FUTURE PROSPECTS FOR THE CLINICAL TRANSFUSION OF PIG RED BLOOD CELLS

Abstract

Transfusion of allogeneic human red blood cell (hRBCs) is limited by supply and compatibility between individual donors and recipients. In situations where the blood supply is constrained or when no compatible RBCs are available, patients suffer. As a result, alternatives to hRBCs that complement existing RBC transfusion strategies are needed. Pig RBCs (pRBCs) could provide an alternative because of their abundant supply, and functional similarities to hRBCs. The ability to genetically modify pigs to limit pRBC immunogenicity and augment expression of human ‘protective’ proteins has provided major boosts to this research and opens up new therapeutic avenues. Although deletion of expression of xenoantigens has been achieved in genetically-engineered pigs, novel genetic methods are needed to introduce human ‘protective’ transgenes into pRBCs at the high levels required to prevent hemolysis and extend RBC survival in vivo. This review addresses recent progress and examines future prospects for clinical xenogeneic pRBC transfusion.

Introduction

On a global scale, the demand for blood products has been unmet for decades. In the USA, there is at times an inadequate supply of blood to meet the demand, resulting in clinical harm [1]. This delicate balance between supply and demand is affected by a plethora of intrinsic and extrinsic factors, including (i) increased incidence of acute blood loss, (ii) illness-based restrictions on blood donation, e.g., Creutzfeldt-Jakob disease in Europe, or a high incidence of human immunodeficiency virus in some countries [2]; (iii) an increased number of transfusion-dependent patients who have become heavily alloimunized for whom compatible RBCs can be difficult to procure [3, 4]. As another example, during the SARS-CoV-2 (COVID-19) pandemic, the donor pool underwent major changes, causing severe blood shortages internationally [5–7]. Current RBC replacement therapy options for patients for whom no compatible blood is available is confined to hemoglobin-derived products, which are of limited efficacy, require an emergency use investigational new drug approval and are difficult to obtain [8]. As a direct result of these limitations, heavily alloimmunized patients can experience significant complications that result in increased morbidity and mortality [4, 9–11]. While innovative therapies are being explored to develop substitutes for human RBCs (hRBCs), e.g., stem cell-based therapies, these approaches will be challenging to scale up to meet the current clinical demands [12].

The first clinical blood transfusions in humans in 1667 were xenotransfusions. However, due to a lack of understanding of RBC biology and the immunological basis and possible consequences of RBC incompatibility, xenotransfusion practices were associated with complications and fell out of favor. In the middle of the 19th century, xenotransfusion was rescued from oblivion and fervently defended, but Ponfick and Landois stressed the potentially harmful effects of inter-species transfusion [13]. Xenotransfusion was then largely abandoned.

In recent years, considerable progress has been made in the experimental transplantation of gene-edited pig organs transplanted into immunosuppressed nonhuman primates (NHPs). Specific examples include reports of >1 year survival of life-supporting pig kidneys [14–16] and >6 months survival of orthotopically-placed pig hearts [17–19] in NHPs. Based on this experimental progress, gene-edited pig organ transplantation in brain-dead or living human recipients has been carried out. Although largely unsuccessful, these studies received significant public attention and encouraged the transplant community [20–22].

The main immunologic barriers to successful xenotransplantation are related to the presence of anti-pig natural antibodies in primates that bind to the three known xenoantigens on pig cells (Table1) as well as the molecular incompatibilities that preclude the regulation of complement-mediated intravascular hemolytic responses [23]. We now have available pigs in which expression of all three xenoantigens has been deleted (triple-knockout [TKO] pigs). We also have the ability to boost expression of protective genes, such as complement-regulatory proteins (CD46, CD55, CD59) on porcine somatic and vascular cells [14, 24]; such protective gene products may not only be important in enhancing baseline RBC survival, but may be especially important in settings where active complement activation may contribute to antibody-independent RBC removal [25, 26]. The generation of TKO pig red blood cells (pRBCs) that likewise express complement regulatory proteins would be predicted to help reduce complement-mediated lysis. These bioengineered pRBCs could be a viable alternative for hRBCs in the clinical setting. In this brief review, we report relevant in vitro and in vivo studies that take us closer to the introduction of clinical pRBC transfusion.

Table 1:

Carbohydrate xenoantigens that have been deleted in genetically-engineered pigs

Carbohydrate (Abbreviation)	Responsible enzyme	Gene-knockout pig
1. Galactose-α1,3-galactose (Gal)	α1,3-galactosyltransferase	GTKO
2. N-glycolylneuraminic acid (Neu5Gc)	Cytidine monophosphate-N-acetylneuraminic acid hydroxylase (CMAH)	CMAH-KO
3. Sda	β−1,4N-acetylgalactosaminyltransferase	β4GalNT2-KO

Potential clinical applications of xenotransfusion

Patients who have experienced major trauma associated with significant blood loss either in civilian or military settings, particularly when group O blood is not immediately available, may be candidates for a “universal donor” bioengineered pRBC transfusion [2]. Patients with sickle cell disease (SCD) and other hematological states who develop alloimmunization to repeated hRBC exposure could also be potential candidates for pRBCs [4, 9–11, 27]. The incidence of immunization to pRBCs in humans is unknown, but these xenogeneic cells are probably at least as immunogenic as hRBCs; and likely more so. For example, the rate of RhD-alloimunization after transfusion of at least one RhD-positive RBC unit is 11%–50% [28]. Importantly, there is current evidence to suggest that sensitization to pRBCs is not detrimental to a subsequent allotransfusion [27, 29]. Similarly, allosensitized patients with SCD (anti-RBC antibodies in allosenstized SCD patients are all IgG, not IgM) may demonstrate no IgG binding to TKO pRBCs (Figure 1). These data indicated that, in alloimmunized SCD patients, anti-hRBC IgG did not cross-react with TKO pRBCs, providing a viable therapeutic option for these patients [27].

Figure 1: Plasma IgG binding from patients with sickle cell disease (SCD) or healthy controls (HC) to triple-knockout (TKO) pRBCs.

Only one (8.3%) HC sample showed IgG binding to TKO pRBCs. There was no significant difference in IgG binding between the three groups. IgG binding of HC to human blood type O RBCs is shown as the negative control. On the y axis, the dotted red line represents the lowest measurable limit of binding, below which no binding is considered to exist (rGM: IgG 1.1). (Reproduced with permission from Yamamoto et al, ref[27]).

Patients with immunosuppressed states, such as patients with HIV/AIDS, become alloimmunized at a far lower rate when compared to routinely transfused adult patients [3]. For example, in a retrospective examination of RhD-incompatible transfusions in an HIV/AIDS patient population, not a single patient was shown to become alloimmunized to the RhD antigen [30, 31]. These data suggest that patients receiving immunosuppressive therapy may also be candidates for pRBCs in an emergency if hRBCs are not available, as their rates of alloimmunization can be lower when compared to that in healthy adults. In a two-center retrospective study evaluating adult patients taking immunosuppressants, the relative risk of alloimmunization against hRBCs was 0.55 compared to patients not taking immunosuppressants [32].

Overcoming the immune barriers to xenotransfusion

Excluding initial clinical attempts and a few experiential approaches studying RBC xenoimmunization conducted within the last century [33], the initial experimental attempts at xenotransfusion with pRBCs were limited by pRBC immunogenicity and our lack of ability to genetically engineer the source pig. The two main immunological barriers were, and remain, the innate and adaptive arms of the immune system. The innate immune system barriers include (i) the naturally-occurring ‘preformed’ IgM antibodies directed to specific epitopes on pRBCs [34–36]. Binding of IgM to pRBCs results in complementmediated lysis of pRBCs [34, 37] similar to that seen in some cases of human ABO-incompatible blood transfusion [38], and (ii) phagocytosis of pRBCs [34, 36] and (iii) coagulation and platelet activation potentially augmented by molecular incompatibilities [39]. The adaptive arm of the immune system consists of a T cell-mediated response leading to the generation of new anti-pRBC xenoantibodies [27] (Figure 2).

Figure 2: IgM/IgG binding of capuchin monkey serum to TKO pRBCs before and after xenotransfusion, followed by allotransfusion.

Briefly, twenty-five percent (25%, estimated to be 15mL/kg) of the total blood volume (60mL/kg) of two capuchin monkeys (C9 and C11) was replaced by TKO pig RBCs (xenotransfusion) or by RBCs from four capuchin monkeys. Five days after xenotransfusion, IgM binding to TKO pRBCs increased, especially in C11, and reached a peak on day 7 (C11) or day 14 (C9). IgG binding also increased seven days after xenotransfusion, especially in C11. Fourteen days post-xenotransfusion, IgG binding remained at a high level until euthanasia (day 28). These data suggested that T cell-mediated immune responses lead to the generation of anti-TKO pRBC xenoantibodies. Day 49 after xenotransfusion, allotransfusion was done in this experiment. There was no-cross-reactivity with alloantigens. (Reproduced with permission from Yamamoto et al, ref[27]).

The innate immune system

In vitro studies

In infancy, humans develop natural (preformed) antibodies directed to three carbohydrate epitopes present on pRBCs alpha Gal (αGal), Neu5Gc, and Sd^a (Table1, Figure3) [40], the major antigen being αGal, a terminal oligosaccharide similar to the human ABO blood group saccharides, especially blood group B (Figure 4). The development of anti- αGal antibodies is believed to be similar to that of the anti-A/B antibodies, and occurs in infancy as a response to colonization of the gastrointestinal tract by various bacterial and viral flora that express the same glycans as those expressed on pRBCs [41–43].

Figure 3: Correlation of human serum antibodies to WT and TKO pRBCs with age.

Relative geometric mean (rGM) binding and age correlation of human serum IgM (A,C) and IgG (B,D) antibodies to WT (top) and TKO (bottom) pRBCs. The dotted lines indicate no IgM or IgG binding. (Note the great difference in the scale on the Y axis between the top and bottom figures.) (Reproduced with permission from Li Q et al, ref[105])

Figure 4: Carbohydrate structures of human ABO RBCs and α-Gal on pig RBCs.

Pig RBCs express Gal oligosaccharides that are similar in structure to human blood type B oligosaccharide, except for the fucose side-arm.

Old World NHPs (e.g., baboons, rhesus monkeys, cynomolgus monkeys) were originally used as surrogates for humans to study pRBC transfusion [34, 36, 44]. However, a limitation of Old World NHPs is that they express N-glycolylneuraminic acid (Neu5Gc) (Table1), while humans do not. If the gene for the enzyme, cytidine monophosphate-N-acetylneuraminic acid hydroxylase (CMAH) (Table 1), is knocked out in the pig (resulting in deletion of expression of Neu5Gc), as in TKO pigs, this appears to expose a new xenoantigen (or xenoantigens) on the pig cells (the so-called “fourth xenoantigen”). (The structure of this antigen is hitherto unknown, though it is known to be a glycan.) As a result, Old World NHPs have antibodies against TKO pig cells [45–47], but New World NHPs (e.g., capuchin, squirrel, and spider monkeys) do not have antibodies against TKO pRBCs, and therefore more closely mimic many humans in this respect [46, 47]. However, other factors may be playing a role in the destruction of TKO pRBCs in Old World NHP sera. There is a very strong cytotoxic effect against pig cells (Figure 5), and this could possibly be related to factors other than expression of the fourth xenoantigen, e.g., complement activity, and possibly to molecular incompatibilities [48].

Figure 5: Correlation of human (n=9) and baboon (n=72) serum IgM (left) and IgG (right) antibody binding with serum complement-dependent cytotoxicity (CDC, at 50% serum concentration) to TKO pPBMCs.

In both humans and baboons, there was a significant increase in cytotoxicity as IgM and IgG antibody binding to TKO pig peripheral blood mononuclear cells (PBMCs) increased. In baboons, however, cytotoxicity was high whether IgM binding was high (e.g., 80% cytotoxicity at a relative geometric mean (rGM) of 8) or relatively lower (e.g., 75% at a rGM of 2). (**p<0.01). (Reprinted with permission from Yamamoto T, et al. ref [46]).

A major breakthrough in xenotransfusion/xenotransplantation occurred in 2003 when pigs were genetically engineered to delete the gene responsible for the generation of αGal epitopes, α1,3-galactosyltransferase (Table 1) [49–51]. Subsequently, knockout of the enzymes involved in the production of Neu5Gc [52–54] and Sd^a [55, 56] were knocked out (Table 1). In vitro studies demonstrated that there was a significantly greater lysis of wild-type (WT, i.e., genetically-unmodified) pRBCs than of ABO-incompatible (ABO-I) hRBCs and of GTKO pRBCs (p<0.01). Furthermore, ABO-incompatible hRBCs sustained significantly greater lysis than GTKO pRBCs (p<0.01) [36], but there was significantly greater lysis of GTKO pRBCs than of ABO-compatible hRBCs (Figure6A). When TKO pRBCs were exposed to human serum, no lysis was observed (as with ABO-compatible hRBCs) (Figure6B) [27, 40]. Wang et al carried out several studies evaluating pRBC viability in vitro. Human IgM binding was (i) 3-fold less to GTKO pRBCs than to WT pRBCs, and (ii) 227-fold less to double-KO (GTKO/CMAH-KO) pRBCs than to GTKO [57]. Furthermore, (iii) human IgG binding was 27-fold less to GTKO/CMAH-KO pRBCs than to GTKO pRBCs, and (iv) 2-fold less to GTKO pRBCs than to WT pRBCs. When evaluating complement-dependent cytotoxicity (CDC), the results of WT, GTKO and double-KO (GTKO/CMAH-KO) pRBCs were 1.5, 0.64 and 0.07. Human antibody binding to TKO pRBCs was further significantly reduced [44]. Our own previous evaluation of IgM and IgG binding of sera from kidney transplant waitlist patients (n=19) to TKO pRBCs showed almost no IgM binding (17/19, 89.5%) and no IgG binding (Figure6C) [27]. These reports indicated that (i) deletion of expression of the three xenoantigens increased pRBC survival and (ii) TKO pRBCs and ABO-compatible hRBCs showed similar levels of IgG and IgM binding. Building on these data, we recently evaluated TKO pRBCs as potential sources for transfusion in alloimmunized patients with SCD [27]. The in vitro data were promising as plasma from neither alloimmunized nor non-alloimmunized SCD patients bound IgG/IgM to, or induced CDC of, TKO pRBCs (Figure6C). TKO pigs are therefore the optimal source of pRBCs for xenotransfusion identified to date. The first barrier of immediate pRBC lysis by human serum has therefore been eliminated.

Figure 6: (A) Human serum CDC of ABO-compatble (ABO-C), ABO-incompatible (ABO-I) hRBCs and of wild-type (WT) and GTKO pRBCs.

(A) Human sera (50%) of blood types O (n=10), A (n=9), B (n=8), and AB (n=4) were tested for CDC of ABO-C, ABO-I, WT, and GTKO RBCs. There was significantly greater lysis of WT than of ABO-I and GTKO RBCs (p<0.01). ABO-I RBCs sustained significantly greater lysis than GTKO RBCs (p<0.01). There was significantly greater lysis of GTKO than ABO-C RBCs (p<0.01). **p<0.01. (Reproduced with permission from Long et al, ref [36]).

(B) Pooled human serum complement-dependent cytotoxicity (hemolysis) of WT, GTKO, and TKO pRBCs. Briefly, RBCs were incubated with diluted serum for 30 min at 37°C. After washing, RBCs were incubated with rabbit complement (Sigma; final concentration 20%) for 150 min at 37°C. After centrifugation, supernatant was collected, and hemolysis was evaluated using a Multi-Label Microplate Reader (Perkin Elmer Victor3). The absorbance of each sample at 541 nm was measured. Cytotoxicity of the same serum to autologous human O RBCs was tested as a control. (Reproduced with permission from Yamamoto et al, ref [40]).

(C) Serum IgM/IgG binding of (i) kidney transplant waitlist patients (n=19), (ii) pooled human plasma, and (iii) sensitized baboon serum to TKO pRBCs. The great majority of human sera have no antibodies (positive IgM 2/19 [10.5%], IgG [0%]) that bind to TKO pRBCs. The dotted line indicates no antibody binding. The sensitized baboon [B6414] sample is shown as a positive control. The assay was carried out x3. Results are expressed as mean+/−SD. (Reproduced with permission from Yamamoto et al, ref [27]).

In vivo studies

Potential mechanism of complement-dependent hemolysis after pRBC transfusion in primates

The current major barrier to enable pRBC xenotransfusion is likely complement. If serum antibody binding to TKO pRBCs is present (e.g., because of expression on pRBCs of other currently unidentified glycan antigens), the cells may be hemolyzed by complement. Activation of the complement system serves two general roles: (i) opsonization of target cells by C3b, which facilitates as a phagocytosis signal, and (ii) recruitment of the C5–9 and formation of the membrane attack complex (MAC), facilitating cell lysis. Of the three complement pathways, it is more likely that the classical and possibly the alternate pathways are the more relevant targets [58, 59]. However, the lectin pathway cannot definitively be ruled out as non-contributory to pRBC lysis [60–63]

The classical pathway is believed to be the major complement cascade responsible for pRBC lysis in vivo, and likely acts similarly to that of hRBC lysis in autoimmune hemolytic anemia [58]. When IgM and IgG to pRBCs have been elicited, the antigen-antibody complex activates C1q, turns on the C3 and C5 convertases, and leads to the generation of the MAC [64]. Also, C3b generated by the classical and alternate pathways can opsonize the pRBCs for macrophage-mediated destruction in the liver and/or the spleen. Targeting this pathway would be the most appropriate candidate (by a C1 and/or C3 and/or C5 inhibitor) to prolong pRBC survival in vivo.

Cells have developed a complex system to protect against complement attack by expressing several membrane-bound complement-regulatory proteins, e.g., decay accelerating factor (DAF or CD55), CD59, and membrane co-factor protein (MCP, CD46). CD59, is a GPI-anchored protein preventing membrane attack complex (MAC) assembly and consequent cell lysis. Individuals lacking CD59 suffer from paroxysmal nocturnal hemoglobinuria, a condition manifested by intravascular hemolysis, reduced hematopoiesis, and elevated risks of venous thrombosis.

Potential approaches to prevent complement-mediated lysis

Complement inhibition

It is likely that all three complement pathways work in concert to facilitate clearance of microbes and non-self-tissues and cells (e.g., transplants or pRBCs) [63]. Inhibiting the complement cascade at multiple points may be most effective and beneficial. For example, a combination of C3 and C5 inhibitors would reduce the formation of C3b and the MAC. Reduced levels of C3b would reduce both pRBC opsonization and the phagocytosis of pRBCs by liver and spleen macrophages. Inhibiting the complement pathway at only the C1 level (thus reducing the generation of C3 convertase) by IVIg is successful in increasing pRBC survival in vitro [37]. In an in vivo study, a naive baboon administered IVIG showed significantly decreased C3a levels and the baboon’s serum no longer lysed WT pRBCs in vitro (Figure 7) [37]. These data suggest that, if C3 activation can be fully targeted and inhibited, pRBCs may show prolonged survival after transfusion into humans of NHPs.

Figure 7: Serum cytotoxicity of baboon serum against WT pRBCs after i.v. administration of IVIg (FLEBOGAMMA) (2g/kg).

(A)Baboon serum cytotoxicity (final concentration 50%) immediately before IVIg (day 0), immediately after IVIg (2mg/kg) (day 0), and on days 1 and 13 after IVIg against WT pRBCs was measured. Cytotoxicity was significantly decreased immediately after IVIg infusion but was recovering on day 1. The dotted line represents cut-off value (7%), below which cytotoxicity is considered negative. Results are expressed as mean+/−SD (*p<0.05, **p<0.01) (Reproduced with permission from Yamamoto et al, ref [37]).

(B)Immediately after IVIg, lysis of WT pRBCs was completely inhibited. (Reproduced with permission from Yamamoto et al, ref [37]).

(C)Complement activity (as measured by C3a levels) after IVIg administered to a naive baboon. The plasma C3a level immediately (day 0) after IVIg administration was significantly decreased [p<0.05] but was recovering by day 1. Results are expressed as mean+/−SD (*p<0.05, **p<0.01). (Modified from Yamamoto et al, ref [37]).

Many complement inhibitors are currently in various phases of clinical trials or are already FDA-approved [65, 66] (Table2). Frequently transfused patients generate IgM and IgG antibodies to human leukocyte antigen (HLA) class I epitopes expressed on transfused allogeneic platelets, leading to platelet destruction [67, 68]. A successful pilot clinical trial of complement inhibition to alleviate platelet refractoriness was recently completed [66] (NCT02298933).

Table 2:

Recent or current clinical trials of novel inhibitors of complement.

#	Condition(s)	Inhibitor	Alternate name	Type	Target	Company, Sponsor, Collaborator	Phase	NCT Number
1	wAIHA	ANX005		ab	C1q	Annexon	2	NCT04691570
2	PNH	Danicopan	ALXN2040	sm	FD	Alexion	3	NCT05389449
3		Iptacopan	LNP023	sm	FB	Novartis	3	NCT04820530
4		ARO-C3		RNAi	C3	Arrowhead	1	NCT05083364
5		Nomacopan	Coversin, REV-675	pro	C5	AKARI	2	NCT03427060
6		BCX9930		sm	FD	BioCryst	2	NCT05116787
7		KP104		ab-pro	C5 & FH	Kira	2	NCT05476887
8		Pegcetacoplan	APL-2	pep	C3	Apellis	2	NCT03593200
9	Transplant, kidney, complement mediated rejection	conestat alfa, IVIG	rhC1INH	pro	C1	Pharming Tech B.V., U Wisconsin	2	NCT01035593
10	AHE	Cinryze	C1INH-nf	pro	C1 (C1r and C1s)	Shire, Takeda	3	NCT00462709
11	aHUS	OMS721	OMS620646	ab	MASP 2	Omeros	3	NCT03205995
12		CCX168	Avacopan	sm	C5aR1	ChemoCentryx	2	NCT02464891
13	GPA & MPA	Vilobelimab	CaCP29 IFX1	ab	C5a	InflaRx GmbH	2	NCT03712345
14	IgA nephropathy	Cemdisiran	ALN-CC5	RNAi	C5	Alnylam	2	NCT03841448

ab-antibody; ab-pro-antibody fusion protein; aHUS-atypical hemolytic uremic syndrome; C1r-complement component 1r; C1s-complement component 1s; C5aR1-complement C5a Receptor 1; FB-complement factor B; FD-complement factor D; FH-complement factor H; GPA-granulomatosis with polyangiitis; HAE-acute hereditary angioedema; HMV-healthy male volunteers; IVIG-intravenous immunoglobulin; MASP2-mannan-binding lectin serine protease 2; MPA-microscopic polyangiitis; sm-small molecule; pdC1INH-plasma-derived human C1-inhibitor concentrate; pep-peptide; PNH-paroxysmal nocturnal hemoglobinuria; pro-protein; rhC1IHC-recombinant human C1-inhibitor; RNAi-Ribonucleic acid (RNA) interference; warm AIHA-autoimmune hemolytic anemia-

Eculizumab (a C5-inhibitor) shows promise to prevent platelet refractoriness, and is also effective at blocking the CDC of hRBCs associated with paroxysmal nocturnal hemoglobinuria [69] and atypical hemolytic uremic syndrome [70]. After a single eculizumab infusion, responding patients showed a reduction in platelet requirement and higher post-transfusion platelet numbers for 14 days. Similar benefits may be achieved when eculizumab is administered at the time of pRBC transfusion. Studying eculizumab in vivo in NHPs is impossible because it does not recognize NHP C5 [71]. However, Adams et al. successfully utilized another anti-C5 antibody, tesidolumab, in the pig-to-rhesus kidney xenotransplant model [72].

Multiple C3 inhibitors are currently (or have been) in clinical trials (Table2). A C3 inhibitor, Cp40, prevented complement activation and demonstrated reduced cell damage and preserved function of WT pig hearts perfused ex vivo with human serum [73]. C1 esterase inhibitors have been shown to reduce complement activation and delayed allograft function in recipients of deceased donor kidneys [74]. C1 esterase inhibitors may prove effective in xenotransfusion. Further vivo studies of complement inhibitors in pRBC transfusion in NHPs are required.

Transgenic expression of human complement-regulatory proteins

Based on studies of pig organ and cell xenotransplantation in NHPs, antibody-mediated rejection can be prevented or reduced by the transgenic expression of one or more human complement-regulatory proteins on the cells of the pig graft, e.g., hCD55 (decay-accelerating factor), hCD46 (membrane cofactor protein), or hCD59 (MAC-inhibiting protein) [24, 75, 76]. For example, failure of WT pig organs exposed in vivo to human or NHP blood generally occurs within minutes or hours [75, 77–79], whereas kidneys from pigs expressing a single human complement-regulatory protein (hCD55) have functioned for up to 90 days [80].

To overcome the species-specificity of complement-regulatory proteins, transgenic pigs expressing human complement-regulatory proteins have been produced, but in many cases expression has not been demonstrated on pRBCs [27, 36, 81]. However, in one early report α- or β-globin promoter-driven expression of hCD55 and CD59 was achieved on pRBCs [82]. In vitro assays demonstrated a 50–82% reduction in pRBC lysis by baboon serum. Transgenic mice carrying human β-globin genes had been produced previously [83, 84]. The successful expression of human complement-regulators (and possibly other human ‘protective’ proteins) on pRBCs is currently the major contribution needed to advance this field.

Reducing phagocytosis of pRBCs

Direct phagocytosis and antibody-dependent cell-mediated cytotoxicity (ADCC) are other innate immune barriers to successful xenotransfusion [27]. pRBCs do not provide the necessary inhibitory signals to prevent phagocytosis by human macrophages [85–87]. A specific example is the lack of CD47/SIRPα (signal regulatory protein α) signaling on pRBCs [85–91]. Human CD47 interaction with human SIRPα inhibits phagocytosis. The expression of pig membrane protein CD47, e.g., on pRBCs, does not have the same effect as human CD47 due to species incompatibility [86]. Transgenic expression of human CD47 on pRBCs can reduce phagocytosis and could prolong pRBC survival. Current techniques of gene-editing have not been employed to express human CD47 on pRBCs. If this became possible, it would be a major contribution to advance this field. Expression of hCD47 on hematopoietic progenitor cells in pigs increased engraftment in a murine model of pig-to-human cell transplantation [91].

Other options for suppressing macrophage activity by gene-editing exist. For example, expression of HLA-E and/or HLA-G inhibits macrophage-mediated cytotoxicity [89, 92]. If the current technological limitations prevent the expression of human CD47 or other anti-phagocytic proteins on pRBCs, clinically-approved drugs like dasatinib and bosutinib may have some effect [93]. These drugs were originally developed as inhibitors of protein-tyrosine kinases Bcr-Abl and Src, and also target salt-inducible kinase 2 found on macrophages. These compounds appear to induce features of regulatory macrophage phenotypes and have anti-inflammatory properties. By inducing macrophages of a regulatory phenotype, direct phagocytosis may be inhibited, possibly leading to prolonged survival of TKO pRBCs. However, further investigation is needed.

In addition to the better-known roles of the erythrocyte in the transport of oxygen and carbon dioxide, the concept that the RBC is involved in the transport and release of ATP has been evolving. Membrane proteins involved in the release of ATP from erythrocytes now appear to include members of the ATP binding cassette (ABC) family [94]. Enzymatic removal of ATP and UTP (by apyrase or the expression of ectopic CD39) abrogated the ability of apoptotic cell supernatants to recruit monocytes in vitro and in vivo [95]. Therefore, human CD39 on pig RBCs may reduce phagocytosis by monocytes.

Summary of relevant xenotransfusion studies in vivo

Owing to the difficulty of establishing an animal model capable of reproducibly evaluating long-term survival of pRBCs, in vivo studies have been limited (Table3). Despite this, in vivo survival of pRBCs has improved over time. Initially, treating the pRBCs with α-galactosidase (to remove αGal antigens) reduced in vitro binding of baboon or human antibodies, and inhibited serum CDC [34]. When non-enzymatically-treated pRBCs were transfused into baboons, the survival of the pRBCs was <5 min, whereas α-galactosidase-treated pRBC were still detected at 120 min [34]. Treatment with cobra venom factor (CVF) (to deplete complement) increased pRBC detection to 24 hours, suggesting that the complement system is a major barrier to achieving successful pRBCs transfusion. When CVF therapy was combined with inhibition of anti-αGal antibody by the infusion of a bovine serum albumin (BSA)-Gal conjugate, α-galactosidase-treated pRBCs survived for to >72 hours. Survival of >72 hours was also observed when α-galactosidase-treated pRBCs were infused into a baboon that received CVF and medronate liposomes (impairing phagocytic activity) [34]. Survival was also extended when very large numbers of pRBCs were transfused (300% of the baboon’s total number of RBCs). Despite this, the pRBCs were removed by the spleen, which was enlarged 2–3 times its normal size, was severely congested, and histologically demonstrated marked follicular hyperplasia [96].

Table 3:

Relevant in vivo xenotransfusion studies of pig RBCs

#	Authors	Year	Pig	NHPs	Modification	Survival	Ref
1	Eckermann, et al.	2004	WT pigs	baboons	(−)	<5 min	[34]
					AGL	2 hours
					CVF	24 hours
					AGL+CVF	72 hours
					AGL+CVF+ medronate liposome	72 hours
2	Dor et al.	2004	GTKO pigs WT pigs	baboons baboons	(−)	5 min	[96]
					soluble Gal conjugate +CVF	15 min
3	Tan et al.	2006	WT pigs	Rhesus monkeys	AGL+SPA-mPEG	12 hours	[97]
					AGL+SPA-Mpeg +immunosuppressive	40 hours
4	Yamamoto et al.	2021	TKO pigs	Capuchin monkeys	(−)	5–7 days	[27]

AGL-α-galactosidase; CVF-cobra venom factor; WT-wild type; GTKO-α1,3-galactosyltransferase knockout; SPA-mPEG-succinimid propionate-linked methoxypolyethyleneglycol; TKO-triple knockout

These combined results suggested that, at that time, the next logical steps to prolong pRBCs survival in vivo were (i) deletion of expression of xenoantigens on pRBCs, (ii) further depletion of anti-pig antibodies, (iii) complement inhibition, and (iv) inhibition of phagocytic activity and possibly (v) inhibition of coagulation-platelet activation.

Succinimid propionate-linked methoxypolyethyleneglycol can be used to camouflage non-αGal antigens (Table3) [97]. Modification with α-galactosidase and succinimid propionate-linked methoxypolyethyleneglycol increased the survival of pRBCs in a rhesus monkey to 12 hours after transfusion. While immunosuppressive therapy would likely not be effective if given with a pRBC transfusion, it did increase pRBC survival to 40 hours when α-galactosidase and succinimide propionate-linked methoxypolyethyleneglycol-modified pRBCs were transfused into a rhesus monkey [97].

We recently demonstrated that TKO pRBCs survived in non-immunosuppressed NHPs for 5–7 days, despite the fact that the NHPs had some anti-TKO pig IgM antibodies prior to xenotransfusion [27] (Table3). Longer survival of pRBCs would likely be achieved if NHPs were selected that did not have anti-TKO pig IgM/IgG antibodies, and/or when anti-complement and/or anti-phagocytic medications are administered.

In patients with clinical transfusion-dependent conditions, e.g., SCD or thalassemia, the transfusion of pRBCs may need to be accompanied by a short course of immunosuppressive therapy to prevent the potential formation of new human anti-pRBC antibodies, i.e., to prevent sensitization to the pRBCs [98, 99].

The challenges of expressing human regulatory proteins on pRBCs

A major challenge facing xenotransfusion is our frequent inability to date to induce expression of human transgenes on pRBCs as, with perhaps one exception [82]. No approach to achieve this has yet been demonstrated to be fully successful [2]. The expression of human transgenes, such as human complement-regulatory proteins and anti-phagocytic proteins, might allow pRBCs to escape destruction by human complement and phagocytosis. Gene manipulation in fetal fibroblasts, somatic cell nuclear transfer, and embryo implantation may resolve this issue [45, 75] . CRISPR/Cas9 gene editing may enable transgenes to be inserted at specific genomic locations during embryonic development, which might drive the expression of human transgenes on pRBCs [100]. Hemoglobin-driven transgenic approaches have been used to express human proteins on the surface murine RBCs [101–103], but this same approach has not be employed on pRBCs. If using the pig endogenous promoters fails, other promoters need to be explored. Nanoparticle-mediated gene delivery may also be an option. For example, nanoparticles have been succesfully used to to deliver and express therapeutic levels of anti-tumor genes (e.g., interleukin-12), in cancer cells [104].

Future directions for clinical xenotransfusion

There is substantial evidence that TKO pRBCs are important for successful clinical xenotransfusion because a significant percentage of humans do not have antibodies to TKO pRBCs [27, 44, 47, 105]. Direct phagocytosis due to species incompatibility also needs to be resolved [27]. However, only few in vivo studies have been carried out to date (Table 3). In order to achieve clinically successful xenotransfusion, TKO pRBCs that express human ‘protective’ proteins would be a major step forward. Until this has been achieved, though not ideal, the transfusion of TKO pRBCs combined with the administration of an appropriate complement inhibitor and/or a short course of immunosuppressive therapy, e.g., a single injection of an agent that blocks the CD40/CD154 T cell costimulation pathway or prevents B cell activation, may improve the survival of the cells sufficiently to be of clinical therapeutic value. Such an approach may be particularly warranted in patients with life threatening anemia for whom no compatible allogeneic hRBCs can be obtained.

Practice points

1. In 2022, the successful completion of the first xeno-to-human heart transplant galvanized the xenotransplantation community by demonstrating direct clinical utility of transgenic xeno-biotechnology.

2. Transgenic pig red blood cells can supplement human blood products for transfusion support in transfusion dependent patients (ex: Sickle cell disease), unexpected blood shortages (ex: SARS-CoV-2 (COVID-19) pandemic), and during trauma, when human RBCs are not immediately available.

Research Agenda

1. Develop efficient, ethical, and financially responsible extraction, storage, and transport infrastructure for pig red blood cells.

2. Reduce, and when possible eliminate, in vivo complement and macrophage mediated cytotoxicity to prolong pig red blood cell survival using available inhibitors.

3. Genetically engineer pig red blood cells to express additional human proteins, including anti-complement proteins, to increase pig red blood cell survival in vivo.

Abbreviations

CDC

complement dependent cytotoxicity

Gal

galactose-α1,3-galactose

h

human

HLA

human leukocyte antigen

IVIg

intravenous immunoglobulin

Neu5Gc

N-glycolylneuraminic acid

NHP

nonhuman primate

P

pig

RBC

red blood cell

SCD

sickle cell disease

SIRPα

signal-regulatory protein α

TKO

triple knockout

WT

wild-type