Sphingosine kinases and sphingosine-1-phosphate are critical for transforming growth factor beta-induced extracellular signal-regulated kinase 1 and 2 activation and promotion of migration and invasion of esophageal cancer cells

Abstract

Transforming growth factor beta (TGFbeta) plays a dual role in oncogenesis, acting as both a tumor suppressor and a tumor promoter. These disparate processes of suppression and promotion are mediated primarily by Smad and non-Smad signaling, respectively. A central issue in understanding the role of TGFbeta in the progression of epithelial cancers is the elucidation of the mechanisms underlying activation of non-Smad signaling cascades. Because the potent lipid mediator sphingosine-1-phosphate (S1P) has been shown to transactivate the TGFbeta receptor and activate Smad3, we examined its role in TGFbeta activation of extracellular signal-regulated kinases 1 and 2 (ERK1/2) and promotion of migration and invasion of esophageal cancer cells. Both S1P and TGFbeta activate ERK1/2, but only TGFbeta activates Smad3. Both ligands promoted ERK1/2-dependent migration and invasion. Furthermore, TGFbeta rapidly increased S1P, which was required for TGFbeta-induced ERK1/2 activation, as well as migration and invasion, since downregulation of sphingosine kinases, the enzymes that produce S1P, inhibited these responses. Finally, our data demonstrate that TGFbeta activation of ERK1/2, as well as induction of migration and invasion, is mediated at least in part by ligation of the S1P receptor, S1PR2. Thus, these studies provide the first evidence that TGFbeta activation of sphingosine kinases and formation of S1P contribute to non-Smad signaling and could be important for progression of esophageal cancer.

FIG. 1.

Activation of MAP kinases by TGFβ and S1P. OE33 cells were treated with 80 pM TGFβ for the indicated times (A), with the indicated concentrations of S1P (μM) for 15 min (B), with 50 nM S1P for the indicated length of time (C), or with either 80 pM TGFβ or 50 nM S1P for 0, 15, and 30 min (D). Cells were lysed, and equal amounts of lysates were separated by SDS-PAGE. Activation of ERK1/2, p38, and Smad3 was determined by immunoblotting with phospho-specific antibodies as indicated. Blots were stripped and reprobed with ERK1/2, p38, or Smad3 antibodies to demonstrate equal loading.

FIG. 2.

TGFβ and S1P stimulate chemotactic migration (chemotaxis) and invasion to similar extents. (A) Transwell migration assays were performed as described in Materials and Methods. Medium without or with TGFβ (80 pM) or S1P (50 nM) was added to the lower chamber as indicated. Data are means ± standard errors of the means from three separate experiments. Asterisk, P < 0.05. (B) Invasion assays were performed with Matrigel-coated membranes as described in Materials and Methods. Medium, without or with TGFβ (80 pM) or S1P (50 nM), was added to the lower chambers as indicated. Data are means ± standard errors of the means from three separate experiments. Asterisk, P < 0.05.

FIG. 3.

Chemotactic migration and invasion induced by TGFβ or S1P is ERK1/2 dependent. (A) OE33 cells were incubated for 15 min in the presence or absence of TGFβ (80 pM), S1P (50 nM), and the MEK inhibitor PD98059 (50 μM) as indicated. Cells were lysed, and activation of ERK1/2 was determined by immunoblotting equal amounts of lysates with anti-phospho ERK1/2. Blots were stripped and reprobed with anti-ERK1/2 as a loading control. (B) Transwell invasion assays were performed as described in Materials and Methods. PD98059 (50 μM) was added to the upper chambers, and either TGFβ (80 pM) or S1P (50 nM) was in the lower chamber. OE33 cells were transduced with either control or dominant negative MEK (DN-MEK) adenoviral vectors (C and D) or transfected with control or ERK1 siRNAs (E and F). (C and E) Cells were analyzed by Western blotting as in panel A, except that blots were also stripped and reprobed with actin as a control for equal loading and transfer. Chemotactic migration (D) and invasion (F) assays were performed as described in Materials and Methods. Data are means ± standard errors of the means from three separate experiments. Asterisk, P < 0.05.

FIG. 4.

ERK1/2 activation, chemotactic migration, and chemotactic invasion are G_i dependent (A to C). OE33 cells were pretreated for 2 h with 100 ng/ml pertussis toxin as indicated. (A) Cells were treated with TGFβ (80 pM) or S1P (50 nM) for 15 min. Cell lysates were separated by SDS-PAGE and analyzed for pERK1/2 and total ERK1/2 by Western analysis. Chemotactic migration (B) and invasion (C) assays were performed as described in Materials and Methods. Data are means ± standard errors of the means from triplicate cultures. Asterisk, P < 0.05.

FIG. 5.

The SphK inhibitor DMS blocks TGFβ-induced ERK1/2 activation and cell motility. (A) OE33 cells were incubated for 15 min in the absence or presence of TGFβ (80 pM) and DMS (5 μM) as indicated, and ERK1/2 activation was determined by sequential immunoblotting with anti-phospho ERK1/2 and anti-ERK1/2. Chemotactic migration (B) and invasion (C) assays were performed as described in Materials and Methods. TGFβ (80 pM) and DMS (5 μM) were added to the lower and upper chambers, respectively. Data are means ± standard errors of the means from three separate experiments. Asterisk, P < 0.05.

FIG. 6.

SphK1 and SphK2 have different roles in TGFβ-induced chemotactic migration and invasion. OE33 cells were transfected with control, SphK1, or SphK2 siRNA as indicated. (A) RNA was isolated and reverse transcribed, and SphK1, SphK2 and GAPDH levels were measured by QRT-PCR. SphKs are normalized to GAPDH. Data are expressed as the change with respect to control siRNA. (B) Total cell lysates were assayed for the expression of the indicated proteins by immunoblotting with anti-SphK1 and anti-SphK2 antibodies. Blots were also probed with anti-p65 to show equal loading. Chemotactic migration (C) and invasion (D) were determined, and data were expressed as the change in the ratio of migrating cells in TGFβ treated to untreated cells for each group of transfectants. The results are means ± standard errors of the means from three separate experiments. Asterisk, P < 0.05. OE33 cells (E) or SEG1 cells (F) were transfected with the indicated siRNAs and cultured without or with TGFβ (80 pM) as indicated. The activation of ERK1/2 was determined by immunoblotting with pERK1/2 and ERK1/2 as loading controls.

FIG. 7.

TGFβ activates SphK to generate S1P. (A) OE33 cells were treated with TGFβ (80 pM) for the indicated time. Activation of SphK1 was determined by immunoblotting with anti-phospho Ser225. Blots were stripped and reprobed with anti-SphK1 antibody to confirm equal loading and transfer. (B and C) OE33 cells were treated with TGFβ for the indicated times and lysed, and SphK1 (B) and SphK2 (C) activities were determined with isoenzyme-specific assays. SphK activity is expressed as pmol/min/mg protein. (D) S1P mass levels were measured in duplicate cultures of 4 × 10⁵ cells as described in Materials and Methods and expressed as pmol. Data are means ± standard errors of the means. Asterisk, P < 0.05.

FIG. 8.

S1PR₂ receptor activation is involved in TGFβ-induced ERK1/2 activation, as well as in chemotactic migration and invasion. (A) S1P receptor mRNA expression in OE33 cells. RNA was isolated and reverse transcribed, and S1PR₁ to S1PR₅ and GADPH levels were measured by QRT-PCR and normalized to those of GAPDH. Data are expressed relative to S1PR₁ mRNA. (B, C, and D) OE33 cells were treated for 15 min without or with TGFβ (80 pM), S1P (50 nM), VPC23019 (10 μM), and JTE013 (JTE013, 1 or 10 μM) as indicated. Activation of ERK1/2 was determined by immunoblotting with pERK1/2 antibody. Blots were stripped and reprobed with anti-ERK1/2 antibody to demonstrate equal loading and transfer. Chemotactic migration (E) and invasion (F) of OE33 induced by 80 pM TGFβ or 50 nM S1P were determined in the absence or presence of VPC23019 (10 μM) and the indicated concentrations of JTE013. Data are means ± standard errors of the means from three separate experiments. Asterisk, P < 0.05.

FIG. 9.

S1PR₂ ligation is involved in TGFβ-induced ERK1/2 activation, as well as in chemotactic migration and invasion. OE33 cells were transfected with siControl or siS1PR₂ RNAs as indicated. (A) RNA was isolated and reverse transcribed, and S1P₂ and GAPDH levels were measured by QRT-PCR. Data are expressed as changes after normalization to GAPDH. (B) OE33 cells were treated with TGFβ (80 pM) or S1P (50 nM) for 15 min. Cell lysates were separated by SDS-PAGE and analyzed for pERK1/2 and total ERK1/2 by Western analysis. Chemotactic migration (C) and invasion (D) assays were performed as described in Materials and Methods.

FIG. 10.

Model of cross talk between TGFβ and S1P signaling pathways. See text for more details.