Is there a role for natural antibodies in rejection following transplantation?

Abstract

Antibody-mediated rejection continues to hinder long-term survival of solid organ allografts. Natural antibodies (Nabs) with polyreactive and autoreactive properties have recently emerged as potential contributors to antibody-mediated graft rejection. This review discusses Nabs, their functions in health and disease, their significance in rejection following kidney, heart and lung transplantation and their implication in serum reactivity to key antigens associated with rejection. Finally, potential Nabs effector mechanisms in the context of transplantation are explored.

INTRODUCTION

Solid organ transplantation is often the only recourse for end-stage organ failure. However, the limited survival of transplanted organs, especially in the long-term, still poses a challenge. Recent data obtained from the nationwide Organ Procurement and Transplantation Network indicate that the one-year survival rate is approximately 90% following kidney transplantation although this number drops to roughly 76% after 5 years. For other organs, such as for lung transplants, the outcome can be far worse with survival rates of approximately 50% after 5 years posttransplant.

One of the main barriers to the longevity of transplanted organs is the development of immune responses directed against the allograft. Both the cellular and humoral arms of the immune system are implicated. On the whole, cellular responses are now well controlled by improved immunosuppressive drugs. The humoral arm, however, is still poorly responsive to immunosuppression and continues to impact long-term outcomes through a complication known as antibody-mediated rejection (AMR).

Antibodies targeting the allogeneic donor graft are thought to contribute to greater than 60% of late failure in solid organs transplants.¹ These are predominantly donor-specific human leukocyte antigen (HLA) antibodies (DSA) that can either predate or develop after transplantation (de novo DSA). Antibody-mediated rejection can also occur in the absence of donor-specific antibodies and despite negative B- and T-cell crossmatches,^2–5 indicating that other types of antibodies are implicated in humoral rejection. Earlier studies in kidney⁶ and cardiac rejection patients⁷ suggested the presence of antibodies that were reactive to antigenic structures distinct from HLA on the graft vascular endothelium. More recently, several of these structures have been identified. Non-HLA antibodies typically target autoantigens expressed on graft donor cells including major histocompatibility complex class I chain-related molecule A (MICA),⁸ angiotensin II type 1 receptor (AT1R),⁹ endothelin-1 type A receptors,¹⁰ vimentin¹¹ and cardiac myosin.¹² These non-HLA antibodies have been attributed pathogenic potential in the context of kidney, heart and lung transplantation.

Often included in the group of non-HLA antibodies are “natural” antibodies (Nabs), however, their unique characteristics largely differ from that of specific autoantibodies. Most notable Nabs include antibodies to ABO blood group antigens or xenoantigens such as α-(1,3)-galactose (α-Gal) and N-glycolylneuraminic acid (Neu5Gc) and are responsible for hyperacute rejection in the context of ABO incompatibility or xenotransplantation, respectively. More recently, studies from our group revealed the development of Nabs following ABO-compatible allotransplantation. However, their impact on the outcomes of transplantation is still unclear. This review will summarize the current evidence supporting a role for Nabs in mechanisms of rejection and identify key areas of uncertainty in need of further investigation.

NATURAL ANTIBODIES AND POLYREACTIVITY

Nabs were recognized over 50 years ago as a group of immunoglobulins implicated in innate defense and homeostasis in humans and other mammalian species.^13,14 They are of IgM, IgG and IgA isotypes and their repertoire profiles within an individual are relatively consistent over time.^15,16 Most Nabs are polyreactive,^14,17 meaning that they are capable of binding to multiple and apparently unrelated antigenic structures. Their inherent polyreactive properties enable them to carry out important functions in health and disease.^18,19

Antigenic determinants of Nabs shape their roles in health and disease

Polyreactive Nabs have been reported to maintain healthy host tissue by binding to structures on apoptotic and senescent cells and facilitating their removal, thereby preventing unwanted inflammation. They also bind to various bacterial antigens including phosphorylcholine and lipopolysaccharide (LPS)^19–21 and viral proteins such as hemagglutinin,²² forming a first-line defense mechanism against pathogens. Nabs are also able to control the damaging effects of oxidative stress, a consequence of normal cellular processes, by recognizing oxidized epitopes on low-density lipoprotein.²³ As such, they are protective against atherosclerosis, a chronic inflammatory disease of the vascular wall.

In contrast, Nabs are also implicated in autoimmune and inflammatory reactions. In lupus, polyreactive DNA-binding antibodies may also react to the N-methyl-D-aspartate receptor.²⁴ Reactivity to DNA deposited in glomeruli^25,26 and N-methyl-D-aspartate receptor expressed in neurons²⁴ leads to target cell injury and mediates nephritis and central nervous system disease in lupus, respectively.

Detection of Nabs

Polyreactive Nabs are difficult to define based on their reactivity profile. Perhaps the most direct way to appreciate the breadth of such reactivity is by studying monoclonal polyreactive antibodies. We and others have generated a number of clones from healthy donor blood that bind to multiple targets including LPS, double-stranded and single-stranded DNA, insulin, undefined intracellular structures in Hep-2 cells, apoptotic cells, the oxidation-specific epitope (OSE) malondialdehyde (MDA) as well as numerous proteins in a HEK293 epithelial cell lysate. Monoclonal Nabs generated from solid organ recipient blood or graft infiltrates can also exhibit similar polyreactivity.^27,28 One important observation, however, is that no structure appears to be recognized by all monoclonal Nabs. In itself, this broad binding profile and absence of a universally shared target antigen is an obstacle for designing tests to identify all Nabs in clinical specimens.

Conventional methods of identifying polyreactive Nabs use immunoassays to assess reactivity to common autoantigens such as DNA, LPS or insulin.^29,30 Monoclonal antibodies binding to multiple antigens are considered polyreactive. Another approach takes advantage of the known reactivity of Nabs to OSEs generated on self-molecules by highly reactive lipid degradation products.³¹ Studies have used OSE-modified proteins, such as MDA-modified bovine serum albumin, in immunoassays in heart and kidney transplant to detect Nabs.^28,32,33 In a similar vein, researchers have measured binding to apoptotic cells as a means to detect Nabs in the transplantation setting.^27,34 Another study used a synthetic molecule, dinitrophenol, in their detection assays reasoning that as this molecule is not present in the environment serum antibodies reacting to it must be polyreactive.³⁵

Generation of Nabs

Most of the information on the generation of Nabs comes from mouse studies where B1 B cells from the peritoneal and pleural cavities are the main source. These innate-like cells are present at birth, are self-replenishing, show restricted immunoglobulin heavy chain variable usage and mainly produce germline IgM³⁶ although mutated Nabs have also been observed.³⁷ Subsequent studies indicated that splenic marginal zone B cells and bone marrow precursors also contribute to Nabs generation.^38,39

In humans, the presence of IgM,^40,41 IgG^16,42,43 and IgA⁴⁴ Nabs is established.^31,45 Both germline and somatically mutated sequences coding for polyreactive Nabs have been described.^46,47 The exact source of Nabs has not yet been unequivocally identified, however, B cells that produce polyreactive antibodies and express polyreactive receptors are highly abundant in circulation^30,48,49 and in the gut mucosa.^50,51

Nabs levels can fluctuate depending on age, disease status, and trauma.^16,52 Nabs-producing B cells are likely to be stimulated to secrete antibody through B-cell receptor (BCR)-dependent and -independent mechanisms.^35,39,43,53 While IgM Nabs are abundant at the steady-state, we have shown IgG Nabs to be elevated in chronic kidney graft rejection and following ventricular assist device implantation in heart transplant recipients.^27,32 A possible scenario is that innate-like B cells producing IgM Nabs in normal conditions would undergo class switching to produce IgG Nabs following antigen encounter or in inflammatory situations.

EFFECTOR AND IMMUNOREGULATORY FUNCTIONS OF NABS IN HEALTH AND DISEASE

The protective role of Nabs against pathogens, inflammation and autoimmunity is well-documented. Nabs recognize pathogen-associated molecular patterns to exert antibacterial^19,54 and antiviral responses.²² IgM Nabs protect against unwanted inflammation, such as in atherosclerosis and autoimmunity, by binding to and preventing the accumulation and subsequent immune responses to apoptotic cells⁵⁵ and harmful cell debris.⁵⁶ Nabs may also block pathogenic autoantibodies⁵⁷ and modulate antibody responses. Conversely, the same Nabs effector mechanisms may also contribute to inflammatory responses in pathogenic autoimmune conditions. Below are a few key Nabs functions in health and disease that could also point to their role after solid organ transplantation (Figure 1).

Figure 1.

Effector and immunoregulatory functions of Nabs in health, disease and transplantation

Complement-dependent mechanisms

Activation of the complement component C1q leads to downstream deposition of C3b on target cells. C3b serves as a signal that promotes the efficient ingestion of opsonized cells by phagocytes. Complement activation can also result in the formation of the membrane attack complex and subsequent lysis of target cells. The binding of Nabs to apoptotic cells^55,58,59 or pathogens^22,60 recruits C1q and triggers complement activation, leading to their clearance. Mice that are deficient in Nabs exhibit defective apoptotic cell clearance by macrophages³¹ and dendritic cells⁶¹ and are more susceptible to infection.²² These complement-dependent functions can also be pathogenic in conditions conducive to acute tissue damage, such as ischemia reperfusion or spinal cord injury, where the exposure of injury-associated neoantigens are recognized by Nabs to mediate complement-driven damage.^62–64

Fc receptor-mediated mechanisms

Nabs are also able to mediate their effects via Fc receptor interactions. Nabs recognizing bacteria opsonized by serum lectins were shown to co-localize with monocyte Fc receptors.⁵⁴ In a Brucella abortis infection model, passive transfer of Nabs to mice lacking B cells and antibodies contributed to the eliminate the bacteria.⁶⁵ This protective effect of Nabs was mediated by Fc-receptor signaling as mice deficient in Fcγ receptors demonstrated enhanced bacterial persistence.⁶⁵

Induction of apoptosis and virus neutralization

In cancer, Nabs recognizing modified carbohydrate structures on malignant cells have anti-tumor reactivity and have been implicated in tumor surveillance. These aberrant oligosaccharides are found on tumor cell glycoproteins such as mucins,⁶⁶ decay acceleration factor⁶⁷ as well as glycolipids (reviewed in⁶⁸). Several such Nabs isolated from patients with different cancers were capable of inducing apoptotic^69,70 or necrotic signaling in tumor cells.⁷¹ A direct action of Nabs in viral protection is also evident as Nabs are capable of binding and directly neutralizing viruses such as influenza^72,73 and HIV in certain circumstances.^74,75

Nabs modulate antibody responses

Nabs also exhibit modulatory properties and influence the development of B-cell immunity. For instance, Nabs can enhance IgG responses against the T cell-dependent antigen, keyhole limpet hemocyanin (KLH). Mice deficient in secreted IgM were shown to produce less KLH-specific IgG and these IgG were of lower affinity.⁷⁶ Similarly, Nabs have been shown to play a part in mouse protective IgG responses against viral infection.⁷⁷ In these latter studies, the absence of IgM Nabs abrogated antiviral IgG. In contrast, Nabs can also regulate B-cell development and the generation of autoimmunity. Mice that lacked soluble IgM Nabs had elevated IgG autoantibodies to double stranded (ds) DNA, single stranded DNA and nuclear antigens⁷⁸ and their B2 cells displayed altered development and BCR signaling.^79,80 Additionally, the cells that produce Nabs themselves exert immunomodulatory functions. In recent years, a population of B cells with regulatory properties termed B10 B cells or Breg has been identified. In mice, these cells appear to share similarities with the innate-like Nabs-producing B1 B cell population. These cells produce the anti-inflammatory cytokine IL-10^81,82 and express polyreactive BCR.⁸³ Following appropriate stimulation, they can produce polyreactive antibodies akin to Nabs.⁸³

ASSOCIATION BETWEEN NABS AND TRANSPLANT REJECTION

Antibodies to blood group antigens and xenoantigens have long been known to promote harmful rejection and precipitate graft loss of ABO-incompatible as well as certain cross-species transplants. These humoral barriers to transplantation are due to natural antibodies targeting carbohydrate structures on the transplanted organs. Early studies demonstrated that xenoreactive antibodies are by nature polyreactive.⁸⁴ The same appears to be true for blood group-specific antibodies, over 80% of which display polyreactivity.⁸⁵ Conversely, later studies showed that polyreactive Nabs can also be xenoreactive,⁸⁶ further supporting the idea that xenoantibodies, ABO antibodies and Nabs belong to the same category of antibodies.

Beyond xenotransplantation and ABO incompatibility, polyreactive Nabs were also investigated in the context of human allogeneic transplantation. We highlight here the most recent studies implicating Nabs in rejection of kidney, heart and lung allografts.

Kidney

Among the earliest reports hinting at a possible contribution of polyreactive antibodies to rejection are studies using protein microarrays to assess the reactivity profile of sera from patients undergoing chronic AMR of kidney allografts. These studies revealed a marked increase in serum reactivity to numerous self-antigens in the rejectors compared to nonrejectors.⁸⁷ The breadth of the reactivity profiles was unexpected and puzzling as it suggested the development of a significant autoimmune response to self-antigens in these patients. B-cell clones were subsequently generated from the blood of patients experiencing chronic rejection. Some of the clones secreted polyreactive monoclonal antibodies binding to dsDNA, LPS, insulin⁸⁸ as well as apoptotic cells.²⁷ These polyreactive monoclonal antibodies also reacted to multiple targets on protein microarrays (manuscript in preparation), suggesting that the broad serum reactivity profiles observed in patients with chronic AMR⁸⁷ could be due to polyreactive Nabs rather than a multitude of monospecific antibodies. A later study by our group correlated higher levels of pretransplant Nabs with worse kidney graft survival in a single-center cohort of 300 patients.³⁴ These polyreactive antibodies were identified as almost exclusively IgG1 and IgG3 and demonstrated the ability to activate complement, resulting in C4d deposition in vitro. These latter findings pointed to their potential in vivo capabilities. A subsequent blinded study investigated Nabs reacting to the OSE MDA in both pretransplant and posttransplant serum samples obtained from a cohort of over 600 well-characterized kidney transplant patients. Posttransplant sera were either protocol samples obtained at one year or for-cause samples collected within the first year. A significant association was found between the development of Nabs in the first-year posttransplant period and graft loss.³³ This association was independent of DSA as revealed by multivariable analysis. Strikingly, Nabs generation observed in the year following transplant was also associated with higher grades of Banff-defined histological lesions specific to AMR, including C4d deposition, microvascular inflammation, transplant glomerulopathy, interstitial inflammation and tubulitis and arteriosclerosis.³³ Lastly, the overall outcomes were worse in patients with both Nabs and DSA, suggesting an additive detrimental effect of these two types of antibodies on kidney graft survival.

Heart

Cardiac allograft vasculopathy (CAV), characterized by intimal thickening and lumen narrowing of the main coronary arteries, is a major cause of graft loss following heart transplantation. CAV has been associated with intra-graft immune infiltrates^28,89–91 containing B- and T-cell clusters together with macrophages and antibody-secreting plasma cells. The plasma cells in these clusters actively secreted IgG and, more rarely, IgM.^28,91 Studies by Huibers et al. reported that some of these cells produce DSA in the context of human CAV,⁹² suggesting a role in local alloresponses. Chatterjee et al. further interrogated the reactivity profile of graft-infiltrating B cells by generating over 100 EBV-immortalized B-cell clones from three explanted heart grafts with CAV.²⁸ No HLA-reactive clones were found in these studies. However, in all cases, approximately half of the clones were polyreactive, i.e. reactive to apoptotic cells, MDA, insulin, dsDNA, LPS and cardiolipin.²⁸ These results suggest an intriguing role for innate B cells and locally secreted polyreactive antibodies in CAV pathogenesis. While their exact functions are currently unknown, it is possible that innate B cells and Nabs promote innate immune responses involving the activation of M2 macrophages. As B cell clusters are also detected during chronic rejection in other organs, it is possible that polyreactive clones and Nabs secreted in situ may also be important in these situations.

The development of Nabs recognizing apoptotic cells and OSE has also been reported following ventricular assist device implant in patients bridged to transplant.³² In this study, elevated Nabs levels at time of transplant were associated with primary graft dysfunction, a serious complication of unclear etiology that represents the leading cause of mortality in the first month following transplant. These findings suggest that very early complications following heart transplant may also be connected to Nabs or to the inflammatory processes that lead to the production of these antibodies.

Lung

In lung transplants, both alloimmunity and autoimmunity, particularly to k-α-tubulin and collagen V, have been implicated in rejection and chronic graft dysfunction.^93–95 The involvement of Nabs however, is not so clear. One report looked at pretransplant polyreactive antibodies targeting different types of apoptotic cells.⁹⁶ The authors found that polyreactive antibodies were elevated in patients with end-stage lung disease and remained high during posttransplant follow-up. However, no correlation to chronic or acute rejection was observed in this study. The discrepant findings between lung and kidney transplant rejection may reflect differences in rejection mechanisms. Alternatively, since most end-stage lung disease patients already had high Nabs levels pretransplant it may be difficult to find an association between these levels and transplant outcome.

POLYREACTIVE ANTIBODIES AND CROSS-REACTIVITY TO IMPORTANT TRANSPLANT ANTIGENS

Studies described in the previous section report situations where Nabs and Nabs-producing innate B cells were directly associated with rejection episodes and graft loss. However, their implication in posttransplant immune reactions may be far broader than what our current knowledge indicates. Below are two instances suggesting that polyreactive Nabs can be responsible for certain serum reactivity wrongly attributed to antigen-specific antibodies. If confirmed, these examples would support the idea that Nabs are more common than initially appreciated but are not always detected as such.

HLA

Advances in laboratory testing methods now allow for the highly sensitive assessment of serum reactivity to specific HLA antigens. Such high sensitivity gives rise to the detection of nonspecific⁹⁷ or cross-reactive antibodies to nondonor specific antigens, presumably due to shared epitopes between different HLA.⁹⁸ A recent study by Gao et al, (2016) puts forward the reasoning that polyreactive antibodies could also contribute to serum reactivity to HLA. The authors examined monoclonal Nabs secreted by B-cell clones derived from transplant recipients for binding to self-antigens and HLA.⁹⁹ Some polyreactive monoclonal Nabs that bound to multiple self-antigens or structures including LPS, dsDNA, insulin, apoptotic cells and HEK293 cell lysate, also recognized multiple HLA class I antigens coated onto Luminex beads routinely used in immunogenetics laboratories. Moreover, reactivity to HLA could be reduced in some serum samples, especially those with high PRA, by adsorption with apoptotic cells, indicating cross-reactivity between HLA on the beads and apoptotic cell determinants.⁹⁹ These findings strongly suggested that polyreactive Nabs can account for part of serum reactivity to HLA initially attributed to specific antibodies.

AT1R

Preexisting and de novo antibodies to AT1R on endothelial cells are implicated in destructive non-HLA responses against kidney allografts.⁹ Several groups have correlated these antibodies with increased risk of rejection after kidney,^100–102 heart¹⁰³ and lung transplantation.¹⁰⁴ Angiotensin II type 1 receptor antibodies directly affect endothelial cells and smooth muscle cells via extracellular signal-regulated kinase 1/2 signaling, leading to increased binding of nuclear factor κB and proinflammatory chemokine expression.⁹ These antibodies also appear to act in synergy with DSA.¹⁰³ Ventricular assist device support prior to heart transplant significantly increased AT1R antibody levels.¹⁰⁵ Intriguingly, a recent study using cardiac transplant recipient and healthy donor serum revealed the extent of cross-reactivity between anti-AT1R antibodies, xenoantibodies to Neu5Gc or heterophile antibodies one could also classify as Nabs. These observations raise the possibility that common methods to test for AT1R antibodies may also detect cross-reactive Nabs.¹⁰⁶

POSSIBLE NABS EFFECTOR MECHANISMS AND CONCLUDING REMARKS

Collectively, the studies presented above provided converging evidence supporting a role for Nabs in the outcome of solid organ transplants. These studies, however, have not demonstrated a causal link between Nabs and rejection. Several possible effector mechanisms could be envisaged whereby Nabs could influence immune reactions following transplantation and impact on graft survival. Below, we put forward the two most likely mechanisms (Figure 1).

Complement activation

Complement activation is a potential mechanism of Nabs-mediated graft damage. As shown previously, monoclonal Nabs produced by B-cell clones derived from kidney transplant recipients’ specimens have the capacity to activate the complement cascade in vitro resulting in C3d and C4d deposition on target cells.²⁷ Moreover, studies in heart and kidney transplants showed that serum IgG Nabs are predominantly IgG1 and IgG3, the two main complement-activating IgG subclasses.^33,34 These polyclonal Nabs could also activate complement in vitro.³⁴ Another observation connecting Nabs and complement comes from a recent study by See et al. Following kidney transplantation, patients who developed Nabs in the first year had increased C4d deposition in biopsies compared to those who did not develop Nabs, even in the absence of DSA.³³ These studies uncover Nabs as plausible contributors to complement activation and graft injury.

Sensing ligands from damaged allografts

Danger-associated molecular patterns (DAMPs) are markers of “sterile” tissue injury or cell stress that are recognized by pattern recognition receptors (PRR), the best known being Toll-like receptors on innate immune cells. The ability to sense these danger signals are critical for responses to damaged self.¹⁰⁷ The relevance of DAMPs in solid organ transplantation has emerged in recent years, with increasing evidence that immune responses mediated by DAMPS can impact graft inflammation, fibrosis and alloimmunity.¹⁰⁸ Damage to the allograft, from ischemia and reperfusion injury to alloimmune responses, results in the release of DAMPs.¹⁰⁹ Both animal and clinical studies showed enhanced levels of DAMPs-related elements such as nucleic acids,^110,111 high mobility group box 1 protein¹¹² and extracellular components^113,114 in kidney, heart and lung transplants. These DAMPS trigger signaling downstream of PRRs and facilitate allograft injury by promoting alloreactive adaptive responses,^109,115 fibrosis and up-regulation of major histocompatibility complex expression.^116,117

Since DAMPs include a broad range of molecules that are known Nabs targets, including DNA and modified biological molecules such as oxidation epitopes,¹¹⁸ Nabs may be considered as soluble PRRs. It is plausible that Nabs recognize and bind DAMPs released from stressed graft endothelium and amplify their potency by opsonizing them with complement molecules or addressing them to Fc-expressing innate immune cells. Furthermore, graft DAMPs may also activate infiltrating innate B cells through their polyreactive BCR leading to Nabs secretion in situ. Alternatively, the presence of IgG Nabs in graft rejection may signify a protective response, similar to that seen for IgM Nabs in atherosclerosis.²³

In conclusion, while the existence of Nabs has been acknowledged for decades, their role in potentiating tissue damage in transplantation is only beginning to emerge. Recent evidence supports a significant, nonredundant role of Nabs in solid organ transplant rejection. Yet, uncertainty remains over their causal link to graft outcomes and mechanisms of action. Further investigations are now necessary to determine their exact function and confirm their importance in transplant rejection.

Abbreviations

AMR

antibody-mediated rejection

AT1R

angiotensin II type 1 receptor

BCR

B-cell receptor

CAV

cardiac allograft vasculopathy

DAMP

danger-associated molecular pattern

DNA

deoxyribonucleic acid

DSA

donor-specific antibodies

HLA

human leukocyte antigen

LPS

lipopolysaccharide

MDA

malondialdehyde

MICA

major histocompatibility complex class I chain-related molecule A

Nabs

natural antibodies

OSE

oxidation-specific epitopes

PRR

pattern recognition receptors