Silencing of GM3 synthase suppresses lung metastasis of murine breast cancer cells

67NR and 4T1 (kindly provided by Dr Fred R Miller at Karmanos Cancer Institute) are mammary tumor cell lines characterized by enhanced lung metastatic potential that were derived from a single, spontaneously arising mouse mammary tumor [25]. The nonmetastatic 67NR cells fail to leave the primary site, whereas the 4T1 cells can metastasize to lung, liver, bone, and brain via the hematogenous route. They were cultured in Dulbecco's modified Eagles medium supplemented with 10% fetal calf serum (DME-10), 1 mmol/l mixed nonessential amino acids, and 2 mmol/l L-glutamine.

The dual promoter small interfering RNA (siRNA) expression vector pMEHMpuro was constructed based on a self-inactivating murine stem cell virus plasmid, namely pMSCVpuro. First, the HindIII restriction site of pMSCVpuro was deleted by digesting the plasmid with HindIII, blunting the 3' recessed ends and re-ligating. The enhanced green fluorescent protein (EGFP) gene was then inserted into the BglII/XhoI sites of HindIII restriction site deleted pMSCVpuro vector; the resulting construct was named pMSCVpuro-EGFP. Then, the dual promoter siRNA expression cassette was constructed as described previously [26] and inserted into the XhoI/EcoRI-digested pMSCVpuro-EGFP plasmid to form pMEHMpuro. In order to generate siRNA vectors, three pairs of oligonucleotides, corresponding to different regions of murine GM3S encoding gene, were annealed and inserted into BglII/HindIII double digested pMEHMpuro: siGM3S1, 5'-AGC TTA AAA AGT AAG GTT GAA CAG TGC GCC TTT TTA-3' and 5'-GAT CTA AAA AGG CGC ACT GTT CAA CCT TAC TTT TTA-3'; siGM3S2, 5'-AGC TTA AAA AGA CTG CCT TCG ACA TCC TTC TTT TTA-3' and 5'-GAT CTA AAA AGA AGG ATG TCG AAG GCA GTC TTT TTA-3'; and siGM3S3, 5'-AGC TTA AAA AGT GTG ACC ACA GAG ACC AAG TTT TTA-3' and 5'-GAT CTA AAA ACT TGG TCT CTG TGG TCA CAC TTT TTA-3'. This process yielded the vectors pMEHM-siGM3S1 (siGM3S1), pMEHM-siGM3S2 (siGM3S2), and pMEHM-siGM3S3 (siGM3S3), respectively. In addition, annealed oligonucleotides (5'-AGC TTA AAA AGT TCC GTA TGT TGC ATC ACC TTT TTA-3' and 5'-GAT CTA AAA AGG TGA TGC AAC ATA CGG AAC TTT TTA-3') that did not match any known mouse gene was inserted into BglII/HindIII double digested pMEHMpuro, yielding pMEHM-siMock. The full-length GM3S encoding gene was amplified from the total RNA of 4T1 cells using RT-PCR and inserted into XhoI/EcoRI double digested pMSCVpuro, resulting in the GM3S expression vector pMSCV-GM3S. All of the resulting constructs were confirmed by DNA sequencing.

4T1 cells were transfected with various GM3S-targeting RNA interference and mock plasmids. 67NR cells were transfected with GM3S and pMSCVpuro vectors. Transfection was carried out using LipofectAMINE 2000 (Invitrogen, Carlsbad, CA, USA), and the transfected cells were cultured in the medium supplied with 4 μg/ml puromycin for 2 weeks. The obtained cells were named 4T1-siMock, 4T1-siGM3S1, 4T1-siGM3S2, 4T1-siGM3S3, 67NR-Control and 67NR-GM3S.

Cells were lysed in lysis buffer (50 mmol/l Tris-HCl [pH 7.4], 150 mmol/l NaCl, 1% NP40, 1 mmol/l EDTA, 1 mmol/l Na₃VO₄, 10 mmol/l NaF) containing a protease inhibitor cocktail (Roche, Nutley, NJ, USA). Protein samples (50 μg) were separated by 12% SDS-PAGE and transferred to Immobilon-P membranes (Millipore, Bedford, MA, USA). Antibodies to phosphorylated and total Akt (Cell Signaling, Beverly, MA, USA), phosphorylated (Ser380/Thr382/383) and total PTEN (Cell Signaling), NFAT1 (Santa Cruz, CA, USA), and GAPDH (Santa Cruz) were used, with detection by ECL-detecting reagent (Amersham Biosciences, Buckinghamshire, UK). Quantification of the blots was conducted using Image-Pro Plus software (version 6.0; Media Cybernetics, Bethesda, MD, USA).

The flow cytometry procedure was as described previously [27]. Briefly, cells were trypsinized and incubated with anti-GM3 or anti-GD3 monoclonal antibodies (Seikagaku Corporation, Tokyo, Japan) and then labeled with rat anti-mouse IgM-P-RE (Southern Biotechnology Associates, Birmingham, AL, USA). A total of 1 × 10⁴ labeled cells were analyzed using a cytometer (Becton Dickinson, Franklin Lakes, NJ, USA).

Cells were seeded 1 × 10⁴ per well in a 96-well plate. Cells were allowed to grow for 24 hours. Then, 20 μl of 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) (5 mg/ml) was added to each well. After 4 hours of incubation at 37°C, cells were lysed by addition of 200 μl dimethylsulfoxide. Absorbance was measured at 570 nm using a Rainbow microplate reader (Tecan, Groding/Salzburg, Austria).

Cells (5 × 10³ to 1 × 10⁴) were suspended in 1 ml top agar medium (DME-10 supplied with 0.4% agar), in the presence or absence of phosphoinositide-3 kinase (PI3K)-specific inhibitor LY294002 (Merck, Nottingham, UK). The cell suspension was then overlaid onto 1.5 ml bottom agar medium (DME-10 supplied with 0.8% agar) in six-well tissue culture plates in triplicate. DME-10, with or without LY294002, was added to the plates every 3 days as a feeder layer. On day 12, the number of colonies was counted in six random fields at 40× magnification.

Cell migration was assayed using Transwell chambers (6.5 mm; Corning, New York, USA) with 8 μmol/l pore membranes. The lower chamber was filled with 600 μl NIH-3T3 conditioned medium containing 20 μg/ml fibronectin (BD Biosciences, Bedford, MA, USA) with or without 2 μmol/l LY294002. Cells (5 × 10⁴) were suspended with 100 μl upper medium (Dulbecco's modified Eagles medium with 1% fetal calf serum) and plated into the upper chamber with or without 2 μmol/l LY294002. After 16 hours, the number of cells appearing by crystal violet staining on the undersurface of the polycarbonate membranes was scored visually in five random fields at 100× magnification using a light microscope.

For invasion assays, the upper face of the membrane was covered with 70 μl Matrigel (1 mg/ml; BD Biosciences). The invasion assay procedure was the same as for the migration assay, except that the incubation time of the experiment was prolonged to 24 hours.

These procedures were performed as described previously [25, 28] with minor modifications. Female BALB/c mice, aged 8 to 10 weeks, were used in the experiment. In brief, mice (six to eight per group) were anesthetized with sodium pentobarbital (50 mg/kg body weight), and tumor cells (5 × 10⁵) in 10 μl DME-10 were injected into the mammary gland. The weight of the primary tumors and the number of metastatic nodules on the lung surface were evaluated 30 days after the tumor cells injection. The animals were housed and cared for in accordance with the guidelines established by the National Science Council of Republic of China. The lung tissues (six lungs per group) from in vivo experiments described above were fixed in 10% neutral buffered formalin and embedded in paraffin, and then thick sections (4.0 μm; three sections per lung tissue) were cut and stained with hematoxylin and eosin.

We found that the expression of GM3S mRNA was greater in the highly metastatic 4T1 tumor cells than in the nonmetastatic 67NR cells (Figure 1a). These observations led us to regard GM3S to be a potential candidate for orchestrating metastasis in breast cancer. To determine whether GM3S played an important role in breast cancer malignancy, expression of GM3S in the highly metastatic 4T1 cell line was silenced by RNA interference. The expression of the GM3S siRNAs siGM3S2 and siGM3S3, but not that of siGM3S1, drastically reduced the expression of GM3S mRNA in 4T1 cells (Figure 1b). The results of flow cytometry analysis indicate that GM3S knockdown significantly suppressed the expression of the ganglioside GM3 (Figure 2a) and the more complex ganglioside GD3 (Figure 2b). On the other hand, GM3S was over-expressed in 67NR cells. The expression of GM3S was upregulated (Figure 1c) and the levels of expression of GM3 (Figure 2c) and GD3 (Figure 2d) gangliosides were enhanced. These GM3S-reconstructed cells enabled us to identify the effects of GM3S expression on the malignant properties of breast cancer cells, and these cell lines also appeared to be useful for analyzing the roles played by gangliosides in breast cancer cells.

To investigate whether GM3S expression affected cell proliferation and anchorage-independent growth, MTT and soft agar colony assays were conducted. Silencing of GM3S in 4T1 cells and over-expression of GM3S in 67NR cells did not affect the proliferation (Figure 3a,b). The results were also verified by cell counting experiments (data not shown). Interestingly, GM3S knockdown in 4T1 cells significantly reduced the number of colonies and slightly decreased the colony size (Figure 3c), whereas 67NR cells over-expressing GM3S formed much larger colonies than did control cells, and the number of colonies was slightly increased (Figure 3d). These results indicated that endogenous GM3S was important for maintaining the anchorage-independent growth of breast tumor cells.

To investigate the role played by GM3S in breast cancer cell metastasis, we conducted migration and invasion assays in vitro. As shown in Figure 4a and 4b, the migration and invasion abilities were inhibited by silencing of GM3S in 4T1 cells. Over-expression of GM3S in noninvasive 67NR cells enhanced their ability to migrate (Figure 4c). However, over-expression of GM3S was unable to transform noninvasive 67NR cells into invasive ones (data not shown). These findings suggest that GM3S expression is essential but not sufficient for breast cancer cell migration and invasion.

Based on the roles played by GM3S expression in migration/invasion and anchorage-independent growth of breast cancer cells described above, we next examined the effects of GM3S silencing on tumor formation and metastasis of breast cancer cells. Consistent with the results in vitro, GM3S silencing in 4T1 cells had no effect on primary tumor weight (Figure 5a) but it dramatically reduced the number of visible metastatic nodules on the lung surface of tumor-bearing mice (Figure 5b,c). Histologic analyses confirmed that the number of micrometastatic lesions was significantly reduced in lungs of mice carrying 4T1-siGM3S2 and 4T1-siGM3S3 tumors (Figure 5d).

To investigate the molecular mechanism underlying GM3S-mediated migration and invasion, some metastasis-associated signaling pathways were assayed by immunoblotting. We found that suppression of GM3S significantly enhanced the phosphorylation of Akt at two key residues, namely Thr308 and Ser473 (Figure 6a). Further studies demonstrated that the PI3K specific inhibitor LY294002 (2 μmol/l) restored Akt phosphorylation to the level exhibited by 4T1-siMock cells (Figure 6b) and partially compensated for the inhibition of cell migration/invasion induced by the GM3S knockdown (Figure 6c).

PTEN is a major negative regulator of the PI3K/Akt signaling pathway. In the present study, GM3S inhibition suppressed the expression and phosphorylation of PTEN in 4T1 cells (Figure 7a,b). These findings suggest that the inhibition of PTEN expression at least partially contributed to activation of the PI3K/Akt pathway in 4T1 cells in which GM3S had been knocked down.

We further investigated the downstream effectors of PI3K/Akt pathway. GM3S knockdown inhibited NFAT1 expression (Figure 8a), and inhibition of the PI3K/Akt pathway induced NFAT1 expression (Figure 8b). In conclusion, as illustrated in Figure 9, GM3S knockdown suppressed PTEN expression and resulted in activation of Akt signaling, which reduced the expression of NFAT1 and inhibited cell migration and invasion.

To test whether activation of the PI3K/Akt pathway was responsible for the anchorage-independent growth affected by GM3S knockdown in 4T1 cells, we inactivated the PI3K/Akt pathway with LY294002. The colony forming abilities of both GM3S-knockdown 4T1 cells and the mock transfected cells were unaffected by inhibition of the PI3K/AKT pathway (Figure 10), indicating that the anchorage-independent growth caused by GM3S was not associated with the PI3K/Akt pathway. Extra pathway(s) may exist to compensate for the PI3K/Akt pathway with respect to anchorage-independent survival of breast tumor cells.