Matrigel and type I collagen were purchased from BD Biosciences (Bedford, MA, USA) and Sigma-Aldrich Chemical Co (St. Louis, MO, USA), respectively. Hema-3 was purchased from Fisher Scientific (Middleton, VA, USA). Rabbit polyclonal GST-π and GST-α antibodies were purchased from EMD Biosciences (La Jolla, CA, USA). β-actin monoclonal antibody was purchased from Sigma-Aldrich Chemical Co. The polyclonal antibody to TIMP-2 was purchased from R & D Systems (Minneapolis, MN, USA). Phospho(Thr²⁰²/Tyr²⁰⁴)-extracellular signal-regulated kinase 1/2 (ERK1/2), ERK1/2, phospho(Thr¹⁸⁰/Tyr¹⁸²)-p38, p38, phospho(Thr¹⁸³/Tyr¹⁸⁵)-c-Jun N-terminal kinase (JNK), and JNK antibodies were purchased from Cell Signaling (Beverly, MA, USA).

JS-K 20 and JS-43-126 21, a JS-K analog that does not release NO, were prepared as previously described. Stock solutions of JS-K and JS-43-126 were prepared in dimethylsulfoxide and were stored at -20°C. The structures of JS-K and JS-43-126 are presented in Figure 1.

The human MDA-MB-231 breast cancer cell line was obtained from American Type Cell Culture (Manassas, VA, USA). The MDA-MB-231 cell line is an estrogen-independent, highly metastatic human breast cancer cell line. Breast cancer commonly metastasizes to the skeletal system. MDA-MB-231/F10 (F10) is a bone metastatic derivative of MDA-MB-231 cells selected in vivo by repeated intracardiac injections of the MDA-MB-231 cells into female nude mice until no micrometastases were detected histologically in any organs other than bone 22. The F10 cell line was kindly provided by Dr Toshiyuki Yoneda (The University of Texas Health Science Center, San Antonio, TX, USA).

Breast cancer also commonly metastasizes to lymph nodes. Elevated COX-2 expression in invasive breast tumor is associated with lymph node metastasis 232425. MCF-7/COX-2 cells are estrogen-dependent MCF-7 cells stably transfected with plasmids encoding the human COX-2 gene 26. The parental MCF-7 cells are poorly invasive but the MCF-7/COX-2 cells are highly invasive 27.

The MDA-MB-231 and F10 cell lines were cultured in DMEM/F12 (Invitrogen, Carlsbad, CA, USA) supplemented with 5% heat-inactivated FBS) (Invitrogen) at 37°C under 5% carbon dioxide in a humidified incubator. MCF-7/COX-2 cells were continuously cultured in DMEM/F12 medium containing 5% FBS and 500 μg/ml antibiotic G418 (Invitrogen).

Protein lysates (50 μg) from untreated exponentially growing MDA-MB-231, F10, and MCF-7/COX-2 breast cancer cells were loaded onto 15% polyacrylamide gels to determine the expression of GST-π and GST-α. The MDA-MB-231 cells (3 × 10⁵ cells), F10 cells (3 × 10⁵ cells), and MCF-7/COX-2 cells (4 × 10⁵ cells) were plated in T25 flasks in 5 ml DMEM/F12 medium supplemented with 5% FBS. The next day, cells were treated with JS-K (0.5 and 1.0 μM) for 24 hours. Protein lysates (30 μg) were loaded onto 12% polyacrylamide gels to determine the activity and expression of ERK1/2, p38, and JNK mitogen-activated protein kinases. Proteins were electrophoresed and electrotransferred as described previously 10. Membranes were incubated with the appropriate antibodies. β-actin was used as a loading control. Protein bands were visualized by enhanced chemiluminescence (Kirkgaard & Perry Laboratories, Gaithersburg, MD, USA). Images were scanned and quantified by an Alpha Innotech densitometer using the Alpha Imager application program (San Leandro, CA, USA).

The MDA-MB-231 cells (3 × 10⁵ cells), F10 cells (3 × 10⁵ cells), and MCF-7/COX-2 cells (4 × 10⁵ cells) were plated in T25 flasks in 5 ml DMEM/F12 medium supplemented with 5% FBS. The next day, cells were treated with JS-K (0.5 and 1.0 μM) or JS-43-126 (0.5 and 1.0 μM) for 72 hours. The medium was recovered, centrifuged for 5 minutes, and concentrated using spin columns with 10-kDa-cutoff filters (Millipore, Bedford, MA, USA). Total NO was determined in the conditioned concentrated supernatants by quantifying nitrite, the stable end product of NO oxidation, spectrophotometrically using a colorimetric nonenzymatic nitric oxide assay kit (Oxford Biomedical Research, Oxford, MI, USA) as described previously 9. Cell growth was determined by total live cell counts using trypan blue exclusion. Nitrite values were normalized for total cell counts and expressed as picomoles per 10⁶ cells. The experiments were performed in triplicate wells.

One hundred microliters of Matrigel (0.7 mg/ml) were added to each well of 96-well plates. The MDA-MB-231 cells, F10 cells, and MCF/COX-2 cells (1 × 10⁴ cells resuspended in 100 μl medium) were added to Matrigel-coated wells. The next day, JS-K (0, 0.5 and 1.0 μM) was added to cells in pentaplicate wells. After 3 days of incubation, cell proliferation was determined by the Promega (Madison, WI, USA) Celltiter 96^® AQ_ueous nonradioactive cell proliferation assay. The CellTiter 96^® AQ_ueous Assay is composed of 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium (MTS) and the electron coupling reagent phenazine methosulfate. MTS is reduced by live cells into a formazan product, whose absorbance can be read at 490 nm. The quantity of formazan product is directly proportional to the number of living cells in culture. The absorbance of the formazan product was read within 2 hours after the MTS/phenazine methosulfate dye addition. Cell proliferation was expressed as the percentage of untreated cells. The experiments were repeated twice.

The effect of JS-K on breast cancer invasion was determined in vitro using modified Boyden chamber assays as previously described 28. Briefly, six-well plate transwell inserts with 8-μm pore-size polycarbonate filters (Fisher Scientific) were coated with Matrigel (0.7 mg/ml) or type I collagen (20 μg/ml) in cold serum-free DMEM/F12 medium and were placed at room temperature for 40 minutes. Cells were trypsinized, resuspended in serum-supplemented media, and were then counted.

Cells were then washed three times with serum-free medium. The MDA-MB-231 cells (3 × 10⁵ cells), F10 cells (3 × 10⁵ cells), and MCF-7/COX-2 cells (4 × 10⁵ cells), resuspended in 500 μl, were added into the Matrigel-coated or the collagen-coated transwell inserts and were incubated for 72 hours in the absence or presence of JS-K (0.5 and 1.0 μM) or JS-43-126 (0.5 and 1.0 μM). The lower chambers were filled with 2 ml DMEM/F12 medium supplemented with 5% FBS. After incubation, noninvading cells on the upper surface of the filter were removed with cotton swabs. Cells that had invaded through the pores onto the lower side of the filter were fixed, stained with Hema-3, and photographed. The invaded cells were counted in five fields for each filter under a light microscope at 40× magnification. The invasiveness of the cells was expressed as the mean number of cells that had invaded to the lower side of the filter. The experiments were performed in triplicate wells and were repeated twice.

To determine the importance of TIMP-2 in JS-K-mediated anti-invasive effects, TIMP-2 activity was blocked with a neutralizing antibody (R & D Systems). The MDA-MB-231, F10, and MCF-7/COX-2 cells were treated with 1 μM JS-K in the presence or absence of the anti-TIMP-2 antibody (2.5 μg/ml) for 72 hours in a Matrigel invasion assay. The experiments were performed in triplicate wells and were repeated twice.

The MDA-MB-231 cells (3 × 10⁵ cells), F10 cells (3 × 10⁵ cells), and MCF-7/COX-2 cells (4 × 10⁵ cells) were plated in T25 flasks in 5 ml DMEM/F12 medium supplemented with 5% FBS. The next day, cells were treated with JS-K (0.5 and 1.0 μM) or JS-43-126 (0.5 and 1.0 μM) for 24 hours. The medium in each flask was then replaced with serum-free medium and the flasks were incubated for an additional 24 hours. The medium was recovered, centrifuged for 5 minutes, and concentrated using spin columns with 10-kDa-cutoff filters. The medium collected was used for the matrix metalloproteinase (MMP) array and to determine the expression of TIMP-2.

The expression of MMPs and TIMPs in the conditioned medium supernatant was qualitatively screened using a human MMP array kit (RayBiotech, Norcross, GA, USA). The array allows for the simultaneous detection of seven MMPs (MMP-1, MMP-2, MMP-3, MMP-8, MMP-9, MMP-10, MMP-13) and three TIMPs (TIMP-1, TIMP-2, TIMP-4). Images were scanned using an Alpha Imager application program.

The concentration of TIMP-2 in the conditioned medium supernatant was determined using a TIMP-2 ELISA kit (EMD Biosciences). The concentration of TIMP-2 was normalized to the cell number and was expressed as nanograms per milliliter per 10⁶ cells. The experiments were performed in triplicate wells and were repeated three times.

For statistical analysis of the invasion experiments, the Shapiro-Wilk test was first performed to assess the normality of assumption data. Given that the data were normally distributed, two-sample t tests were performed for each of the three cell lines to compare the number of invading cells for the untreated group with the number of invading cells for each dose of JS-K and JS-43-126. The number of invading cells was also compared between the two doses of JS-K and JS-43-126. Two-sample t tests were also used to compare the number of invading cells for the group treated with JS-K with the number of invading cells for the group treated with JS-K in the presence of the anti-TIMP-2 antibody for each cell line.

The significance level for each individual comparison (0.05/3 = 0.0167) was adjusted by the Bonferroni method to account for multiple testing within each cell line to achieve an overall significance level of 5%. All analyses were performed using SAS software (SAS Institute, Inc., Cary, NC, USA). The same statistical analyses were used to compare the NO and TIMP-2 levels of untreated cells with those treated with JS-K or JS-43-126 as appropriate.

JS-K is activated to release NO by GST enzymes 15; the expression of GST-π and GST-α in MDA-MB-231, F10, and MCF-7/COX-2 breast cancer cells was therefore determined. The MDA-MB-231 and F10 cells expressed GST-π and GST-α, but GST-π was the predominant isoform (Figure 2). MCF-7/COX-2 cells expressed GST-α but not GST-π (Figure 2).

NO levels were determined in untreated and JS-K-treated MDA-MB-231, F10, and MCF-7/COX-2 cells to confirm drug activation. The NO production was significantly increased (P < 0.05) in the three cell lines as a result of JS-K treatment (Figure 3).

The NO levels were 2.1-fold and fourfold higher (P < 0.05) in MDA-MB-231 cells treated with 0.5 and 1 μM JS-K, respectively (Figure 3). The NO levels were increased (P < 0.05) 5.8-fold and 6.1-fold at the 0.5 and 1 μM concentrations of JS-K in F10 cells, respectively (Figure 3). Although the two concentrations of JS-K did not differ (P > 0.05) in the NO levels produced, the NO levels of JS-K-treated F10 cells were significantly higher (P < 0.05) in comparison with untreated cells (Figure 3). The NO levels were increased (P < 0.05) 4.9-fold and sevenfold in MCF-7/COX-2 cells at the 0.5 and 1 μM concentrations of JS-K, respectively (Figure 3). JS-K can therefore be activated to release NO by breast cancer cells. In contrast, NO production was not different (P > 0.05) between untreated cells and those treated with JS-43-126 for each of the three cell lines (Figure 3).

The invasion of cancer cells through basement membranes is an essential step in cancer metastasis. Matrigel is a solubilized basement membrane preparation extracted from the Engelbreth-Holm-Swarm mouse sarcoma, a tumor rich in extracellular matrix proteins. The major component of Matrigel is laminin. Matrigel has been used by numerous groups to assay the invasive activity of tumor cells across the basement membrane 293031.

Matrigel invasion assays were performed to determine the effect of JS-K on the invasiveness of breast cancer cells across the basement membrane. Untreated MDA-MB-231, F10, and MCF-7/COX-2 cells displayed a high invasive capacity on Matrigel (Figure 4a). In all cell lines, JS-K significantly (P < 0.05) reduced the number of invasive cells (Figure 4a, b). The number of invaded MDA-MB-231 cells was decreased (P < 0.05) 37% and 85% at the 0.5 and 1 μM doses of JS-K, respectively (Figure 4b). The number of invaded F10 cells was reduced (P < 0.05) 63% and 76% by the 0.5 and 1 μM doses of JS-K, respectively (Figure 4b). The two doses of JS-K, however, did not have significantly different (P > 0.05) anti-invasive effects in F10 cells (Figure 4b). In MCF-7/COX-2 cells, JS-K reduced (P < 0.05) the number of invaded cells 49% and 75% at the 0.5 and 1 μM doses of JS-K, respectively (Figure 4b). In contrast, the invasiveness of the three cell lines was unaffected by treatment with JS-43-126 (Figure 4b). JS-K can therefore decrease breast cancer invasion across Matrigel, and this is dependent on NO production.

JS-K has been shown to induce growth inhibition in cancer cells 1516171819. We determined the effects of JS-K on the proliferation of breast cancer cells grown on Matrigel, in order to mimic the conditions used in the Matrigel invasion assays. The 0.5 and 1.0 μM doses of JS-K induced < 20% growth inhibition in any of the breast cancer cell lines (Figure 4c). JS-K-mediated decreases in the Matrigel invasion assays were therefore not the result of growth inhibition.

Bone is the most prevalent site of first distant relapse of breast cancer, with as many as 85% of patients with advanced breast cancer suffering from bone metastases 32. Type I collagen is the most abundant protein within the bone, making up > 90% of the total protein within this site 33. Type I collagen has been used to assay the invasive activity of tumor cells across the bone matrix 3435. A type I collagen invasion assay was performed to determine whether JS-K may inhibit the invasiveness of breast cancer cells across the bone matrix. The conditions for the collagen invasion assay were identical to those of the Matrigel invasion assay, except that type I collagen was used to coat the transwell insert. The MDA-MB-231 and F10 cells displayed a high invasive capacity on type I collagen (Figure 4d), but MCF-7/COX-2 cells did not (data not shown). JS-K did not reduce the invasiveness of breast cancer cells across type I collagen-coated insert (Figure 4d). These data indicate that JS-K can block breast cancer cells from invading through Matrigel but not through type I collagen, suggesting that JS-K can block breast cancer invasion through the basement membrane but not through the bone matrix.

MMPs, which are involved in the degradation of the basement membrane, are essential to the invasive process. In contrast, TIMPs regulate the activity of MMPs and protect the basement membrane from proteolysis. A human MMP array was performed to screen the effects of JS-K on MMP and TIMP production. The array profiles for JS-43-126-treated cells were similar to those of untreated cells (Figure 5a). In contrast, the most consistent effect observed in the arrays of the three cell lines as a result of JS-K treatment was an increase in the production of TIMP-2 (Figure 5a). To confirm the JS-K-mediated increase in TIMP-2 levels that were observed in the MMP arrays, TIMP-2 ELISAs were performed. In MDA-MB-231 cells, TIMP-2 levels were increased (P < 0.05) 1.9-fold and threefold at the 0.5 and 1 μM doses of JS-K, respectively, while TIMP-2 was increased (P < 0.05) 1.5-fold and 7.2-fold in F10 cells at the same doses (Figure 5b). In MCF-7/COX-2 cells, TIMP-2 was increased (P < 0.05) only at the higher dose of JS-K (Figure 5b). TIMP-2 was increased (P < 0.05) twofold in MCF-7/COX-2 cells at the 1 μM concentration of JS-K (Figure 5b). These data indicate that TIMP-2 may be the major, but not the only, target of JS-K.

Next we determined the importance of TIMP-2 in the anti-invasive effects of JS-K. To do this, TIMP-2 activity was blocked with a commercially available neutralizing antibody and the effect of JS-K on the invasiveness of MDA-MB-231, F10, and MCF-7/COX-2 cells across Matrigel was determined. At the concentration used, the anti-TIMP-2 antibody had no effect on invasion (Figure 5c, d). JS-K decreased the invasiveness of all cell lines across Matrigel; however, blocking TIMP-2 activity significantly (P < 0.05) suppressed the anti-invasive effects of JS-K (Figure 5c, d). In comparison with untreated MDA-MB-231 cells, JS-K decreased (P < 0.05) the number of invaded cells by 72% and 37% in the absence and presence of the anti-TIMP-2 antibody, respectively (Figure 5d). The number of invaded F10 cells was 72% and 40% lower (P < 0.05) relative to untreated cells when treated with JS-K alone or in combination with the anti-TIMP-2 antibody, respectively (Figure 5d). JS-K decreased (P < 0.05) the number of invaded MCF-7/COX-2 cells by 65%, but in the presence of anti-TIMP-2 the number of invaded cells was decreased (P < 0.05) by 30% (Figure 5d). These data indicate that TIMP-2 is an important mediator of the anti-invasive activity of JS-K across the Matrigel basement membrane.

Mitogen-activated protein kinase pathways, which have been shown to regulate TIMP-2 363738, are activated by JS-K 1617. We therefore determined whether these pathways were involved in JS-K-mediated TIMP-2 production. In all cell lines, p38 phosphorylation (indicative of activity) was unaffected by the 0.5 μM concentration of JS-K (Figure 6). The 1.0 μM concentration of JS-K decreased p38 phosphorylation by approximately 27%, 62%, and 70% in MDA-MB-231 cells, F10 cells, and MCF-7/COX-2 cells, respectively (Figure 6). At 0.5 and 1.0 μM concentration, JS-K decreased ERK1/2 phosphorylation in F10 cells by 36% and 57%, respectively (Figure 6). In contrast, JS-K did not affect ERK1/2 phosphorylation in MDA-MB-231 cells or MCF-7/COX-2 cells (Figure 6). The phosphorylation of JNK was not affected by JS-K in any cell line (Figure 6).