Expression of transcriptional factors that repress E-cadherin transcription

It is well-known that E-cadherin is a hallmark of EMT and that EMT is driven by Snail, ZEB, and Twist transcriptional factors (5, 6). Thus, we analyzed the levels of mRNA and protein of these transcriptional factors in mock, wild-type, and mutated Trop-2 expressing cells by DNA microarray analyses (Fig. 2A) and immunoblotting (Fig. 2B). Similar levels of Snail1 mRNA and protein were expressed in these cells. ZEB1 mRNA was markedly increased in SE cells, and Twist1 mRNA was elevated in WT and SE cells. E-cadherin protein was decreased only in SE cells, suggesting that ZEB1 mainly regulates the expression of E-cadherin, but there may be two pathways to upregulate the transcription of ZEB1 and Twist1 through downstream events after Trop-2 phosphorylation. The reason why E-cadherin did not decrease in WT cells (Fig. 1, C and D) despite the up-regulation of Twist1 (Fig. 2, A and B) is discussed later.

Open in a separate window

Figure 2

Expression of transcriptional factors to down-modulate E-cadherin transcription.A, the levels of ZEB1, Twist1, and Snail1 mRNAs in the four cell types were determined by DNA microarray analysis (n = 1), and that in Mock cells was taken as 1. B, lysates prepared from the four cell types were subjected to SDS-PAGE and blotting. The blotted membrane was cut into three pieces containing proteins with high, intermediate, and low molecular masses, followed by detection of ZEB1, β-actin, and Snail1 proteins with each antibody, respectively. Because ZEB1, Snail1, and β-actin proteins were detected using the same blotted membrane, the β-actin image is shown as a loading control of both ZEB1 and Snail1 proteins. Separately, β-actin and Twist1 proteins were detected through the same procedure as described above. This β-actin image is shown as a loading control of the Twist1 protein. The intensities of the bands were determined. That in Mock cells was normalized as to β-actin, and the value was taken as 1. (means ± S.E., n = 3, p < 0.01, n.s.: not significant).

Although Trop-2 is expressed broadly in cancers (14, 26, 27, 28, 29), little is known about Trop-2–mediated signaling. In addition, it remains to be determined whether or not Trop-2 phosphorylation is related to signaling.

Enhanced cleavage of Trop-2 through its phosphorylation and accumulation of β-catenin in the nucleus

It is reported that Trop-2 is cleaved through regulated intramembrane proteolysis (RIP), which is performed by the TNF-α-converting enzyme (TACE) and subsequently by γ-secretase, resulting in shedding of its extracellular domain and nuclear translocation of its intracellular peptide (30). In order to detect the intracellular peptide and investigate whether or not Trop-2 phosphorylation is related to the cleavage of Trop-2, wild-type Trop-2, and phosphorylation-blocked Trop-2 tagged with FLAG at the N-terminal end were further modified by adding a HA tag at the C-terminal end. FLAG-and HA-tagged wild-type Trop-2 and phosphorylation-blocked Trop-2 were immunoprecipitated from lysates of FLAG- and HA-tagged wild-type Trop-2 (WT-HA) and phosphorylation-blocked Trop-2 (SA-HA) expressing HCT116 cells, respectively, and subjected to SDS-PAGE, followed by immunoblotting and detection with anti-HA- or FLAG-tag antibodies. A broad band with a molecular mass of 50~75 kDa was detected with both antibodies in both WT-HA and SA-HA cells (Fig. 3A, lanes a, b, c, d). Other bands with molecular masses of ~40 kDa, ~20 kDa, and ~13 kDa were detected with anti-HA antibodies but not with anti-FLAG antibodies in WT-HA cells (Fig. 3A, lanes a, b), whereas these bands were hardly detected in SA-HA cells with either anti-FLAG-tag antibodies or anti-HA-tag antibodies (Fig. 3A, lanes c, d). These results suggest that the smaller three bands in WT-HA cells (Fig. 3A, lane b) were C-terminal peptides produced through cleavage of Trop-2 after its phosphorylation. In relation to Trop-2 phosphorylation, lysates of WT-HA cells were also analyzed by immunoblotting using anti-phospho-Trop-2 antibodies. Trop-2 C-terminal fragments corresponding to two bands with molecular masses of ~40 kDa and ~13 kDa were detected, but the ~20-kDa band was not found probably due to its undetectable level (Fig. 3A, lane e). This result suggests that the phosphorylation of Trop-2 occurred before but not after the cleavage of Trop-2, this being consistent with the result that phosphorylation of Trop-2 may play an important role in the cleavage of Trop-2. In this context, furthermore, we investigated the relationship of Trop-2 phosphorylation with its cleavage. PKC-dependent phosphorylation of Trop-2 was demonstrated in the treatment of WT cells with PMA and PKC inhibitor Gö6983 as described previously (23). After treatment with PMA or 4α-PMA (control) for several hours, cell lysates were subjected to SDS-PAGE, followed by immunoblotting. Cleavage of Trop-2 was chased by the detection of a ~13 kDa fragment with anti-HA antibodies. Cleavage of Trop-2 was enhanced by PMA treatment, peaking at 6 h (Fig. 3B). After treatment with Gö6983 or DMSO (control), the WT-HA cells were cultured in the presence of PMA or 4α-PMA for an additional 6 h, followed by detection of the Trop-2 fragment as described above. Cleavage of Trop-2 was clearly reduced on the treatment with PKC inhibitor Gö6983, indicating that Trop-2 phosphorylation initiated or enhanced its cleavage (Fig. 3C). It has been reported that the intracellular fragment of Trop-2 associated with β-catenin plays a role in a signaling event, leading to the promotion of self-renewal and initiation of prostatic intraepithelial neoplasia (30). To investigate whether or not phosphorylation of Trop-2 affects its association with β-catenin, Trop-2 was immunoprecipitated from lysates of wild-type and mutated Trop-2 expressing HCT116 cells, and then subjected to SDS-PAGE, followed by immunoblotting and detection with anti-β-catenin antibodies. Similar levels of β-catenin were coimmunoprecipitated with Trop-2 in these cells, suggesting that the interaction of Trop-2 with β-catenin is not regulated through its phosphorylation (Fig. S1). Next, the distributions of the HA-tagged Trop-2 fragment and β-catenin in WT-HA cells were observed immunochemically after treatment with PMA or 4α-PMA. Both the HA-tagged Trop-2 fragment and β-catenin were accumulated in nuclei on the treatment with PMA (Fig. 3D). In the merged image, β-catenin colocalized with the HA-tagged Trop-2 fragment in the nuclei can be observed as white dots in Figure 3D. White dot–positive cells in the nuclei were counted, and the percentage of positive cells in total cells is shown as a histogram (Fig. 3E). PMA treatment increased the β-catenin colocalized with the HA-tagged Trop-2 fragment in nuclei. These results indicate that nuclear transport of the Trop-2 fragment associated with β-catenin was enhanced by increased cleavage due to Trop-2 phosphorylation.

Open in a separate window

Figure 3

Enhanced cleavage of Trop-2 through its phosphorylation and accumulation of β-catenin in the nucleus.A, FLAG- and HA-tagged Trop-2 were immunoprecipitated from lysates of HCT116/WT-HA (WT-HA) cells (lanes a, b) and HCT116/SA-HA (SA-HA) cells (lanes c, d), and then subjected to SDS-PAGE, followed by immunoblotting and detection with anti-FLAG-tag or HA-tag antibodies. HA-tagged Trop-2 was also immunoprecipitated from lysates of WT-HA cells and subjected to SDS and immunoblotting, followed by detection with anti-phosphorylated Trop-2 antibodies (lane e). B, HA-tagged Trop-2 was immunoprecipitated from lysates of WT-HA cells treated with PMA (100 ng/ml) or with 4α-PMA for the indicated times and then subjected to SDS-PAGE, followed by immunoblotting and detection with anti-HA-tag antibodies. The intact Trop-2 and intracellular fragment with a molecular mass of ~13 kDa were demonstrated. The ratio of HA-tagged ~13-kDa fragment to HA-tagged intact Trop-2 was calculated, and the value for the cells treated with PMA or 4α-PMA for 0 h was taken as 1 (means ± S.E., n = 3, p < 0.01, n.s.: not significant). C, WT-HA cells were treated with Gö6983 (1 μM) or DMSO (control) for 2 h and subsequently with PMA (100 ng/ml) for an additional 6 h, and the intact molecule and ~13-kDa fragment were detected as described above. The ratio of HA-tagged ~13 kDa fragment to HA-tagged intact Trop-2 was calculated, and the value for the control experiment (DMSO) was taken as 1 (means ± S.E., n = 3, p < 0.05). D, WT-HA cells were treated with PMA (100 ng/ml) or 4α-PMA (control) for 6 h and then immunostained with anti-HA-tag (green) and anti-β-catenin (red) antibodies. Nuclei were stained with DAPI (blue), and a representative photograph is shown. Scale bars, 10 μm. E, the nuclear transport of HA-tagged Trop-2 fragment and β-catenin was calculated by counting the white dots (β-catenin colocalized with the Trop-2 fragment) per five randomly selected fields in merged images in Figure 3D. The ratio (%) of white dot positive cells per total cells was calculated (means ± S.E., n = 5, p < 0.01). F, β-catenin in the four cell types was immunostained with anti-β-catenin antibodies (red). Nuclei were stained with DAPI (blue), and a representative photograph is shown. Scale bars, 10 μm. G, the ratio (%) of nuclear β-catenin positive cells per total cells in five randomly selected fields was calculated and is shown as a histogram (means ± S.E., n = 5, p < 0.01).

Furthermore, the accumulation of nuclear β-catenin, which is a final down-stream event of Trop-2 mediated signaling, may be critical to indicate the significance of Trop-2–mediated signaling. Recently, Zhao et al. (31) also reported that Trop-2 increased the accumulation of β-catenin in the nucleus to accelerate metastasis in gastric cancer. Thus, we tried to compare the levels of nuclear β-catenin by staining with anti-β-catenin antibodies among four cell types (Fig. 3, F and G). It was noted that the levels of nuclear β-catenin in Mock cells and SA cells were negligible, low in WT cells, and markedly high in SE cells (Fig. 3, F and G). Negligible levels of nuclear β-catenin in Mock cells and phosphorylation-blocked SA cells are consistent with the report that HCT116 cells have low levels of nuclear β-catenin (32). These results suggest that Trop-2–mediated signaling may occur weakly in WT cells and prominently in phosphomimetic SE cells and be blocked in SA cells. In fact, PMA treatment of WT cells increased ZEB1 expression and decreased E-cadherin expression. In contrast, Twist1 was hardly enhanced on the treatment with PMA (Fig. S2).

β-Catenin–dependent expression of ZEB1

It has been reported that nuclear β-catenin binds to members of the TCF/LEF transcription factor family, and β-catenin/TCF4 binds to the human ZEB1 promoter, triggering its transcription (32). The formation of the β-catenin/TCF4 complex was confirmed by their coimmunoprecipitation in WT cells (Fig. 4A, left and center). Association with the Trop-2 fragment with β-catenin and TCF4 was also confirmed by coimmunoprecipitation with anti-HA-tag antibodies in WT-HA cells (Fig. 4A, right). ZEB1 was highly expressed in SE cells, but not in other Trop-2–expressing cells (Fig. 2, A and B). Thus, SE cells were treated with a β-catenin inhibitor (XAV939, PNU74654) for 24 h, and then the level of ZEB1 was determined by SDS-PAGE and immunoblotting of cell lysates. About 30% of ZEB1 protein decreased on the treatment with the β-catenin inhibitors (Fig. 4B). The decreases of the β-catenin and TCF4 proteins on the treatment with each siRNA were confirmed (Fig. 4C). Knockdown of β-catenin and/or TCF4 caused about a 50% decrease in the ZEB1 protein level (Fig. 4D). An equivalent decrease of ZEB1 promoter activity was confirmed by luciferase assay (Fig. 4E). It was also noted that the expression of E-cadherin increased on the knockdown of these factors (Fig. 4F). The reason why the level of E-cadherin protein was not elevated significantly in the treatment of SE cells with β-catenin siRNA is discussed later.

Open in a separate window

Figure 4

β-Catenin-dependent expression of ZEB1.A, β-catenin (left) or TCF4 (center) was immunoprecipitated from lysate of WT cells, and coimmunoprecipitated TCF4 or β-catenin was detected by SDS-PAGE and immunoblotting. (right) Immunoprecipitates from lysates of WT-HA cells with anti-HA-tag antibodies were subjected to SDS-PAGE, followed by immunoblotting and detection of β-catenin and TCF4. B, lysates of SE cells treated with β-catenin inhibitors or DMSO (control) as described under Experimental Procedures were subjected to SDS-PAGE, followed by immunoblotting, and detection of ZEB1. β-Actin served as a loading control. The densities of bands were determined, and the ratio of ZEB1 to β-actin was calculated. That in SE cells treated with DMSO was taken as 1 (means ± S.E., n = 3, p < 0.01). C, SE cells were treated with β-catenin and/or TCF4 siRNAs or control siRNA for 24 h, and the cell lysates were subjected to SDS-PAGE, followed by immunoblotting and detection of β-catenin and TCF4. β-Actin served as a loading control. D, lysates of SE cells treated with β-catenin and/or TCF4 siRNAs as described above were subjected to SDS-PAGE, followed by immunoblotting, and detection of ZEB1, and the densities of bands were determined. The ratio of ZEB1 to β-actin was calculated and that in SE cells treated with control siRNA was taken as 1. β-Actin served as a loading control (means ± S.E., n = 3, p < 0.01). E, SE cells were treated with β-catenin and/or TCF4 siRNAs as described above, and ZEB1 promoter activity was determined by using luciferase assay as described under Experimental Procedures. The value for SE cells treated with control siRNA was taken as 1 (means ± S.E., n = 3, p < 0.05, p < 0.01). F, lysates of SE cells treated with β-catenin and/or TCF4 siRNAs as described above were subjected to SDS-PAGE, followed by immunoblotting and detection of E-cadherin. The densities of bands were determined, and the ratio of E-cadherin to β-actin was calculated. The value in SE cells treated with control siRNA was taken as 1. β-Actin served as a loading control (means ± S.E., n = 3, p < 0.01).

STAT3-dependent expression of Twist1

We also found that the levels of Twist1 mRNA in WT and SE cells were considerably higher than those in Mock and SA cells (Fig. 2, A and B). This finding suggests that Twist1 may be induced by downstream events through Trop-2 phosphorylation. Twist1 is transcriptionally induced by activation of STAT3 and mediates the STAT3 oncogenic function (33). In this context, we analyzed the levels of STAT3 protein and mRNA in wild-type and mutated Trop-2–expressing HCT116 cells. Although the levels of the STAT3 protein and mRNA were similar in these cells, phosphorylated (705Tyr) STAT3 was significantly elevated in WT and SE cells compared with Mock and SA cells (Fig. 5, A and B). It was also revealed that the phosphorylated STAT3 protein is dimerized and translocated to up-regulate the expression of Twist through its binding to the second proximal STAT3 binding site on the Twist promoter (33). In fact, the knockdown of STAT3 decreased the expression of Twist1 and reciprocally enhanced the expression of E-cadherin (Fig. 5C). Because similar levels of STAT3 were phosphorylated despite different signaling intensities between WT and SE cells, this STAT3-dependent pathway may be regulated by a STAT3 tyrosine kinase.

Open in a separate window

Figure 5

Phosphorylated STAT3-dependent expression of Twist1.A, the levels of STAT3 mRNA in four cell types were determined by DNA microarray analysis (n = 1) and that in Mock cells was taken as 1. The levels of STAT3 protein in the four cell types were determined as described in Figure 1D, and the ratio of STAT3 to β-actin was calculated. The value in Mock cells was taken as 1. β-Actin served as a loading control (means ± S.E., n = 3, n.s.: not significant). B, phosphorylated STAT3 protein was detected with anti-phosphorylated STAT3 antibodies using the same blotted membrane as described in (A) after stripping of anti-STAT3 antibodies. Because STAT3, phosphorylated STAT3 and β-actin proteins were detected using the same blotted membrane, the β-actin image is shown as a loading control of both STAT3 (A) and phosphorylated STAT3 (B) proteins. The ratio of phosphorylated STAT3 to β-actin was calculated, and the value in Mock cells was taken as 1 (means ± S.E., n = 3, p < 0.01, n.s.: not significant). C, WT cells were treated with STAT3 siRNA or control siRNA as described under Experimental Procedures. The levels of Twist1 and E-cadherin were determined as described in Figure 1D. The ratio of Twist1 and E-cadherin to β-actin was calculated, and the value in WT cells treated with control siRNA was taken as 1. β-Actin served as a loading control (means ± S.E., n = 3, p < 0.05, p < 0.01).

Binding of galectin-3 to Trop-2 and subsequent signaling

A small part of Trop-2 was phosphorylated in WT cells, as reported in our previous paper (23). Cleavage of Trop-2 and transport of β-catenin to nuclei were detected in WT cells as described above. We speculated that ligation of an endogenous ligand with Trop-2 initiates this signaling and that there are two ways to elevate the phosphorylation of Trop-2. One is the growth-factor receptor-mediated phosphorylation of Trop-2. Expression of various growth factor receptors in HCT116 cells was revealed by DNA microarray analysis (Fig. S3A). Ligation of growth factors to their receptors is expected to activate receptor tyrosine kinase, resulting in activation of PKC. We examined the phosphorylation and subsequent cleavage of Trop-2 when WT cells were treated with various growth factors including EGF, FGF, TGF-β, HGF, and IGF-1. However, no growth factors exhibited a detectable effect on Trop-2 cleavage (Fig. S3B). Another possibility is that Trop-2 is phosphorylated directly on the binding of unknown extracellular ligands. Although we searched for a specific endogenous protein that binds to the ectodomain of Trop-2, no specific ligand was detected, as discussed later. Next, we examined the binding of galectin-3 to Trop-2, because Trop-2 is a highly glycosylated membrane protein, and Trop-2 and galectin-3 exhibited similar distribution patterns in various human cancer tissues (34). Lysates of WT cells were incubated with galectin-3-Sepharose in the presence or absence of 50 mM lactose. Galectin-3 binding proteins were examined on SDS-PAGE and immunoblotting. FLAG-tagged Trop-2 was detected in the galectin-3 binding proteins and its binding was completely abolished by lactose, indicating specific binding of galectin-3 to Trop-2 (Fig. 6A). Colocalization of Trop-2 and galectin-3 on the cell surface was observed immunochemically, suggesting the binding of galectin-3 to Trop-2 on the cell surface (Fig. 6B). After incubation of WT cells with galectin-3 for 30 min, FLAG-tagged Trop-2 was immunoprecipitated from the cell lysates, and then subjected to SDS-PAGE, followed by immunoblotting and detection of phosphorylated Trop-2 with anti-phosphorylated Trop-2 antibodies. The binding of galectin-3 to Trop-2 elevated the phosphorylation of Trop-2 (Fig. 6C). Subsequent cleavage of Trop-2 was also enhanced at 2~6 h after treatment of the cells with galectin-3 (Fig. 6D). Next, we investigated whether or not the accumulation of nuclear β-catenin in four cell types increases through the binding of exogenously added galectin-3 to Trop-2 (Fig. 6, E and F). Expectedly, nuclear β-catenin in WT cells increased on the treatment with galectin-3 compared with that in WT cells without galectin-3 treatment, as shown in Figure 3F. In addition, the levels of nuclear β-catenin in SA cells and SE cells hardly changed regardless of the galectin-3 treatment. Those in SA cells were still negligible, probably due to the blocking of Trop-2–mediated signaling in Trop-2 phosphorylation-blocked cells. High levels of nuclear β-catenin in SE cells were maintained, suggesting that the maximum level of signaling occurs at all times in Trop-2 phosphomimetic cells.

Open in a separate window

Figure 6

Trop-2 mediated signaling triggered by binding of galectin-3 to Trop-2.A, lysates of WT-HA cells were incubated with galectin-3-Sepharose in the presence or absence of 50 mM lactose at 4 °C for 3 h. Galectin-3 binding proteins were subjected to SDS-PAGE, followed by immunoblotting and detection of FLAG-tagged Trop-2. B, galectin-3 (green) and FLAG-tagged Trop-2 (red) in WT-HA cells were immunostained as described under Experimental Procedures. Scale bars, 10 μm. C, WT-HA cells were treated with or without galectin-3 (40 μg/ml) for 30 min, and the cell lysates were subjected to SDS-PAGE, followed by immunoblotting, and detection of phosphorylated Trop-2 and FLAG-tagged Trop-2. The densities of bands were determined, and the ratio of phosphorylated Trop-2 to FLAG-tagged Trop-2 was calculated. The value in control cells was taken as 1 (means ± S.E., n = 3, p < 0.01). D, WT-HA cells were treated with galectin-3 (40 μg/ml) for the indicated times. HA-tagged Trop-2 and its fragment were immunoprecipitated from the cell lysates and subjected to SDS-PAGE, followed by immunoblotting and detection of the intact Trop-2 and ~13-kDa fragment. The ratio of HA-tagged ~13 kDa fragment to HA-tagged intact Trop-2 was calculated, and the value in control experiment (PBS) was taken as 1 (means ± S.E., n = 3, p < 0.01, n.s.: not significant). E, β-catenin in four cell types treated with galectin-3 (80 μg/ml) for 24 h was immunostained with anti-β-catenin antibodies (red) as described in Figure 3F. Nuclei were stained with DAPI (blue), and a representative photograph is shown. Scale bars, 10 μm. F, the ratio (%) of nuclear β-catenin positive cells per total cells in five randomly selected fields was calculated and is shown as a histogram (means ± S.E., n = 3, p < 0.01, n.s.: not significant). G, WT-HA cells were cultured in the presence or absence of galectin-3 (40 μg/ml) for 24 h, and the levels of ZEB1 and E-cadherin protein were determined as described in Figure 1D. The ratio of ZEB1 and E-cadherin to β-actin was calculated. The value in control cells was taken as 1. β-Actin served as a loading control (means ± S.E., n = 3, p < 0.01). H, migration of WT-HA cells treated with or without galectin-3 (80 μg/ml) as described above was evaluated by means of Transwell assays. A representative photograph of migrated cells is shown. The numbers of migrated cells per field were determined and is shown as a histogram. Scale bar, 100 μm (means ± S.E., n = 3, p < 0.05).

We also investigated the effect of galectin-3 on the expression of ZEB1 and E-cadherin. Expectedly, expression of ZEB1 increased and, reciprocally, that of E-cadherin clearly decreased (Fig. 6G) at 24 h after incubation of the cells with galectin-3. Eventually, galectin-3–treated cells exhibited higher cell mobility compared with control cells (Fig. 6H).

In this context, we performed the same experiment using SA cells. Exogenously added galectin-3 had no effect on the expression of ZEB1 and E-cadherin (Fig. S4), consistent with the result that β-catenin was hardly accumulated in nuclei in SA cells (Fig. 6E). Actually, exogenously added galectin-3 enhanced ZEB1 transcription and subsequent down-modulation of E-cadherin in WT cells. It would be interesting to see whether or not endogenous galectin-3 could reduce E-cadherin expression. We observed four cell types cultured for 7 days microscopically (Fig. S5A). WT cells and SE cells appeared to be in contact more loosely compared with SA cells. Next, the expression of E-cadherin and cell surface galectin-3 was examined and compared using WT cells cultured for 1 and 7 days. Consistently, the level of E-cadherin in WT cells cultured for 7 days decreased compared with that in WT cells cultured for 1 day (Fig. S5B). WT cells cultured for seven cells were more densely stained with anti-galectin-3 antibodies compared with WT cells cultured for 1 day (Fig. S5C), suggesting that Trop-2–mediated signaling was elevated through accumulated cell surface galectin-3 in WT cells cultured for 7 days.

Change in E-cadherin expression in MCF-7 cells and DU145 cells treated with PMA, galectin-3, or Trop-2 siRNA

Trop-2 was immunoprecipitated from lysates of MCF-7 cells and DU145 cells and then subjected to SDS-PAGE and immunoblotting, followed by detection with anti-Trop-2 antibodies. A band with a molecular mass of about 50 kDa was detected (Fig. S6A). Expression of Trop-2 in MCF-7 cells and DU145 cells was also demonstrated by flow cytometry (Fig. S6B). Next, lysates were prepared from MCF-7 cells and DU145 cells treated with PMA or 4α-PMA and then subjected to SDS-PAGE and immunoblotting. Phosphorylated Trop-2, ZEB1, Twist1, E-cadherin, and β-actin were detected. Although the expression of Twist1 hardly changed, the level of ZEB1 was enhanced, and, reciprocally, the expression of E-cadherin was down-modulated on the treatment with PMA (Fig. 7A).

Open in a separate window

Figure 7

Change of E-cadherin expression in MCF-7 cells and DU145 cells treated with PMA, galectin-3, or Trop-2 siRNA.A, the levels of phosphorylated Trop-2, ZEB1, Twist1, and E-cadherin were determined as described in Figure 1D. MCF-7 cells and DU145 cells were treated with 4α-PMA or PMA (100 ng/ml), and after 24 h, lysates were prepared to determine the levels of phosphorylated Trop-2, ZEB1, and Twist1. Similarly, the levels of E-cadherin were determined using lysates prepared after 48 h. β-Actin served as a loading control. The densities of bands were determined, and the ratios of ZEB1, Twist1, and E-cadherin to β-actin were calculated. The values in MCF-7 cells and DU145 cells treated with 4α-PMA were taken as 1 (means ± S.E., n = 3, p < 0.05, p < 0.01, n.s.: not significant). B, lysates of MCF-7 cells and DU145 cells were incubated with galectin-3-Sepharose in the presence or absence of 50 mM lactose, and galactose binding proteins were subjected to SDS-PAGE, followed by immunoblotting and detection with anti-Trop-2 antibodies. C, lysates of MCF-7 cells and DU145 cells treated with or without galectin-3 (40 μg/ml) for 24 h were subjected to SDS-PAGE and immunoblotting, followed by detection of phosphorylated Trop-2, ZEB1, Twist1, and β-actin. β-Actin served as loading control. The densities of bands were determined, and the ratios of ZEB1 and Twist1 to β-actin were calculated. The values in galectin-3 non-treated cells were taken as 1 (means ± S.E., n = 3, p < 0.01, n.s.: not significant). D, lysates of MCF-7 cells and DU145 cells treated with or without galectin-3 (40 μg/ml) for 24 h were subjected to SDS-PAGE and immunoblotting, followed by detection with anti-E-cadherin and anti-β-actin antibodies. β-Actin served as a loading control. The densities of bands were determined, and the ratio of E-cadherin to β-actin was calculated. The values in galectin-3 non-treated cells were taken as 1 (means ± S.E., n = 3, p < 0.05). E, MCF-7 cells and DU145 cells were treated with Trop-2 siRNA or control siRNA for 24 h, and cultured in the presence of galectin-3 (40 μg/ml) for 24 h. The cell lysates were subjected to SDS-PAGE, followed by immunoblotting and detection of Trop-2, ZEB1, E-cadherin, and β-actin. β-Actin served as a loading control. The densities of bands were determined, and the ratios of ZEB1 and E-cadherin to β-actin were calculated. The values in each cell treated with control siRNA was taken as 1 (means ± S.E., n = 3, p < 0.01).

The binding of galectin-3 to Trop-2 expressed in MCF-7 cells and DU145 cells was investigated by using galectin-3-Sepharose as described in Figure 6A. Trop-2 was detected by SDS-PAGE and immunoblotting in the galectin-3 binding proteins in both cells (Fig. 7B). Trop-2 phosphorylation and expression of ZEB1 and Twist1 in MCF-7 cells and DU145 cells were also investigated by immunoblotting after treatment with exogenously added galectin-3. ZEB1 but not Twist1 increased in association with the enhancement of Trop-2 phosphorylation (Fig. 7C). Expectedly, the expression of E-cadherin was clearly down-modulated after treatment with galectin-3 (Fig. 7D).

Furthermore, in order to confirm the biological significance of Trop-2, we tried to knockdown Trop-2 in MCF-7 cells and DU145 cells. Decreased Trop-2 expression in MCF-7 cells and DU145 cells was confirmed by SDS-PAGE and immunoblotting. MCF-7 cells and DU145 cells treated with Trop-2 siRNA or control siRNA were cultured in the presence of galectin-3 for 24 h, and then cell lysates were subjected to SDS-PAGE and immunoblotting, followed by detection of Trop-2, ZEB 1, E-cadherin and β-actin. Downregulation of ZEB1 and reciprocal upregulation of E-cadherin were also demonstrated associated with Trop-2 siRNA treatment of MCF-7 cells and DU145 cells (Fig. 7E).

Detection of FLAG-tagged Trop-2 in the liver and lung of nude mice bearing intraperitoneally or subcutaneously injected three cell types

Next, to examine the cell behavior of the three cell types in vivo, these cells were injected into nude mice intraperitoneally or subcutaneously. Intraperitoneally injected cells gave rise to bloody ascites regardless of being wild-type or mutated Trop-2 expressing cells, however, the appearance of nude mice was extremely different between SA and the other two cells (Fig. 8A). Livers and lungs were harvested at 4 weeks after tumor inoculation, and immunoprecipitation of FLAG-tagged Trop-2 from tissue extracts and its detection were performed as described under Experimental Procedures section. FLAG-tagged Trop-2 was detected in two livers and four lungs of five WT cell-inoculated nude mice, and in two livers and three lungs of four SE cell-inoculated nude mice, whereas FLAG-tagged Trop-2 was not detected in either liver or lung of nude mice inoculated with SA cells (Fig. 8B).

Open in a separate window

Figure 8

Detection of FLAG-tagged Trop-2 in liver and lung of nude mice bearing intraperitoneally or subcutaneously injected three cell types.A, SA, WT, and SE cells (5 × 10⁶ cells) were injected intraperitoneally into nude mice, and after 1 month, the appearance of tumor-bearing mice was observed. Scale bar, 10 mm. B, livers were harvested from the tumor-bearing mice. Trop-2 was immunoprecipitated from the liver extracts, which were subjected to SDS-PAGE, followed by immunoblotting and detection of FLAG-tagged Trop-2. With respect to WT and SE cells, typical samples of FLAG-tagged Trop-2-detected or non-detected nude mice are shown. The occurrence of FLAG-tagged Trop-2-detected liver was 2/5, 2/4, and 0/3 of WT, SE, and SA cell-inoculated mice, respectively. C, three cell types, as described in Figure 8A, were injected subcutaneously into the backs of nude mice. After 2 months, livers and lungs were harvested, and detection of FLAG-tagged Trop-2 in the livers and lungs was performed as described above. Although FLAG-tagged Trop-2 was not detected in liver, FLAG-tagged Trop-2 in lung was detected in 3/5, 1/5, and 0/5 of WT, SE, and SA cell-inoculated mice, respectively. D, WT cells (5 × 10⁶ cells) were injected intraperitoneally into nude mice. Starting 1 day after tumor cell injection, nude mice received daily intraperitoneal injections of Gö6983 (0.25 mg/mouse) or DMSO. After 4 weeks, the mice were euthanized, and their livers and lungs were harvested. Thereafter, FLAG-tagged Trop-2 was detected in the extracts of livers and lungs. Histograms show the numbers of FLAG-tagged Trop-2-detected nude mice.

Subcutaneously injected cells formed tumors apparently at similar rates among these cells. At 8 weeks after subcutaneous injection, the nude mice were euthanized, and then their primary tumors, livers, and lungs were harvested. FLAG-tagged Trop-2 was detected in some lungs but not in livers (Fig. 8C). FLAG-tagged Trop-2 was detected in three lungs and one lung of five nude mice inoculated with WT cells and SE cells, respectively. However, even in the lungs or livers in which FLAG-tagged Trop-2 was detected, metastatic colonies were not observed histologically, probably due to the undetectable size of tumor cell clusters. In contrast, no FLAG-tagged Trop-2 was detected in the lungs of nude mice inoculated with SA cells (n = 5). In any case, SA cells seem to be unable to move to secondary sites, suggesting that phosphorylation of Trop-2 may be linked to the acquisition or enhancement of metastatic potential.

Therefore, we examined the effect of PKC inhibitor Gö6983 on the mobility of WT cells. WT cells were injected intraperitoneally into nude mice. Starting 1 day after tumor cell injection, nude mice received daily intraperitoneal injection of Gö6983 (0.25 mg/mouse). After 4 weeks, FLAG-tagged Trop-2 was detected in the liver and lung tissues as described above. FLAG-tagged Trop-2 in the liver was detected in nine of nine control mice and two of nine of Gö6983-treated mice. FLAG-tagged Trop-2 in lung was detectable in seven of nine of control mice and three of nine of Gö6983-treated mice, indicating that Gö6983 inhibited the mobility of WT cells effectively (Fig. 8D).

Cells

A human colorectal cancer cell line, HCT116 cells, a human breast cancer cell line, MCF-7 cells, and a human prostate cancer cell line, DU145 cells, were obtained from the American Type Culture Collection (Manassas). HCT116 cells, their Trop-2 transfectants, and MCF-7 cells were cultured in DMEM containing 10% heat-inactivated FBS, 4 mM L-glutamine, 100 units/ml penicillin, and 100 μg/ml streptomycin. DU145 cells were cultured in RPMI1640 containing 10% heat-inactivated FBS, 4 mM L-glutamine, 100 units/ml penicillin, and 100 μg/ml streptomycin. HCT116 cells expressing FLAG-tagged wild-type Trop-2 (WT cells) and Trop-2 mutated with Ala or Glu instead of Ser322 (SA or SE cells), and mock cells (Mock cells) were prepared and cultured as described previously (23). The four cell types (Mock, WT, SA, and SE cell) were prepared as follows unless otherwise described. Confluent cells were dispersed with trypsin treatment. After washing with a culture medium, cells were seeded and cultured for 1 day.

HCT116 cells expressing wild-type and phosphorylation-blocked Trop-2 with FLAG and HA tags at the N- and C-terminals (WT-HA cells, SA-HA cells), respectively, were prepared as follows: FLAG-tagged wild-type Trop-2 and phosphorylation-blocked Trop-2 plasmids, as described previously (23), were modified by introducing a HA tag at the C-terminus using a KOD-Plus-Mutagenesis Kit (TOYOBO), and the following primers: 5′-GTT CCA GAT TAC GCT TAG GTA CCC GGC GGG GCA GGG GAT G-3′ and 5′-ATC GTA TGG GTA CAA GCT CGG TTC CTT TCT CAA CTC C-3′. After sequencing and purification, the plasmids were transfected into HCT116 cells using FuGENE HD Transfection Reagent (Promega) according to the manufacturer’s instructions, and then selected with 600 μg/ml G418.

DNA microarray analysis

The mRNA levels of several genes were determined by DNA microarray analysis as described previously (23).

Flow cytometry

MCF-7 cells and DU145 cells were incubated with mouse anti-Trop-2 antibodies or the mouse isotype IgG (IgG2a). After incubation with fluorescein isocyanate (FITC)-conjugated secondary antibodies, the cells were analyzed by flow cytometry.

Immunoprecipitation and immunoblotting

Cell extracts were prepared basically as described previously (23). In brief, subconfluent cells were lysed with a solubilizing solution (25 mM Tris-HCl, pH7.5, 150 mM NaCl, 5 mM EDTA, 1% Triton X-100 (TX-100)) containing protease and phosphatase inhibitor cocktails (Nakarai Tesque), and subsequently sonicated. After centrifugation at 20,000g at 4 °C for 10 min, the supernatants were used as cell extracts. To detect Trop-2 in lung and liver tissues harvested from tumor-bearing nude mice, lungs and livers were cut into small pieces, and after washing with PBS, tissue extracts were prepared as described above. Trop-2 in the cell extracts was immunoprecipitated using anti-FLAG M2 Magnetic Beads (Sigma-Aldrich) or anti-HA-tag mAb-Magnetic Agarose (Medical and Biological Laboratories). Trop-2 in the tissue extracts was immunoprecipitated using goat anti-Trop-2 antibodies and protein G-SepharoseTM4 Fast Flow (GE Healthcare). The immunoprecipitates and cell extracts were subjected to SDS-PAGE, followed by immunoblotting. The target proteins were detected as bands by chemiluminescence, and the densities of the bands were determined with Image J software (National Institutes of Health).

Immunocytochemistry

Wild-type or mutated Trop-2 expressing HCT116 cells were stained immunocytochemically with anti-FLAG-tagged Trop-2, anti-E-cadherin and anti-galectin-3 antibodies as described previously (23).

When an intracellular antigen was detected, cells were fixed with methanol at −20 °C for 40 min, and then washed with PBS. After permeabilization with PBS containing 1% TX-100 at room temperature for 20 min, the cells were treated with methanol at −20 °C for 40 min. After blocking with PBS containing 5% BSA and 0.1% TX-100, the cells were incubated with primary antibodies (rabbit anti-HA tag and mouse anti-β-catenin antibodies), and then stained with fluorescence-conjugated secondary antibodies and DAPI. Finally, the stained cells were observed under a confocal laser-scanning fluorescence microscope (Leica) at a magnification of ×630.

Immunohistochemistry

Immunohistochemical staining was performed as described previously (54). Briefly, sections of paraffin-embedded tumor tissues were de-paraffinized in xylene and then de-hydrated through graded ethanol series. The sections were heated for 10 min at 105 °C in an autoclave in Target Retrieval Solution (DAKO). The sections were then incubated with 3% hydrogen peroxide to block endogenous peroxidase activity. The specimens were incubated with rabbit anti-phosphorylated Trop-2 antibodies, goat anti-Trop-2 antibodies, and mouse anti-E-cadherin antibodies, and then successively with each HRP-conjugated secondary antibody. Specimens of gastric cancer tissues (20 cases) were obtained from cancer patients in accordance with a protocol approved by the Osaka City University Ethics Committee (approval number 924). Written informed consent for research was obtained from the patients. All human aspects of this study abide by the Declaration of Helsinki principles.

Treatment with inhibitors

Cells were treated with β-catenin inhibitors (20 μM XAV939, Cayman Chemical) or (100 μM PNU74654, Enzo Life Science) for 24 h, and with PKC inhibitor (1 μM Gö6983, Sigma-Aldrich) or DMSO for the indicated times.

Treatment with phorbol 12-myristate 13-acetate (PMA)

Cells were treated with 100 ng/ml PMA or its analog, 4α-PMA, for the indicated times.

Treatment with siRNAs

siRNAs of β-catenin, TCF4, STAT3 and control siRNA (Silencer Select Negative Control #1) were purchased from Ambion. Treatment of cells with siRNAs of β-catenin (5 nM), TCF4 (5 nM), or STAT3 (15 nM), and control siRNA was performed as described previously (23). The sequences of the siRNAs were as follows: β-catenin, 5′-CUGUUGGAUUGAUUCGAAAtt-3′ and 5′-UUUCGAAUCAAUCCAACAGta-3′; TCF4, 5′-GAUGGAAGCUUACUAGAUUtt-3′ and 5′-AAUCUAGUAAGCUUCCAUCtg-3′; STAT3, 5′-GCCUCAAGAUUGACCUAGAtt-3′ and 5′-UCUAGGUCAAUCUUGAGGCct-3′.

siRNAs of Trop-2 and its control siRNA (Control siRNA-A) were purchased from Santa Cruz Biotechnology. Treatment of the cells with three to five Trop-2 target-specific 19 to 25 nt siRNAs and control siRNAs was performed according to the manufacturer’s instructions.

Construction of a luciferase vector and dual luciferase assays

A luciferase construct was prepared basically as described previously (55). Briefly, the promoter region of ZEB1 (−917~+56) was amplified from the genomic DNA of HCT116 cells using the following primers: 5′-ATA GGT ACC GCC TGT GGA TAC CTT AGC TC-3′ and 5′-CTG AAG CTT CGC TTG TGT CTA AAT GCT CG -3′. After cloning and digestion with KpnI/HindIII, the fragments were introduced into PicaGene Basic Vector 2 (Toyo Ink: ZEB1 vector).

Dual luciferase assays were performed using an empty vector, the ZEB1 vector, and the pRL-TK vector (Tokyo Ink) as described previously (55).

Binding of galectin-3 to Trop-2

Galectin-3 bound Sepharose was added to the lysates of WT cells, MCF-cells, and DU145 cells, followed by incubation at 4 °C for 3 h in the presence or absence of 50 mM lactose. After washing with PBS, proteins bound to beads were subjected to SDS-PAGE, followed by immunoblotting and detection of Trop-2.

Treatment with galectin-3

After washing WT cells, MCF-7 cells, and DU145 cells with serum-free DMEM containing 50 mM lactose and serum-free DMEM successively, the cells were treated with 40~80 μg/ml of galectin-3 in the serum-free medium for the indicated times.

Migration assay

Migration assay was performed as described previously (23). Briefly, WT-HA cells (4 × 10⁴ cells) treated with or without galectin-3 were suspended in serum-free medium containing 0.2% BSA and then seeded into the upper chamber of a Transwell (24-well culture plate, pore size 8.0 μm; Corning Inc) precoated with fibronectin. After filling the lower wells with serum-containing medium, the cells were incubated for 20 h. The chambers were fixed with methanol and then stained with hematoxylin. The migration was quantified by counting the migrated cells in five randomly selected fields at a magnification of ×100.

Detection of FLAG-tagged Trop-2 in lung and liver

Cells (5 × 10⁶ cells) were injected into 6-week-old BALB/c nu/nu male mice intraperitoneally and subcutaneously, and at 4 and 8 weeks after inoculation, respectively, lungs and livers were harvested, and used for biochemical detection of FLAG-tagged Trop-2. When the effect of Gö6983 was examined, nude mice intraperitoneally inoculated with WT cells as described above were treated daily by intraperitoneal injection of Gö6983 (0.25 mg/mouse/day). Mice were handled in accordance with the guidelines of the animal committee of Kyoto Sangyo University.

RT-qPCR

Extraction of total RNAs and RT-qPCR were performed using TRIzol reagent and QuantStudio3 Flex Real-time PCR system (Thermo Fisher Scientific), respectively. Gene specific primers for E-cadherin (One step TB Green PrimeScript RT-PCR Kit II) were obtained from Takara Bio. The PCR conditions for gene amplification were as follows: one cycle (42 °C for 5 min and 95 °C for 10 s) and 40 cycles (95 °C for 5 s and 60 °C for 30 s). Data were normalized with GAPDH. The sequences of the primers used were E-cadherin: 5′-GTCACTGACACCAACGATAATCCT-3′ (sense) and 5′- TTTCAGTGTGGTGATTACGACGTTA-3′ (antisense); GAPDH: 5′- TGCACCACCAACTGCTTAG-3′ (sense) and 5′- GGATGCAGTGATGATGTTC-3′ (antisense).