Background
Epithelial ovarian cancer (EOC) is the most lethal gynecological malignancy. Most ovarian cancer patients are treated with cytoreductive surgery followed by combination chemotherapy. Although the majority of patients respond to first-line treatment, most will relapse and overall survival rates for patients with advanced EOC remain disappointing [
1‐
3]. As such, it remains critical to develop an effective therapy to reduce the relapse and achieve long-term remissions.
In recent years, the study of cancer stem cells (CSCs) has resulted in an improved understanding of carcinogenesis. CSCs are proposed to represent the fundamental driving force of tumor development, initiation, invasion, and metastasis, as well as recurrence [
4]. Conventional anti-cancer treatments (e.g. chemotherapy and radiation) can often transiently shrink tumors, but these therapies fail to target and kill CSCs, leading to relapse, and ultimately death [
5]. As such, it is necessary to develop an effective stratagem targeting CSCs. Evidence to date has demonstrated that ovarian CSCs are tumorigenic and chemo-resistant. Elimination of ovarian CSCs, therefore, might help improve the sensitivity of chemotherapy drugs and reduce the tumor relapse [
6,
7].
Considering that the immune system plays a pivotal role in tumorigenesis, immunotherapy has been established as an important component of many cancer treatment regimens, which could destroy cancer cells both
in situ and within the metastatic compartment [
8‐
10]. In addition, immunotherapy using antibodies (Abs) targeting tumor-specific antigens expressed on CSCs can selectively kill CSCs, while sparing the normal counterpart [
11]. Furthermore, tumor vaccines have also showed promising preliminary data in targeting CSCs. The prerequisite for the induction of an effective antitumor immune response by tumor vaccine is the effective uptake of this vaccine by professional antigen-presenting cells (APC). It was reported that the addition of α-gal epitope to MUC1
+ pancreatic carcinoma whole-cell vaccine could enhance presentation to APC and induce immune responses against not only differentiated cancer cells but also CSCs [
12].
The α-gal epitope is a glycoconjugate present on cell membranes of non-primate mammals, prosimians and New World monkeys, but not in humans. However, the corresponding human anti-Gal antibody was found to be present in high titer in the serum of every normal individual studied [
13] and is continuously produced as an immunological response to antigenic stimulation by bacteria of the normal flora [
14]. It is reported that α-gal epitope specific IgG, IgM, IgD, and IgA titers remained unvaried over longer time periods in healthy subjects [
15]. Tumor cells engineered to express α-gal epitopes were able to bind anti-Gal and to be destroyed by this antibody in an experimental animal model [
16].
Consistent with other studies [
6,
7], our previous work demonstrated that ovarian epithelial cancer cells cultured in serum-free medium could form spheroid cells, which are cancer stem-like cells that have the characterization of CSCs and can be distinguished from differentiated ovarian cancer cells [
17‐
19]. Herein, we hypothesized that biosynthesis of α-gal epitopes to ovarian cancer spheroid cells could effectively induce Abs production against ovarian cancer stem-like cells. Using α1,3GT knockout mice, we further investigated the immune response induced by vaccines expressing α-gal epitopes against both differentiated ovarian cancer cells and cancer stem-like cells.
Methods
Cell culture
All cell lines were obtained from Shanghai Cell Bank of Chinese Academy of Sciences (Shanghai, China). 293 T cells (Immortalized human embryonic kidney cells) were cultured in Dulbecco’s modified Eagle’s medium (DMEM; Gibco, Grand Island, NY) supplemented with 10 % fetal bovine serum (FBS) in a humidified incubator with 5 CO
2 and 95 % air at 37 °C. Human ovarian cancer cell line SKOV3 cells were maintained in McCoy’s 5A medium (Sigma-Aldrich, Oakville, ON, Canada) supplemented with 10 % FBS. Then the SKOV3 cells were dissociated by 0.02 % trypsin-EDTA and maintained under stem cell conditions as described before [
17‐
19]. In this condition, cancer cells grow as non-adherent spheroid cells. Culture media were changed every 2 days by centrifuging at 800 rpm for 5 min to remove the dead cell debris. Regular culture plates were coated with poly (2-hydroxyethyl methacrylate) (Sigma) before spheroid cell culturing [
17‐
19]. 293 T cells were used for recombinant lentivirus transfection, amplification, and titration.
Construction of recombinant lentivirus vector expressing pig α1,3GT gene
Primers for amplification of α1,3GT coding sequence (α1,3GT CDS) were previously reported by Yu et al. [
20]. Compared with the plasmid profile of pCDH-CMV-MCS-EF1-copGFP (System Biosciences, Mountain View, CA), the amplified sequence of α1,3GT CDS was analyzed for XbaI and BamHI restriction enzymes sites, and confirmed with NEBcutter V2.0. XbaI and BamHI restriction sites were designed into the primers of α1,3GT CDS prior to gene amplification. Both α1,3GT CDS and pCDH-CMV-MCS-EF1-copGFP was digested by XbaI and BamHI and the resulting 1.19 kb fragment of porcine α1,3GT cDNA was subcloned into pCDH-CMV-MCS-EF1-copGFP to construct pCDH-CMV-α1,3GT-EF1-copGFP. Finally the recombinant plasmid was used to produce lentivirus-based vector with three plasmid packaging systems including another two plasmids - pCMV dR8.91 and VSVG [
21]. All three plasmids were co-transfected into packaging 293 T cells using the calcium phosphate method [
21]. High titers of recombinant lentivirus were amplified, purified, and stored. Eventually SKOV3 cells were transfected with lentivirus vector at multiplicity of infection (MOI) of 40, which formed SKOV3-gal cells expressing α-gal epitopes.
Real-time quantitative PCR
After total RNA was isolated from the samples, five hundred nanograms of total RNA from each sample were reverse transcribed using the primescript RT reagent kit (Takara Bio Inc., Shiga, Japan). Real-time PCR was performed on cDNA using SYBR Premix Ex Taq (Takara) on the Mastercycler® ep realplex (Eppendorf, Hamburg, Germany). All reactions were carried out in triplicate using a 25-μl volume. In brief, PCR amplification was carried out using an initial denaturation at 95 °C for 5 min, followed by 40 cycles for 30 s at 95 °C, 30 s at 60 °C, and 30 s at 72 °C. A sample lacking template DNA was used as a negative control. The primers for the genes are provided in Additional file
1: Table S1.
Proliferation and drug resistance assessment
SKOV3 cells and SKOV3-gal cells were seeded at a density of 3 × 103 cells per well in 96-well plates. The MTT assay was employed to count the viable number of cells at 1, 2, 3, 4, 5, and 6 days after seeding cells. A growth curve was plotted for each cell line depicting the viable number of cells versus the duration post infection.
To assess the effect of α-gal epitopes on drug resistance, SKOV3 cells and SKOV3-gal cells were plated in culture medium containing cisplatin (Sigma-Aldrich, 20 μmol) for 24 h and 48 h. Cultures were set up in triplicate. The proliferation conditions were determined by MTT assay with an OD reading at 490 nm. Each experiment was repeated three times under the same experimental conditions.
Both SKOV3 cells and SKOV3-gal cells were incubated in the presence of 1:3, 1:6, 1:9, and 1:18 human serum solution diluted with McCoy’s 5A medium without FBS for twenty-four hours. Human blood samples were collected from three ovarian cancer patients who were categorized as stage III, grade 2–3 serous adenocarcinoma according to the International Federation of Gynecology and Obstetrics (FIGO) classification. The present study was approved by the institutional review boards at Shanghai Jiaotong University and written informed consent was obtained from these patients. SKOV3 cells and SKOV3-gal cells without human serum and FBS were used as negative controls. McCoy’s 5A medium without FBS was used as blank control. Cultures were set up in triplicate. Proliferation was monitored by MTT assay under optical density (OD) reading at 490 nm. The percentage of survival rate was determined as follows: (OD490(sample)- OD490(blank control))/(OD490(negative control)- OD490(blank control)) × 100 %.
Antibody-dependent cell-mediated cytotoxicity (ADCC) assay
The ADCC assay was performed as described by Tanemura et al. [
22] with minor modifications. SKOV3 cells and SKOV3-gal cells were plated in 96-well plates. Subsequently, peripheral blood mononuclear cells (PBMC) were obtained by Ficoll–Hypaque centrifugation of venous blood from ovarian patients and were added to the wells at cell ratios of 1:20 (effector cells: target cells) in the presence or absence of serum from the same patient. The plates were incubated at 37 °C for 72 h and MTT assay was employed to count the number of viable cells. The percentage of cytotoxicity rate was determined as follows: 1-(OD
490(sample)-OD
490(PBMC))/OD
490(target cells).
Tumor cell vaccination
Six α1,3GT knockout mice were obtained from Dr Jung-Ah, Cho (Center for animal resource development, College of Medicine, Seoul National University, Seoul, Korea). The genotypes were confirmed by polymerase chain reaction (PCR) on DNA extracted from the tail of the α1,3GT KO mice [
23], with the wild-type (WT) mice counterpart (C57BL/6) as control. The mice were mated to supply enough mice for the experiment. Prior to the experimental procedure, six-week-old KO mice received 4-weekly intraperitoneal immunization of 100 mg of pig kidney membrane homogenate to induce high titers of anti-Gal Abs as described previously in humans [
22,
24]. Subsequently, the KO mice were vaccinated subcutaneously three times at 1 week intervals with 1 × 10
6 50 Gy–irradiated SKOV3 cells, SKOV3 spheroid cells, SKOV3-gal cells, or SKOV3-gal spheroid cells. One week after vaccination, the mice were assessed for the immune response as described below. All animal procedures were approved by the Institutional Animal Care and Use Committee of Shanghai, and were performed in accordance with the National Research Council Guide for Care and Use of Laboratory.
Enzyme-linked immunosorbent assay (ELISA)
The production of anti-Gal IgG antibody was measured by ELISA. Galalph1-3Gal beta1-4GlcNAc-BSA (Dextra Laboratories, Reading, UK) was attached to microtiter plates at a concentration of 10 μg/ml in carbonate buffer, pH 9.5. The plates were coated at 4 °C overnight. After washing five times, 1 % bovine serum albumin in carbonate buffer was added to the wells to prevent nonspecific binding. The serum samples obtained from vaccinated-KO mice were added at various dilutions with PBS containing 0.3 % BSA. One hour later, horseradish peroxidase (HRP)-linked goat anti-mouse IgG was added. A color reaction was obtained with peroxidase reagent 3.30, 5.50-tetramethylbenzidine (TMB, Sigma-Aldrich), and stopped by the addition of 2 M H2SO4. The optical density was read at 450 nm using a Bio-Rad ELISA reader.
In vivo studies of tumor cell vaccine in an experimental mouse model
5 × 10
6 SKOV3 cells were injected subcutaneously into 6 to 8 week-old female (
n = 8) athymic nude mice to form SKOV3 tumors. After injection, the nude mice were kept in a specific pathogen-free environment and checked for tumor development every two days for 1 month. Simultaneously, the α1,3GT KO mice were pre-immunized with pig kidney fragments and then vaccinated with 50 Gy–irradiated SKOV3 cells, SKOV3 spheroid cells, SKOV3-gal cells, or SKOV3-gal spheroid cells as described above. One week after the last immunization, the tumors formed in nude mice were collected and stored at 4 °C in Hanks’ balanced salt solution supplemented with antibiotics. Within 2 h, the tumor sample was cut into two parts. One part was fixed for histological analysis and the other was cut into multiple 1 mm
3 pieces, which were meticulously implanted under the renal capsule of the left kidney (2 grafts per kidney) of the immunized α1,3GT KO mice (
n = 15 in each group) as described previously by Lee et al. [
25]. Subsequently, on days 4, 8, and 12 after grafting, five mice in each group were euthanized under anesthesia and the kidneys were removed and the grafts were measured
in situ by using a stereoscopic microscope, fitted with an ocular micrometer, and calibrated in ocular units (OMU, 1OOMU = 1 mm) to examine tumor growth. Two perpendicular diameters were measured and the difference in mean tumor diameter was calculated [
26].
Western blotting analysis
The production of Abs toward tumor antigens on SKOV3 cells or SKOV3 spheroid cells was examined by Western blot analysis using vaccinated KO mice sera. The total amount of proteins from the cultured cells was subjected to 10 % SDS-PAGE (50 μg protein each lane) and transferred onto Hybrid-PVDF membranes (Millipore, Bedford, USA). After blocking with 10 % (w/v) non-fat dried milk in TBST (Tris-buffered saline containing Tween-20), the PVDF membranes were washed with TBST at room temperature and incubated with immunized α1,3GT KO mice sera diluted with 5 % (w/v) non-fat dried milk in TBST overnight. Following extensive washing, membranes were incubated with HRP-conjugated goat anti-mouse IgG secondary anti-body (1:5000; Santa Cruz Technology, Santa Cruz, USA). After washing, the immunoreactivity was visualized by enhanced chemiluminescence using ECL kit from Thermo Fisher Scientific.
To confirm whether c-erbB2 protein was existent in the 185 kDa band and 150 kDa band of the lysates of SKOV3 adherent and spheroid cells, Western-blot analysis was performed. Both 5 μl and 10 μl lysate were loaded and reacted with anti-ERBB2 antibody (Ascend Biotechnology, AR0071) and anti-β-actin antibody (Santa Cruz Technology).
Flow cytometric (FCM) analysis
The expression of α-gal epitopes was evaluated on SKOV3-gal cells by flow cytometry. Approximately 1 × 106 parental SKOV3 and α1,3GT transfected cells were suspended in 2 % BSA/PBS and incubated with biotinylated bandeiraea simplicifolia isolectin B4 (BS-IB4) lectins, which specifically binds to α-gal epitopes. Then cells were incubated with PE-Cy5 conjugated streptavidin.
The effect of cancer cell vaccine on the immune system of the KO mice was evaluated by analyzing the changes of CD3 + CD4+ and CD3 + CD8+ T cells in spleen. First the spleen was removed, then cut into pieces and grinded gently through a 200 mesh sterile nylon net. The cell suspension was carefully collected and layered on the Ficoll-Paque, PREMIUM (GE Healthcare Life Sciences, USA). The separated splenic mononuclear cells were incubated with fluorochrome-conjugated antibodies directed at the following CD markers: PE anti-mouse CD3, FITC anti-mouse CD4, and PE-Cy7 anti-mouse CD8 (all purchased from eBioscience, San Diego, CA). Gated CD3 positive events were analyzed for CD8+ and CD4+ T cells. Flow cytometry was performed using a FC500 flow cytometer (Beckman Coulter, Fullerton, CA) and analyzed using Beckman Coulter CXP software.
Mass spectrometry analysis
Proteins were identified by mass spectrometry from bands digested from preparative gels run with 50 μg of protein on each lane and stained with commassie blue staining solution. Picking, destaining, digestion, extraction, and sample preparation of the bands were carried out. The lyophilised samples were diluted with 5 % aqueous acetonitrile (ACN) and 0.1 % formic acid and subsequently analyzed using a Thermo Finnigan LTQ ion trap mass spectrometer (Thermo Finnigan, Bremen, Germany) at the core facility of Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences. Database searches were done using Sequest (TurboSEQUEST v. 27) search engines against the nonredundant IPI human database version 3.68.
RNA interference
To knock down the ERBB2 gene expression, two small interference RNAs (siRNAs) targeting different sites of human ERBB2 mRNA (Genbank accession no. NM_004448.3) were designed and synthesized by GenePharma Co. (Shanghai, China), and a control siRNA that could not target any human mRNA was synthesized as a negative control (N.C. for short). Additionally, normal SKOV3 cells served as an untreated control (Untreated for short). The detailed information of the siRNA was listed in Additional file
2: Table S2.
One day prior to transfection, 1 × 105 SKOV3 cells were seeded in six-well plates and incubated at 37 °C with 5 % CO2 overnight. The next day, cells were grown to about 50 % confluence and the medium was changed for 1.5 ml fresh McCoy’s 5A medium. Then the cells were transfected with 80 nM of siRNA using DharmaFECT reagent (Thermo Fisher Scientific Inc. MA. USA) according to the manufacturer’s instructions. Forty-eight hours after transfection, total RNA was collected for real-time PCR analysis and proteins were collected for western blot analysis. The primers for ERBB2 were as follows: 5’- CGCTTTTGGCACAGTCTACA-3’ (forward) and 5’- TCCCGGACATGGTCTAAGAG-3’ (reverse). GAPDH served as an internal control for normalization and the primers for PCR were as follows: 5’- ATGGGCAGCCGTTAGGAAAG-3’ (forward) and 5’- TGAAGGGGTCATTGATGGCA -3’ (reverse). PCR amplification was carried out using an initial denaturation at 95 °C for 2 min, followed by 40 cycles for 5 s at 95 °C and 34 s at 60 °C.
Statistical analysis
Parametric comparisons of normally distributed values that satisfied the variance criteria were made by unpaired Student’s t-tests. Data that did not pass the variance test were compared with a nonparametric two-tailed Mann–Whitney rank sum test.
Discussion
Tumor escape from immune response is considered as the greatest impediment to the application of effective immunotherapy [
31]. It was reported that a three-stage process was needed for the generation of effector T cells after vaccination with tumor cell vaccine. Firstly, an effective immune response requires tumor-antigen uptake and processing at the site of injection by APCs. Then APCs migrate into the regional draining lymph nodes, where tumor-specific T cells are activated respectively by TAA peptides presented on MHC class I and class II molecules of APCs. Finally circulation of activated T cells leads to tumor cell destruction [
32,
33].
Since the identity of specific TAAs of most types of cancer is not known, the irradiated autologous tumor may be used as a source of vaccinating material. However, tumor cells in cancer patients evolve in a stealthy way lacking identifying markers that label them for uptake by APC, so they can’t be detected by APCs, even in the form of vaccine [
34]. It was reported that autologous tumor cell membrane vaccine processed to express α-gal epitopes by incubating ovarian carcinoma membranes prepared from homogenates of freshly obtained tumors with recombinant α1,3GT could successfully induce anti-tumor vaccination
via the formation of immune complexes by exploiting the natural anti-Gal Ab and its ligand the α-gal epitope [
34]. However, it did not seem that all the TAAs were on the membranes of the cancer cells - some were within the cytoplasm of the cancer cells. Therefore, cytoplasmic TAAs may not be targeted to APC when the autologous tumor vaccine consists of only tumor membranes [
35]. In addition, CSCs occupy only a small size in the whole tumor population, thus the quantity of ovarian CSCs in the tumor membranes may be too small to induce immune responses against CSCs, which might affect the efficiency of the vaccination [
36].
Previous studies showed that ovarian cancer stem-like cells could be enriched by culturing spheroid cells under a serum free stem cell selective system [
17‐
19]. In the present study, we compared SKOV3 spheroid cells and SKOV3 differentiated cells by cDNA microarray and found that SKOV3 spheroid cells have a distinct gene expression profile (Additional file
5: Figure S3). By using this cell model, we obtained intact ovarian cancer stem-like cells expressing α-gal epitopes.
Complement dependent cytolysis (CDC) and ADCC were suggested as the main mechanism that induce xenograft cell destruction and xenograft rejection following the binding of the natural anti-Gal to α-gal epitopes on xenograft cells [
37,
38]. More than 90 % of A375 melanoma cell expression of α-gal epitopes was killed when exposed to human serum [
39]. Similarly, it was reported that 98 % of α-gal-positive MC38 colon carcinoma cells were killed by media containing human serum [
16]. However, Li et al. reported that there is no significant difference in susceptibility to lysis at various concentrations of human serum in three tested human tumor cell lines expressing gal epitope compared with their corresponding parental cells [
40]. Consistent with this report, we incubated SKOV3-gal cells with different dilutions of serum obtained from ovarian cancer patients. However, no significant difference was found with regards to sensitivity to human serum-mediated lysis compared with SKOV3 parental cells. It was suggested that the complement regulatory proteins expressed on the tumor cell surface may inhibit CDC [
41]. We investigated whether SKOV3-gal cells could be killed by ADCC. To this end, PBMC and serum from ovarian cancer patients were separated and then co-cultured with SKOV3-gal cells. The MTT assay showed that SKOV3-gal cells were killed when culturing with both PBMC and serum, implying that anti-Gal antibody in ovarian cancer patients’ serum can bind with SKOV3-gal cells and induce targeting of tumor cells to APCs.
The efficiency of the SKOV3-gal spheroid cell vaccine was tested in α1,3GT KO mice. Sera obtained from mice vaccinated with SKOV3-gal spheroid cells contained some new Abs that reacted with protein from both SKOV3 cells and SKOV3 spheroid cells. Mass spectrometric analysis demonstrated that the protein reacted with the new Abs including a protein c-erbB-2, which was associated with poor prognosis in cancers such as salivary gland adenocarcinoma, breast cancer, and lung cancer [
42‐
44]. Meanwhile, overexpression of c-erbB-2 proteins corresponded with aggressive characteristics of tumors, such as advanced clinical stage, high grade, presence of vascular invasion, and chemotherapy resistance, suggesting potential as a prognostic marker in patients with epithelial ovarian carcinoma [
27‐
29]. In this study, we found c-erbB-2 proteins at the 150 kDa band by mass spectrometry analysis. Further, the 150 kDa was thinner when the SKOV3 cell protein was collected after SKOV3 cells were transfected with siRNAs to knock down ERBB2 and then incubated with serum from SKOV3-gal spheroid cell-immunized mice. It was reported that the expression of the NH
2 terminally truncated ErbB2 (95 kDa) in breast cancer correlated with metastatic disease progression compared with the expression of full-length ErbB2 (185 kDa) [
45]. So c-erbB2 may also be truncated (about 150 kDa) in ovarian cancer and correlate with metastatic disease progression. Future studies are required to identify and characterize this molecule as novel TAA for immunotherapy of ovarian cancer.
It was determined that vaccination with inactivated influenza virus processed by the use of recombinant 1,3GT to express multiple α-gal epitopes could activate 12 to 13.7 % of CD4+ T-cell, whereas no significant activation was induced by virus lacking α-gal epitopes. Meanwhile, about six fold higher CD8+ T-cell responses were elicited by vaccine expression α-gal epitopes [
46]. Herein, we isolated spleen lymphocytes to identify the immune status of the immunized mice, and results showed that spleen lymphocytes from mice immunized with SKOV3-gal spheroid cells had a higher proportion of CD3 + CD4+ T cells, while the proportion of CD3 + CD8+ T cells decreased significantly when compared with non-immunized KO mice. Indeed, MHC class I restricted CD8+ T cells that specifically lyse tumor cells in vitro are frequently measured to document vaccine efficacy and to serve as a surrogate end point in clinical tumor vaccine trials [
12,
47,
48]. However, vaccine strategies that directly enhance the priming of tumor-specific CD4+ T-cells have been shown to augment the systemic rejection of MHC class II negative tumors [
49,
50]. Recently, a study reported that cancer cells pre-treated with paclitaxel and doxorubicin could enhance cancer-specific CD4+ T cell production and thus induce significant apoptosis and a therapeutic anti-tumor immune response in a mouse model [
51]. In this study, CD4+ T cells also mediated significant anti-tumor effector functions in this ovarian CSC vaccination. It seemed that mice immunized with SKOV3-gal cells did not induce a similar immune response as SKOV3-gal spheroid cells did, which may result from the distinct gene expression profile of spheroid cells and adherent cells.
Unfortunately, we failed to establish an ovarian cancer xenograft in the α1,3GT KO mice. 5 × 10
6 SKOV3 cells, which formed tumors in nude mice, were injected subcutaneously into α1,3GT KO female mice, however, no macroscopic tumor developed. It was reported that human tumor xenografts, including tumors that had been established in serial transplantation in athymic nude mouse hosts or primary surgical tumor tissues, implanted under the renal capsule of normal mice were quantifiable in ocular micrometer units for six days and could retain morphological and functional characteristics of the parent tumors [
26,
52]. So we administered injections of SKOV3 cells s.c. into female nude mice to from tumors and then cut them into small pieces (1 mm
3) to graft under the renal capsules of immunized α1,3GT KO mice. Our results showed that the grafted tumors in non-immunized mice continued to grow, while grafted tumors in mice immunized with all four kinds of vaccine (SKOV3 cells, SKOV3 spheroid cells, SKOV3-gal cells or SKOV3-gal spheroid cells) grew slower than non-immunized mice. Furthermore, xenografts in mice immunized with α-gal expressing cells were smaller at 12 days after grafting when compared with grafts in SKOV3 cell- and SKOV3 spheroid cell-immunized mice.
Whalen and his colleagues [
53] had suggested that intra-tumoral injection of α-gal glycolipids is feasible and safe for inducing a protective anti-tumor immune response. It was found to be safe after intra-tumoral injection of 0.1, 1, and 10 mg α-gal glycolipids per tumor. No patients developed clinical signs of toxicity, or any indication of autoimmunity [
53]. Accordingly, immunotherapy by synthesis of alpha-gal epitopes on the tumor cells may be safe also, but still needs further study.
Competing interests
Authors declare that there are no competing financial interests in relation to the work described. There are no non-financial competing interests.
Authors’ contributions
DL participated in the design of the study. XY and ZD conducted the experiments. QZ and QW provided technical expertise and analyzed the data. XY, QZ and QW prepared the manuscript and DL critically reviewed the manuscript. All authors read and approved the final manuscript.