Requirement of the Akt/beta-catenin pathway for uterine carcinosarcoma genesis, modulating E-cadherin expression through the transactivation of slug

Abstract

Uterine carcinosarcomas (UCSs) are considered to represent true examples of the epithelial-mesenchymal transition. Akt plays a key role in the induction of epithelial-mesenchymal transition, but little is known about its involvement in tumorigenesis. Here we examined the functional roles of the Akt/beta-catenin pathway in UCSs. In clinical samples, phospho-Akt (pAkt) expression was found to be significantly increased in mesenchymal compared with epithelial components, exhibiting both positive and negative correlations with nuclear beta-catenin and E-cadherin, respectively. Expression levels of the transcription factor Slug were also significantly up-regulated in the mesenchymal components and strongly correlated with both pAkt and nuclear beta-catenin. In endometrial cancer cell lines, active Akt induced the stabilization of nuclear beta-catenin through the phosphorylation of GSK-3beta, and this, in turn, led to the transactivation of Slug, which was mediated by nuclear beta-catenin. Moreover, Slug overexpression itself caused repression of E-cadherin, with subtle changes in cell morphology. In addition, knockdown of the retinoblastoma gene product (Rb) up-regulated pAkt and repressed E-cadherin, consistent with the in vivo finding of significantly decreased Rb expression in mesenchymal components. These findings suggest that changes in the Akt/beta-catenin pathway, as well as alterations in Rb expression, may be essential for both the establishment and maintenance of phenotypic characteristics of UCSs, playing key roles in the regulation of E-cadherin through the transactivation of the Slug gene.

Figure 1

Expression of Rb and pRb in UCSs. A, Top: H&E staining and immunohistochemistry for Rb, pRb, and Ki-67 are illustrated. The mesenchymal lesion indicated by the arrow is magnified in the inset. Bottom: Nuclear LIs for Rb, pRb, and Ki-67 in epithelial (Ep) and mesenchymal (Me) components of uterine CS (carcinosarcoma), as well as non-CS tumors. B, Left and middle: Comparison of expression levels of several EMT-related molecules for mRNA by RT-PCR (left) and protein by Western blot (middle) between Ishikawa (Ish) and Hec251 cells. E-cad, E-cadherin; FN, fibronectin Right: Ishikawa and Hec251 cells were transfected with Top reporter constructs to determine β-catenin-dependent transcriptional activity. Relative activity was determined based on arbitrary light units of luciferase activity normalized to pRL-TK activity. The activities of the reporter plus the effector relative to that of the reporter plus empty vector are shown as means ± SD. C: Analysis of mRNA (at 24 hours after transfection) (left) and total protein (at 48 hours) (right) in Hec251 cells transfected 30 and 50 nmol/L siRNAs for Rb. A siRNA with no homology to the mammalian genome was used as a negative control (con). Original magnifications: ×100 (A, top); ×400 (A, inset).

Figure 2

Akt regulates expression of E-cadherin and β-catenin. A: Serial sections through a UCS. H&E and immunohistochemistry for E-cadherin, pAkt, and β-catenin are illustrated. The lesions enclosed by boxes in the top panels are magnified in the bottom panels. B: Immunohistochemical (IHC) scores for E-cadherin (left) and LIs for pAkt and nuclear (N) β-catenin (right) in epithelial (Ep) and mesenchymal (Me) components of uterine CS (carcinosarcoma), as well as non-CS tumors. C, Top left: Western blot analysis of E-cadherin (E-cad) was performed with subcellular protein fractions (Cyt, cytoplasmic; Nu, nuclear) extracted from myr-Akt-transfected Hec251 cells. In the pAkt panel, upper bands (indicated by the asterisk) are exogenous myr-Akt, in contrast to the lower bands demonstrating endogenous pAkt. Bottom left: After transfection of myr-Akt into Hec251 cells, staining for pAkt and E-cadherin was performed. Note the lack of E-cadherin expression in cell overexpressing myr-Akt (indicated by arrows). Top right: Analysis of E-cadherin (E-cad) mRNA levels by RT-PCR with total RNA extracted from myr-Akt-transfected Hec251 cells. Bottom right: Hec251 cells were transfected with reporter constructs containing E-cadherin promoter (E-cad Luc), together with expression plasmid for myr-Akt. Relative activity was determined based on arbitrary light units of luciferase activity normalized to pRL-TK activity. The activities of the reporter plus the effector relative to that of the reporter plus empty vector are shown as means ± SD. D, Top left: Western blot analysis of HA-β-catenin, total GSK-3β, and phospho-GSK-3β (pGSK-3β) was performed with subcellular protein fractions (Cyt, cytoplasmic; Nu, nuclear) extracted from myr-Akt-transfected Ish-bcat mo.#25 cells. Bottom left: After transfection of myr-Akt, Ish-bcat#25 cells were treated with cycloheximide (25 μg/ml) for the times shown. Western blotting was performed using anti-HA, anti-pAkt, anti-GSK-3β, and anti-pGSK-3β antibodies. In the pAkt panel, upper bands (indicated by the asterisk) are exogenous myr-Akt, in contrast to lower bands demonstrating endogenous pAkt. Top and bottom right: Relative amounts of HA-β-catenin were calculated by normalization to the signals for β-actin, using NIH Image. Expression levels of HA-β-catenin observed in myr-Akt-transfected cells in the absence of cycloheximide treatment (0 hours) were set as 100%. The experiment was performed in triplicate. Original magnifications: ×100 (A, top); ×400 (A, bottom).

Figure 3

Up-regulation of Slug in UCSs. A, Left: Serial sections through a UCS. H&E staining and immunohistochemistry for Slug and Snail are illustrated. Boxes with dotted and solid outlines enclose epithelial and mesenchymal lesions magnified in a and b, respectively. Right: Nuclear LIs for Slug and Snail in epithelial (Ep) and mesenchymal (Me) components of UCSs. B, Left: Analysis of mRNA levels for Slug and Snail by RT-PCR with total RNA extracted from Hec251 cells transfected with myr-Akt and β-catenin. Right: Hec251 cells were transfected with reporter constructs containing 1.89 kbp of the Slug promoter (Slug Luc), together with expression plasmids for myr-Akt and β-catenin. Relative activity was determined based on arbitrary light units of luciferase activity normalized to pRL-TK activity. The activities of the reporter plus the effector relative to that of the reporter plus empty vector are shown as means ± SD. C, Left: Analysis of E-cadherin (E-cad) mRNA by RT-PCR (top) and protein by Western blot (bottom) in HA-Slug-transfected Hec251 cells. Detection of HA-Slug mRNA was performed using a combination of 5′-HA forward (5′-TACCCATACGATGTTCCAGATTACGC-3′) and Slug mRNA reverse (Table 1) primers. Right: Hec251 cells were transfected with reporter constructs containing the E-cadherin promoter (E-cad Luc), together with a Slug expression plasmid. D: Hec251 cells stably overexpressing HA-Slug. Left: Analysis of E-cadherin expression levels for mRNA by RT-PCR (top) and protein by Western blot (bottom) in three independent stable cell lines (nos. 13, 31, and 50). Right: The subtle changes in the morphology in the stable cell lines. Original magnifications: ×100 (A, top); ×400 (A, bottom).

Figure 4

Transactivation of Slug promoter by β-catenin. A: The various promoter deletion constructs used for evaluating transcriptional regulation of the Slug promoter. Numbers for each construct corresponding to the 5′- and 3′-bp locations within the promoter (relative to the translation start site as +1). B and C: Hec251 cells were transfected with 5′ various deletion constructs of Slug promoter, together with the β-catenin expression plasmid. Relative activity was determined based on arbitrary light units of luciferase activity normalized to pRL-TK activity. The activities of the reporter plus the effector relative to that of the reporter plus empty vector are shown as means ± SD. D: Hec251 cells were transfected with reporter constructs containing 1.89 kbp of the Slug promoter, together with expression plasmids for β-catenin, dominant-negative (DN)-TCF4, and p300.

Figure 5

Schematic representation of possible roles of the Akt/β-catenin pathway in determination of phenotypic characteristics of UCS.