The red blood cell surface galactose antigen (DGalα1→3DGal) that induces hyperacute homotransplant rejection due to an ABO mismatch was identified in the late 1980s (11). During xenotransplantation, hyperacute rejection results in an abnormal increase in immunoglobulin (Ig)M serum levels rather than in IgG levels. This indicates that the recipient’s immune system first recognizes the specific antigens harbored in xenografts.

Due to the emergence of monoclonal antibodies (mAbs) using hybridomas, human anti-swine antibodies waere generated and used to identify significant carbohydrate structures for xenotransplantation (12). Then the α-galactosyltransferase (α-Gal) was found, which is encoded by the α-1,3-galactosyltransferase (GGTA1) gene (13). Other carbohydrate antigens, such as non-fucosylated chondroitin sulfate monolayers and linear antigens, are also found, locating on the surfaces of all porcine vascular endothelial cells. These antigens tightly bind to anti-Gal isogalectin β4 antibodies and specifically bind to natural, human anti-α-Gal antibodies. Gal epitopes are expressed abundantly in the brush margins of proximal convoluted tubules, moderately in distal convoluted tubules, and not at all in renal collecting tubules and glomeruli. A specific antigen-antibody reaction activates the complement system, leading to a powerful cytotoxic effect that leads to hyperacute grafts (14–18). The discovery of the α-Gal antigen was a major breakthrough in xenotransplantation.

Thereafter, considerable efforts were directed toward decreasing hyperacute rejection of kidney xenotransplants by removing anti-porcine antibodies in vitro, short-term infusions of specific carbohydrates (19), or the absorption of anti-xenoantigen antibodies produced in the spleen and kidneys (20). Soluble Gal proteins can partially inhibit human rejection of porcine kidneys. Intravenous infusions of bovine serum albumin-Gal in vivo can essentially maintain the depletion of circulating anti-Gal antibodies and prevent or delay antibody deposition and the acute humoral rejection of pig-to-baboon xenografts, but it might be associated with liver damage (21).

Transgenic technology was established in 1981 using microinjections; and a transgenic mouse model was created in 1982. The first generation of the gene-editing tool, zinc finger nuclease, was introduced during the late 1990s, and another, transcriptional activator-like effector nuclease, was identified in 2009. These gene-editing techniques had a positive global impact on life sciences.

Pigs with α-Gal knockout (α-Gal^-/-, GTKO) are important xenotransplantation models (22–24). In the α-Gal^-/- pigs to baboon kidney xenotransplantation models, most recipients did not develop hyperacute rejection; however, they succumbed to acute humoral rejection. The significantly increased abundance of peripheral anti-non-α-Gal antibodies in recipients suggested that non-α-Gal antigens in kidney xenografts might trigger the production of large amounts of corresponding antibodies. Thereafter, non-α-Gal antigens were recognized as obstacles to α-Gal^-/- pig organ xenotransplantation (25). The α-Gal antigen is crucial for hyperacute rejection, and non-α-Gal antigens play important roles in humoral rejection of xenotransplants. In addition to α-Gal and non-α-Gal, other carbohydrate antigens have a complex spatial distribution in porcine kidneys and are strongly associated with the outcome of porcine kidney xenotransplantation (26).

Non-α-Gal antigens, such as N-glycolylneuraminic acid (Neu5Gc; HD antigen), encoded by the cytidine monophospho-N-acetylneuraminic acid hydroxylase (cMAH) gene have been identified (27, 28). Compared with GGTA1^-/- pig xenotransplantation, humoral rejection is reduced in GGTA1^-/-/CMAH^-/- pigs xenotransplantation (29), implying that the immune heritability of the Neu5Gc antigen potentially plays an important role in pig-human xenotransplantation. The other carbohydrate non-α-Gal antigen, glycosyltransferase, (SD(a) antigen), is encoded by the β-1,4-N-acetyl-galactosaminyl transferase (B4GalNT2) gene (30, 31).

Clustered regularly interspaced short palindromic repeats (CRISPR)-associated protein (Cas9) is a third-generation gene-editing tool. Porcine embryonic fibroblasts with GGTA1^-/-/Gal^-/- were initially created using CRISPR/Cas9 in 2014 (32). Since then, CRISPR/Cas9 has become the preferred means of generating genetically engineered pigs. The serum of many waitlisted patients contained only a minimal number of antibodies that reacted with peripheral blood mononuclear cells from GGTA1^-/-/CMAH^-/-/B4GalNT2^-/- pigs. However, anti-human leukocyte antigen antibodies in some sensitized patients cross-reacted with porcine major histocompatibility complex (MHC) I antibodies (33). Pigs with simultaneous MHC and three antigen (GGTA1/CMAH/B4GalNT2) inactivation have been generated using the CRISPR/Cas method (34). Natural and inducible anti-SDa plays important roles in GTKO pig-to-rhesus monkey xenotransplant rejection, thus providing further support for the notion that Gal and SDa antigens should be simultaneously targeted (35). Exploration of new key non-α-Gal antigens is currently underway (36).

SLAs are being discovered to play an important role in swine innate and adoptive immune responses. In some sensitized kidney transplant-waitlisted patients, some human leucocyte antigen (HLA) antibodies cross-react with SLA class I (37). SLA II is also a xenoantigen (38, 39). And triple (GGTA1, CMAH, B2M) genes modified pigs expressed the SLA I^low phenotype, which effects on immune status and susceptibility to human immune responses (40). In vitro human TNF-α could increase SLA I expression, while human IL-17 could decrease TNF-α-mediated SLA-I upregulation (41), and downregulation of SLA expression decreases the strength of xenogeneic immune responses towards renal tubular epithelial cells (42). These data may support the SLA-silencing strategy application to prevent xenogeneic cellular immune responses.

Attempts to induce immune tolerance in xenografts by multiple low-dose xenoantigen inoculations have been unsuccessful. Transplanting fetal porcine thymus and liver tissues into mice to eliminate T and natural killer cells and removing the thymus induces specific tolerance to porcine antigens (78). The mouse CD4⁺ T cell repertoire developed in implanted pig thymus grafts indicated positive selection by porcine (xenogeneic) MHC antigens and negative selection by both mice (recipients) and porcine MHC; this suggested a high level of tolerant immunocompetence (79–81). Findings of kidney allotransplantation in large animals have indicated that the thymus is essential for rapid and stable immune tolerance (82, 83), implying the potential value of thymus transplants to induce tolerance.

The “thymus kidney” was invented by placing thymus tissues under a kidney quilt to facilitate autologous thymus transplantation. The results suggested that the abundance of peripheral CD4⁺CD45RA⁺ T cells increased steadily from 30 to 150 days after transplanting “thymus kidneys” into athymic micropigs, and recipient pigs had acquired immune tolerance. Vascularized donor thymus tissue can induce rapid and stable immune tolerance in recipients to MHC-unmatched allograft (84–86).

In “thymus kidneys” xenotransplantation models, recipient baboons transplanted with a “thymus kidney” graft from a human CD55 Tg pig survived for 30 days, and live thymic epithelial cells and thymic bodies, including a few baboon lymphocytes, were discovered under the renal capsule and omentum of the baboons. The “thymus kidney” can induce the production of non-responsive donor-specific cells and stable amounts of anti-α-Gal antibodies, thus inducing immune tolerance across the genetic immune barrier (87). Transplanting GTKO pig kidneys with the vascular thymus into baboons significantly extended recipients’ survival (88). Recipient baboons with or without cortisol transplanted with “thymus kidneys” from GTKO micropigs survived for >80 days with no signs of cellular rejection or IgG deposition in the transplants and no loss of the transplanted kidneys, suggesting establishment of donor-specific T cell tolerance (89).

Fetal porcine thymus grafts containing mice thymic epithelial cells implanted into mice improved the development of T cells in the thymus, increased the likelihood that they would develop tolerance to the grafts, and reconstructed the T cell population (90). The method for preparing donor thymus grafts enriched with recipient thymic epithelial cells in large animals (cynomolgus monkeys and micropigs) was established (91). This should induce the tolerance of transplanted solid organs, including the kidneys (92).

Mouse T cell receptor–transgenic T cells can be functionally educated using porcine MHC antigens (93). Human T cells develop normally in porcine thymus grafts and form specific tolerance to porcine MHC in immunodeficient mice (94). However, a mouse with a transplanted porcine thymus would develop analogous autoimmune diseases, in which mouse CD4⁺ T cells play a key role (95). Therefore, the differentiation of host T precursor cells in the porcine thymus should differ from the normal physiological state. The number of Tregs in the athymic mice that were grafted with porcine thymus was close to normal, but the regulatory function was not (96). Moreover, T cell differentiation in humanized mice after bone marrow (BM) transplantation revealed that the positive selection was inadequate (97).

These findings should be helpful for thymus transplantation in large animals. Autologous thymus tissues were co-transplanted with GTKO porcine kidneys in the clinical trial of transplantation in two brain-dead patients (2). The results exceeded expectations; however, the mechanisms of tolerance induction need to be further explored.

Transplanted BM or hematopoietic cells can establish chimera-induced tolerance (98). Long-term survival has been achieved using kidneys co-transplanted with BM (99). Moreover, the role of CD4⁺CD25⁺FoxP3⁺Treg cells in these results cannot be ignored (100–102).

Simon et al. (103) injected large doses of porcine spleen cells into baboons and found that low-level chimera status was maintained for almost 1.5 years, during which the baboons did not get sick. These results suggested that donor leukocyte infusion can be used to induce peripheral tolerance during xenotransplantation. Perhaps infusing BM cells with differentiation potential would be more advantageous for establishing chimera-induced immune tolerance.

Griesemer et al. found that baboons transplanted with GTKO BM alone in vivo developed peripheral chimeras within 28 days, and the abundance of anti-GTKO porcine antibody or porcine-specific cytotoxicity did not increase. However, anti-porcine and other specific antibodies appeared 14 days after transplantation in baboons that were co-transplanted with BM cells and kidneys, and relatively high levels of anti-Gal antibodies were detected when the porcine kidney was rejected (104). These data suggested that BM infusion is associated with a loss of anti-Gal antibodies. To improve chimerism, the infusion method was modified, and the results were successful, the donor pig kidneys in the two groups survived for 47 and 60 days, respectively (105).

The cell- and species-specific CD47/Signal regulatory protein α (Sirp-α) signaling pathway might be involved in clearing cells derived from porcine BM cells in recipients. Porcine BM transferred the human CD47 gene survived much longer in a recipient baboon, and the chimeras prolonged the survival of porcine skin grafts (106).

Although CRISPR/Cas9 technology is widely applied, it has some limitations, such as off-target effects, low delivery efficiency, and the immune heritability of Cas9 protein. Any unexpected changes in the human (or xenograft) genome could result in serious and unintended consequences, including the activation of proto-oncogenes and production of new single nucleotide polymorphisms that can alter cellular behavior. In addition, >60% of the population harbors components of humoral and cellular immune responses to Cas9. Therefore, if sustained, Cas9 expression is required during treatment and the immune response induced by the Cas9 must be considered (108). Improvements in CRISPR/Cas9 technology will be conducive to the long-term outcome of clinical kidney xenotransplantation (109).