DACT2 Is a Candidate Tumor Suppressor and Prognostic Marker in Esophageal Squamous Cell Carcinoma

Esophageal squamous cell carcinoma (ESCC) is one of the most aggressive carcinomas in China, due to its markedly poor prognosis despite significant advancements in multimodal therapies (1–3). In ESCC, the TGF-β pathway is altered, with TGF-β expression being upregulated, resulting in enhanced invasive capability by affecting a series of growth- or invasiveness- related genes (4–8). Thus, blocking TGF-β's function might be an effective therapy for ESCC.

Recently, DACT2, a member of dapper family, has been identified as an antagonist of TGF-β/Nodal signaling and an important factor involved in normal vertebrate development (9–12). In zebrafish, dpr2 (DACT2) suppresses Nodal-mediated mesoderm induction by promoting the lysosomal degradation of TGF-β receptors ALK4 and ALK5 (9). Interestingly, zebrafish dpr2 is essential for convergent extension movement while acting as an enhancer of noncanonical Wnt signaling (13). Unlike zebrafish DACT2, murine DACT2 is expressed in all 3 germ layers during gastrulation, and targets the TGF-β type I receptor ALK5 for degradation. (12). Knockdown of mDACT2 expression in collecting duct cells enhances sensitivity to TGF-β to change cell behavior (10). Dpr2(DACT2)-deficient (Dpr2^−/−) mice showed accelerated re-epithelialization during cutaneous wound healing, which is associated with enhanced response to TGF-β signaling (11). Because of these important functions and the fact that DACT2 gene is conserved in evolution (14–16), DACT2 is believed to play an important role in the development of human cancers. However, little is known about the roles of DACT2 in tumorigenesis, although it has been predicted to have function in a series of cancers (15). In this study, we investigated the DACT2 expression pattern, the functional mechanism, and its prospect as a potential prognostic in ESCC.

Human kidney 293T cells and human ESCC cell lines (EC171, EC8712, KYSE150, KYSE180, and KYSE510) were used. 293T and KYSE cell lines were provided by Prof. Dong Xie (Institute for Nutritional Sciences of Chinese Academy of Sciences, Shanghai China). EC171 and EC8712 cell lines were provided by Prof. Sai-Wah Tsao (Department of Anatomy, University of Hong Kong, Hong Kong, China). KYSE cells were cultured in RPMI-1640 medium (HYCLONE). EC cell lines were cultured in 199 medium (HYCLONE). The 293T cell line was cultured in Dulbecco's modified Eagle medium (HYCLONE). All media were supplemented with 10% FBS and antibiotics (100 U/mL penicillin and 100 mg/L streptomycin). Cells were grown under a 5% CO₂ atmosphere at 37°C. All 5 human ESCC and 293T cell lines were tested and authenticated by short tandem repeat DNA profiling (Land Huagene Biosciences Co., LTD), and the authentication results are shown in Supplementary Fig. S1.

For DACT2 expression vector construction, the DACT2 coding region was amplified from EC8712 cells by using the following primers: 5′-AAGCTTATGTGGACGCCGGGCGGAC-3′ and 5′-GGTACCGTCACCATGGTCATGACCTTCAGGGC-3′. The PCR product was then subcloned into the pEGFP-N1 plasmid (Clotech) or pcDNA3 (Invitrogen) plasmid vector.

For transient transfection assays, plasmids were transfected with Lipofectamine 2000, and cells were harvested after 48 hours for further analysis.

Stable transfection process was conducted as described (17). For stable cell line generation, KYSE150 cells were plated into 35 mm dishes at least 24 hours before transfection to achieve 60% to 80% confluence per well. pEGFP/DACT2 or control vectors were transfected with Lipofectamine 2000 (Invitrogen) following the manufacturer's instructions. Transfected cells were selected in G418 (600 μg/mL) for 1 month. Pooled transfectants were obtained and expanded for further analysis.

The technique used was described previously (18). The primary antibodies used in this study are summarized in Supplementary Table S1. Antigen–antibody complexes were detected by Western blotting luminol reagent (Santa Cruz Biotechnology).

Transfected cells were trypsinized, counted with a cell counter (Bio-Rad), and then 500 cells were inoculated in each well of 6-well plates. Cultures were maintained for 3 weeks, and cells were then fixed, stained, and photographed.

For cell-cycle analysis by flow cytometry, transfected cells were fixed with 70% ethanol overnight at 4°C. Cell pellets were incubated in PBS containing 0.1% Triton-X 100 for 10 minutes at room temperature, and then were treated with RNase (50 μg/mL) for 10 minutes and propidium iodide (PI; 5 μg/mL) for 30 minutes, respectively. DNA content was analyzed by flow cytometry (BD FACSAria II). FlowJo 7.6 software was used to determine the cell-cycle phases.

The experimental protocols of cell migration and invasion assays were described previously (7). Transfected cells were counted with a cell counter (Bio-Rad), and 3 × 10⁴ cells were seeded for the cell migration assay and 5 × 10⁴ cells for the invasion assay.

Stably transfected pEGFP/DACT2 and pEGFP/vector cells (3.0 × 10⁶ cells/flank) were injected into 8-week-old male nude mice. The resulting tumors were measured twice a week, and tumor volume (mm³) was calculated using the formula: 1/2 × length × width × width. Tumors were harvested 18 days after injection and individually weighed before fixation. Data are presented as tumor volume (mean ± SD) and tumor weight (mean ± SD). Statistical analysis was conducted via SPSS using the Student t test. Tumor tissue was analyzed by hematoxylin and eosin (H&E), and DACT2 (1:100; Abcam) and cytokeratins 5/6 (ready to use, clone numer, CK5/6. 007; ZSGB-Bio) were detected by immunohistochemistry.

The tumor specimens used in this study were described previously (19). Briefly, tumor specimens were obtained from 168 patients with primary ESCC at Shantou Central Hospital (Guangdong, China) from 2000 to 2006. The follow-up study was conducted up to December 31, 2012. We retrieved information, including sex, age, stage of disease, and histopathological factors, from the patient hospital charts. Patient data are summarized Supplementary Table S2. The patients who experienced severe postoperative complications, who died of other tumors or other causes were excluded. The pathologic features of the specimens were classified on the basis of the seventh edition of the tumor–node–metastasis (TNM) classification of the International Union against Cancer. The local ethics committee approved the study.

Tissue microarray construction was as described previously (19). The primary antibody was rabbit polyclonal antibody against DACT2 (1:100; Abcam). In assays of negative controls, the primary antibody was omitted.

The evaluation of immunostaining was also described (19). A score range of ≤6 or >6 was marked as negative or positive, respectively. The correlations between DACT2 protein expression levels and clinicopathologic variables were analyzed using Kendall tall-b test. Survival curves were estimated by the Kaplan–Meier method and compared by the log-rank test.

Cignal Finder cancer 10-pathway Reporter Array (SABiosciences) was used on the basis of the manufacturer's instructions. Briefly, cells were inoculated in 96-well plates and grown to 50% to 80% confluence. Then, cells were transfected with 100 ng of each dual luciferase Cignal transcription factor-responsive reporter constructs and 200 ng of pcDNA3/DACT2 or pcDNA3 constructs by using with Lipofectamine 2000 (Invitrogen) following the manufacturer's instruction. After transfection, cells were incubated for 48 hours and harvested in Passive Lysis Buffer (Promega). The luciferase reporter activity of the lysates was measured using Dual-Luciferase Reporter Assay System (Promega) according to the manufacturer's recommendations. Statistical analysis was conducted with SPSS using the Student t test.

For F-actin analysis via flow cytometry, cells were fixed with 70% ethanol overnight at 4°C. Cell pellets were incubated in PBS containing 0.1% Triton-X 100 for 15 minutes at room temperature, and then were treated with phalloidin–coumarin (2 μg/mL; Sigma –Aldrich) for 30 minutes. F-actin content was analyzed by flow cytometry (BD FACSAria II). FlowJo 7.6 software was used to analyze the F-actin content of each group.

The F-actin and G-actin ratio was determined using the G/F-actin in vivo assay kit (Cytoskeleton, Inc.) according to the manufacturer's instructions. In brief, cells were homogenized in the lysis and F-actin stabilization buffers followed by centrifugation for 1 hour at 100,000 × g at 37°C for F-actin separation. The supernatants containing G-actin were collected, and the pellets containing F-actin were resuspended in F-actin depolymerizing solution. Equal amounts of both supernatant (G-actin) and the resuspended pellet (F-actin) were subjected to Western blotting with the use of anti-actin antibody.

Transfected cells grown on coverslips were treated with or without TGF-β1 (5 ng/mL, Invitrogen) for 10 minutes after starvation, and then were fixed with 3.7% paraformaldehyde in PBS for 10 minutes, rinsed with PBS, permeabilized in 0.1% Triton X-100 for 5 minutes. After washing, nonspecific binding was blocked by incubation of cells with blocking solution (10% donkey serum in PBS) for 30 minutes. Next, cells were incubated with a primary antibody overnight at 4°C, followed by donkey anti-mouse immunoglobulin G (IgG; DyLight 488) or donkey anti-rabbit IgG (DyLight 488; Jackson) for 30 minutes at 37°C. Finally, the cells were incubated with 100 nmol/L Acti-stainTM 555 phalloidin (Cytoskeleton, Inc) and then with 4′, 6-diamidino-2-phenylindole (Sigma). Cells were analyzed by Olympus FV1000 confocal microscope (Olympus).

First, we assessed the expression level of DACT2 in ESCC cell lines, using 293T cells transfected with DACT2 plasmid as positive control. The results showed that DACT2 was weakly detected in the SHEE and KYSE180 cell lines, but rarely detected in EC171, KYSE150 cell lines, as well as 293T cell (Fig. 1A). Then, for functional investigation of DACT2, the low-expressing cell lines, such as KYSE150 and KYSE180, were selected for high exogenous DACT2 expression (Fig. 1B) and further analysis. Next, we monitored the effect of DACT2 on the capacity of ESCC cell colony formation. In contrast to the control, DACT2-transfected cells displayed dramatically fewer colonies (Fig. 1C). Furthermore, we examined the effects of DACT2 on cell-cycle progressions in ESCC cells (Fig. 1D). Enhanced DACT2 expression blocked ESCC cell-cycle progression at the G₂/M phase. These results above suggested that DACT2 could inhibit the proliferation of ESCC cells.

To determine whether DACT2 may play a role in cell migration and invasion, a Transwell migration assay and Matrigel invasion chambers were used in this study. Our results showed that DACT2-transfected cells had lower migration and invasion rates compared with control cells (Fig. 1E and F), suggesting that DACT2 inhibits the invasiveness of ESCC cells.

We further tested whether DACT2-expressing ESCC cells could show attenuated tumorigenesis in an in vivo model. Control and DACT2-transfected KYSE150 cells were injected into 10 nude mice in each group. Seven days after injection, tumor development in both control and DACT2 groups was measured. Tumors in the control group (mean size: 74 ± 37 mm³) were much larger than those in the DACT2-expressing group (mean size: 47 ± 22 mm³). In addition, DACT2-expressing tumors displayed significantly slower growth rates. At the end of observation (18 days), the tumors of DACT2-expressing tumors grew to 159 ± 131 mm³ in size (0.145 ± 0.099 g in weight), whereas tumors in the control group grew to 953 ± 315 mm³ (0.600 ± 0.222 g in weight; Fig. 2A–C). Similar results were observed when the experiment was terminated 8 days later (Supplementary Fig. S2). H&E staining of the xenograft tumor tissue showed that the tumor cells of control group were undifferentiated with a higher nuclear to cytoplasmic ratio compared with DACT2-expressing tumors. Furthermore, we detected CK5/6, a marker for poor differentiation (20), using immunohistochemistry. CK5/6 immunohistochemistry gave diffuse positive staining in the control tumors, but no staining in the DACT2-expressing tumors (Fig. 2D). These data show that DACT2 contributes to growth inhibition and enhanced differentiation of tumor cells.

To explore the clinical significance of DACT2 expression profiles in ESCC, we immunohistochemically stained for DACT2 in patient's ESCC tumor tissue. DACT2 expression was detected in nontumor esophageal epithelium (Fig. 3A and B). In cancerous tissues, poorly differentiated tumors (Fig. 3C and D) displayed less intense DACT2 staining than those of advanced differentiation (Fig. 3E and F). Further statistic analysis showed that DACT2 expression significantly correlated with differentiation in ESCC (Table 1; P = 0.008). Moreover, the 5-year survival rate in patients with DACT2-positive tumors was much higher than that in patients with DACT2-negative tumors (54.9% for DACT2-positive patients and 45.1% for DACT2-negative patients; P = 0.045; Fig. 4). These results indicated that DACT2 might be a potential prognostic marker of ESCC.

Previous studies showed that DACT2 inhibits TGF-β activity via lysosomal inhibitor-sensitive degradation of the TGF-β receptor (9, 12). However, whether a similar mechanism exists in ESCC remains unclear. We used a Cignal Finder 10-pathway Reporter Array to examine cellular signaling mechanisms that may be affected by DACT2 in ESCC cells. Only TGF-β signaling was altered by DACT2 overexpression (Fig. 5A; the ratio of vector vs. DACT2 = 2.42; P < 0.05). Further Western blotting analysis also showed that elevation of DACT2 expression reduced p-Smad 2/3, an index of TGF-β activity, but not other signaling pathway factors in ESCC cells (Fig. 5B). To further confirm the antagonism of TGF-β signaling by DACT2, cells were treated with TGF-β1 after cotransfection with a SMAD response element reporter gene and the DACT2/vector plasmid. The results showed that TGF-β1 stimulated the expression of the SMAD reporter gene in ESCC cells, and DACT2 interfered with TGF-β1–mediated induction of the SMAD-dependent reporter gene (Fig. 5C). Moreover, when treated with proteasome inhibitor such as MG132 and lysosomal inhibitors such as NH₄Cl and chloroquine, the TGF-β–dependent activity recovered (Fig. 5D). These results suggested that DACT2 remarkably suppressed the activity of TGF-β signaling in ESCC via both proteasome and lysosomal pathways.

TGF-β signaling was found to alter a series of cytoskeleton-related genes expressions, which could induce cytoskeleton recombination that could subsequently promote/inhibit cell migration (21–26). In cytoskeleton rearrangement, F-actin reorganization is the essential step (27, 28). Thus, we investigated whether DACT2 affects F-actin rearrangement. Phalloidin–coumarin staining indicated that DACT2-restored expression increased the F-actin contents (Fig. 6A). The subsequent F-actin isolation and immunoblot analysis also showed that DACT2 increased F-actin and decreased G-actin molecules (Fig. 6B). In comparing with the control group, cells with upregulated DACT2 displayed fewer filopodia. When treated with TGF-β1, the filopodia were induced in control cells but not in the DACT2-expressing cells (Fig. 6C). These results indicated that DACT2 was involved in breaking the balance between F-actin and G-actin content, and inhibiting the filopodia formation induced by TGF-β1.

To explore the molecular mechanism of F-actin rearrangement induced by DACT2, we first tested the presence of cofilin and ezrin–redixin–moesin (ERM), which played a crucial role in F-actin rearrangement (29, 30). Our results showed that DACT2 could reduce ERM, phosphorylated ERM (pERM), and phosphorylated cofilin (p-cofilin), and increased cofilin (Fig. 6D). Further immunofluorescence results showed that restored DACT2 expression could increase cofilin and decrease p-cofilin. In addition, DACT2 could cause the rearrangement of both cofilin and p-cofilin location, in accordance with F-actin distribution (Fig. 6E). These results suggested that DACT2 alters the level or/and the location of a series of cytoskeletal-related proteins, resulting in F-actin accumulation.

DACT2 has been found to have play important functions in embryo development and wound healing in animal (9–12). However, roles for DACT2 in humans have rarely been known. In keratinocytes, downregulation of DACT2 enhances cell migration but has no effect on proliferation (11). In collecting duct-derived cells, DACT2 downregulation causes a migration defect and partial epithelial–mesenchymal transition (10). In lung cancer cells, restoration of DACT2 expression causes cell arrested in the G₀/G_lphase (31). We show that elevation of DACT2 expression inhibits not only cell proliferation, in which the cells arrest in G₂/M phase, but also inhibits the invasiveness of ESCC cells (Fig. 1). Various reports show that the DACT2 has distinct roles in different cells, indicating the tissue type- or species-specific roles of DACT2 function. These actions of DACT2 were also confirmed in vivo. Restored expression of DACT2 significantly reduces the ability of tumor cells to grow as xenograft tumors in nude mice (Fig. 2). In ESCC, increased expression of DACT2 correlates with well differentiation status (Figs. 2D and 3; Table 1); and a higher survival rate is observed in patients with DACT2-expressing tumor (Fig. 4). We therefore propose that DACT2 serves as a protective factor against the progression of ESCC. The mechanism by which decreased expression of DACT2 in poorly differentiated ESCC tissue occurs is unknown. In this process, epigenetic regulations might be involved as reflected in methylation of the DACT2 promoter in colorectal cancer cells and lung cancer cells (31, 32).

In animals ranging from fish to mice, the function of DACT2 as a negative regulator of TGF-β/Nodal signal pathway is conserved in evolution (9, 11, 12). In the mouse model, DACT2 has no detectable effects on canonical Wnt signaling and has minor effects on noncanonical Wnt signaling only when DACT2 is massively overexpressed (12). However, little is known of its roles in human cancer, such as ESCC. Recent studies showed that DACT2 could suppress Wnt signaling by inhibiting T-cell factor/lymphoid enhancer factor in lung cancer (31). However, in our study, DACT2 significantly decreases pSmad2/3 binding to its response element, but has no effect on the activators or transcriptional regulatory elements of the canonical Wnt pathway or other pathways (Fig. 5A–C and Supplementary Fig. S3), suggesting possible tissue-type specificity of function. Furthermore, DACT2 inhibits TGF-β activity via lysosomal inhibitor-sensitive degradation of ALK4 and/or ALK5, the TGF-β receptors (9, 12). However, as shown in our study, TGF-β signaling activity could be recovered by treating with both proteasome and lysosomal inhibitors (Fig. 5D), suggesting that DACT2 might affect TGF-β signaling activity by both proteasome and lysosomal pathways in ESCC. On the basis of recent studies, both proteasome and lysosomal pathways are involved in TGF-β signaling regulation, which targets not only the receptors but also the important mediators of TGF-β signaling (33–36). Thus, in ESCC system, new mechanism by which DACT2 suppresses TGF-β signaling pathway might exist. However, it is still unclear which receptor or/and mediator of TGF-β signaling pathway is involved.

TGF-β can cause the morphologic changes of cell via actin cytoskeleton reorganization (37–39). In mIMCD3 cells, knockdown of DACT2 altered actin organization to change the morphogenetic behavior (10). Therefore, we became curious as to whether there was any connection between DACT2 and F-actin. In ESCC cells, DACT2 could alter the equilibrium between F-actin and G-actin, disturb actin depolymerization, and inhibit filopodia formation (Fig. 6A–C), consistent with an ability of DACT2 to modify actin cytoskeleton organization. In actin cytoskeleton rearrangement, a series of molecular events are involved. The TGF-β signaling pathway has an essential role in regulation of cofilin and ERM (25, 40, 41). Cofilin can determine the cytoskeleton dynamics by its phosphorylation status. Phosphorylation of cofilin by TGF-β/RhoA/ROCK/LIMK pathway could inhibit ability of cofilin to promote filament disassembly (29, 40, 41). Nevertheless, cofilin can also promote actin polyzerization, indicating that it serves as a dynamic component of the steering wheel of cells (42). Besides cofilin, ERM family members play an important role in actin rearrangement. Members of the ERM family, which can be activated by Rho kinase/phosphatidylinositol 4,5-bisphosphate (PI(4,5)P2), act both as linkers between actin cytoskeleton and plasma membrane proteins, and as signaling transducers in responses involving cytoskeletal remodeling (30, 43, 44). Our data indicate that DACT2 might alter the conversion between cofilin and p-cofilin, and decrease both ERM and pERM in ESCC (Fig. 6D). In addition, DACT2 could rearrange cofilin and p-cofilin (Fig. 6E). These data suggest that DACT2 might cause the disruption of F-actin/G-actin equilibrium by altering the level or/and location of TGF-β–dependent cytoskeletal-related proteins in ESCC cells.

In summary, we show that decreased DACT2 protein expression correlate with poor differentiation status and poor survival of patients with ESCC. Restoration of DACT2 expression significantly inhibits not only proliferation and invasiveness of ESCC cells in vitro, but also the tumorigenicity in vivo. DACT2 expression suppresses TGF-β signaling, and leads to F-actin rearrangement induced by cofilin and ERM proteins. Taken together, we propose that DACT2 served as a prognostic marker by inhibiting tumor cell malignancy in ESCC.

No potential conflicts of interest were disclosed.

Conception and design: J. Hou, M. Guo, E.-M. Li, L.-Y. Xu

Development of methodology: J. Hou, L.-D. Liao, Y.-M. Xie, F.-M. Zeng, C.-X. Yang, J. Shen, M. Guo, E.-M. Li, L.-Y. Xu

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): J. Hou, Y.-M. Xie, L.-Y. Li, Q. Zhao, M. Guo, E.-M. Li, L.-Y. Xu

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): J. Hou, B. Chen, M. Zhu, Q. Zhao, M. Guo, E.-M. Li, L.-Y. Xu

Writing, review, and/or revision of the manuscript: J. Hou, M. Guo, E.-M. Li, L.-Y. Xu

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): J. Hou, Y.-M. Xie, F.-M. Zeng, X. Ji, B. Chen, T. Chen, X.-E. Xu, M. Guo, E.-M. Li, L.-Y. Xu

Study supervision: J. Hou, M. Guo, E.-M. Li, L.-Y. Xu

The authors thank Profs. Stanley Lin and Junhui Bian for manuscript revision.

This work was supported by grants from the National Basic Research Program (973 Program No. 2012CB526608 and No. 2010CB912802), the National High Technology Research and Development Program of China (No. 2012AA02A503), the Natural Science Foundation of China-GuangDong Joint Fund (No.U0932001), and the National Science Foundation of China (No. 81172264).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.