The highly invasive MDA-MB-231 cell line, the poorly invasive luminal A MCF-7 cell line and the normal-like MCF-12A cell line, were purchased from the American Type Culture Collection (Manassas, VA, USA) and were used for the purposes of the present study. All breast tumor cells were maintained at 80% confluence in Dulbecco's Modified Eagle's medium (DMEM, Gibco; Thermo Fisher Scientific, Inc., Waltham, MA, USA) supplemented with 10% fetal bovine serum (FBS, Gibco; Thermo Fisher Scientific, Inc.), 2 mM glutamine (Sigma-Aldrich; Merck-Millipore, Darmstadt, Germany), 10 mM HEPES buffer (Sigma-Aldrich; Merck-Millipore), and penicillin (100 IU/ml)-streptomycin (100 g/ml) (Beyotime Institute of Biotechnology, Haimen, China) at 37°C in a humidified atmosphere (5% CO₂).

siRNA-ppGalNAc-T1 and siRNA-ppGalNAc-T2 were purchased from Shanghai GenePharma Co., Ltd. (Shanghai, China) and were generated by subcloning multiple ppGalNAc-T1 and ppGalNAc-T2 cDNAs, respectively (Table I). MDA-MB-231 breast tumor cells were seeded onto a 6-well plate at a density of 2×10⁵ cells and cultured until they reached 80% confluence for transfection. Cells were transfected with 5 ml siRNA-ppGalNAc-T1 and/or siRNA-ppGalNAc-T2 vectors, using Lipofectamine™ 2000 (Invitrogen; Thermo Fisher Scientific, Inc.) according to the manufacturer's protocol. The 5 µl empty vector, pEGFP-C1 (GenePharma Co., Ltd. Shanghai, China), was transfected into the MDA-MB-231 cells and used as a control. Transfection efficiency was determined by fluorescence microscopy (Zeiss GmbH, Jena, Germany) according to the fluorescence intensity of the control group. The optimal transfection system volume was determined using the GAPDH interference effect with different times and ratios.

Total RNA was extracted from 1.5×10⁶ MDA-MB-231 cells and 1.5×10⁶ MDA-MB-231 cells that transfected with siRNA using TRIzol reagent (Invitrogen; Thermo Fisher Scientific, Inc.) and treated with DNase I (Ambion; Thermo Fisher Scientific, Inc.) according to the manufacturer's protocol. RNA concentration and quality was measured using an ultraviolet spectrometer at a wavelength ratio of 260/280 nm. cDNA was synthesized from 1 µg total RNA using an RNA reverse transcription diagnostic kit (Takara Bio, Inc., Otsu, Japan). PCR reaction mixtures (30 µl) were amplified with 40 cycles of denaturation at 95°C for 60 sec, annealing at 56°C for 60 sec and extension at 72°C for 120 sec. In addition, RT-qPCR reaction mixtures (25 µl) were amplified using 40 cycles of denaturation at 95°C for 60 sec, annealing at 95°C for 15 sec and extension at 60°C for 60 sec. RT-qPCR was performed using an ABI PRISM 7500 Sequence Detection system (Applied Biosystems; Thermo Fisher Scientific, Inc.) using SYBR Green Realtime PCR Master Mix (Toyobo Co., Ltd., Osaka, Japan). Gene expression was quantified using the 2^−∆ΔCq method (20) following normalization to the expression of GAPDH. The primer sequences used for RT-qPCR analysis are presented in Table II.

MDA-MB-231 cells (1.5×10⁶) and MDA-MB-231 cells (1.5×10⁶) that transfected with siRNA were harvested and homogenized with lysis buffer, consisting of 50 mM Tris-HCl (pH 7.3), 150 mM NaCl, 1 mM EDTA, 1% Triton X-100 and 1% protease inhibitor cocktail (Beyotime Institute of Biotechnology). Total protein (15 µg) from each well was boiled for 10 min in 17% Laemmli buffer (Beyotime Institute of Biotechnology) and separated via 10% SDS-PAGE, before samples were transferred to a polyvinylidene difluoride (PVDF) membrane using a semi-dry technique. The membranes were blocked for ~1.5 h with 5% skim milk in Tris-buffered saline (TBS), containing 10 mM Tris, 150 mM NaCl (pH 7.9) and 0.05% Tween-20 (Beyotime Institute of Biotechnology). The membranes were then incubated with the appropriate primary antibodies [Anti-MMP14 antibody (cat. no. ab51074) was purchased from Abcam Co., Ltd., Cambridge, UK. GAPDH antibody (cat. no. AF0006) was purchased from Beyotime Institute of Biotechnology. Dilution, 1:1,000] at 4°C overnight. Following washing with TBS Tween-20 buffer, the cells were incubated with the appropriate horseradish peroxidase (HRP)-conjugated IgG-AP secondary antibody (Beyotime Institute of Biotechnology. Dilution, 1:1,000) at room temperature for 90 min. After washing in TBS Tween-20 buffer, detection and autoradiography were performed using a Chemiluminescence Western Blotting kit (Roche Diagnostics GmbH, Mannheim, Germany), according to the manufacturer's instructions. Blots were quantified using Quantity One software version 4.62 (Bio-Rad Laboratories, Inc., Hercules, CA, USA).

Whole-cell extracts were harvested and homogenized as aforementioned. Total protein (15 µg) from each well was boiled for 10 min in Laemmli buffer, separated via 10% SDS-PAGE, and then transferred onto a PVDF membrane using a semi-dry technique. The membranes were blocked for ~1.5 h with 3% bovine serum albumin (BSA) in phosphate-buffered saline (PBS) at room temperature, before they were incubated with biotinylated Maackia Amurensis Lectin II (MALII; dilution, 1:500) in PBS with Tween-20 buffer (PBST) for 1 h at room temperature. After washing in PBST buffer, the membranes were incubated with HRP-streptavidin (dilution, 1:1,000; Beyotime Institute of Biotechnology) for 1 h at room temperature and then washed with PBST buffer. Detection and autoradiography were performed using a Chemiluminescence Western Blotting kit (Roche Diagnostics GmbH), according to the manufacturer's protocol.

Cells were digested with 0.25% trypsin (Gibco; Thermo Fisher Scientific, Inc.) and harvested in 1.5 ml Eppendorf tubes. Cell density was adjusted to 2×10⁶/ml with lectin buffer (10 mM Hepes, 0.15 M NaCl, 0.08% sodium azide, 0.1 mM Ca²⁺ and 0.01 mM Mn²⁺, pH 7.5), before they were stained with 10 µg/ml MALII (Sigma-Aldrich; Merck-Millipore) in PBS (containing 0.5% BSA and 0.05% sodium azide) at 37°C for 120 min. The cells were then washed three times with PBS and incubated with phycoerythrin-avidin (Sigma-Aldrich; Merck-Millipore) at 37°C for 60 min. Following incubation, the cells were collected and fluorescence intensity was measured using a FACScan flow cytometer (BD Biosciences, Franklin Lakes, NJ, USA) and analyzed with BD CellQuest software version 7.5.3 (BD Biosciences).

1.5×10⁶ MDA-MB-231 cells and 1.5×10⁶ MDA-MB-231 cells that transfected with siRNA Cells were harvested and homogenized with lysis buffer as described in Western blot analysis. The primary rabbit monoclonal antibody against MMP14 (cat. no. ab51074, Abcam, Cambridge, UK) was added to the cell lysate and incubated at 4°C overnight. A total of 30 ml protein A agarose beads (Beyotime Institute of Biotechnology) was incubated with the antigen-antibody complex at 4°C for 2–4 h. The immunoprecipitation products were washed five times with cold PBS. Finally, the beads were washed with 5 µl PBS, boiled, and gently centrifuged at 2,000 × g for 2 min at room temperature, and the recovered samples were run on an 8% SDS-PAGE for lectin blotting using the aforementioned methods.

Drug resistance was measured by using the Cell Counting Kit-8 (CCK-8; C0038, Beyotime Institute of Biotechnology) assay. MDA-MB-231 cells and 1.5×10⁶ MDA-MB-231 cells that transfected siRNA-ppGalNAc-T1 and siRNA-ppGalNAc-T2, were plated onto 96-well plates (Corning Life Sciences, Tewksbury, MA, USA) at a density of 2×10⁵ and incubated with 0.1, 1, 10 and 100 ng/µl docetaxel (DOC; Jiangsu Aosaikang Pharmaceutical Co., Ltd., Nanjing, China) and 0.5, 2.5, 5 and 10 ng/µl gemcitabine hydrochloride (GEM; Qilu Pharmaceutical Co., Ltd., Jinan, China) anticancer drugs for 48 h. Cells were then treated with 100 ml CCK-8 (Beyotime Institute of Biotechnology) and incubated at 37°C and 5% CO₂ for 4 h. The absorbance value, at a wavelength of 450 nm, was measured using a microplate reader (Thermo Fisher Scientific, Inc.,). Each group contained three wells and was prepared in triplicate. The drug concentrations that resulted in 50% growth inhibition (the half maximal inhibitory concentration, IC₅₀ value) were determined.

In order to evaluate the invasion ability of the cells transfected with siRNA-ppGalNAc-T1 and siRNA-ppGalNAc-T2, Transwell plates (Corning Life Sciences) with a filter diameter of 6.5 mm and a pore size of 8.0 µm were employed. According to the manufacturer's instructions, cells (density, 1×10⁵) were plated into the upper chambers of 8-µm Transwells in serum-free medium, and incubated for 24 h to generate serum-starved cells. Following serum starvation, the culture medium was decanted and 500 µl DMEM was added to the upper chamber, while 500 µl DMEM plus 10% FBS was introduced to the lower compartment. The cells and Matrigel (included in the Transwell plates) were incubated at 37°C for 24 h, before washing twice with pre-heated PBS. The filters were subsequently fixed with 1 ml 1% paraformaldehyde for 15 min at room temperature. Following removal of the paraformaldehyde solution with PBS, the cells and Matrigel were stained with 500 µl eosin staining solution (Beyotime Institute of Biotechnology) for 20 min at room temperature, and then rinsed twice with distilled water. Cell penetration through the membrane was quantified by counting the number of cells that had diffused through the membrane in 10 fields of view (magnification, ×200) using a light microscope. Experiments were performed in triplicate.

Results are expressed as the mean ± standard deviation of triplicate experiments. Statistically significant differences were determined using SPSS version 16.0 software (IBM SPSS, Armonk, NY, USA). The independent samples t-test was performed to compare the data collected from independent samples. P<0.05 was considered to indicate a statistically significant difference.

In order to determine the expression level of glycosyltransferase enzymes in breast cancer cells, the expression levels of ppGalNAc-T1, T2, T3, T4, T7, T10 and T11 in highly invasive [estrogen receptor (ER)-negative] MDA-MB-231 cells and poorly invasive (ER-positive) MCF-7 cells, as well as in normal-like MCF-12A mammary cells. The mRNA expression level of these genes was detected by RT-qPCR analysis. As shown in Fig. 1A, the expression of ppGalNAc-T1, T2, T7 and T10 in MDA-MB-231 cells was significantly higher when compared with MCF-7 and MCF-12A cells (T1: P-value, MDA-MB-231 vs. MCF-7=0.035, MDA-MB-231 vs. MCF-12A=0.0096; T2: MDA-MB-231 vs. MCF-7=0.022, MDA-MB-231 vs. MCF-12A=0.0082; T7: MDA-MB-231 vs. MCF-7=0.028, MDA-MB-231 vs. MCF-12A=0.0075; T10: MDA-MB-231 vs. MCF-7=0.031, MDA-MB-231 vs. MCF-12A=0.0062). In addition, the expression of α2,3 sialic acid was detected by lectin blot analysis. As shown in Fig. 1B, α2,3-sialic acid was expressed in breast cancer cells and its expression in MDA-MB-231 cells was markedly higher when compared with MCF-7 and MCF-12A cells. Upregulation of glycosyltransferase enzymes promotes O-glycosylation, which may, in turn, lead to the increased production of terminal α2,3-sialic acids in highly invasive breast cancer cells (21). Considering the strong catalytic activity of glycosyltranferase enzymes on peptide substrates of ppGalNAc-T1 and -T2, the present study assessed the importance of ppGalNAc-T1 and -T2 in O-linked glycosylation initiation. In addition, the inhibitory effects of downstream sugar chain extension on tumor invasion and the metastatic potential of breast cancer cells was examined.

Of the identified members of the MMP family, MMP2 is an enzyme that has been reported to directly participate in the degradation of the extracellular matrix (ECM). By contrast, MMP14 is involved in tumor invasion and metastasis through activation of pro-MMP2 (22). In order to investigate the expression of metastasis-associated genes in breast cancer cells, the present study measured the expression of the MMP2, MMP9, MMP14, MMP15, TGFβ and VEGF genes in MDA-MB-231, MCF-7, and MCF-12A cell lines by RT-qPCR analysis. Fig. 2A shows that MMP14, TGFβ, and VEGF expression levels were significantly increased in highly invasive MDA-MB-231 cells, when compared with MCF-7 and MCF-12A cells. Upregulation of MMP14, TGFβ, and VEGF is associated with tumor invasion, metastasis and angiogenesis (23). No alterations in the expression levels of MMP2 among these cell lines were observed. Out of all MMP proteins examined, only MMP14 was significantly upregulated in MDA-MB-231 breast cancer cells at the mRNA and protein levels (Fig. 2A and B). These results suggest that MMP14 may serve an important role in the invasion and metastasis of breast cancer cells.

To further explore the role of ppGalNAc-T1 and ppGalNAc-T2 in breast tumors, highly invasive MDA-MB-231 cells were transiently transfected with siRNA-ppGalNAc-T1 and for siRNA-ppGalNAc-T2. The transfection efficiency was 72% and was determined by fluorescence microscopy (Fig. 3A). The optimal transfection system volume was determined using the GAPDH interference effect (Fig. 3B). ppGalNAc-T1 and ppGalNAc-T2 mRNA levels in MDA-MB-231 cells were then measured using RT-qPCR analysis. The expression levels of ppGalNAc-T1 and ppGalNAc-T2 were significantly downregulated following transfection with siRNA-ppGalNAc-T1-1327 and siRNA-ppGalNAc-T2-519 vectors, respectively, when compared with control vectors (Fig. 3C, P=0.0071, P=0.0093). These results demonstrated the optimal concentration and interference fragments required to silence the expression of ppGalNAc-T1 and ppGalNAc-T2 in MDA-MB-231 cells. The results presented were used to further examine the role of ppGalNAc-T1 and ppGalNAc-T2 in breast cancer cells.

ppGalNAc-Ts catalyze the attachment of the first GalNAc monosaccharide to the polypeptide during the initiation of O-linked glycosylation. The final step of O-linked glycosylation is terminal sialic acid acidification, and the metabolic flux through the sialic acid synthesis pathway is associated with substrate availability (11). Thus, the downregulated expression of ppGalNAc-T1 and ppGalNAc-2 by RNAi may reduce terminal sialic acidification. Fig. 4A demonstrates that the terminal sialic acid in the RNAi groups was downregulated when compared with the control group. In addition, downregulation of ppGalNAc-T1 in MDA-MB-231 cells resulted in a lower density of sialylated proteins when compared to that of the siRNA-ppGalNAc-T2 groups and the co-interference groups. The findings presented in the current study indicated that the sialic acid structure of O-glycans in glycoproteins generally decreased due to the reduction in substrate following the downregulation of ppGalNAc-T1. ppGalNAc-T1 serves an important role in O-glycosylation of MMP14 due to the presence of potential O-glycosylation sites at the hinge region of MMP14, which was demonstrated by the results of the lectin blot analysis of α2,3 sialic acids, including the 65 kDa protein of MMP14 proteins (Figs. 1B and 2B). Based on these results, the authors hypothesized that an α2,3 sialylation event had occurred at the hinge region of MMP14. Fig. 4B demonstrated the detection of α2,3 sialic acids on MMP14 using an immunoprecipitation assay, and the expression of α2,3 sialic acids in the RNAi groups was downregulated. The siRNA-ppGalNAc-T1 group exhibited a lower density of α2,3 sialic acids when compared with the siRNA-ppGalNAc-T2 and co-transfection groups. These results suggest that α2,3 sialylation occurs in the hinge region of MMP14, and terminal sialylation is regulated by the initiation of O-glycosylation.

Fig. 2 demonstrated that MMP14, VEGF, and TGFβ were upregulated in MDA-MB-231 cells when compared with MCF-7 and MCF-12A cells. MMP14 is a key enzyme involved in tumor invasion and angiogenesis through activation of pro-MMP2 and regulating VEGF (24). Incomplete glycosylation of this glycoprotein stimulates extensive autocatalytic degradation and self-inactivation of MMP14 (19). Due to the decrease in terminal α2,3 sialic acids on MMP14 by transient transfection of siRNA constructs against ppGalNAc-T1 and ppGalNAc-T2, the expression of MMP14, MMP2 and VEGF were measured by RT-qPCR and western blot analyses. Fig. 5A shows that MMP9, VEGF, and TGFβ mRNA expression was downregulated. MMP14 may activate MMP2 at post-transcription level; therefore, the MMP2 mRNA expression remained unchanged. Fig. 5B shows that MMP14 protein expression was downregulated in the RNAi groups, which was consistent with the immunoprecipitation results presented in Fig. 4B. In addition, alterations in the protein expression patterns of MMP2 and VEGF were similar to that of MMP14 in all experimental groups. Thus, decreasing α2,3 sialylation by silencing ppGalNAc-T1 may lead to extensive autocatalytic degradation and self-inactivation of MMP14.

Chemotherapy resistance is a major obstacle for the successful treatment of TNBC with chemotherapeutic agents. The effect of ppGalNAc-T expression on the sensitivity of TNBC cells to anticancer drugs remains elusive. In order to examine its effect in more detail in the present study, the expression of ppGalNAc-T1 and ppGalNAc-T2 was inhibited by RNAi in MDA-MB-231 cells prior to treatment with DOC (0.1–100 µg/ml) or GEM (0.5–10 µg/ml) for 48 h. As shown in Fig. 6A, downregulation of ppGalNAc-T1 or ppGalNAc-T2 resulted in a greater reduction in the proliferation of MDA-MB-231 cells treated with DOC and GEM, when compared cells treated with DOC and GEM alone (P-values, 1 µg/ml DOC, MDA-MB-231-si-T2: MDA-MB-231=0.003, MDA-MB-231-si-T1: MDA-MB-231=0.042; 10 µg/ml DOC, MDA-MB-231-si-T2: MDA-MB-231=0.0026, MDA-MB-231-si-T1: MDA-MB-231=0.038; 100 µg/ml DOC, MDA-MB-231-si-T2: MDA-MB-231=0.041, MDA-MB-231-si-T1: MDA-MB-231=0.040; P-values, 0.5 µg/ml GEM, MDA-MB-231-si-T2: MDA-MB-231=0.031; 2.5 µg/ml GEM, MDA-MB-231-si-T1: MDA-MB-231=0.002; 5 µg/ml GEM, MDA-MB-231-si-T1: MDA-MB-231=0.034, MDA-MB-231-si-T2: MDA-MB-231=0.001; 100 µg/ml GEM, MDA-MB-231-si-T2: MDA-MB-231=0.046, MDA-MB-231-si-T1: MDA-MB-231=0.044). The IC₅₀ values of the drugs were significantly lower in the ppGalNAc-T1 siRNA (IC_DOC50=7.3, P=0.0005 vs. untreated MDA-MB-231 group; IC_GEM50=6.2, P=0.021 vs. untreated MDA-MB-231 group) or ppGalNAc-T2 siRNA (IC_DOC50=3.8, P=0.0002 vs. untreated MDA-MB-231 group; IC_GEM50=5.7, P=0.014 vs. untreated MDA-MB-231 group) group compared with those in the untreated MDA-MB-231 group (IC_DOC50=19.2; IC_GEM50=8.1). Therefore, downregulation of ppGalNAc-T1 or ppGalNAc-T2 in MDA-MB-231 cells resulted in increased chemosensitivity to anti-tumor drugs.

Flow cytometry assays were performed to evaluate the inhibitory effect of ppGalNAc-T1 and ppGalNAc-T2 on terminal α2,3 sialylation. Cells were treated with 10 µg/ml DOC for 48 h. Fig. 6B demonstrated a marked reduction in cell surface α2,3 sialic acids between the control group and the DOC treatment group. In addition, knockdown of ppGalNAc-T1 and ppGalNAc-T2 inhibited sialylation further following DOC treatment. These results indicate that ppGalNAc-T1 and ppGalNAc-T2 may be involved in mediating drug resistance in TNBC cells by regulating the generation of terminal α2,3 sialic acids by O-glycosylation.

The results presented so far, demonstrated that MMP14 expression and associated α2,3 sialylation were affected by the downregulation of ppGalNAc-T1 and ppGalNAc-T2. MMP14 is a crucial membrane enzyme that participates in tumor invasion and migration (24). Therefore, the effect of ppGalNAc-T1 and ppGalNAc-T2 expression on the capacity of breast carcinoma cells to metastasize was investigated in the present study. Transwell invasion chambers coated with Matrigel were used. Fig. 7 shows that the RNAi groups exhibited a significant reduction in invasion capabilities when compared with that of the control groups (P=0.013 <0.05). In addition, the inhibitory effect of the siRNA-ppGalNAc-T1 group was significantly higher when compared with that of the siRNA-ppGalNAc-T2 and co-interference groups (P=0.022; P=0.031). These results suggest that ppGalNAc-T1 may regulate the metastatic behavior of breast cancer cells by regulating the initiation of O-glycosylation. In addition, decreasing α2,3 sialylation through silencing of ppGalNAc-T1 may lead to extensive autocatalytic degradation and self-inactivation of MMP14. Furthermore, the glycosyltransferase ppGalNAc-T1 may serve a more important role in the glycosylation-initiation process than ppGalNAc-T2.