Reduction of the Major Swine Xenoantigen, the α-Galactosyl Epitope by Transfection of the α2,3-Sialyltransferase Gene*

α2,3-Sialyltransferase represents a putative enzyme that reduces the Galα1-3Gal β1-4GlcNAc-R (the α-galactosyl epitope) by intracellular competition with α1,3-galactosyltransferase for a common acceptor substrate. This study demonstrates that the overexpression of the α2,3-sialyltransferase gene suppresses the antigenicity of swine endothelial cells to human natural antibodies by 77% relative to control cells and by 30% relative to cells transfected with α1,2-fucosyltransferase, and in addition, it reduces the complement-mediated cell lysis by 75% compared with control cells and by 22% compared with cells transfected with α1,2-fucosyltransferase. The mechanism by which the α-galactosyl epitope was reduced was also studied. Suppression of α1,3-galactosyltransferase activity by 30–63% was observed in the transfectants with α2,3-sialyltransferase, and mRNA expression of the α1,3-galactosyltransferase gene was reduced as well. The data suggest that the α2,3-sialyltransferase effectively reduced the α-galactosyl epitope as well as or better than the α1,2-fucosyltransferase did and that the reduction of the α-galactosyl epitope is due not only to substrate competition but also to an overall reduction of endogenous α1,3-galactosyltransferase enzyme activity.

This study demonstrates that the overexpression of the ␣2,3-sialyltransferase gene suppresses the antigenicity of swine endothelial cells to human natural antibodies by 77% relative to control cells and by 30% relative to cells transfected with ␣1,2-fucosyltransferase, and in addition, it reduces the complement-mediated cell lysis by 75% compared with control cells and by 22% compared with cells transfected with ␣1,2-fucosyltransferase. The mechanism by which the ␣-galactosyl epitope was reduced was also studied. Suppression of ␣1,3-galactosyltransferase activity by 30 -63% was observed in the transfectants with ␣2,3-sialyltransferase, and mRNA expression of the ␣1,3-galactosyltransferase gene was reduced as well.
The data suggest that the ␣2,3-sialyltransferase effectively reduced the ␣-galactosyl epitope as well as or better than the ␣1,2-fucosyltransferase did and that the reduction of the ␣-galactosyl epitope is due not only to substrate competition but also to an overall reduction of endogenous ␣1,3-galactosyltransferase enzyme activity.
Xenotransplantation has been proposed as the most promising procedure to alleviate the world-wide shortage of organs for transplantation (1)(2)(3). However, in a vascularized organ, a xenograft between discordant species, such as swine to human (4), hyperacute rejection occurs within minutes due to the destruction of endothelial cells, a process that is mediated by fixation of natural antibodies and/or complement activation (5). Initial attempts to prevent hyperacute rejection have focused on modification of the donor xenograft via the expression of human complement regulatory proteins (6 -10). This strategy has proven to be very useful in prolonging swine grafts survival (11)(12)(13)(14).
The purpose of this study is to investigate whether the antigenicity of swine endothelial cells is affected by ␣2,3ST, an enzyme that catalyzes the sialylation of N-acetyllactosamine (Gal ␤1-4GluNAc ␤1-) to form ␣2,3-sialyl N-acetyllactosamine, the precursor of the sialyl Lewis X structure (35) and to assess the mechanisms by which the ␣-galactosyl epitope is reduced.
In the case of the sialyltransferase family, three types of sialyltransferases (26 -31, 36) are known to participate in the intracellular competition with ␣1,3GT for the common acceptor substrate. Our previous study, using ␤-D-mannoside ␤-1,4-Nacetylglucosaminyltransferase III, which catalyzes the ␤-1,4 addition of N-acetylglucosamine to the ␤-linked mannose of the trimannosyl core of N-linked sugar chains, demonstrated that N-linked sugars were the main source of xenoantigenicity (37)(38)(39)(40)(41). Therefore, in the present study, we employed ST3Gal III, which catalyzes the transfer of sialic acids to a lactosamine structure, mainly to N-linked terminal sugar residues. Technologies, Inc.) (9).
Construction of Plasmids-A cDNA of mouse ␣2,3ST (ST3Gal III) was subcloned into the site of pCXN2 (42), which is a ␤-actin promoter and a cytomegalovirus enhancer with a neomycin-resistant gene. The plasmid was separately transformed into Escherichia coli C600 and amplified using standard techniques (43). The cDNA of human ␣1,2FT, a gift from Dr. John B. Lowe (University of Michigan), was also subcloned into pCXN2.
Expression of cDNAs-The cDNAs (20 g) were introduced into MYP-30 by lipid-mediated DNA transfection with lipofectamine (Lipo-fectAMINE Reagent, Life Technologies, Inc.) (53). Transfected MYP-30 was maintained in complete medium for several days in an atmosphere of humidified 5% CO 2 at 37°C. The cells were then transferred to a complete medium containing 0.4 mg/ml G418 (Life Technologies, Inc.) for selection (9). Expression of plasmids was confirmed by high performance liquid chromatography as described below.
␣2,3ST activity was assayed in a reaction mixture that contained 0.1 M cacodylate buffer, pH 6.8, 0.01 M MnCl 2 , 0.45% Triton X-100, 10 mM CMP-sialic acid (Sigma). 10 l of 50 M substrate and 15 l of cell lysate were added to 25 l of this solution. The mixture was then incubated at 37°C for 3 h (27).
The assay for 1,2FT activity employed 50 mM potassium phosphate buffer, pH 6.1, containing 0.2% Triton X-100, 2 mM GDP-fucose, 25 mM phenyl-D-galactoside, and 10 mM ATP. To 25 l of this solution was added 10 l of 50 M substrate followed by 15 l of cell lysate. The assay mixture was then incubated at 37°C for 3 h.
␣1,3GT activity was assayed in a reaction mixture containing 10 mM HEPES, pH 7.2, 20 mM UDP-galactose, 10 mM MnCl 2 , 33 mM NaCl, 3 mM KCl. 10 l of 50 M substrate and 15 l of cell lysate were added to this mixture, and the mixture was then incubated at 37°C for 3 h.
The enzyme reactions were quenched by boiling for 5 min. The samples were then centrifuged at 12000 ϫ g for 5 min, and an aliquot of each supernatant was subjected to high performance liquid chromatography analysis, using a TSK-gel ODS-80TM column (4.6 ϫ 250 mm). The reaction products were eluted with 20 mM acetate buffer, pH 4.0, containing 0.01% 1-butanol at flow rate of 1.0 ml/min at 55°C and were monitored with a fluorescence spectrophotometer (Shimadzu, model RF-10AXL) using excitation and emission wavelengths of 320 and 400 nm, respectively. The specific activity of the enzyme is expressed as moles of product produced per hour of incubation per mg of protein.
Protein concentrations were determined with a BCA protein assay kit (Pierce) using bovine serum albumin as a standard.
Mixed Cell Lysates Assay-20 g of the plasmid that carried the cDNA of ␣2,3ST or ␣1,2FT was transfected into COS 7 cells individually, using an electroporation method (250 V/0.4 cm and 960 microfarads). The transfected COS 7 cells were incubated in Dulbecco's modified Eagle's medium for 2 days and harvested. After two washes with PBS followed by sonication, 15 l of each cell lysate from the ␣2,3STor ␣1,2FT-transfected COS 7 cells was mixed with the same volume of that from the parental SECs, and the ␣1,3GT activity of the mixed cell lysate was then assayed, using the same methods as described above. As a control, 15 l of PBS was mixed with the same volume of cell lysate from parental SECs, and the ␣1,3GT activity of this mixed cell lysate was also measured.
Flow Cytometry-Parental SECs and transfectants were incubated with various dilutions of normal human pooled serum (NHS) at 4°C for 1 h, washed, and then incubated with 1.25 mg of fluorescein isothiocyanate-conjugated anti-human Ig (Cappel) as a second antibody for 1 h at 4°C. Stained cells were analyzed with a FACSCalibure flow cytometer (Becton Dickinson). The direct fluorescence of cell-surface carbo-hydrate epitopes was also examined with an fluorescein isothiocyanateconjugated IB4 lectin (Honen Co. Ltd., Tokyo, Japan) that binds the ␣-galactosyl epitope.
Lactate Dehydrogenase (LDH) Assay-The assay was performed according to a modified version of the method of Korzeniewski and Callewaert (48), using a Kyokuto MTX LDH kit. The transfected cells were plated at 2 ϫ 10 4 cells/well in 96-well trays 1 day prior to assay. The next morning, the wells were washed twice in serum-free Dulbecco's modified Eagle's medium to remove the LDH that is present in fetal calf serum and incubated in several concentrations of NHS diluted with Dulbecco's modified Eagle's medium. The plates were incubated for 2 h at 37°C, and the released LDH was measured. The percentage cytotoxicity was calculated using the formula: where E is the experimentally observed release of LDH activity from the target SECs, N is the LDH activity in each concentration of NHS, S is the spontaneous release of LDH activity from target SECs incubated in the absence of NHS, and M is the maximal release of LDH activity, as determined by the addition of 2% Nonidet P-40 (49).
Western Blotting-Total cell lysates (3 g) from parental or transfected SECs were subjected to 12% SDS-PAGE under reducing conditions using the methods of Laemmli (50) and then transferred electrophoretically onto a nitrocellulose membrane (Schleicher & Schuell). The membrane was blocked in PBS containing 3% bovine serum albumin and incubated for 1 h with 0.2% NHS. After washing, the blots were incubated with horseradish peroxidase-avidin complex (Vector) and developed using an ECL detection system (Amersham Pharmacia Biotech).
Lectin Blotting-Parental and transfected cell products were also tested by lectin blot analysis, using Griffonia simplicifolia I (IB4). The cell lysate (3 g) was subjected to 12% SDS-PAGE under reducing conditions and then transferred electrophoretically onto a nitrocellulose membrane (Schleicher & Schuell). The blots were blocked in PBS containing 0.05% Tween 20 and 3% bovine serum albumin and incubated for 1 h with biotinylated 10 g/ml IB4.
Northern Blotting-Total RNA was isolated from parental SECs and transfectants with TORI-ZOL (Biotec Laboratories, Inc., Houston, TX) and separated by electrophoresis (20 g/lane). The probe used for hybridyzation was the PCR product that was generated from a swine cDNA library (Porcine Liver 5Ј-STRETCH cDNA Library; CLON-TECH), using the primers 5Ј-AAGACCATCGGGGAGCACAT-3Ј and 5Ј-GGAGAAGTAGCCAGAGTAAT-3Ј (51,52), and labeled by a Multi DNA labeling kit (Amersham Pharmacia Biotech). Hybridization signals on the Northern blots were evaluated by scanning the imaging plates with a Bioimage Analyzer (Fuji Photo Film Co., Ltd.).

RESULTS
␣2,3ST and ␣1,2FT Activity of Parental SECs and Transfectants-Four positive stable clones of ␣2,3ST and several positive clones of ␣1,2FT were established. Of the ␣1,2FT transfectants, one had moderate activity, whereas the other had a higher activity. These were used as controls in comparison with the ␣2,3ST transfectants. The enzyme activities of these ␣2,3ST and ␣1,2FT transfectants are shown in Table I  with human natural antibodies in NHS, the ␣2,3STor ␣1,2FTtransfected SECs had a decreased reactivity. The percentage of reduction of xenoantigenicity to human natural antibodies was 42-77% in the ␣2,3ST transfectants and 32-66% in the ␣1,2FT transfectants, as judged by mean fluorescence intensity (Fig. 1). The ␣-galactosyl epitope, as judged by IB4 lectin binding was approximately 27-81% down-regulated in the ␣2,3ST transfectants and 5-61% down-regulated in the ␣1,2FT transfectants. ␣2,3ST appeared to be more effective in the down-regulation of xenoantigenicity to human natural antibodies and to the ␣-galactosyl epitope than ␣1,2FT (Fig. 2).
From the results of SEC transfectants, B-11,␣1,2FT was very effective in reducing the xenoepitope of SECs. However, ␣2,3ST was also quite effective, and it appeared to be more efficient in down-regulating the xenoepitope of SECs. An inhibition of over 70% of cytotoxicity was observed in the ␣2,3ST transfectants A-4, B-11, and A-2 (Fig. 3B).
Western and Lectin Blotting-Western and lectin blotting were performed in order to analyze the alterations of reactivity to human natural antibodies and the IB4 in ␣2,3ST transfectants. Evaluation of these blot profiles revealed that proteins with molecular masses under 66 kDa derived from the ␣2,3ST transfectants had reduced reactivity to NHS, especially IgG, compared with the parental SECs. However, no detectable differences were observed between the ␣2,3ST and ␣1,2FT transfectants.
Similar to Western blotting patterns, proteins with molecular masses under 66 kDa in both transfected cells reduced reactivity to IB4, as evidenced by lectin blotting (Fig. 4).
␣1,3GT Activities of Parental SECs and Transfectants-To assess the influence of the transfected gene on intrinsic ␣1,3GT and ␤1,4GT activity, enzyme activities were measured. Parental SECs and the mock samples contained relatively high ␣1,3GT and ␤1,4GT activities. Whereas the ␣2,3ST and ␣1,2FT transfectants had an elevated ␤1,4GT activity, the intrinsic activity of the ␣1,3GT was clearly present in both the ␣2,3ST and ␣1,2FT transfectants (Table II). The degree of down-regulation of ␣1,3GT activity in each transfectant was less pro-nounced than that of antigenicity to NHS and to the complement mediated lysis.
The Influence of an Excess of Other Transferases on ␣1,3GT Measurement in SECs-To ascertain whether or not overexpressed ␣2,3ST or ␣1,2FT down-regulates the ␣1,3GT value of each transfected SECs, the lysates of parental SECs were mixed with several COS 7 cell lysates that contained excess ␣2,3ST or ␣1,2FT activities, and ␣1,3GT enzymatic activities were then estimated and compared with those in parental SECs with PBS.
Initially, the enzymatic activities of ␣1,3GT, ␣2,3ST, and ␣1,2FT in parental SECs, parental COS, and COS transfectants with ␣2,3ST or ␣1,2FT were determined, and the results are shown in Table III. The ␣1,3GT enzymatic activities of parental SECs with PBS, parental COS, or COS transfectants were then determined. The results clearly show that the ␣1,3GT activity in this assay system is not affected by other glycosyltransferases, which are capable of competing with ␣1,3GT for the same substrate (Table III).
mRNA Expression of ␣1,3GT in the Transfectants (Northern Blotting)-Northern blot analysis was performed in order to determine the alterations in mRNA production. The introduction of the ␣2,3ST and ␣1,2FT genes into SECs decreased the amount of ␣1,3GT mRNA in each transfectant (Fig. 5). A 30 -40% suppression of ␣1,3GT mRNA production was observed in the high expression clones A-2, A-4, B-11, and B-11FT. These data parallel the decreased ␣1,3GT enzyme activity.

DISCUSSION
The amelioration of antigenicity of SECs by overexpression of ␣2,3ST was examined as a possible approach to provide a permanent solution for overcoming hyperacute rejection in clinical xenotransplantation.
As expected, the down-regulation of xenoantigenicity was quite obvious for the case of the ␣2,3ST transfectants as judged by flow cytometric analysis and a cytotoxicity assay. However, no measurable differences between ␣2,3ST and ␣1,2FT transfectants were observed. Therefore, Western and lectin blot analysis of the ␣2,3ST transfectants was carried out. However, both ␣2,3ST and ␣1,2FT also reduced the the extent of reactivity to human IgG and IgM, and they reduced the reactivity to ␣-galactosyl epitopes in nearly the same manner (Fig. 4). Interestingly, the changes in the bands for these transfectants were more evident for IgG and IB4 lectin than for IgM blotting. This is consistent with data reported by Galili et al. (15)(16)(17). Nearly the same results were obtained for the ␣1,2FT transfectants.
It might be generally thought that ␣1,2FT is effective in modulating N-linked sugars as well as O-linked sugars and glycolipids, whereas ␣2,3ST;ST3GalIII mainly affects N-linked sugars (31). However, at the present time, there are no specific data to support this hypothesis. Therefore, it is possible that ␣2,3ST;ST3GalIII might have affected O-linked sugars and glycolipids, as well as N-linked sugars, on the SECs. On the contrary, the results in this study are consistent with findings in our previous study of SEC transfectantants with ␤-D-mannoside ␤-1,4-N-acetylglucosaminyltransferase III, which acts on an N-linked sugar (41). A possible explanation for this is that the ␣2,3ST;ST3GalIII and ␣1,2FT gene acts largely on the same xenoepitopes of N-linked sugars.
In addition, regarding ␣1,2FT-transfected SECs, Sepp et al.  Parental SECs and transfectant cell membrane glycoproteins were separated by electrophoresis in a 12% SDS-polyacrylamide gel and transferred onto a nitrocellulose membrane. The membrane was incubated for 1 h with 10% NHS and stained. After washing, the blots were incubated with horseradish peroxidase-avidin complex and developed using an ECL detection system. A typical Western blot pattern of IgG and IgM is shown. Lectin blot analysis of whole-cell lysates from the parental SECs and transfectants was also performed. Cell lysates from SECs, ␣2,3ST, and ␣1,2FT transfectants were separated by SDS-PAGE and transferred onto a nitrocellulose membrane. The blots were probed with biotinylated IB4 lectin.  (32) reported only a 40 -50% decrease in the expression of the ␣-galactosyl epitopes in their PLECT cell model, which contained ␣1,2FT. However, Sandrin et al. (33) reported an approximately 70% decrease in the ␣-galactosyl epitopes of LLC-PK1 in their study. In this study, the data relative to ␣1,2FTtransfected SECs are in good agreement with the latter study. It is generally accepted that transfection of ␣1,2FT or ␣2,3ST decreases the ␣-galactosyl epitope as the result of the intracellular enzymatic competition for a common acceptor substrate with ␣1,3GT. In this study, we also analyzed the mechanism by which the ␣-galactosyl epitope is reduced. As expected, the data in this study clearly showed that the ␣-galactosyl epitopes were reduced in both the ␣1,2FT and ␣2,3ST transfectants. However, to our surprise, the data provided evidence that the enzymatic activities of ␣1,3GT in the transfectants were remarkably low as compared to those of parental SECs and mock controls (Table II).
In each enzymatic activity assay in this study, excess substrate was used to avoid the effect of artifacts resulting from overexpressed glycosyltransferases. The reliability of each value for enzymatic activity was confirmed by the fact that the value in enzymatic assay of ␣1,3GT was not affected by adding a high dose of exogenous ␣1,2FT and ␣2,3ST (Table III).
Furthermore, Northern blotting was carried out to verify that this reduction of ␣1,3GT activity was the result of mRNA production. The results relative to the mRNA levels of ␣1,3GT correspond to those of ␣1,3GT enzymatic activity in each transfectant. However, the rate of down-regulation of ␣1,3GT activity in each transfectant was milder than that of antigenicity to NHS and changes in the complement dependent cell lysis. Therefore, the alteration of the ␣-galactosyl epitopes cannot be explained solely by the the down-regulation of mRNA levels of ␣1,3GT. We therefore conclude that transfections of the ␣1,2FT and ␣2,3ST genes into SECs reduced the ␣-galactosyl epitopes as the result of both intracellular enzymatic competition and the reduction of intrinsic activity of ␣1,3GT due to the low expression of mRNA.
Regarding the ␤1,4GT activity, the up-regulation of the enzyme activities of these transfectants is shown in Table II. However, because the gene code of this molecule in swine has not yet been reported, it is impossible to verify the mRNA levels of this enzyme. To understand the mechanism will require more detailed investigations into the influence of, for example, ␣1,3GT and ␤1,4GT activities, message formation, gene transcription, and so forth, by the overexpression of other glycosyltransferase genes.