL-6, and GM-CSF) 24-26 and immune regulation via contact-dependent and -independent mechanisms (e.g., IL-10 or TGFβ secretion). [27][28][29] Indeed, regulatory B cells have been implicated in the maintenance of transplant tolerance by several studies. Clatworthy et al reported a higher incidence of acute rejection in kidney transplant patients who received anti-CD20 therapy at induction compared to controls. 29 Numerous groups have described a B-cell signature in peripheral blood (including an enrichment of transitional or naïve B cells) in operationally tolerant (OT) or rejection-free patients. 28,[30][31][32][33][34] The extent to which this finding reflects differences in maintenance immunosuppression has been questioned, 35-37 but B cell genes remain in the consensus gene expression signature of OT, even after adjusting for differing immunosuppression. 38

| PRINCIPLES OF B CELL TARGETING IN NON-SENSITIZED VERSUS SENSITIZED PATIENTS
The therapeutic goal in non-sensitized patients is the prevention of naïve alloreactive B cell activation, as well as subsequent plasmablast formation and GC response ( Figure 1). In sensitized patients, the treatment target also includes alloreactive memory B cells and plasma cells. Notably, on-going exposure to alloantigen results in continued alloreactive B-cell activation with further refinement of existing DSA affinity in GCs and the generation of new DSA by bystander B cell activation and via epitope spreading as the allograft is progressively damaged (chronic ABMR). 39-42 Hence, there is substantial overlap in the therapeutic targets in acute vs chronic ABMR.
Ideally, alloreactive B cells should be targeted selectively in antigen-specific manner, while sparing regulatory B cells and pathogen-protective B cells. Therapeutic expansion of regulatory B cells represents a promising indirect approach to this conundrum. Unfortunately, current treatments generically target B cells regardless of their function or specificity, with inevitable undesirable side effects.

| SUMMARY OF HISTORIC AND MOST RECENT STRATEGIES
The lack of robust evidence for a standard of care in ABMR prompted The Transplantation Society to assemble an international working group to define consensus treatment recommendations, mostly based on expert opinion rather than evidence. 43 This expert group suggested the combination of plasmapheresis (PLEX), IVIG and steroids for early (<30 days post-transplant) and late active AMBR in patients with preformed DSAs, and for chronic ABMR, optimization of maintenance immunotherapy (e.g., target trough tacrolimus level > 5 ng/mL). Adjunctive therapies such as rituximab, complement inhibition (e.g., eculizumab or C1 esterase inhibitors) or protease inhibitors may be considered in patients at high risk of rapid allograft loss. 43 In AMBR with de-novo DSA, provision of sufficient maintenance immunosuppression was recommended along with concomitant treatment of TCMR (e.g., with anti-thymocyte globulin). The evidence for additional therapy in this context was viewed as inadequate. 43 The rationale for combining IVIG with PLEX is based not only on its ability to facilitate the removal of DSAs (by enhancing IgG turnover when occupying neonatal Fc receptors) 44 but also on the modulatory effects of IVIG on immune cells (e.g., via binding surface inhibitory receptors such as FcγRIIb or CD22). 45 Despite varying protocols for these two treatment modalities, several studies have reported benefit for at least short-term outcomes in acute ABMR. 46-49 There is a little convincing evidence of long-term benefit, possibly reflecting the limited effects of PLEX and IVIG on DSA-producing cells. Hence, B cell and plasma-cell depleting agents, such as rituximab or bortezomib, respectively, have been added to PLEX and IVIG.
Rituximab, a chimeric anti-CD20 depleting monoclonal antibody, has been successfully used for autoimmune and hematological conditions. However, in ABMR, a systematic review of rituximab utility could not draw any strong conclusions because of high heterogeneity and low power of available studies. 50 Indeed, a recent multicentre RCT failed to show benefit for rituximab when added to standard of care in active ABMR and it was associated with more opportunistic infections. 51 However, 8/19 patients in the control arm also received rituximab as "rescue therapy", undermining any definitive conclusions. A similar outcome was reported in a Spanish multicentre RCT testing IVIG plus rituximab in chronic ABMR. 52 Both studies were underpowered (N = 38 and N = 25, respectively) and weakened by a limited follow-up (up to 1 year). 51,52 Two ongoing UK-based RTCs TAR:GET1 (NCT03994783) and MOT-AMR (NCT04496037) aiming to recruit 170 patients in total with a 4-year follow-up should enable robust conclusions. As noted previously, nonselective B-cell depletion with rituximab in ABMR may have counterproductive effects on regulatory B cells.
Bortezomib, a first-generation proteasome inhibitor (PI), has been used both in the context of acute ABMR and desensitization, with variable success. In uncontrolled studies, bortezomib combined with standard treatment modalities was associated with higher response in early compared with late acute ABMR (87.5% vs 53.8%; P = .13), suggesting higher sensitivity of newly formed plasmablasts to PI therapy compared to long-lived nicheresident plasma cells. 53 In a prospective iterative desensitization trial, bortezomib reduced HLA antibody levels in 38 out of 44 patients for up to 10 months. 54 Although the reduction of immunodominant HLA antibodies was stronger with higher bortezomib dosing, this was limited by drug toxicity (mainly peripheral neuropathy). 54,55 A similar desensitization trial with carfilzomib, a secondgeneration PI, showed a better safety profile, even at higher doses with a complete absence of neurotoxicity, and produced a more robust depletion of HLA antibodies both as monotherapy and in combination with PLEX. 56 For late ABMR, a recently published RCT (BORTEJECT) comparing bortezomib monotherapy to placebo failed to show any impact on eGFR decline, DSA reduction or histological molecular signatures at 24 month follow-up, in spite of significant toxicity. 57 However, the BORTEJECT study used only two cycles of bortezomib up to 3 months apart, and patient numbers were also relatively limited (only 18 patients completed both cycles of bortezomib). 57

| NOVEL STRATEGIES TO TARGET B CELL ACTIVATION/GC RESPONSE
As outlined previously, naïve alloimmune B cell activation is typically dependent on two signals -engagement of the BCR and its co-receptors (signal 1) and costimulation from helper T cells (signal 2), which is also essential for a successful GC reaction ( Figure 2).
Co-stimulation blockade currently represents one of the most promising interventions in kidney transplantation. 7-year follow-up data from the phase 3 BENEFIT and BENEFIT-EXT RCTs showed better graft survival and lower de-novo DSA formation in patients treated with belatacept (CTLA4-Ig fusion protein blocking CD28-CD80/86 co-stimulation) when compared with cyclosporine. 58,59 Leibler et al showed that belatacept not only interferes with B cell-Tfh cell crosstalk but also has direct B-cell intrinsic (i.e., T cell-independent) effects, such as reduced plasmablast differentiation, BLIMP-1 expression and STAT3 activation. 60 Subsequently, in the BELACOR trial, an open, multicentre, prospective pilot study with a retrospective control group, Leiber et al assessed kidney transplant outcomes and susceptibility to ABMR in patients with preformed DSAs treated with maintenance belatacept, or a CNI. 61 The belatacept group (N = 49) had 0% ABMR incidence at 12 months compared to 5.81% (P = .12) in controls (N = 73) and a significantly higher proportion of belatacept-treated patients had complete disappearance of class II DSAs (81% vs 42%, P = .001). However, patients on belatacept had a more than four times higher frequency of acute TCMR (P = .003) but similar median eGFR at 12 months. 61 A significantly higher rate of acute TCMR in belatacept-treated patients was also reported in the BENE-FIT trial and this likely reflects its limited action on cytotoxic CD8 + T cells that lack CD28 expression. 58 Hence, the combination of belatacept with a CNI, for early posttransplant period, has been suggested. 62 Co-stimulation blockade has also been combined with PIs for desensitization. Kwun et al showed in a nonhuman primate model that although bortezomib monotherapy significantly depleted bone-marrow plasma cells, it failed to reduce DSAs and induced a rapid compensatory GC response in sensitized animals. 63 In contrast, co-stimulatory blockade efficiently suppressed the GC reaction and de-novo DSA production in non-human primates. 64 Hence, the combination of PI with belatacept and anti-CD40 mAb was tested in the same desensitization model and achieved both a significant reduction in pre-transplant DSA and prolonged graft survival, without ABMR. 65,66 To reduce cumulative drug toxicity and the overall immunosuppressive burden, a follow-up study treated animals with a second-generation PI, carfilzomib, in combination with belatacept only. This produced similar graft survival but reduced viral infections. 67 Aside from belatacept, anti-CD40 and anti-CD154 (ie, CD40L) monoclonal antibodies, alternative agents to target the immunological synapse (e.g., OX40 or ICOS) are yet to be tested in the context of solid organ transplantation.
The GC reaction is modulated by a plethora of locally secreted cytokines, including IL-21, IL-6, or B-cell survival factor BAFF. 68,69 In mice, BAFF-deficiency was associated with prolonged heart allograft survival 70 and BAFF blockade in combination with rapamycin induction resulted in long-term survival of islet allografts. 71 We, and others, reported that elevated circulating BAFF was associated with both the development of DSA and an increased risk of ABMR in human kidney recipients. 72,73 Subsequently, we showed in an experimental medicine, phase 2 RCT (N = 28) that belimumab (anti-BAFF mAb) added to the standard immunosuppression in the first 6 months post-kidney transplantation in non-sensitized recipients did not increase infection risk while reducing activated memory B cells and circulating plasmablasts, inhibiting de-novo IgG formation and increasing circulating regulatory B cells. 26 The surrogate end-points in this study also raised the possibility that belimumab may have utility in sensitized patients and provided a solid foundation for a larger RCT in the future. 26 Several drugs have been developed to block B cell activation by disrupting BCR or its co-receptors, the signaling of which is essential for triggering memory B-cell recall responses. Imlifidase (IgG-degrading enzyme of Streptococcus pyogenes, IdeS), a highly specific protease cleaving human IgG at its hinge region, has been primarily trialed for DSAs degradation as part of desensitization protocols (see below). However, Järnum et al showed ex vivo that IdeS also cleaves BCRs on the surface of human memory B cells, temporarily inhibiting their activation and F I G U R E 2 Cellular/molecular targets of immunotherapeutics interfering with B cell activation and germinal centre reaction. Left panel-numbers identify specific cellular/molecular target. Right panel-therapeutic agents for each cellular/molecular target subsequent differentiation into plasma cells. 74 Epratuzumab, a humanized monoclonal antibody against CD22, an inhibitory BCR co-receptor, is able to terminate BCR activating signals, including via the recruitment of phosphatases and down-regulation of CD19. 75,76 Although it does not induce complement-(CDC) or cellulardependent cytotoxicity (ADCC), it leads to a significant reduction in peripheral B cells (in particular, naïve and transitional). 77,78 A recent meta-analysis suggested epratuzumab as a safe treatment option for moderate-tosevere SLE, but its utility in solid organ transplantation is yet to be explored. 79 Obexelimab (XmAb5871), an anti-CD19 humanized antibody that simultaneously engages FcγRIIB, is another promising agent harnessing inhibitory BCR co-receptor that could be of benefit in transplantation. 80 Recent trials in autoimmune conditions such as SLE, rheumatoid arthritis and IgG4 disease suggest potential utility in reducing humoral responses. [81][82][83] The ultimate method to prevent activation of alloreactive naïve or memory B cells is their depletion, which has been traditionally achieved through the administration of polyclonal anti-thymocyte globulin (ATG), monoclonal alemtuzumab (anti-CD52, CAMPATH-1H) or rituximab (anti-CD20). One reason that rituximab has shown no substantial benefit in ABMR and desensitization (aside from under-powered RCTs) may be that B cell depletion is incomplete, particular within tissues. 84,85 Obinutuzumab, a glycoengineered type 2 anti-CD20 humanized monoclonal antibody, was developed to supersede rituximab by enhancing the capacity for ADCC and B cell depletion, including in secondary lymphoid organs. 86,87 NOBILITY, a phase 2 RCT trial, has recently showed efficacy and safety of obinutuzumab in proliferative lupus nephritis, which contrasts with the earlier results of EXPLORER and LUNAR trials that failed to show efficacy for rituximab in lupus. 88-90 THEORY, a phase 1b open-label trial, showed good tolerability of obinutuzumab plus IVIG in highly sensitized patients with end-stage kidney disease and induced profound depletion of both circulating and lymph node B cells, but had limited effects on HLA antibodies. 91 To extend B cell depletion to plasmablasts and potentially a substantial fraction of plasma cells, CD19 targeting therapies have been developed (Figure 3). In a phase 2/3 RCT, a humanized anti-CD19 monoclonal antibody, inebilizumab, significantly reduced autoantibodies and relapse rates in patients with neuromyelitis optica. 92 Tolerability and safety of this drug is currently being evaluated in highly sensitized patients awaiting kidney transplantation (NCT04174677). Blinatumomab, a monoclonal CD19 bi-specific T-cell engager antibody capable of inducing cytolytic synapse between B cells and cytotoxic T cells has been used for acute lymphoblastic leukemia and may also have utility for B cell depletion in ABMR. 93 Finally, F I G U R E 3 Action sites of immunotherapeutics targeting plasma cells and alloantibody effector function. Left panelnumbers identify specific cellular/molecular target. Right panel-therapeutic agents for each cellular/molecular target chimeric antigen receptor (CAR) directed T cell therapy targeting CD19 represents potentially another attractive approach to B cell depletion in ABMR. Although it has revolutionized the treatment of B-cell malignancies, it is often associated with significant side-effects such as cytokine release syndrome and hypogammaglobulinaemia. 94 Ultimately, a CAR could be designed to express a particular HLA antigen enabling selective binding to alloreactive B cells and their subsequent inhibition (by engaging Tregs) or death (via cytotoxic CD8 + T cells), while sparing bystander regulatory and protective B cells. 95 Such a selective approach has been successfully applied to eliminate autoreactive B cells in experimental pemphigus. 96 6 | NOVEL STRATEGIES TO TARGET PLASMA CELLS Current plasma cell therapies induce depletion either directly, by targeting surface molecules or proteasome activity, or indirectly, by disrupting their survival niches ( Figure 3).
In addition to anti-CD19 antibodies, monoclonals directed against CD38, B-cell maturation antigen (BCMA) and CD138 have been also developed to deplete plasma cells. Daratumumab, a fully human anti-CD38 antibody, has shown treatment efficacy in multiple myeloma, [97][98][99] eliminating plasma cells in bone marrow via CDC and ADCC, but it also reduced CD38 + regulatory T and B cells and myeloid-derived suppressor cells. 100,101 This additional immuno-modulatory effect may limit its use in the context of ABMR or autoimmunity where induction of immune tolerance is desirable. Indeed, Kwun et al showed a significant DSA reduction and prolonged renal graft survival in sensitized rhesus macaques treated with daratumumab, but this was followed by a rebound in DSA and TCMR. 102 Nevertheless, several dose-escalation studies are currently underway to test safety and efficacy of daratumumab (or chimeric mouse/human isatuxumab) in sensitized patients awaiting kidney or heart transplantation (NCT04088903, NCT04204980, NCT04294459). Unlike CD38, BCMA is expressed only on late memory B cells and all plasma cells where it provides an essential survival signal after binding APRIL (and to a lesser extent BAFF). 103 Currently, anti-BCMA targeting (including bispecific antibodies, CAR-T cells or antibody-drug conjugates) is driving a new era in the treatment of multiple myeloma with encouraging results from early phase trials and more than 70 other studies recruiting. 104 Indatuximab ravtansine is an anti-CD138 antibody-drug conjugate that has showed promising results in inhibiting myeloma cell growth in relapsed or refractory disease. 105 Targeting CD138 with a similar approach or using a bi-specific T cell engager is yet to be tested in solid organ transplantation.
As outlined already in previous chapters, PIs bortezomib and carfilzomib have been actively explored for desensitization strategies both in monotherapy and in combination with other modalities. Ixazomib is a new second-generation PI, that is, unlike its predecessors, administered orally and is currently being evaluated in IXADES phase 2 trial for its safety and efficacy in desensitization of kidney transplant candidates (NCT03213158). Apart from drug toxicity, PI therapy is also limited by the development of plasma-cell drug resistance through a compensatory overexpression of immunoproteasome, which is a specialized (structurally different) form of proteasome present in immune cells and inflamed tissues. 106 Selective immunoproteasome inhibitors are an attractive alternative to traditional constitutive proteasome inhibitors because of their lower toxicity and additional immunomodulatory effects. 107 Recently, Li et al showed that the selective immunoproteasome inhibitor ONX0914 prevented chronic allograft nephropathy in rats by eliminating plasma cells and reducing DSA. Interestingly, this inhibition did not only activate the unfolded protein response but also supressed plasma-cell survival factors in the bone marrow (e.g., IL-6 or APRIL). 108,109 The plasma cell survival niche is a complex microenvironment providing essential cell-cell interactions (e.g., CXCR4-CXCL12) and secreted cytokines such as BAFF/ APRIL and IL6). Plerixafor is a small-molecule reversible CXCR4 inhibitor that has been used for more than a decade as a bone-marrow stem cell mobilizer. Woodle et al recently reported preliminary data supporting plerixafor efficacy to mobilize plasma cells, leading to their apoptosis and enhanced effectiveness of bortezomib in highly sensitized patients. 110,111 There has also been emerging interest in IL-6/IL-6 receptor blockade in ABMR. IL-6 is a pleiotropic cytokine that fuels various aspects of alloimmune humoral responses. 112 IL-6 induces maturation of Tfh with subsequent differentiation of activated naïve B cells into plasma cells in the GC and also serves as a key survival factor for pre-existing long-lived DSA-producing cells. [113][114][115] Moreover, in return, DSAs trigger IL-6 production in allograft endothelium, ultimately leading to obliterative vasculopathy and perpetuating the alloimmune reaction. 116,117 Tocilizumab, a humanized anti-IL6 receptor-blocking antibody, provided a significant benefit in kidney graft survival at 6-years when given monthly to 65 patients with chronic refractory ABMR compared to a historic cohort of 39 controls receiving rituximab plus IVIG. 118 Chandran et al reported the efficacy of tocilizumab in a prospective RCT to reduce subclinical kidney allograft inflammation while inducing circulating Tregs (N = 11 in both arms). 119 Vo et al achieved a 50% transplant rate with tocilizumab, with no observed ABMR up to 6 months post-transplant in 10 patients unresponsive to standard desensitization protocols. 120 While a larger RCT utilizing tocilizumab is still awaited, clazakizumab (anti-IL-6 IgG1 monoclonal antibody) is currently being tested in a multicentre placebo-controlled phase 3 RTC (IMAGINE) in chronic active ABMR (estimated completion in 2028, NCT03744910). This follows early-phase studies demonstrating safety and efficacy of clazakizumab to reduce DSAs and stabilize kidney allograft function. 121,122 Finally, BAFF blockade was also trialed as monotherapy in sensitized patients but had no clinically meaningful efficacy. 123 Kwun et al reported prevention of DSA formation and prolonged allograft survival in non-human primate model of ABMR treated with atacicept (a recombinant fusion protein containing soluble TACI receptor). 124 However, interest in this treatment strategy has waned, after atacicept failed to show efficacy in several autoimmune conditions. 125 APRIL/BAFF interact with three different receptors on plasma cells, and this requires consideration when attempting their blockade, which may necessitate combination with other anti-plasma cell therapies.

| NOVEL STRATEGIES TO TARGET ALLOANTIBODY EFFECTOR FUNCTION
Alloantibody effector functions are derived from both their Fab (e.g., proliferation stimulus to allograft vasculature) and Fc region (complement activation and Fcreceptor-mediated modulation of immune cells activity, including phagocytosis and cytotoxicity) (Figure 3). 126 Complement activation (predominantly via lectin and classical pathways) is a key player in allograft ischaemiareperfusion injury (IRI) and ABMR. [127][128][129] Recently, Cippa et al proposed a causal link between these two pathological states where IRI-driven tissue damage uncovers cryptic alloantigens and ultimately leads to alloreactive B-cell activation with de-novo DSAs production. 42 Preventing IRI by therapeutic complement inhibition is being actively explored in solid organ transplantation. A phase 1/2 RCT (N = 70) reported a significant benefit of C1 esterase inhibitor (C1-INH) for 3.5-year allograft survival when given intraoperatively and 24 hours post-transplant. 130,131 In contrast, a large multi-centre registration RCT (PROTECT, N = 288) did not meet its primary endpoint to justify perioperative administration of eculizumab (a C5 inhibitor) to prevent delayed graft function, although long-term follow data were not collected (NCT02145182). In a cohort of pediatric kidney transplant recipients, a single pre-transplant dose of eculizumab led to significantly better graft function up to 3 years later, but was associated with an unacceptably high rate of early graft losses during a flu-like infection (all in non-vaccinated children). 132 Two small early-phase studies showed either histological or clinical improvement of kidney graft function 6 months after ABMR treatment with C1-INH either in addition to standard of care or in combination with IVIG, respectively. 133,134 Two larger multicentre phase 3 RCTs testing C1-INH in acute ABMR are underway (NCT03221842, NCT02547220). Lefaucheur et al showed that ABMR driven by complement-binding DSAs had a specific histological and transcriptional signature and its incidence at 3 months could be reduced with prophylactic eculizumab, unlike AMBR in patients with noncomplement binding DSAs. 135 In parallel, Viglietti et al reported that the presence of complement-binding DSAs 3 months after standard ABMR treatment initiation (IVIG, PLEX and rituximab) was an independent predictor of poor allograft outcome, further justifying the role of complement blockade in AMBR treatment strategies. 136 The efficacy and safety of eculizumab to prevent ABMR in sensitized kidney recipients was recently shown in an openlabel single-arm phase 2 study (N = 80). 137 A case series (N = 15) of early active ABMR treated with eculizumab plus PLEX as primary (rather than salvage) therapy also showed an impressive increase of median eGFR with no graft loss at minimum of 12-months follow-up. 138 The development of complement-targeting therapies is a fast evolving field and there are many new agents, which will require rigorous testing in transplantation (e.g., ravalizumab or anti-C1s antibody BIVV009). 139,140 Although the current evidence suggests efficacy for complement inhibitors in the treatment of ABMR, combination with DSA-reducing strategies to mitigate other non-complement-related DSA effector functions will likely be required. Imlifidase (IdeS) cleaves IgG and eliminated all complement-binding DSAs (IgG only) within 1 hour of administration, with residual Fab fragments incapable of inducing CDC or ADCC. 141,142 Two openlabel phase 1/2 trials (in US and Sweden) reported successful HLA-incompatible kidney transplantation in 24 of 25 patients 4 to 6 hours after IdeS administration. 143 In contrast to the Swedish group, the US patient cohort received IVIG plus rituximab 7 to 14 days after transplant, which prevented DSA rebound. Of note, imlifidase also cleaves IVIG or therapeutic monoclonal antibodies, the administration of which should therefore be delayed for at least for 5 days. Three cases of ABMR were reported in the Swedish cohort (N = 11) within a few weeks, and two in the US at 2 and 5 months. 143 Allograft survival was 92% at 2 years, with only one new ABMR case despite several patients having high DSA. 144 Imlifidase (IdeS) is now approved by the European Medicine Agency (with accelerated FDA approval in process) for DSAs degradation in highly sensitized kidney transplant recipients. An open-label phase 2 RCT evaluating imlifidase for ABMR treatment in kidney transplantation is currently recruiting (NCT03897205).

| CONCLUSION
In recent years, there has been a considerable expansion of novel therapeutics (often repurposed from oncology or autoimmunity) that may target humoral alloimmune responses, now regarded as the leading cause of longterm allograft loss. Currently, rigorous evaluation of these agents in phase 3 RCTs is lacking, precluding regulatory approval. Innovative trial design, utilizing novel prognostic markers and surrogate end-points will be required to enable trials with a feasible size and duration to test these agents, providing hope that humoral alloimmunity will ultimately be better controlled.

CONFLICT OF INTEREST
The authors have declared no conflicting interests.

DATA AVAILABILITY STATEMENT
Data sharing is not applicable to this article as no new data were created or analysed in this study.