NMuMG epithelial cells were obtained from the American Type Culture Collection (Manassas, VA, USA) and cultured in Dulbecco’s modified Eagle’s medium (DMEM; HyClone, Logan, UT, USA) containing 10% fetal bovine serum (HyClone), 10 μg/ml insulin (Sigma-Aldrich, St. Louis, MO, USA) and 1X penicillin/streptomycin (Gibco-BRL, Carlsbad, CA, USA) at 37°C in 5% CO₂.

All samples from the patients with breast cancer and the healthy controls were collected at the First Affiliated Hospital of Xi’an, Jiaotong University (Xi’an, China). The patient characteristics are shown in Table I. The present study followed the tenets of the Declaration of Helsinki for the Use of Human Subjects. The present study was approved by the Ethics Committee of Jiangnan University (Wuxi, Jiangsu, China) and written informed consent was obtained from all patients and healthy donors.

The following primers were used to amplify the gene for mouse Smad3: Smad3p28, forward 5′-GGAATTTCATATTCGTCCATCCTGCCC-3′, NdeI and Smad3cDNA, reverse 5′-CCGCTCGAGACCCGCTCCCTTTACTCCTA-3′, XhoI. The following primers were used were used to amplify the gene for mouse Smad4: Smad4p28, forward 5′-GGAATTTCATATGGACAATATGTCTATAACAAATAC-3′, NdeI and Smad4cDNA, reverse 5′-GGAATTCCTGAGATCTCAGTCTAAAGGCT-3′, EcoRI. The polymerase chain reaction (PCR) fragments were inserted into the expression vector pET-28a (+) to generate pET28-Smad3 and pET28-Smad4 and introduced into E. coli BL21 (DE3) for protein expression. The recombinant His₆-tagged Smad3 and Smad4 proteins produced by isopropyl β-D-1-thiogalactopyranoside induction were purified on a Ni²⁺-NTA spin column (Roche Diagnostics, Basel, Switzerland).

EMSAs were performed using a 2nd Generation DIG Gel Shift kit (Roche Diagnostics) according to the manufacturer’s instructions. The probes were amplified using PCR and labeled with digoxigenin (DIG) at the 3′-terminal end. The probes, proteins and poly(deoxyinosinic-deoxycytidylic) in binding buffer [100 mM Hepes, pH 7.6; 5 mM EDTA; 50 mM (NH₄)₂S0₄; 5 mM DTT; Tween 20, 1% (w/v); 150 mM KCl; Roche Diagnostics, Basel, Switzerland] were mixed and incubated at 25°C for 30 min prior to adding 5 μl loading buffer. The protein-DNA complex and free DNA were separated on native 5% polyacrylamide gels and transferred onto nylon membranes. The membranes were treated according to the manufacturer’s instructions and the bands on the membranes were detected using a Chemi Doc XRS chemiluminescent imaging system (Bio-Rad, Hercules, CA, USA).

ChIP assays were performed, as described previously (18). In brief, NMuMG cells were grown in regular medium (DMEM medium containing 10% fetal bovine serum) for 24 h, following which the medium was replaced by regular medium containing 2 ng/ml TGFβ and the cells were incubated for 48 h at 37°C and then fixed in DMEM containing 1% formaldehyde. ChIP was performed using the anti-Smad4 antibody. Anti-Pol II antibody immunoprecipitated with a Pol II-actin promoter complex was used as a positive control. DNA was extracted and pellets were purified using a Quick DNA purification kit (Beijing CoWin Biotech. Co., Beijing, China). The DNA template (2 μl) was used for PCR with the following primers: β3GalT4pL, forward 5′-GGTGTGTTAGGGGACTGGT-3′ and reverse 5′-TGGACTGTGCAGCCTGAT-3′ for the β3GalT4 promoter; pNon, forward 5′-CTGAGGGTCTTGAGGGTGAG-3′ and reverse 5′-CTCTTCCTCCTGGGAAAACC-3′ for the nonspecific DNA fragment and pactinp forward 5′-TCAATCTCGCTTTCTCTCTCG-3′ and reverse 5′-CAA CGAAGGAGCTGCAAAG-3′ for the actin promoter.

The protein coding regions of the genes were amplified by PCR using the following primers: Smad3cDNA, forward ‘5-CCCAAGCTTGCCACCATGTCGTCCATCCTGCCC-3′, HindIII, and Smad3cDNA, reverse for the smad3 gene. Smad4cDNA, forward ‘5-GGGGTACCCCCTTGAACAAATGGACAATATGT-3′, KpnI, and Smad4cDNA, reverse for the smad4 gene. The products were digested and ligated into vector pcDNA3.1 (Invitrogen Life Technologies, Carlsbad, CA, USA). The constructed plasmids were transfected into NMuMG cells using Lipofectamine 2000 (Invitrogen Life Technologies). Stable transfectants were selected by screening with the antibiotic G418 and confirmed by western blot analysis.

Cells were harvested and lysed in radioimmunoprecipitation assay buffer. Protein lysates were analyzed by SDS-PAGE and western blot analysis, as described previously (10).

RNA was extracted using an RNApure Tissue kit (Beijing CoWin Biotech, Co.). RNA samples were treated with DNase I and assessed by PCR to rule out chromosomal DNA contamination. Each RNA sample was reverse transcribed using ReverTra Ace-α-^® (Toyobo, Shanghai, China). Semiquantitative and quantitative RT-PCR analysis were performed to determine the transcription levels of various genes using the following primers: β3GalT4real, forward 5′-CTCTTCCTCCTGGGAAAACC-3′ and reverse 5′-CTGAGGGTCTTGAGGGTGAG-3′ for the β3GalT4 gene and Tubulinreal, forward 5′-ATCTACCTGTCGGAGCATGG-3′ and reverse 5′-GCCTCCCGATCTATGATGTC-3′ for the Tubulin gene. The following thermocycler conditions were used for RT-qPCR: 95°C for 10 min, 40 cycles of 95°C for 10 sec and 60°C for 1 min in a 15 μl reaction system using an UltraSYBR mixture (Beijing CoWin Biotech, Co.). The DNA products were analyzed using CFX manager software (version 3.0.1224.1015; Bio-Rad).

GSL extraction, high performance thin layer chromatography (HPTLC) analysis and immunostaining were performed, as described previously (19). Cells were harvested following washing with PBS, extracted with 2 ml isopropanol/hexane/water (55:25:20) by sonication for 30 min and centrifuged at 1082 × g for 5 min. The extracts were dried in a nitrogen stream. Redissolved GSLs were incubated in 0.1 M NaOH in methanol at 40°C for 2 h to hydrolyze the phospholipids and then neutralized with 1 M HCl. The hydrolyzed phospholipids were removed by hexane and the remaining solution was evaporated and dissolved in 1 ml distilled water. The solution was applied to a Sep-Pak C18 cartridge (Varian Medical Systems, Palo Alto, CA, USA) and washed with water. The total GSLs were eluted with 2 ml chloroform/methanol (2:1) and stained using 0.5% orcinol in 2 M sulfuric acid. Gg4 bands were then immunostained using mAb TKH7 on HPTLC plates (EMD Biosciences, Inc., San Diego, CA, USA).

Data were analyzed using a two-tailed two-sample t-test assuming equal variance using GraphPad Prism 5 software (GraphPad Software Inc., La Jolla, CA, USA). P≤0.05 was considered to indicate a statistically significant difference.

Our previous study demonstrated that the transcription levels of β3GalT4 were decreased during TGFβ-induced EMT in NMuMG cells (9). To determine the expression of β3GalT4 in breast cancer, the expression of β3GalT4 at the mRNA level in different clinical samples was determined using RT-qPCR. As shown in Table I, a significant reduction in the expression of β3GalT4 was detected in the patients with breast cancer compared with the healthy control subjects. This suggested that decreased expression of the β3GalT4 gene is often present during breast cancer progression.

The Smad3 and Smad4 proteins have a ‘Mad-homology 1’ domain, a DNA-binding module stabilized by a bound zinc atom, at the N-terminus (11). Smad3 and Smad4 can bind to DNA directly via the SBE (16). A previous study of the β3GalT4 promoter (17) revealed a Smad4 binding site between positions -788 and -795 (5′-GTCTAGAC-3′) relative to the β3GalT4 transcriptional start point (Fig. 1A). In the present study, EMSAs and ChIP assays were performed to identify the interactions between Smads and the β3GalT4 promoter.

EMSAs were performed using full-length recombinant His₆-Smad3 and -Smad4 proteins expressed in E. coli. The probe used was a 237 bp DNA fragment of the β3GalT4 promoter, amplified using the forward and reverse β3GalT4pL primers and labeled with DIG. The probe was clearly retarded on the EMSA gel by the Smad3/4 complex and became more notable as the quantity of Smad3 increased (Fig. 1B). This finding indicated that Smad4 alone bound to the probe with low affinity, whereas the Smad3/4 complex enhanced the retardation more specifically. To eliminate nonspecific binding, unlabeled probes were used in a competitive EMSA assay. The retarded band was eliminated completely by a 100 or 200-fold excess of unlabeled β3GalT4 promoter probe, whereas an unlabeled nonspecific probe, amplified by the forward and reverse pNon primers had no effect (Fig. 1C). Xia et al (17) reported that a DNA fragment (5′-GTCTAGAC-3′) of the β3GalT4 promoter is a Smad4 binding site. To assess the relative contribution of the SBE, EMSAs were performed using an unlabeled probe containing either the intact SBE (β3GalT4p) or a mutated sequence lacking SBE (β3GalT4pm). The affinity of Smad3/4 for β3GalT4pm was eliminated completely compared with β3GalT4p (Fig. 1D). These findings indicated that the SBE is essential for Smad3/4 binding activity.

ChIP assays are commonly used to determine DNA-protein binding regions. In the present study, NMuMG cells were treated with TGFβ and formaldehyde was used to fix the cross-linking between the Smad3/4 proteins and their DNA target. The cross-linked DNA was extracted and fragmented by sonication and an anti-Smad4 antibody was used to screen the DNA fragments attached to the Smad4 protein. The forward and reverse β3GalT4pL primers used in the above EMSAs were also used in the ChIP assays. PCR products of the correct size were obtained from the input DNA and the immunoprecipitated DNA, whereas no such PCR bands were detected in the control experiments, which used the same DNA fragment as the template, but without specific antibody addition. Negative control bands were amplified only with input DNA using the non-specific primers used in the EMSAs (Fig. 1E). Antibodies directed to Pol II, an enzyme responsible for transcription of protein-coding genes, were used as positive controls to ensure the accuracy of the ChIP results. Forward and reverse pactinp primers were used to detect the PCR band. Taken together, the findings from the EMSAs and ChIP assays indicated that the Smad3/4 complex bound specifically to the β3GalT4 promoter directly through the SBE.

Our previous study demonstrated that the β3GalT4 gene, responsible for the expression of Gg4, is downregulated during the EMT process in NMuMG cells (9). Smads are key factors in this process. In the present study, to clarify the association between Smad3/4 and β3GalT4, Smad3 and Smad4 overexpression cells were constructed. NMuMG cells containing empty vector pcDNA3.1 were described as ‘mock’.

The two transfectants, when cultured in normal DMEM, exhibited a flattened epithelial morphology similar to that of the mock (Fig. 2A), indicating that neither Smad3 nor Smad4 affected the expression of genes that control cell morphology. Exogenously added TGFβ activated the TGFβ/Smads signaling pathway and altered cell shape, with the transfectants and mock cells converted to a fibroblastic morphology (Fig. 2A).

Reduced expression of the epithelial marker E-cad and increased expression of the mesenchymal markers N-cad and vimentin are characteristic processes of EMT. Cells overexpressing Smad3 and Smad4 demonstrated a markedly enhanced expression of vimentin and N-cad and a reduced expression of E-cad and β-catenin, which forms a complex with E-cad stabilizing the cell-cell junction (Fig. 2B and C). TGFβ treatment intensified the enhanced and reduced expression of the EMT markers mentioned above. These findings suggested that the genes controlling these markers are regulated by Smads and involved in the TGFβ/Smads pathway via different regulatory mechanisms.

To determine whether the β3GalT4 gene is positively or negatively regulated by Smad3/4, the transcription levels of β3GalT4 in two transfectants, as above, and mock were evaluated by semiquantitative and quantitative RT-PCR using the forward and reverse β3GalT4real primers. TGFβ treatment caused a reduction in β3GalT4 expression (Fig. 3A and B) and the degree of reduction was greater in the transfectants compared with the mock (Fig. 3B). These findings suggested that, following activation of the TGFβ/Smads signaling pathway by TGFβ, the Smads complex is translocated into the nucleus, binds to the β3GalT4 promoter and downregulates its transcription.

In our previous study, the mRNA levels of Gg4 and β3GalT4 were reduced during TGFβ-induced EMT and the reduction in Gg4 was associated with decreased expression of E-cad and β-catenin (9). In the present study, total GSL fractions from transfectants and mock, with or without TGFβ treatment, were prepared and analyzed using HPTLC. The level of Gg4 expression decreased in parallel with the levels of E-cad and β-catenin (Fig. 3C and D). These findings suggested that the Smad3/4 complex affects the expression of Gg4 by suppressing mRNA levels of β3GalT4 and that Gg4 interacts closely with the E-cad/β-catenin complex to stabilize the cell-cell junction.