The α-Gal epitope has a special terminal carbohydrate structure in the form of Gala1, 3Galb1-4GlcNAc-R, which is confirmed by the study of two structures of the major glycolipids in rabbit red cell membranes: Ceramide trihexoside (Galα1-4Galα1-4Glc-Cer) and ceramide pentahexoside (Galα1-3Gaα1-4GlcNAcα1-3Galα1-4Glc-Cer) (3). The amount of α-Gal epitope exhibits significant differences in different organs. High concentrations of the α-Gal epitope were observed in all the hepatic cells. The largest quantity of α-Gal epitope was observed in the proximal convoluted tubules in the kidney and the α-Gal epitope were presented in smaller quantities in the distal tubules and in the glomeruli. No α-Gal epitope was identified in the collecting ducts. In addition, the α-Gal epitope is expressed on numerous molecules on platelets and it is predominantly found on fibrinogen, von Willebrand's factor, α2 integrin and β3 integrin. On endothelial cells, >20 glycoproteins carry α-Gal epitope, some of which have been identified as von Willebrand's factor, DM-GRASP, and the α1, αv, α3/α5, β1, β3 integrins (25). The expression of the α-Gal epitope varies in different species; for example, it is expressed on the red cell membrane glycolipids of rabbits and cows, the cell surface glycoproteins of mouse Ehrlich ascites cells, lymphoma cells and fibroblasts, on mouse laminin and on bovine, canine and porcine thyroglobulin.

The α-Gal epitope is synthesized by the transferase α1,3GT, and similar to numerous other glycosylating enzymes, the α1,3GT resides within the Golgi of the cell (5). The α1,3GT is believed to be the only enzyme to synthesize Gal(1,3)Gal. However, the rat iGb3 synthase is also capable of synthesizing Gal(1,3)Gal as the glycolipid structure iGb3, showing that the rat expresses two distinct α1,3GT (26), α1,3GT and iGb3 synthase. Therefore, in rats there exists two separate glycosylation pathways to synthesize Gal(1,3)Gal. The α1,3GT is one of the α1,3GTs, and this enzyme is active in the Golgi apparatus of cells and transfers galactose from the sugar-donor UDP-gal to the acceptor N-acetyllactosamine residue (Galα1-4GlcNAc-R) on carbohydrate chains of glycolipids and glycoproteins, to form the α-Gal epitope (27).

Additionally, the distributions of the α1,3GT gene have a significant difference in various species (28). The α1,3GT gene of murine is composed of nine exon sequences, and only six exons: Exons 4–9 (based on the numbering of mouse α1,3GT), are translated to α1,3GT mRNA, which is ~1,500 base pairs (bp) (GenBank: M26925.1) (5). In Sus scrofa (pig), the α1,3GT gene resides on chromosome 1, and α1,3-GT mRNA is 1,269 bp (GenBank: L36152.1) (8). The α1,3GT mRNA completed cds for bovine are 1,828 bp (GenBank: J04989.1) (29). The coding sequence of bovine α1,3GT predicates a membrane-bound protein with three distinct structural features: A large, potentially glycosylated COOH-terminal domain (346 amino acids), a single transmembrane domain (16 amino acids), and a short NHz-terminal domain (6 amino acids) (29). By the analysis of the panel of human-rodent somatic cell hybrids, the nonfunctional, processed pseudogene and the human homologue gene represented by HGT-10 have been established to be located on human chromosomes 12 and 9 (GenBank: J05421.1) (6,30). The locus of α1,3GT mRNA for certain primates are about the same locus as those of Gorilla gorilla (GenBank: M73304.1), Pongo pygmaeus (GenBank: M73305.1) and Macaca mulatta (GenBank: M73306.1) (31).

The expression levels of the α-Gal epitope were assessed by measuring the binding of the anti-Gal antibody with the α-Gal epitope-specific lectin Bandeiraea simplicifolia IB4 in various cells. The distributions of α-Gal epitope, anti-Gal antibody and the activities of α1,3GT in various species have a significant difference. As shown in Table I, the α-Gal epitope is abundant on red blood cells and nucleated cells of non-primate mammals, prosimians and New World monkeys (such as marmoset, squirrel monkey and spider monkey); however, it is absent in ancestral primates such as humans, apes and Old World monkeys. Due to the continual exposure of α-Gal epitope on the surface of entric bacteria and other pathogens, and the lack of α-Gal epitope on humans and Old World primates, they are not immunotolerant to the α-Gal epitope, and therefore produce anti-α-Gal epitope antibody throughout the whole life (32,33). As many as 1% of circulating B cells are capable of generating the anti-α-Gal epitope antibody, which is analogous to the mechanism by which the human ABO blood group antibodies are produced. Anti-Gal antibody is not generated at birth [except maternal-derived immunoglobulin G (IgG)]; however, it is developed during the first few months of life. The anti-Gal antibody isotypes of IgM, IgG, and IgA have been identified (34–37), and the anti-Gal antibody in humans is encoded by several heavy-chain genes, which are primarily within the VH3 immunoglobulin gene family (38). When anti-Gal antibody specifically binds with the α-Gal epitope, the binding site of anti-Gal antibody resembles a 'pocket' for the α-Gal epitope with a size corresponding with the free trisaccharide Gala1-3Galb1-4GlcNAc. The affinity of the anti-Gal antibody to the free trisaccharide Gala1-3Galb1-4GlcNAc is ~7-fold higher compared to the disaccharide Gal1,3Gal. The similarity in binding of anti-Gal antibody to α-Gal epitope on carbohydrate chains of glycoproteins (mannose in the fourth position) and glycolipids (no mannose in the carbohydrate chains) further implies that the pocket size is limited to a trisaccharide length of the antigen.

According to the evolutionary tree of mammalian species (Fig. 1), it is clear that α1,3GT and α-Gal epitope have been conserved in New World monkeys (evolving on the South American continent, which was separated from the African continent for >35 million years) and in lemurs (prosimians residing in the Madagascar, a land mass isolated from the African continent for >60 million years), which suggests that α1,3GT appeared early in mammalian evolution, and prior to the divergence of marsupials and placental mammals (>125 million years ago) (39). However, following the geographical separation between the South American and African continents, ancestral Old World primates were subjected to a selective evolutionary pressure, which resulted in the inactivation of the α1,3GT gene. Northern blot results reveal that no transcripts of the α1,3GT gene were identified in the higher primate species. Notably, a variety of full-length transcripts were determined by sensitive polymerase chain reaction in the tissues of rhesus monkeys, orangutans and humans. Five crucial mutations were delineated in the coding region of the human and rhesus, and three crucial mutations were identified in the orangutan, any one of which could be responsible for inactivation of the α1,3GT gene. Therefore, the α1,3GT gene exhibits a unique evolutionary characteristics.

Starting from the Old World primates, the α-Gal epitope has been interrupted during the course of evolution. Partial sequences of the α1,3GT gene, which governs the synthesis of the α-Gal epitope, were detected in the human genome and were found to be the corresponding pseudogenes. An experiment was performed using the bovine α1,3GT cDNA as a probe to isolate two non-overlapping clones (HGT-2 and HGT-10) from a human genomic DNA library (40). HGT-2 contains a putative coding region, containing multiple frame shift mutations and nonsense codons in all three reading frames, which precludes the synthesis of a functional enzyme. The coding sequences of HGT-10, which is located on human chromosomes 12 and 9, contain nonfunctional and processed pseudogene. Comparison of the nucleotide sequence of the human α1,3GT pseudogene with the corresponding different species (Fig. 2), the evolutionary loss of α-Gal epitope attributes to point mutations in the coding region of the α1,3GT gene. The human α1,3GT pseudogene resides on chromosome 9, and its sequence was found to contain two point mutations, which results in a frame shift mutation and a premature stop codon (13). Sequencing of this exon in the chimpanzee shows the same two mutations, whereas the sequence from gorilla only includes the first frame-shift mutation, which generates a premature stop codon. The common mutations are identified among the α1,3GT sequences of Old World monkey. These findings in various primates suggest that the evolutionary inactivation in ancestral Old World primates following the divergence of apes and monkeys is estimated to have occurred ~20–25 million years ago. The evolutionary inactivation of the α1,3GT gene resulted in the loss of immune tolerance to the α-Gal epitope and the appearance of the anti-Gal antibody. Once the α-Gal epitope was eliminated, primates could produce the anti-Gal antibody, possibly by means of defending against pathogens, which express the α-Gal epitope. Therefore, the inactivation of α1,3GT gene is generated by several deletions, which ultimately encodes premature stop codons. The observed inactivation of primate α1,3GT, along with distinct geographical boundaries, suggests that it was due to an infectious disease, which is endemic to the connected continents of the Old World.

Xenotransplantation is one of the most important methods to solve the problem of the organ shortage in humans. Those cells, tissues or organs from different species are known as xenografts, which are widely used for xenotransplantation (44). For example, the heart valve substitutes and the patch materials are applied to a variety of cardiovascular diseases (45). However, the major barrier to the clinically successful xenotrans-plantation is the lack of effective therapies aimed at eliminating antibody and complement. Therefore, subsequent to knocking out the α1,3GT gene, the α-Gal epitope cannot be produced, and hyperacute rejection can be overcome (46).

As the whole organs of the heart is not the ideal choice for heart xenotransplantation, heart valve substitutes and patch materials are generally applied to a variety of cardiovascular disease (45), and the major application is the bioprosthetic heart valves (BHVs) (47,48). Currently, the most commercially available BHVs are made from porcine aortic valve or bovine pericardium (49). They are conventionally cross-linked with glutaraldehyde (GA) to impart tissue stability, to reduce antigenicity and to maintain tissue sterility (50).

However, GA-fixed xenografts are prone to dystrophic calcification after long-term implantation in humans, which is one of the limiting factors affecting the longevity of bioprostheses (47). The mechanism of the calcification of GA-fixed xenografts is complex and is not completely understood. It is supposed that the degeneration of bioprostheses may begin with the penetration of immunoglobulines into the valve-matrix. Subsequently, macrophages are deposited on the valve surface and erythrocytes penetrate into the valve. Finally, collagen breakdown and calcification occur. Thus, the deposition of immunoglobulins represents the initial immunological trigger for calcification in bioprostheses (50). Another proposed mechanism responsible for the calcification of xenografts is an immune response to the xenografts (51). According to the method of anti-calcification treatments, bovine pericardial tissues were divided into eight groups. Prepared tissues were subcutaneously implanted into the α-gal KO mice and wild-type mice. The result demonstrated that except the high concentration GA-treated group, calcium contents of anticalcification-treated groups were significantly lower compared to the control group (45). These findings suggest a possible role of immune response in the calcification of xenografts, and α-Gal epitope on the BHV contributes to the immune response (52–54). The α-Gal epitope on BHV may be removed completely using recombinant α-galactosidase and by performing a decellularization procedure without modifying the mechanical properties of the tissues (25,55,56). There is no difference in the mechanical properties of BHV following the application of the methods (57,58). In addition, a series of 7 heterotopic heart transplantations were performed from cluster of differentiation 46-positive transgenic pigs to baboons using a combination of therapeutic agents mainly targeted at controlling the synthesis of anti-pig antibodies. This study reported that the longest median survival of pig hearts transplanted into baboons was 96 days (59). However, there also are numerous problems regarding BHV transplantation, in pig-to-primate or human BHV. Although hyperacute rejection can be prevented, the next rejection phenomenon unique to xenotransplantation, known as delayed xenograft rejection or acute vascular rejection, will occur (60). In addition, new immunological problems arise, for example, except the α-Gal epitope, there exist non-gal carbohydrate antigens and the identification of their elusive antigens as neutrophils, macrophages, natural killer cells and T cells (61).

Studies on the xenotransplantation of heart and kidney from α1,3GT KO pigs to monkeys demonstrate that pig hearts could live for ≤6 months, pig kidneys for 8–32 days, and pig kidneys containing vascularized thymic tissue for up to 83 days (62). Therefore, transplantation of organs from α1,3GT KO pigs to humans becomes clinically acceptable. However, in addition to the α-Gal epitope, the non-gal carbohydrate antigen is another obstacle in kidney transplantation. A study of total N-glycans from membrane glycoproteins, derived from specific pathogen-free miniature pig kidneys, was carried out, and the result showed that α-gal was the most abundant compound among the neutral pig N-glycans. However, the various N-glycan structures of Neu5Gc and Gala1-3Lex epitope as new non-gal carbohydrate antigens were identified by mass spectrometry (46,63). Therefore, non-gal carbohydrate antigens may be the next complication in the transplantation from animals to humans.

Tumor cells express a variety of tumor-specific antigens in animals and in humans. Some of the tumor antigens are common to numerous types of tumors. Additionally, as certain antigens may also be expressed on normal cells, the immune response caused by these common tumor antigens is usually extremely low (64). However, tumor antigens are specific to the tumor type and to the patient. Therefore, the autologous tumor cells could be considered a suitable source for the production of the vaccine. In order to induce the expression of α-Gal epitope by the tumor cells, chemical, biochemical and genetic engineering can be applied (8). As an example of the chemical and biochemical methods, autologous intact tumor cells from hematological malignancies, or autologous tumor cell membranes from solid tumors are processed to express α-Gal epitope by the incubation with neuraminidase, recombinant α1,3GT and uridine diphosphate galactose (9). Additionally, the expression of the α-Gal epitope can also be achieved by the transduction of recombinant adenovirus pAd-sGT, expressing pig α1,3GT genes, into the tumor cells (65).

The adenoviral vector-mediated transduction of the pig α1,3GT gene into human tumor cells, such as malignant melanoma A375, stomach cancer SGC-7901, and lung cancer SPC-A-1, has been reported. These studies demonstrated that α-Gal epitope did not increase the sensitivity of human tumor cells to human complement-mediated lysis (66). Studies on the immune response to tumor vaccines have indicated that, to achieve effective activation of tumor-specific naive T cells, the vaccinating antigens must be effectively taken up by antigen-presenting cells (APCs). Subsequently, the vaccinating antigens should be transported from the vaccinating tumor membranes to the regional lymph nodes, and the tumor-associated antigen (TAA) peptides will be processed and presented by APCs (67–69). The autologous TAA peptides activate tumor-specific T cells within the lymph nodes, eliciting an immune response, which would eradicate the residual tumor cells following completion of standard therapy (70).

In addition, the tumor cells can also express the α-Gal epitope by the transfection of α1,3GT genes through gene engineering. Following genetic manipulation, the binding ability of the tumor cells with human IgG and IgM is increased. However, the sensitivity of human tumor cells to human complement-mediated lysis is not enhanced. In a cancer-related study, studies on the immune response of tumor vaccines have indicated that a major prerequisite for the formation of tumor vaccines is their effective uptake by APCs and the transportation of these APCs to the regional lymph nodes where TAAs can be processed and presented (21,71). In detail, TAAs activate the immune response, and the anti-Gal antibody would specifically bind to α-Gal epitope. The Fc portion of the anti-Gal antibody binds with the Fcα receptors on APCs and induce the effective uptake of the vaccine by the APCs (Fig. 4) (72). Subsequent to transporting the vaccinating tumor membranes to the regional lymph nodes, TAA peptides can be processed and presented. The autologous TAA peptides activate tumor-specific T cells within the lymph nodes, which causes an immune response and it will eradicate the residual tumor cells that remain following the completion of standard therapy (73).

With regard to the application of tumor vaccines, MUC1 was overexpressed and aberrantly glycosylated in the majority of human ductal pancreatic carcinoma (74). The α1,3GT KO mice used as the model were pre-immunized with pig kidney and transplanted with B16F10 melanoma cells, and were transfected with the MUC1 expressing vector. Compared with the mice vaccinated with MUC1 alone, the tumor growth was significantly inhibited and the survival time was significantly improved in mice vaccinated with pancreatic cancer cells expressing α-Gal epitope. Based on this study, it will be a novel strategy for the treatment of pancreatic cancer (75,76). Another study reported the treatment of 11 patients with advanced solid tumors with intratumoral injection of 0.1, 1 or 10 mg α-gal glycolipids. There was no dose-limiting toxicity and no clinical or laboratory evidence of autoimmunity, or any other toxicity. Few patients had an unexpectedly long survival. Therefore, it is feasible and safe for inducing a protective antitumor immune response (77). The mice vaccinated with irradiated melanoma cells, which express a-Gal epitope, were followed by a challenge with live parental melanoma cells, and resulted in protection with vaccines for ≥2 months in one-third of the mice. The proportion of survival mice protected with the vaccines doubled when they were immunized twice with irradiated melanoma cells expressing α-Gal epitope, and challenged with the life span of those of the control mice. A further histological study on the developing tumors in the challenged mice, which were immunized with melanoma cells expresssing α-Gal epitope, demonstrated the extensive infiltration of T lymphocytes and macrophages (73). Chiefly, anti-gal can be applied to the cancer immunotherapy, and breast cancer and brain tumors may be treated by autologous tumor vaccines in the future.

Viral vaccines include the flu and HIV vaccines. Similar to the tumor vaccine, targeting the anti-gal-mediated complex to the APCs is essential (78). The Flu vaccine is mainly comprised of the viral envelope glycoprotein 'hemagglutinin' (HA), which may be served as a useful example for the application of anti-Gal into vaccines. In the majority of strains of flu virus, HA has 6–8 N-linked carbohydrate chains, the majority of which are of the complex type. This study demonstrates a method for the expression of α-Gal epitope on the influenza virus HA by recombinant α1,3GT (79). To generate the flu vaccine expressing α-Gal epitope, sialic acid is removed from viral glycoproteins following its digestion by viral neuraminidase, and the terminal of the carbohydrate chains of HA was residual N-acetyllactosamine. Subsequently, α1,3GT transfers sugar from the sugar donor UDP-gal to the residual N-acetyllactosamine (64). Therefore, α-Gal epitope could be added to all carbohydrate chains of HA. In the study of the α1,3GT KO mice, the control group of the KO mice were immunized with the original inactivated flu virus lacking α-Gal epitope, and another group of the KO mice were immunized with 1 µg of inactivated flu virus vaccine expressing the α-Gal epitope. The result demonstrated that the anti-flu virus antibody and T cell response in the group that was immunized with the virus vaccine expressing α-Gal epitope was ~100-fold higher compared to the control group. Furthermore, the survival time of KO mice with the vaccine expressing α-Gal epitope infected with live virus was ~9-fold longer compared to the KO mice immunized with the vaccine lacking this epitope (21). Therefore, the immunogenicity of flu vaccines is increased by targeting anti-Gal antibody mediated complex into APCs.

In addition, the HIV vaccine is another notable viral vaccine. The main glycoprotein on the surface of HIV is gp120, which has 24 N-linked glycosylation sites of the carbohydrate chains, and the terminal structure of the gp120 carbohydrate chains is SA-Galα1-4GlcNAc-R. Following the incubation with neuraminidase, UDP-gal and α1,3GT, gp120 can be capped with the α-Gal epitope. Studying the HIV vaccine on KO mice, it demonstrated that the anti-gp120 antibody and T cell responses were increased ~100-fold following the immunization of gp120 with α-Gal epitope (80).

Chiefly, the immunogenicity of the flu and HIV vaccines is increased by expressing the α-Gal epitope. This method can be used for any microbial, soluble or particulate vaccine. However, the recombination does not affect the immunogenic epitopes of the vaccine.

The interaction between the anti-Gal antibody and α-Gal epitope activates complement system, which induces rapid recruitment of macrophages. Wound healing requires the help of the local recruitment and the activation of macrophages. The activated macrophages generate cytokines, such as vascular endothelial growth factor, interleukin-1, platelet-derived growth factor and colony-stimulating factor. These cytokines promote the healing of injuries. However, if the α-Gal epitope was expressed on cells with inserted α-gal glycolipids, the binding of anti-Gal antibody with α-Gal epitope may lead to cell lysis and a pro-inflammatory response. To avoid this pro-inflammatory response, the α-Gal epitope is expressed on cells in the form of nanoparticles and subsequently there will be no insertion into the cells (38). These nanoparticles were generated from phospholipids, cholesterol and α-gal glycolipids. An experiment regarding full-thickness wounds (20×20 mm) with tattooed borders was carried out on the back of α1,3GT KO pigs. Treated wounds with α-gal nanoparticles exhibited more macrophages compared to the control wounds (saline-treated) in the same pig and at the same time, and it was observed that angiogenesis was increased in the wounds with α-gal nanoparticles. No keloid formation or increase in scar formation was observed in pigs treated by α-gal nanoparticles (81,82). Another study demonstrated that in contrast to the wild-type mice, the healing time of excisional skin wounds treated with α-gal liposomes in α1,3GT KO mice was twice as fast as that of control wounds. Additionally, the time of scar formation in α-gal liposome-treated wounds is much less compared to those wounds with physiological healing (83).

In addition, the anti-Gal antibody can be used on the skin burns healing as well. Efficacy of treatment in accelerating burn healing is demonstrated in pairs of burns in α1,3GT KO mice. It has been demonstrated that 3 days later, the α-gal liposomes treated burns contained 5-fold as many neutrophils as those of the control burns, and macrophages were found only in α-gal liposome-treated burns. Six days later, 50–100% of the surface area of the α-gal liposome-treated burns were covered with regenerating epidermis (re-epithelialization), and almost no epidermis was identified in the control burns (84).

Therefore, the extensive recruitment of macrophages by α-gal nanoparticles that have interacted with anti-Gal antibody will effectively accelerate wound healing, skin burns and tissue regeneration.

In the same way of accelerating wound healing, skin burns healing and tissue regeneration, α-gal nanoparticles may also be applied to the regeneration on ischemic myocardium and injured nerves. Combined with the regeneration on ischemic myocardium and injured nerves, the mechanisms of these treatments have numerous similarities. Firstly, the mechanism of regeneration of ischemic myocardium and injured nerves requires the recruitment of macrophages, which will secrete cytokines/growth factors. One of the differences between them is that the regeneration on ischemic myocardium requires the recruitment of stem cells, which received cues from the adjacent healthy cells, the microenvironment and the extracellular matrix to differentiate into cardiomyocytes, which regenerate the tissue and restore its physiological activity. However, the regeneration of nerves requires re-growth of multiple sprouts from injured axons. These sprouts will reconnect across the lesion and grow into the distal axonal tube of the damaged neurons. In addition, they have a default mechanism for tissue repair simultaneously. When the recruitment of macrophages is delayed, leading to no important cytokines/growth factors regeneration, the useful technique would be removal, which is irreversible.

The treatments for regeneration of ischemic myocardium and injured nerves by α-gal nanoparticles are theoretical possibilities at present, and the experiment is expectantly being carried out on mammalian animal models (38).