CD47: a new player in phagocytosis and xenograft rejection

Abstract

Organ transplantation is limited by the availability of human donor organs. The transplantation of organs and tissues from other species (xenotransplantation) would supply an unlimited number of organs and offer many other advantages for which the pig has been identified as the most suitable source. However, the robust immune responses to xenografts remain a major obstacle to clinical application of xenotransplantation. The more vigorous xenograft rejection relative to allograft rejection is largely accounted for by the extensive genetic disparities between the donor and recipient. Xenografts activate host immunity not only by expressing immunogenic xenoantigens that provide the targets for immune recognition and rejection, but also by lacking ligands for the host immune inhibitory receptors. This review is focused on recent findings regarding the role of CD47, a ligand of an immune inhibitory receptor, signal regulatory protein alpha (SIRPα), in phagocytosis and xenograft rejection.

Introduction

A persistent shortage of human organs and inexhaustible waiting lists continue to result in many people dying while awaiting transplantation which has made it necessary to search for new alternatives.^{1, 2} Cross-species transplantation (xenotransplantation) has an enormous potential to solve the significant need for organs, tissues and cells for clinical transplantation. Although transplants across discordant species barriers are subjected to vigorous immunological rejection because of the greater phylogenetic differences, the increasing availability of genetically engineered pigs is enabling progress to be made in pig-to-non-human primate experimental models.^{3, 4, 5} Hyperacute rejection caused by the presence of natural antibodies against Gal-α1-3-Gal (α1,3Gal) can now be effectively prevented by using α1,3Gal transferase knockout (KO) pigs.^{3, 4, 5, 6, 7} However, non-α1,3Gal-specific anti-pig antibodies produced either prior to (natural antibodies) or after transplantation present an additional barrier to acceptance of transplanted organs from α1,3Gal transferase KO pigs by primates.^{8, 9} Moreover, coagulation dysfunction between the pig and primate is proving to be another major problem, and this can result in life-threatening consumptive coagulopathy.¹⁰ This complication is unlikely to be overcome until pigs expressing a human ‘antithrombotic' or ‘anticoagulant' gene, such as thrombomodulin, tissue factor pathway inhibitor or CD39, become available.¹⁰

The cellular innate immune responses, including those mediated by natural killer cells^{11, 12, 13} and macrophages, have also been shown to contribute to the rejection of porcine xenografts. The interspecies incompatibility of CD47, a ligand for signal regulatory protein alpha (SIRPα), has been implicated in xenograft rejection. This review summarizes recent data regarding the role of CD47 in phagocytosis and xenograft rejection.

CD47 and SIRPα: an important intercellular communication system

The cells in the body have the need to communicate with each other in order to regulate all their functions and events that go from the development of a cell to death. There are several different communication systems that the body uses in its different organs and systems. The immune system for instance has made use of ‘The CD47–SIRPα communication system' among others, for a number of functions it has, such as phagocytosis, apoptosis and antigen recognition etc.^{14, 15}

Integrin-associated protein CD47

CD47 is a member of the immunoglobulin (Ig) superfamily and it is a cell surface glycoprotein ubiquitously expressed. It has a highly glycosylated transmembrane protein with a V-type Ig-like extracellular domain in its N-terminus, a hydrophobic five putative membrane-spanning segments and a short cytoplasmic tail.¹⁶ CD47 was originally identified for its interaction with the αvβ3, αIIbβ3 and α2β1 integrins, regulating integrin function and cellular responses to the arginine–glycine–aspartic acid-containing extracellular matrix proteins, and thus named integrin-associated protein.¹⁷ The Ig domain of CD47 is required for both functional and physical interaction not only with its associated integrins, but also with its ligands thrombospondin and SIRPα. CD47 also has the ability to transduce signals into the cell after interaction with thrombospondin or integrins; hence engagement can potentially give two-way signaling leading to biological activities such as cell spreading, migration or even apoptosis.^{17, 18, 19}

SIRPα

SIRPα (also known as SHPS-1, BIT, CD172a or p84) belongs to a family of membrane proteins involved in the regulation of leukocyte function and is a ligand of CD47¹⁸ whose interaction has been shown to mediate cell–cell adhesion.²⁰ SIRPα is expressed on myeloid cells, including macrophages, granulocytes, myeloid dendritic cells (DCs), mast cells and their precursors, including hematopoietic stem cells.^{21, 22}

SIRPα has three Ig-like domains in its extracellular region and two immunoreceptor tyrosine inhibitory motifs in its cytoplasmic region, which upon ligand binding become phosphorylated and mediate recruitment and activation of the dual Src homology region 2 domain-containing phosphatase (SHP)-2 and SHP-1 to the membrane, and thereby negatively modulates intracellular signaling initiated by various cell surface receptors that promote tyrosine kinase activity or from growth-factor receptors.²³ SHP-1 and -2 can, in turn, also dephosphorylate specific protein substrates and thereby regulate cellular functions, generally in a negative fashion.

CD47 and SIRPα regulating phagocytosis

Macrophages are essential to both innate and adaptive immune responses. They initiate the innate immune response by recognizing pathogens via a variety of receptors with specificity for pattern-recognition molecules.^{24, 25} Macrophages also express inhibitory receptors that control macrophage activity, including phagocytosis as is the case of SIRPα,^{23, 26} which recognizes CD47 on the surface of the cell as a ‘marker of self' and upon binding phagocytosis is inhibited.¹⁴ Being this one of the reasons why macrophages mainly engulf foreign cells such as transplanted cells, but not self.

Evidence of this phenomenon came from the observations that CD47-deficient circulating red blood cells were cleared rapidly by splenic macrophages because this inhibitory signal was lacking, and binding of the blood cells to the macrophages was sufficient to trigger a phagocytic signal.¹⁴ This elimination was abrogated when macrophages were removed by treatment with macrophage-toxic clodronate liposomes. In addition, the use of blocking antibodies to SIRPα resulted in unopsonized CD47⁺ erythrocytes phagocytosed to the same extent as CD47^−/− erythrocytes. The use of mutant mice for SIRPα supported these findings.²⁷

This effect was not only observed in erythrocytes but also in platelets. Studies have shown that mice expressing a SIRPα mutant that lacks most of the cytoplasmic domain manifest thrombocytopenia, which apparently results from an increased rate of clearance of circulating platelets. Furthermore, peritoneal macrophages from these mice exhibit an enhanced phagocytic response.²⁸ Because of the rapid clearance of donor lymphoid cells, CD47^−/− mouse T cells fails to induce graft-versus-host disease in a fully major histocompatibility complex-mismatched allogeneic hematopoietic cell transplantation model, in which injection of a similar number of CD47^+/+ T cells induces lethal graft-versus-host disease.¹⁵ It was found that DCs are more efficient than macrophages in engulfing CD47^−/− T cells.¹⁵ These observations provide strong evidence that CD47 expression is a critical indicator for determining whether lymphohematopoietic cells will survive or be cleared.

Phagocytes have also the ability to discriminate between viable/healthy and apoptotic/foreign/abnormal cells. The human body has to somehow find the way to get rid of a different array of cells, including those produced in excess, altered, damaged or aged cells in order to keep homeostasis. One of such mechanisms is apoptosis where cellular changes mark these cells for further removal by phagocytes after the cells have died, in an efficient and immunologically silent manner. This process has to be accurate enough to spare neighboring healthy cells that are around. The cell surface protein CD47 on healthy cells and its engagement of a phagocyte receptor, SIRPα, appears to constitute a key ‘don't eat me signal', being this one of the different systems used by the immune cells namely phagocytes to discriminate healthy from the unwanted/aged/dying cells. Although the relationship between apoptosis and CD47 is incompletely defined, several studies suggest that the progression of apoptosis is linked to the reduction of cellular CD47 levels independent of the apoptotic stimulus.²⁹ Although the mechanism underlying the reduction of CD47 on the surface of apoptotic cells is unknown, it is believed that protease cleavage³⁰ could be a possibility; as a result, the apoptotic cells will no longer provide an inhibitory signal for phagocytosis via SIRPα which will facilitate their uptake by macrophages.

However, the role of SIRPα may be more complex than previously anticipated and the molecule may not only provide negative signals. For instance, it has been shown that engagement of SIRPα by CD47 in macrophages can promote the production of nitric oxide via the SIRPα-associated Janus kinase.³¹ In addition to the regulatory functions in macrophages, interactions between CD47 on T cells and SIRPα on DCs have been shown to downregulate, in a bidirectional fashion, DC and T-cell activation.³²

The level of CD47 regulates cell clearance

CD47 increases transiently in hematopoietic cells (returning to baseline levels within days) during mobilization or after a strong inflammatory stimulus.³³ It was suggested that this transient upregulation is necessary to prevent the clearance of normal, healthy hematopoietic cells during their recirculation from the bone marrow to blood or extramedullary sites during a stress response, since only low CD47 levels are thought to be sufficient to prevent phagocytosis.^{33, 34} during a steady state. However, during an infection for example in where LPS come into play, macrophages get activated probably due to downregulation of SIRPα, leading to a decreased threshold for phagocytosis.³⁵ Hence as a protective mechanism, mobilizing cytokines and inflammatory stimuli cause CD47 to be transiently upregulated just prior to and during their migratory phase, in order to avoid engulfment by those avidly activated macrophages.³³ Therefore, it has been hypothesized that the level of CD47 on the cells determines the probability that they are engulfed in vivo.

Recent studies have shown that mouse and human myeloid leukemia stem cells upregulate CD47 expression as an escape mechanism, which was associated with worse overall survival prognostic in patients, contributing to the pathogenesis and progression of the disease due to engagement of CD47 with SIRPα leading to inhibition and evasion of phagocytosis.^{36, 37} On the other hand, a body of evidence has demonstrated the usefulness of monoclonal antibodies directed against CD47 in different types of hematopoietic and solid malignancies in vivo and in vitro, promoting phagocytosis or apoptosis of the desire population and thus effective elimination of the tumorigenic cells.^{36, 38, 39, 40, 41, 42}

CD47 and xenograft rejection

Macrophages have been shown to engulf porcine red blood cells in an antibody- and complement-independent manner,⁴³ and to mediate the rejection of porcine islet xenografts in both rodents^{44, 45, 46} and primates.⁴⁷ Although host macrophages can be activated by cytokines produced by xenoreactive T cells, these powerful effects of xenogeneic macrophages could also be accounted for by the combined ability of certain xenogeneic antigens to activate macrophages,^{48, 49} while important inhibitory interactions are not effective.^{50, 51}

The CD47–SIRPα interaction which is species specific has been implicated as a critical regulator of xenotransplantation rejection in several cross-species transplants.⁵⁰ The rapid rejection of CD47-deficient hematopoietic cells by macrophages in syngeneic wild-type (WT) mice demonstrated a critical role for CD47 as a self-protection (or ‘don't eat me' marker). This finding also suggested that the rapid rejection of xenogeneic cells by recipient macrophages could have resulted from the inability of donor CD47 to interact with SIRPα on the recipient macrophages. This question was first addressed in a pig-to-mouse xenogeneic combination,⁵⁰ in which macrophages were found to mediate robust rejection of porcine hematopoietic cells.⁵² It was found that porcine CD47 could not induce SIRPα tyrosine phosphorylation in mouse macrophages and that blocking SIRPα with anti-mouse SIRPα monoclonal antibody (P84) did not affect the engulfment of porcine cells by mouse macrophages. These findings demonstrated the lack of cross-reaction between porcine CD47 and mouse SIRPα.⁵¹ Importantly, transgenic expression of mouse CD47 on porcine cells was found to elicit strong protection against phagocytosis by mouse macrophages both in vitro and in vivo.⁵⁰

It has been reported that recombinant human SIRPα proteins bound to porcine CD47.⁵³ However, the protein-binding assay does not necessarily reveal whether porcine CD47 could deliver inhibitory signaling via SIRPα to human macrophages. In fact, the demonstrated inability of porcine red blood cells to stimulate SIRPα tyrosine phosphorylation in human macrophages indicated a lack of functional interaction between porcine CD47 and human SIRPα.⁵¹ In agreement with this possibility, porcine cells transfected with human CD47 showed markedly reduced sensitivity to in vitro phagocytosis by human macrophages.⁵¹

SIRPα polymorphisms

It has been shown that immunodeficient mice on the non-obese diabetic background allowed better hematopoietic engraftment than did other strains with equivalent immunodeficiency-related mutations. The mouse SIRPα gene exhibits significant coding region polymorphism. This variation takes place mainly in the CD47-interacting V-like Ig domain, which in turn results in a differential binding capacity of human CD47 with the non-obese diabetic SIRPα polymorphic variant showing strong binding.⁵⁴ In contrast to SIRPα, CD47 seems to not be subjected to much polymorphic variation.

In order to understand the ligand, species and/or strain specificity of SIRP family receptors, studies have determined the crystal structure of the ligand-binding domain of murine SIRPα (mSIRPα LBD), for the specific CD47 recognition. Folding topology revealed that mSIRPα LBD adopts an I2-set Ig fold, but its overall structure resembles IgV domains of antigen receptors, but with an extended loop structure (C'E loop), forming a dimer interface in the crystal, consistent with the human SIRPα LBD structure⁵⁵ with the exception that human SIRPα has an additional beta-strand D.

Site-directed mutagenesis studies of mSIRPα identified three residues Ala65, Phe56 and Lys95 critical for specific recognition of mCD47 included in the C'E loop and regions corresponding to complementarity-determining regions of antigen receptors. The C′E loop region containing Ala65 is likely involved in ligand recognition, but it is unclear whether Ala65 directly interacts with the ligand or contributes to maintaining the conformation of mSIRPα however, it is known that the ligand binding domain of SIRPα has substantial sequence diversity even among mouse strains and that Ala65 is a hot spot of polymorphism.⁵⁶ The 129/sv strain type of mSIRPα has Ala65, while Ala65 in Balb/c and C57BL/6 strains is replaced by a Thr.⁵⁷

On the other hand, human SIRPα has a Ser at this site. The ligand-binding activity was destroyed by site-directed mutagenesis of Ser to Asp in this position of human SIRPa/SHPS-1, confirming the critical role of this site.⁵⁷ Thus, it is likely that the residue equivalent to Ala65 of mSIRPa is critical in determining the species, strain and/or isoform specificity in ligand recognition. These results suggest that the variable complementarity-determining region-like loop structures in the binding surface of SIRPα are generally required for ligand recognition in a manner similar to that of antigen receptors, which may explain the diverse ligand-binding specificities of SIRP family receptors.

Most of the previous studies regarding CD47-SIRPa interspecies incompatibility and xenograft rejection have focused on hematopoietic grafts, and little is known about its role in non-hematopoietic grafts. We have recently shown that at least for thymic epithelial cells CD47 is not an impediment for graft survival. Fetal thymus from syngeneic WT or CD47 KO donors were transplanted in thymectomized, T cell-depleted WT and CD47 KO mice. Transplantation of CD47 KO mouse thymus led to T-cell recovery in WT mice, showing a normal distribution of thymocyte subsets and long-term survival of mouse thymic epithelial cells. These results demonstrate that, unlike hematopoietic cells, CD47 KO mouse thymus can survive and function in WT mice.⁵⁸ However, it is possible that CD47 may still play an important role in the rejection of non-hematopoietic cellular xenografts. In support of this possibility, we have recently observed that liver parenchymal cells lacking CD47 do not survive as their WT counterpart does (Navarro-Alvarez and Yang, manuscript in preparation). Thus, the role of CD47 in xenograft rejection may differ for different types of xenografts.

Concluding remarks

The ubiquitously expressed cell surface molecule CD47 serves as an important ‘marker of self' for macrophages and DCs, and its signaling through SIRPα prevents inappropriate self-phagocytosis. The lack of cross-species interaction in the CD47–SIRPα pathway largely accounts for the vigorous rejection of xenogeneic hematopoietic cells by macrophages. However, the role for CD47 in non-hematopoietic xenograft rejection has been less clear. While CD47 expression is not required for survival or function of thymic epithelial cells when grafted as a tissue,⁵⁸ lacking CD47 on donor liver parenchymal cells induces rapid activation of innate immune cells and promotes graft rejection after hepatocyte transplantation (Navarro Alvarez et al., manuscript in preparation). Thus, the role of CD47 in xenograft rejection seems to differ between different types of xenografts, and further studies are needed to confirm this hypothesis.