INTRODUCTION

Gal α(1, 3) Gal (gal epitope) is a carbohydrate epitope, which is produced in large amount on the cells of pigs, mice and New World monkey (monkey of South America) by the glycosylation enzyme Gal β1,4GlcNAc3-α-D-galactosyltransferase [α(1, 3)GT; EC2.4.1.51]1. This enzyme is active in the Golgi apparatus of cells and transfers galactose from the sugar-donor uridine diphosphate galactose (UDP-galactose) to the acceptor N-acetyllactosamine residue (Gal β1-4GlcNAc-R ) on carbohydrate chains of glycolipids and glycoproteins, to form gal epitope (Gala1, 3Galb1-4GlcNAc-R) 1.

Gal epitope is completely absent in humans, apes, and Old World monkeys (monkey of Asia and Africa) because their genes encoding α(1, 3)GT have become inactivated in the course of evolution2,3. Since humans and Old World primates lack the gal epitope, they are not immunotolerant to it and produce anti-gal epitope antibodies (anti-Gal) throughout life in response to antigenic stimulation by gastrointestinal bacteria4 and as many as 1% of circulating B cells are capable of producing this antibody5.

Anti-Gal forms a major immunological obstacle for xenotransplantation of nonprimate mammalian vascularized organs, such as porcine organs, into human recipients6. The binding of anti-Gal to gal epitopes expressed on glycolipids and glycoproteins on the surface of endothelial cells in donor organs leads to activation of the complement cascade and hyperacute rejection, and also plays important role in occurrence of complement-independent delayed xenograft rejection6. Anti-Gal also prevents the use of retroviral vectors propagated in mouse cells for gene therapy, because natural anti-Gal binds to gal epitopes on the viral envelope glycoproteins and induces the destruction of these viruses7.

In cancer research, studies on the immune response of tumor vaccines have indicated that a major prerequisite for the success of tumor vaccines is their effective uptake by antigen-presenting cells (APCs) and transportation of these APCs to the draining lymph nodes, where the processed and presented tumor-associated antigens (TAAs) activate tumor-specific naive T cells8. Galili et al9 proposed that the immunogenicity of autologous tumor vaccines in human maybe augmented by engineering vaccinated tumor cells to express gal epitopes on membranes. Subsequent in situ binding of natural anti-Gal IgG molecules to these epitopes would result in the formation of immune complexes that target tumor vaccines for uptake by APCs via the interaction of the Fc portion of anti-Gal with Fcgreceptors on APCs. This hypothesis has been tested in vivo as well as in vitro10,11.

At present, a number of studies have suggested that human recombinant adenoviruses (rAdvs) are efficient vectors for the delivery of cloned genes to a variety of tumor cells and tissues12. Therefore, in the current study, a recombinant replication-defective adenovirus that expresses the pig α(1, 3) GT gene was constructed. We evaluated the expression of gal epitope on the surfaces of several human tumor cell lines after adenoviral vector transduction.

MATERIALS AND METHODS

Cell culture

All cell lines were from Type Culture Collection of Chinese Academy of Sciences (Shanghai, China). The 293 cells (Immortalized human embryonic kidney cells) and COS cells were cultured in Dulbecco's modified Eagle medium (DMEM; Gibco.BRL) supplemented with 10% heat-inactivated newborn bovine serum (NBS) in a humidified incubator with 5% CO2 and 95% air at 37 °C. Human lung cancer cells (SPC-A-1), melanoma cells (A375), and stomach cancer cells (SGC-7901) were maintained in RPMI-1640 (Gibco.BRL) supplemented with 10% NBS. 293 cells were used for recombinant adenovirus transfection, amplification, and titration.

Construction of recombinant adenoviral vector expressing pig α(1, 3)GT gene

Plasmid pcDNA3-sGT110013 was digested by Xho I and blunted with Klenow enzyme, and then digested by BamH I. the resulting 1.19 kb fragment of pig α(1, 3)GT cDNA was subcloned into BamH I-EcoR V site in pBluescript II KS+/− to produce pKS-sGT, and then subcloned into the Hind III-Not I site in pAdCMV(s)-BGHpA containing 0 to 17 map units (mu) human type 5 adenovirus genome with a human CMV promoter, multicloning site and bovine growth hormone polyA signal inserted in the E1 region of viral genome to obtain the recombinant plasmid pAd-sGT (Fig 1). This plasmid was cotransfected into 293 cells along with a plasmid pJM1714 containing most of the rightward sequences (3.7 to 100 mu) of human type 5 adenovirus genome with a partial deletion in the E3 region. The recombinant replication-defective adenovirus Ad5sGT was rescued by homologous recombination (Fig 1). The replication-deficient recombinant adenoviruses, Ad5β gal and Ad5null, used as the control throughout the study, have been previously described15. High titers of recombinant adenoviruses were amplified, purified, titered, and stored as previously described 16.

Figure 1
figure 1

Construction of recombinant adenovirus Ad5sGT expressing pig α (1, 3)GT gene The recombinant plasmid pAd-sGT was constructed by inserting pig a(1, 3)GT cDNA into the multicloning site of a shuttle vector pAdCMV(s)-BGHpA which harbored a cytomegalovirus promoter (CMV) and a bovine growth hormone polyA signal (poly A). The recombinant adenovirus Ad5sGT was generated by homologous recombination after cotransfecting 293 cells with pAd-sGT and a virus-rescuing vector pJM17.

Full size image

Characterization of adenoviral vector expressing pig α(1, 3)GT gene

The genomic DNA extracted from Ad5sGT-infected 293 cells was digested with Hind III and the fragments were analyzed by agarose gel electrophoresis and Southern hybridization. COS cells were infected with Ad5sGT at multiplicity of infection (MOI) of 40. Flow cytometric analysis and the direct fluorescence of cell surface carbohydrate epitope on Ad5sGT-infected COS cells were performed with fluorescein isothiocyanate (FITC)-conjugated Bandeiraea simplicifolia isolectin B4 (BS-IB4) lectins (Sigma) specific for gal epitope.

Southern hybridization

Southern blot analysis was performed according to the standard procedures using pig α(1, 3)GT cDNA released from pcDNA3-sGT1100 as probe. The probeswere labeled with α-32P dATP using Random primed DNA labeling kit (Boehringer Mannheim, GmbH, Germany). DNA fragments separated by agarose gel electrophoresis were transferred onto Hybond-N membranes (Amersham) by capillary transfer method and were subsequently hybridized with the radiolabeled probes.

Flow cytometric analysis

A375, SGC-7901 and SPC-A-1 cells were detached from the tissue culture flasks 48 h after Ad5sGT infection by brief treatment with 5 mM EDTA in phosphate buffered saline (PBS, pH 7.4), centrifuged for 10 min at 250 g in the presence of 10%NBS, and resuspended at 106 cells/ml in cold (4 °C) PBS, 1% bovine serum albumin (BSA), 0.1% NaN3 (Sigma). 100 ml of cells in triplicate were incubated with the primary antibody for 1 h at 4 °C, and then washed three times with cold PBS/BSA/NaN3 before being resuspended in the same buffer containing second antibody (FITC-conjugated rabbit anti-human IgG and IgM antibodies, DAKO) at the dilutions recommended by the manufacturer. Following incubation at 4 °C for 1 h in the dark, the samples were again washed three times with cold PBS/BSA/NaN3 and resuspended in 100 μl of ice-cold PBS, 1% formaldehyde. FITC-conjugated Ulex europaeus I (UEA I), Vicia villosa agglutinin (VVA), Arachis hypogaea agglutinin (PNA), Glycine max agglutinin (SBA), and BS-IB4 lectins were used at 5 μg/ml in PBS/NBS/NaN3. A375, SGC-7901 and SPC-A-1 cells were labeled for 1 h at 4 °C in the dark and then washed and fixed as above. Each individually labeled sample was analyzed separately by flow cytometry using a Becton Dickinson FACScan cytometer.

Efficiency of recombinant adenovirus infection

Bacterial β-galactosidase (β-gal) gene (lacZ gene) expression was used as a marker of viral infection efficiency. COS, A375, SGC-7901 and SPC-A-1 cells were tested using Ad5β gal containing E. coli lacZ gene. Exponentially growing cells were seeded in duplicate in 24-well tissue culture plates at a density of 4×104 cells/well and infected with Ad5 β gal virus at different MOIs at the same time. After 48 h cells were fixed with 0.5% glutaraldehyde and then stained with X-gal solution (Sigma). Positive cells infected with Ad5βgal expressed bgal activity and developed blue in color. The results were analyzed qualitatively by visualization and captured by photomicrography.

C3c binding assay

A375, SGC-7901 and SPC-A-1 cells were washed and then incubated in 200 ml of normal human serum (NHS) at concentration of 10% for 10 min at 37 °C, washed again, and then incubated with FITC-conjugated rabbit anti-human C3c antibody (Dakopatts, Denmark) at 1:50 dilution at 4 °C for 30 min. Stained cells were analyzed using a FACScan cytometer.

Pig α(1, 3)GT gene expression at different postinfection time points

A375, SGC-7901 and SPC-A-1 cells were seeded in tissue culture flasks and infected with Ad5sGT at a MOI of 20 for A375, a MOI of 40 for SGC-7901, and a MOI of 45 for SPC-A-1. Cells were collected at 1, 2, 3, 4, 5, 6 and 7 d from the time the culture was initiated and stained with FITC-conjugated BS-IB4 lectins. Flow cytometric analysis was performed to detect the expression of gal epitope on the cell surface.

Determination of the effect of normal human serum (NHS) on human tumor cells

Human tumor cells A375, SGC-7901 and SPC-A-1 were seeded in tissue culture flasks and infected with Ad5sGT at MOIs as described above. After 48 h, cells were incubated in the presence of 10%, 20% or 40% NHS (blood group B) at 37 °C for 30 min, and then stained with 0.4% trypan blue. The numbers of viable and nonviable cells were counted and recorded as the percentage of viable cells (viable cells/total cells counted).

Effect of gal epitope expression on cell growth in vitro

A375, SGC-7901 and SPC-A-1 cells were seeded at a density of 2×103 cells/well in 96-well plates in octuple after infection with Ad5sGT at MOIs as described above. The MTT assay was employed to account the viable number of cells at 1, 2, 3, 4, 5, 6 and 7 d after Ad5sGT infection. A growth curve was plotted for each cell line depicting the viable number of cells versus the duration of postinfection. Cell growth assay was repeated for three times under the same experimental condition.

Statistical analysis

Data in this study were expressed as means and standard deviations and analyzed by Student's t test.

RESULTS

Characterization of adenoviral vector expressing pig α(1, 3)GT gene

Upon hybridization to pig α(1, 3)GT cDNA probe, a specific hybrid band was detected in Hind III-digested genomic DNAs from 293 cells infected by the recombinant adenovirus expressing pig α(1, 3)GT gene (Ad5sGT) (Fig 2A), corresponding to about 5.3 kb fragment of viral DNA as predicted. These results demonstrated the incorporation of pig a(1, 3)GT cDNA into the adenoviral genome.

Figure 2
figure 2figure 2figure 2

Characterization of recombinant adenovirus Ad5sGT (A) The genomic DNA extracted from Ad5sGT-infected 293 cells was digested with Hind III. The resultant fragments were gel-separated and hybridized to a pig α(1, 3)GT cDNA probe in Southern hybridization. From left, Lane 1, genomic DNA from control adenoviral vector Ad5null-infected 293 cells; lane 2, 3, and 4, genomic DNA from Ad5sGT-infected 293 cells; lane 5, a DNA ladder as a control and size marker (lDNA/Eco91 I). (B) Cell surface staining of recombinant adenovirus infected COS cells. FITC-conjugated BS-IB4 lectin staining of the surface of COS cells after Ad5null infection (a), and Ad5sGT infection (b). Results are representatives of at least 4 experiments. × 280 (C) Flow cytometric analysis of gal epitope expression on the surface of COS cells after Ad5sGT infection using FITC-conjugated BS-IB4 lectin specific for gal epitope. Ad5null infection was shown as bold line (control), and Ad5sGT infection was shown as filled histogram.

Full size image

Since COS cells were demonstrated previously that no gal epitope was expressed on the cell membrane17, these cells were therefore used in this study to determine the expression of pig α(1, 3)GT gene incorporated in the adenoviral vector Ad5sGT. To verify transgene expression, COS cells were infected with Ad5sGT at 40 MOI. At 48 h after infection, the expression of α(1, 3)GT gene in Ad5sGT-infected COS cells was detected in both the direct immunofluorescence of cell surface carbohydrates (Fig 2B) and flow cytometric analysis (Fig 2C) using the FITC-conjugated BS-IB4 lectin specific for gal epitope. These data display the efficacy of pig α(1, 3)GT gene expression in various cell types, driven by the CMV promoter contained in this adenoviral vector.

Recombinant adenovirus infection

Following infection with Ad5βgal, COS, SPC-A-1, SGC-7901 and A375 cells exhibited different efficacies of recombinant adenovirus infection. However, more than 90% of infectivity was achievable for all these cell lines at different optimum MOIs (Fig 3). The optimum MOI was 40 for COS and SGC-7901 cells, 20 for A375 cells, and 45 for SPC-A-1 cells.

Figure 3
figure 3

β-gal gene expression 48 h after infection with Ad5 bgal in COS cells at MOI of 40 (A), in A375 cells at MOI of 20 (B), in SGC-7901 cells at MOI of 40 (C), and in SPC-A-1 cells at MOI of 45 (D). ×140

Full size image

Expression of gal epitope on human tumor cells

To determine whether α(1, 3)GT expressed from Ad5sGT could effectively synthesize gal epitope on the membrane of human tumor cells, three tumor cell lines were analyzed. The results revealed that all the tumor cells infected with Ad5sGT expressed high level of gal epitope recognizable by BS-IB4 lectin on the cell membrane. In contrast, original tumor cells and Ad5null-infected tumor cells did not express gal epitope, as indicated by a lack of interaction with BS-IB4 lectin (Fig 4).

Figure 4
figure 4

Expression of gal epitope on human tumor cells. Human tumor cells A375 (A), SGC-7901 (B) and SPC-A-1 (C) were infected with Ad5sGT at corresponding optimum MOIs and collected at 48 h, stained with FITC-conjugated BS-IB4 lectin. Flow cytometric analysis was carried out using a fluorescence-activated cell sorter (FACScan, Beckton Dickinson). Mock infection was shown in broken line (control), Ad5null infection in bold line (control), and Ad5sGT infection in filled histogram.

Full size image

Binding of human IgG and IgM natural antibodies to human tumor cells and complement activation after Ad5sGT infection

To determine whether expression of α(1, 3)GT gene in human tumor cells could result in binding of human natural antibodies to cell surface, SGC-7901, SPC-A-1 and A375 cells were infected with Ad5sGT and Ad5null, respectively, and then incubated with 10% normal human serum (NHS). Rabbit anti-human IgG and IgM antibodies were used as second antibodies to detect the binding of human Igs to the cell surface. Flow cytometry was performed as described in Materials and Methods. As shown in Fig 5A, the tumor cells expressing gal epitope exhibited a significant increase in the level of human Igs (IgG and IgM) binding when compared with Ad5null-infected cells which did not express gal epitope.

Figure 5
figure 5

Binding of human IgG and IgM natural antibodies to human tumor cells and their complement activation after recombinant adenovirus infection. Human tumor cells A375, SGC-7901 and SPC-A-1 were infected with Ad5null and Ad5sGT respectively and collected at 48 h postinfection. (A) The binding of human natural antibodies. Cells were incubated with 10% normal human serum (NHS) as the first antibody and then with FITC-conjugated rabbit anti-human IgG and IgM antibodies as second antibodies. Stained cells were analyzed by flow cytometry (Ad5null, Ad5sGT). (B) C3c-binding assay. Cells were incubated in 10% NHS for 10 min at 37 °C, and then washed, examined for deposition of the C3 component of complement by staining with FITC-labeled rabbit anti-human C3c antibody, and analyzed by flow cytometry. Data shown in Fig 5 (A) and (B) represent means and standard deviations of triplicate mean fluorescence intensity (MFI) values for flow cytometry histograms.

Full size image

Next we determined the effect of increased binding of human Igs on human complement activation. The tumor cells from each cell line at 48 h after Ad5sGT infection were incubated in the presence of NHS at concentrations of 10%, and then the deposition of C3, a component of activated complement, was determined as a measure of cell surface complement activation by Flow cytometric analysis. As shown in Fig 5B, C3c deposition was higher on cells expressing gal epitope when compared with control.

Effect of α(1, 3)GT gene expression on expression of cell surface carbohydrates

To investigate whether the gal epitope expression could alter the expression of cell surface carbohydrates, the expression of cell surface carbohydrates, particularly those reacting with UEA-I, VVA, PNA and SBA lectins, on the surface of human tumor cells A375, SGC-7901 and SPC-A-1 was first determined by flow cytometry using FITC-conjugated lectins. The results revealed that all these tumor cell lines expressed the UEA-I, SBA, PNA and VVA lectin binding sites on the membrane at different levels (Tab 1). In the presence of gal epitope on tumor cell surface, the expression of UEA I lectin binding site, known as the blood group H antigen18 which was expressed at the highest level on SGC-7901 cells, became decreased significantly. However, no effect of gal epitope expression on blood group H antigen expression was detected on A375 and SPC-A-1 cells which have rather lower level of blood group H antigen. The appearance of gal epitope also resulted in the decrease of VVA and PNA lectin binding sites on A375 and SGC-7901 cells, but not on SPC-A-1 cells. PNA lectin binding sites on SPC-A-1 cells and SBA lectin binding sites on A375 and SPC-A-1 cells were up-regulated by the expression of gal epitope (Tab 2).

Table 1 Expression of cell surface carbohydrates reacting with UEA I, VVA, PNA and SBA lectins.
Full size table
Table 2 Effect of expression ofa(1, 3)GT gene on the expression of cell surface carbohydrates
Full size table

Effectiveness of gal epitope in inducing destruction of human tumor cells by normal human serum (NHS)

To test the effect of NHS on human tumor cells expressing gal epitope, A375, SGC-7901 and SPC-A-1 cells were infected with Ad5sGT and then incubated in the presence of NHS at different concentrations. Cells without Ad5sGT infection were used as negative controls. As shown in Tab 3, trypan blue exclusion did not detect the significant difference in sensitivity to NHS-mediated lysis between human tumor cells with and without gal epitope.

Table 3 Sensitivity of human tumor cells to normal human serum treatment
Full size table

Expression of gal epitope at different postinfection time points

In all three tumor cell lines, gal epitope levels were analyzed by flow cytometry using FITC-conjugated BS-IB4 lectin and were substantially high at 24 h after Ad5sGT infection. The gal epitope levels progressively increased with time and reached a maximum at 48 h in A375 and SPC-A-1 cells, and at 72 h in SGC-7901 cells. After these time points, gal epitope level decreased with time and became reduced to background level at 7 d after Ad5sGT infection (Fig 6).

Figure 6
figure 6

Gal epitope expression after adenovirus-mediated pig α(1, 3)GT gene transfer in vitro. A375, SGC-7901 and SPC-A-1 cells were infected with Ad5sGT containing pig α(1, 3)GT cDNA at 20, 40 and 45 MOIs respectively. Cells were collected at different time-points and stained with FITC-conjugated BS-IB4 lectin, and then analyzed by flow cytometry (MFI, mean fluorescence intensity).

Full size image

Tumor cell growth in vitro following gal epitope expression

The effect of gal epitope expression on the proliferative capacity of human tumor cells, including SGC-7901, SPC-A-1 and A375, was judged using MTT assay. The results showed that cells expressing gal epitope had the same proliferative capacity as their parent and the Ad5null-infected cells (Fig 7). In addition, the morphology of these human tumor cells remained unchanged following gal epitope expression under light microscope.

Figure 7
figure 7

Human tumor cell growth in vitro following gal epitope expression over 1-7 d in MTT assay. A375, SGC-7901 and SPC-A-1 cells were infected with Ad5null or Ad5sGT at optimum MOIs as determined in previous experiments, and then cultured in 96-well plates. MTT assay was performed on each day, and growth curves were plotted to determine the growth potential of cells expressing gal epitopes. Data represent means of octuple values and standard deviations.

Full size image

DISCUSSION

Data in this study showed that human tumor cells SGC-7901, SPC-A-1 and A375 expressed high level of gal epitope on the membrane after Ad5sGT infection. In parallel, the ability of the binding of human IgG and IgM natural antibodies to gal epitope expressing cells was also enhanced. This suggested that the immune complexes of human tumor cells could be formed easily using human anti-Gal natural antibodies after Ad5sGT transduction. Link et al19 has utilized gal epitope and anti-Gal antibodies reaction as a gene therapy for treatment of human cancer, in which human complement-mediated lysis of human melanoma A375 cells was induced by engineering melanoma cells to express gal epitope followed by incubation with NHS. In this study, we also examined the susceptibility of human tumor cells SGC-7901, SPC-A-1 and melanoma A375 to lysis by normal human serum following gal epitope expression. Cells were analyzed for gal epitope expression by flow cytometry and at the same time were incubated at 37 °C for 30 min in the presence of NHS at 10%, 20% and 40% concentrations respectively, and then the number of viable cells was counted by the method of trypan blue exclusion. However, no significant difference in susceptibility to lysis by NHS was observed in each of three tested human tumor cell lines between gal epitope expressing cells and its corresponding parental cells, although activation of complement was enhanced. Because complement-mediated cell lysis is inhibited by the complement regulatory proteins expressed on the cell surface20, no sensitization of these human tumor cells to NHS may be the results of high level expression of complement inhibitory molecules on these human tumor cells.

The biosynthesis of carbohydrate epitopes in cells is regulated by glycosyltransferases, which are responsible for the addition of carbohydrates to the oligosaccharide chain on glycolipids and glycoproteins in a sequential manner21. The newly introduced carbohydrate epitope had been observed to alter the expression of other carbohydrates. Gorelik et al22 transfected murine melanoma cells, which do not express the gal epitope as well as carbohydrates reacting with SBA, PNA, and VVA lectins, with α(1, 3)GT cDNA. This resulted in the appearance of gal epitopes as well as carbohydrates reacting with SBA, PNA or VVA lectins. Appearance of SBA, PNA and VVA lectin binding carbohydrates in the α(1, 3)GT gene-transfected melanoma cells was suggested to be the results of reduction of cell membrane sialylation and subsequent unmasking of these carbohydrates due to the competition between α(1, 3)GT with α(2, 3) sialyltransferase or α(2, 6) sialyltransferase for the common acceptor N-acetyllactosamine in the Golgi apparatus22. In this study, unlike murine melanoma cells used in Gorelik's report22, human tumor cells SGC-7901, SPC-A-1 and A375 express carbohydrates reacting with SBA, PNA and VVA lectins on their surface but do not express gal epitope due to the inactivation of α(1, 3)GT gene in humans. After transduction with Ad5sGT, expression of carbohydrates reacting with PNA and VVA lectins on SGC-7901 and A375 cells was reduced to different degrees, but SBA and PNA lectin binding carbohydrates on SPC-A-1 cells were increased. The results reflected that expression of gal epitope has different effects on other carbohydrates expression depending on the features of carbohydrates on different cell types.

The α(1, 2) fucosyltransferase gene expression has been reported to resultin a drastic suppression of the gal epitope on mouse cells due to the enzymatic competition between the α(1, 3)GT and α(1, 2) fucosyltransferase for the common acceptor substrate23, 24. In contrast to the mouse cells, SGC-7901, SPC-A-1 and A375 all express the blood group H antigen on the surface due to the α(1, 2)fucosyltransferase activity18, 25. Following gal epitope expression, the blood group H antigen was also reduced significantly on SGC-7901 cells. No changes in blood group H antigen expression on A375 and SPC-A-1 cells following gal epitope expression probably represented the limitation of H antigen which could be replaced by newly introduced gal epitopes. In this study, we also investigated whether expression of gal epitope could affect the growth of human tumor cells in vitro. The results showed that the expression of gal epitope on human tumor cells did not result in changes in cell growth in vitro as determined by MTT assays.