Disruption of TCF/β-Catenin Binding Impairs Wnt Signaling and Induces Apoptosis in Soft Tissue Sarcoma Cells

Soft tissue sarcomas (STS) are malignant tumors of mesenchymal origin and represent around 1% of adult cancers (1). STS are a very heterogeneous group of sarcomas comprising more than 50 different subtypes which can appear anywhere in the body. In case of localized disease, surgical resection with or without radiotherapy and chemotherapy is the standard curative treatment. Unfortunately, STS recur frequently as locally inoperable or metastatic disease, at which point systemic therapy is used to treat patients. For treating advanced-stage STS, chemotherapy consisting of doxorubicin and ifosfamide is the standard treatment with overall response rates of about 25% in the first-line setting (2). Despite this fact, STS are almost invariably fatal making the development of new therapeutic approaches based on personalized medicine with clearly defined molecular targets a primordial necessity.

In this context, current research is focusing on the study of new molecular pathways that provide a better understanding of sarcomagenesis. The Wnt/β-catenin signaling pathway regulates various processes that are important for cancer progression, including tumor initiation, tumor growth, cell senescence, cell death, differentiation and metastasis. Although the role of this signaling pathway is well known in many malignancies such as colorectal cancer (3), there is still limited knowledge of its role in sarcomas. Wnt signaling pathway is involved in development and in the regulation, self-renewal, and differentiation of mesenchymal stem cells (4). The involvement of Wnt signaling in these processes and the knowledge of the mesenchymal origin of sarcomas led to the study of its role in sarcomagenesis. It has been shown that this signaling pathway is activated through the canonical pathway in some types of sarcomas, including leiomyosarcoma, fibrosarcoma, and osteosarcoma, resulting in cell proliferation mainly through expression of the cell-cycle progress regulator CDC25A, which is a Wnt target gene (5). This activation is mainly accomplished by an autocrine loop (5–8) or via crosstalk with other signaling pathways including PI3K/AKT/mTOR pathway (9–11).

The canonical Wnt/β-catenin pathway is activated when a secreted extracellular Wnt ligand binds to a seven transmembrane receptor, Frizzled (FZD), and its coreceptors, low-density lipoprotein receptor-related proteins called LRP5/6. Once FZD is activated, it triggers a cascade of intracellular signals which inactivate the Axin–APC–GSK3β destruction complex that phosphorylates β-catenin, targeting it for degradation (12). When the function of the destruction complex is inhibited, free nonphosphorylated β-catenin accumulates in the cytoplasm, translocates to the nucleus, and binds to the T-cell factor/lymphoid enhancer factor-1 (TCF/LEF) family of transcription factors, thereby inducing canonical gene transcription. Deregulation of Wnt signaling can be driven by upstream or downstream alterations, leading to cancer development or growth. In this line, targeted molecular therapies for a variety of Wnt-activated cancers are now being developed. Furthermore, within the past several years, a number of agents, both small molecules and monoclonal antibodies, have entered clinical trials (13).

In this study, we demonstrate the efficacy of different Wnt inhibitors such as XAV939, which induces the stabilization of AXIN by inhibiting the poly(ADP)-ribosylating enzymes tankyrase 1 and tankyrase 2 (14) or PKF118-310. PKF118-310 disrupts the β-catenin/TCF interaction (15), inducing apoptosis in patient-derived STS cells and established cell lines. On one hand, the Wnt inhibitors studied reduced the expression of important Wnt genes involved in cell-cycle regulation in sarcomas, such as CDC25A (5). On the other hand, treatment of STS with Wnt inhibitors also decreased nuclear β-catenin levels and its mediated transcriptional activity, altogether leading to a subtype-independent inhibition of sarcoma cell growth. When combined with the conventional chemotherapeutic drug doxorubicin, PKF118-310 showed an additive antitumoral effect. Our findings support the impairment of the Wnt pathway as a potential therapeutic treatment for STS.

Experiments were conducted with nine different sarcoma cell lines (Supplementary Table S1), three colorectal cancer cell lines (HT29, HCT116, and SW480), and human mesenchymal stem cells (hMSC; Lonza) as controls. Colorectal cancer cell lines where kindly provided by Dr. Gwendolyn Barceló, IdISPa, Palma de Mallorca, Spain (2016) and hMSC cell line was kindly donated by Dr. Carlos Río, IdISPa, Palma de Mallorca, Spain (2016). All of them were grown according to instructions provided by commercial providers. 93T449 cell line was kindly provided by Dr. Florence Pedeutour (2012) and was established at Hospital de l'Archet, France (16, 17). PharmaMar S.A. (2011) kindly donated cell lines SW684 and SW872 (available from the ATCC). HT-1080, SK-UT-1, and SW982 cell lines were purchased from the ATCC (2011). AW cell line was kindly provided by Dr. Amancio Carnero (2011) and was established at CNIO, Madrid, Spain (18). 93T449 cell line was maintained in RPMI1640 medium (PAA) supplemented with 10% FBS (PAA), 100 U/mL penicillin/streptomycin (PAA), 1% Ultroser (Pall Life Sciences), and 0.5% Fungizone (Invitrogen). SW872, SW684, and SW982 were maintained in Leibovitz medium with l-glutamine (Invitrogen) supplemented with 10% FBS, 100 U/mL penicillin/streptomycin, and 1 mmol/L HEPES (Sigma-Aldrich). HT-1080 was maintained in MEM liquid with Earle Salts medium with l-glutamine (PAA) supplemented with 10% FBS, 100 U/mL penicillin/streptomycin, and 1 mmol/L HEPES. AW was maintained in F-10 Ham medium (Gibco) supplemented with 1% Ultroser, 10% FBS, 100 U/mL penicillin/streptomycin, and 1mmol/L HEPES. SK-UT-1 was maintained in DMEM (PAA) supplemented with 10% FBS, 100 U/mL penicillin/streptomycin, and 1 mmol/L HEPES. Cells were grown in a humidified incubator containing 5% CO₂ at 37°C. XAV939, IWR-1, and PKF118-310 were purchased from Sigma. Doxorubicin was kindly provided by the Pharmacology Department of Son Espases University Hospital.

Two primary cell cultures [CP0024 (2012) and CP0038 (2013)] were derived from resected leiomyosarcomas (Supplementary Table S1). The biopsies were minced in culture medium and then disaggregated by 30-minute incubation with collagenase (Gibco, 221 U/mg). Cells were cultured in RPMI1640 medium (PAA) supplemented with 20% FBS (PAA), 100 U/mL penicillin/streptomycin (PAA). Informed consent was obtained from all patients in accordance with the guidelines of the Ethical Committee of Clinical Investigation (CEIC-IB, Spain) and the Declaration of Helsinki.

A total of 5,000 cells were grown in a Lab-Tek II Chamber Slide (Lab-Tek) chamber, washed three times with a PBS plus 0.2% BSA solution and fixed with 100 μL of a solution of methanol:acetone (1:1), blocked using 300 μL of 5% BSA solution, incubated with the primary antibody β-catenin (D10A8; #8480, Cell Signaling Technology, 1:30) and with the secondary antibody (Alexa Fluor 488 Goat a-R A11008, Invitrogen, 1:150 dilution). To visualize the nuclei, DAPI was used (Fisher Scientific). Samples from two individual experiments performed by duplicate were analyzed with the confocal microscope, ZEISS LSM 710. The ZEN2011 (black edition 64bit) software was used to quantify β-catenin staining intensity, which was scored (score 1) into four categories 0–3 (0 = negative, 1 = intensity <90, 2 = intensity 90–100, 3 = intensity >100). The percentage of β-catenin–positive stained cells was also scored (score 2) into four categories 1–4 (1 = 0%–25%; 2 = 26%–50%; 3 = 51%–75%; 4 = 76%–100%). The level of β-catenin staining was evaluated by IRS (staining average immunoreactive score), which was calculated by multiplying the two scores previously described. On the basis of the IRS, β-catenin staining pattern was defined as negative (IRS: 0), weak (IRS: 1–4), moderate (IRS: 5–8), and strong (IRS: 9–12).

Total RNA was isolated using TRIzol Plus RNA Purification Kit (Ambion) according to the manufacturer's instructions. A total of 300 ng of RNA were reverse-transcribed into cDNA using High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems). Subsequent real-time PCR reactions were performed in duplicate (CFX96 Real-Time System, C1000 Thermal Cycler, Bio-Rad) using the TaqMan probes method. β2-Microglobulin was used as an internal control for normalization. To quantify the level of gene expression changes, the following formula was used: |{2^{ - \Delta \Delta {C_{\rm{t}}}}}$|⁠, where |\Delta \Delta {C_{\rm{t}}} = {C_{{\rm t}_{\rm target}}( {{\rm{control}} - {\rm{sample}}} ) - {C_{\rm{t}}}_{{\rm{reference}}}( {{\rm{control}} - {\rm{sample}}} )$|⁠. The TaqMan probes used were the following: β2-microglobulin (HS99999907), CDC25A (HS00947994), C-MYC (HS01067802), CUL4A (HS00757716), and AXIN2 (HS00610344; Applied Biosystems).

Cell viability was measured using the methylthiazoletetrazolium (MTT) method, which indicates the presence of metabolically active cells, using the CellTiter 96AQueous One Solution (Promega) following the manufacturer's instructions. Briefly, 5,000 cells were seeded in 96-well plates for 24 hours and exposed to increasing concentrations of inhibitors for 48 hours followed by addition of CellTiter reagent. Absorbance at 490 nm was detected using a multiwell scanning spectrophotometer (Synergy H1 microplate reader, Bio-tek). The mean percentage of cell viability relative to vehicle-treated cells was estimated from data of three individual experiments performed by triplicate. The compound concentration resulting in 50% inhibition of cell viability (IC₅₀) was determined using GraphPad software.

The effect of the inhibitors on cell cycle was assessed by flow cytometric analysis. Briefly, 30,000 cells were seeded in 24-well plates. After 24, 48, and 72 hours of treatment, cells were washed in PBS and fixed in 90% ethanol, collected by centrifugation and stained in 500 μL of a mixture of propidium iodide (50 mg/mL; Sigma-Aldrich), and ribonuclease A (50 mg/mL; Sigma-Aldrich) diluted in PBS for 1 hour. Cell populations at different stages of the cell cycle [sub-G₁ peak (apoptosis), G₁, S, and G₂–M] were estimated on the basis of their DNA content in flow cytometer using the BD system Verse BD FACScan and software FACSuit.

Western blot whole-cell extracts were prepared by lysing cells with lysis buffer (1% Nonidet P-40, 20 mmol/L Tris–HCl pH 7.4, 100 mmol/L NaCl, 10 mmol/L NaF, 1 mmol/L Na₃VO₄, and protease inhibitors “Complete”; Roche) on ice for 15 minutes. Nuclear and cytoplasmic fractions were obtained using Nuclear Extract Kit (Active Motif) and following the manufacturer's instructions. The antibodies used were the following: CDC25A (ab989, Abcam), β-actin (#3700), c-Myc (#9402), CUL4A (#2699), β-catenin (#8480), P-β-catenin (Ser552; #2951) from Cell Signaling Technology and α-tubulin (#T9026) from Sigma. The secondary antibodies were: Donkey anti-Rabbit (E 365D5) and Donkey anti-Mouse (E 510KC) from LI-COR. Immunoreactivity intensity of the bands was analyzed with the Odyssey imaging system from LI-COR.

Transient transfection of the TCF reporter system coupled to luciferase was performed using the Lipofectamine Plus (Invitrogen) method. TCF reporter system plasmids (pTOPFLASH and pFOPFLASH), expression plasmids pCMV-APC (APC WT, APC 1309Δ), and the empty vector pcDNA3.1 were kindly donated by Dr. Gabriel Capellà. Luciferase reporter assay was performed using the Dual-Luciferase Reporter Assay System (Promega) according to the manufacturer's instructions. Luciferase activity was quantified by the Luciferase Reporter Assay System (Promega) 24 hours after transfection. Firefly luciferase activity was normalized to the corresponding Renilla luciferase activity. All experiments were performed in triplicates.

Results are expressed as mean ± SEM from n independent experiments. Statistical evaluations were assessed by GraphPad Prism, GraphPad Software, Inc. To detect the difference of quantitative values between the different groups, a ANOVA has been used, along with a Bonferroni test for multiple comparisons. Differences were considered statistically significant at P < 0.05 and were indicated by: ***, P < 0.001; **, P < 0.01; and *, P < 0.05.

To study the role of the Wnt/β-catenin pathway in STS, we first examined the expression and phosphorylation status of the pathway components in a panel of STS cell lines and patient-derived primary cells (Supplementary Table S1). The expression and subcellular localization of β-catenin was examined by immunofluorescence (Fig. 1A). For nuclear localization, a staining average IRS was estimated using immunofluorescence images (Supplementary Table S2). β-Catenin staining was cytoplasmic and nuclear staining (IRS) in STS cells was scored as strong in 75% (6/8), moderate in 12.5% (1/8), and weak in 12.5% (1/8) of cases. Nuclear staining indicated that β-catenin was active and able to transactivate its target genes in most STS cells. Accordingly, protein levels of β-catenin and active phospho-β-catenin (Ser552) were detected in all of the analyzed cell lines as shown by Western blotting (Fig. 1B), being the APC-mutated leiomyosarcoma SK-UT-1 cell line the one that showed the highest level of active phospho-β-catenin (Supplementary Table S3). Wnt target genes were also evaluated (Fig. 1C) including CDC25A, which had been described to drive proliferation of sarcomas (5), CUL4A which is an E3 ubiquitin ligase involved in the Wnt-induced proteolytic targeting (19), C-MYC and AXIN2, which have been described as transcriptionally activated genes in a tissue-independent manner (5). Consistent with β-catenin activation, CDC25A and CUL4A gene levels in STS cells were higher than those of hMSCs, a reference cell line with low levels of endogenous Wnt signaling (20). Moreover, in most of the STS cell lines studied, especially in APC-mutated-SK-UT-1 cells, CDC25A levels were higher than those found in two colorectal cancer cell lines with a strong intrinsic Wnt signaling activity used as positive controls. C-MYC levels were lower than those expressed in the colorectal cancer cells, in which C-MYC is responsible for Wnt-induced proliferation (21, 22), but not in sarcomas as shown by others (5). Likewise, AXIN2 levels in all of the STS cells studied, with the exception of SK-UT-1 cells, were also lower than those expressed in the colorectal cancer cells, being their levels similar to those found in hMSC. Finally, increased basal TCF reporter activity was found in SK-UT-1, HT-1080, and 93T449 cells (Fig. 1D), which was similar to that of human colon cancer cell line SW480, which we had previously reported to be strongly Wnt activated (23). Taken together, these results demonstrate upregulated Wnt signaling in STS cells relative to that observed in hMSCs as evidenced by increased activation of β-catenin, which was able to transactivate Wnt target genes, mainly CDC25A and by increased TCF reporter activity. Interestingly, the lack of correlation between the nuclear (IRS) or total β-catenin levels of the studied sarcoma cells with their histology (Supplementary Tables S3 and S4; Supplementary Methods and Materials) suggests that Wnt activation is a common feature of sarcomas. Moreover, the positive correlation found between CDC25A and CUL4A expression pointed that CUL4A could be an important gene in sarcomas, which deserves further investigation as CDC25A has been reported as a promoter of proliferation in sarcomas. In contrast, a positive correlation was found between C-MYC and AXIN2 low expression levels, suggesting a partial role of these genes in STS.

Altogether our results demonstrate upregulation of the Wnt canonical pathway in several human sarcomas of different histologic subtypes.

Once the activation of Wnt/β-catenin signaling was established in the panel of STS cells, we aimed to investigate the biological effect of small-molecule inhibitors of this signaling pathway PKF118-310, IWR-1, and XAV939 on sarcoma cell viability. Sarcoma cell lines were exposed to increasing doses (0.1–50 μmol/L) of these compounds for 48 h. PKF118-310 selectively disrupts β-catenin/TCF interaction and inhibits its transcriptional activity. IWR-1 and XAV939 are small molecules that stimulate β-catenin degradation by stabilizing AXIN through inhibition of the poly-(ADP)-ribosylating enzymes tankyrase 1 and tankyrase 2 (14, 24). Tankyrase inhibitors IWR-1 and XAV939 were less effective in reducing cell viability of sarcoma cell lines in comparison with PKF118-310 (Fig. 2A), which reduced sarcoma cell viability with IC₅₀ values ranging from 0.21 to 0.57 μmol/L (Fig. 2B; Supplementary Table S1). SK-UT-1, HT-1080, 93T449, and CP0024 cells were more sensitive to treatment compared with AW, SW684, SW872, and SW982 cells, and again no correlation was observed between the inhibitory effect of PKF118-310 and sarcoma histology (Supplementary Tables S3 and S4). In addition, similar results were obtained when cell growth was continuously monitored in the most sensitive sarcoma cell lines under PKF118-310 treatment by using the xCELLigence System (Supplementary Fig. S1 and Supplementary Methods and Materials).

To further characterize the response to these inhibitors, cell-cycle phase distribution was obtained for inhibitor-treated STS cells. For PKF118-310 inhibitor, cytometry data indicated that apoptosis was induced in all analyzed cell lines, as revealed by the percentage of cells with a <2N DNA content (sub-G₁ phase; Fig. 3; Supplementary Fig. S2A). All cell lines showed a dose-dependent increase in the apoptotic fraction after 72 hours of treatment. Liposarcoma 93T449 and leiomyosarcoma CP0024 cells showed the highest levels of apoptosis induction. Treatment of 93T449 and CP0024 cells with 0.50 μmol/L of PKF118-310 resulted in increased cell death, raising 14- and 7-fold the apoptotic cell fraction compared to vehicle-treated cells, respectively. In SK-UT-1 and HT-1080 cell lines, the apoptotic effect of PKF118-310 was more moderate, increasing cell death up to 4-fold at 0.50 μmol/L (Fig. 3). In contrast, neither the apoptotic nor the G₁ fraction was significantly increased upon XAV939 treatment (Supplementary Fig. S2B), which correlated with previous viability analysis (Fig. 2A).

As β-catenin is a transcription factor and its activity is thus dependent on its subcellular localization, we investigated whether compounds used in our study were able to relocate β-catenin to the cytoplasm of STS cells as part of its inactivation process. To this end, SK-UT-1, HT-1080, and 93T449 cells, which had strong nuclear β-catenin staining, were treated with XAV939 (10 μmol/L) and PKF118-310 (0.5 μmol/L) and the intracellular localization of β-catenin was analyzed by immunofluorescence. Results for XAV939 are only shown with the most responsive cell line to this drug, the HT-1080 cells. Following the same line, results for PKF118-310 are shown with these three cell lines, which are representative of the ones that respond to this compound. In untreated cells, β-catenin resided in the cytoplasm and nucleus. Interestingly, XAV939 and PKF118-310 clearly induced a decrease in β-catenin nuclear translocation, being in agreement with our previous data regarding sensitivity to these inhibitors and β-catenin activation (Fig. 4A). In parallel, a significant decrease in nuclear and cytoplasmic β-catenin levels was observed after treatment with PKF118-310 (0.5 μmol/L, 48 hours) as determined by nuclear/cytoplasmic fractionation (Fig. 4B). As expected for the effect of XAV939, which stabilizes AXIN in the destruction complex, there was a decrease in nuclear β-catenin. We found that PKF118-310 compound was also capable of reducing the nuclear localization of β-catenin, suggesting that binding to TCF stabilizes β-catenin nuclear localization in STS.

To further investigate the effects of this reduction on nuclear β-catenin upon treatment with PKF118-310, TCF/β-catenin-mediated transcriptional activity was assessed in STS cells using TOP/FOPflash luciferase reporter assay in comparison with the APC-mutated (truncated at AAs1338) colorectal cancer cell line SW480 (23). Interestingly, APC-mutated-SK-UT-1 cells, that showed similar basal TCF activity to that of SW480 cells (Fig. 1D), had a stronger decrease of TCF/β-catenin-mediated transcriptional activity upon PKF118-310 treatment (Fig. 4C). HT-1080 cells only showed a small decrease in TCF reporter activity after treatment, whereas 93T449 cells showed a similar decrease to that of SW480 cells.

To evaluate the especificity of the used Wnt inhibitors on TCF/β-catenin-mediated transcriptional activity in STS cells, an APC 1309Δ mutated plasmid or a full-length APC (APC WT) plasmid (25, 26), were transfected into SK-UT-1 cells. The latter plasmid was reported to reduce β-catenin levels and to downregulate TCF/β-catenin–mediated transcriptional activity in the colorectal cancer SW480 cell line, which expresses an endogenous mutant APC protein. As shown in Fig. 4D, the expression of APC WT in SK-UT-1 reduced by 78.19% the transcriptional activity of TCF/β-catenin when compared with pcDNA3.1-transfected control cells, whereas the expression of the mutated APC (1309Δ) did not affect the elevated TCF/β-catenin reporter activity found in control SK-UT-1 cells. The fact that the inhibitory effect of PKF118-310 treatment on TCF/β-catenin–mediated transcriptional activity was similar to that observed with the transfection of APC WT in APC-mutated-SK-UT-1 cells indicates that the inhibitory effect of this compound on STS proliferation and cell viability is due to the impaiment of the Wnt pathway.

Taken together, these results support our hypothesis that inhibition of nuclear translocation results in inhibition of β-catenin transcriptional activity, which most probably are necessary requisites for the effective induction of cell death in response to Wnt pathway inhibitors.

The effect of Wnt inactivation on expression of downstream signaling targets was investigated in STS cells. First, a sensitive to XAV939 cell line, HT-1080, and a nonsensitive cell line, SK-UT-1, were treated with this inhibitor and RT-PCR assessment of CDC25A, C-MYC, and CUL4A was performed. Results revealed a time-dependent downregulation for CDC25A and CUL4A gene expression in HT-1080 cells (Fig. 5A), whereas no changes in expression were detected in the nonsensitive SK-UT-1 cell line (Supplementary Fig. S3A). CDC25A levels were strongly downregulated by PKF118-310 treatment in all cell lines (Fig. 5B; Supplementary Fig. S3B), whereas the downregulation effect of this inhibitor on CUL4A expression was less pronounced. In all cell lines, XAV939 and PKF118-310 treatment failed to significantly downregulate C-MYC levels, indicating that it is not a direct transcriptional target of Wnt signaling in sarcoma cells.

Combination of a low dose of PKF118-310 (0.25 μmol/L) with the standard chemotherapeutic agent doxorubicin at different concentrations, used as first-line treatment in metastatic STS, led to an enhanced reduction of STS cell viability. The antiproliferative effects of PFK118-310 and doxorubicin combination were lower in SK-UT-1 cells than in HT-1080 and 93T449 cells (Fig. 6). These results demonstrated that concomitant treatment with a Wnt inhibitor is able to enhance the chemotherapeutic effect of a conventional STS treatment by a mechanism involving inhibition of β-catenin nuclear translocation and, consequently, inhibition of β-catenin/TCF transcriptional activity and CDC25A downregulation.

In recent years, several efforts have been made in characterizing Wnt signaling status in sarcomas, but mostly focusing in specific sarcoma types, especially in osteosarcomas, in which the activation of the pathway is well established (27–29). Nevertheless, knowledge about this pathway in STS is still very limited. However, new -omics studies are pointing to aberrations of members of this pathway in other sarcoma subtypes, such as liposarcomas, but only few results have been reported (30). In contrast to other malignancies, where activation of Wnt signaling is due to genetic aberrations in key pathway components, including CTNNB1 and APC, these alterations are reported to have a low occurrence in STS. This suggests that other mechanisms, such as autocrine loop activation (5), fusion proteins (31–33), or genomic aberrations (30) could be involved. In this study, we demonstrated the constitutive activation of the canonical Wnt/β-catenin pathway in a broad range of STS cell lines and primary cell cultures (liposarcomas, leiomyosarcomas, synovial sarcomas, and fibrosarcomas), and the impact of Wnt signaling inhibition on the growth of these tumoral cells by targeting cytosolic and nuclear components of this pathway. Moreover, to our knowledge, little or no information regarding Wnt signaling inhibition in liposarcomas and leiomyosarcomas has been reported until now, so this is the first study supporting this approach in L-sarcomas (liposarcomas and leiomyosarcomas).

Activation of the canonical Wnt pathway is generally associated with nuclear accumulation of active β-catenin that has escaped from proteasome degradation. In our study, β-catenin displayed a cytoplasmic/nuclear staining using immunofluorescence and in a high percentage of the studied STS cells (75%) nuclear staining was scored as strong, which is in accordance with previous studies (34). Moreover, in all STS subtypes, β-catenin was present in its phosphorylated and transcriptionally active form (Ser552), as shown by others in synovial sarcoma (33). Consistently, expression of Wnt/β-catenin signaling target genes CDC25A and CUL4A was found elevated in STS relative to hMSCs, suggesting that CUL4A could be a novel targetable gene in sarcomas. Moreover, TCF reporter activity was also increased, and its activity in SK-UT-1 cells was similar to that found in the well-studied colorectal cancer cell line SW840. In contrast, C-MYC showed low expression in all the STS cells studied, confirming that CDC25A, but not C-MYC, is a β-catenin/TCF transcriptional target in sarcoma cells, as previously shown (5, 35). Likewise, AXIN2 levels were only found elevated (relative to levels observed in hMSC) in STS cells harboring APC mutations, SK-UT-1 cells, and in the two colorectal cell lines, HT29 and HTC116, with endogenous activation of Wnt pathway, which were used as positive controls. Others have shown elevated AXIN2 expression by Western blot analysis in the fibrosarcoma HT-1080 and the rhabdomyosarcoma cell lines, suggesting that its expression could depend on methodological issues leading to contradictory results in STS (33, 36). Altogether, these results clearly indicate that activated Wnt/β-catenin signaling is a frequent event in sarcomas and is independent of the sarcoma subtype, because sarcoma cells with the same histology had different levels of nuclear β-catenin immunoreactivity.

Once our results had demonstrated the activation of Wnt/β-catenin signaling in sarcoma cell lines, the effect of blocking the Wnt signaling pathway at different levels, by AXIN stabilization of the destruction complex (XAV939 and IWR-1) and by disruption of β-catenin binding to TCF (PKF118-310), was explored. Upon treatment with XAV939 and IWR-1, STS cells displayed only partial inhibition of cell proliferation, cell-cycle changes, and decrease of Wnt target gene expression. The cell line in which XAV939 most effectively inhibited cell proliferation was the fibrosarcoma HT-1080 showing similar antitumoral results to those previously reported by De Robertis and colleagues (35). Only two of the studied STS cell lines harbor mutations, that is, in APC (leiomyosarcoma, SK-UT-1) and CTNNB1 (fibrosarcoma SW684) genes, which could explain the failure of XAV939 to effectively inhibit the viability of these cell lines. In the rest of the cell lines, alternative mechanisms could be responsible for this limited response. In line with this, IHC studies showed elevated AKT and a concomitant upregulation of downstream effectors (e.g., GSK3β and β-catenin) in tumors of patients with STS, especially synovial sarcoma (37, 38) and leiomyosarcoma (39). For this reason, our results show that the disruption of β-catenin/TCF binding with the small-molecule inhibitor PKF118-310 is the best strategy to significantly counteract Wnt/β-catenin dependent signal transduction in vitro in all STS cell lines studied.

At the molecular level, PKF118-310 reduced the levels of β-catenin in the nucleus and β-catenin/TCF-mediated transcriptional activity, leading to decreased mRNA levels of the Wnt target gene CDC25A, but not of C-MYC, in our panel of STS cell lines. Most probably as a consequence, PKF118-310 suppressed STS cell viability by induction of apoptosis. Our results are in agreement with those observed for synovial sarcoma cells (33), hepatocellular carcinoma (40), and hematopoietic malignancies (41), where small molecules were used to inhibit TCF/β-catenin interaction. Moreover, the fact that PKF-110-310 provoked the same decrease in TCF reporter activity as the transfection with wild-type APC of an APC-mutant cell line, that is, SK-UT-1, demonstrates that the inhibition of the Wnt canonical signaling pathway by PKF118-310 is specific. In this line, downregulation of Wnt signaling by using dnTCF4 provoked a reduction of CDC25A levels together with an inhibition of growth in HT-1080 and SK-UT-1 cells (5) similar to what was observed with PKF118-310. In this context, our study is the first showing significant antitumoral effects when PKF118-310 is applied to L-sarcomas and fibrosarcomas, expanding its relevance to the most frequent sarcomas in adults.

Inhibition of the pathway alone is unlikely to result in long-lasting responses due to coactivation of alternative oncogenic pathways (42). For this reason, we also explored the possibility of the combination of PKF118-310 with doxorubicin, as because a combination therapy will probably be the most promising option when targeting this molecular pathway. PKF118-310 simultaneously combined with different concentrations of the conventional therapeutic agent doxorubicin increased its antitumoral effect in an additive manner. Interestingly, Wnt signaling has also been related to the development of drug resistance mechanisms in response to doxorubicin in mantle cell lymphoma (43) and doxorubicin-induced epithelial–mesenchymal transition in hepatocellular carcinoma (44), making it a promising combination therapy for sarcomas. Multidrug resistance (MDR) induced by doxorubicin is mainly caused by high expression of ATP-binding cassette (ABC) transporters, including P-glycoprotein (Pgp) and multidrug resistance-related protein (MRP-1) (45). In fact, Pgp is the product of ABCB1 gene which has been described as a Wnt target gene in chronic myeloid leukaemia (46). The molecular mechanism by which PKF118-310 sensitizes STS cells to doxorubicin should therefore be further explored.

In conclusion, these findings demonstrated that Wnt signaling is involved in STS survival and proliferation. Moreover, these results are highly translational because they show that TCF/β-catenin protein complex, a target of the small-molecule inhibitor PKF118-310, is a possible therapeutic target for STS. Its inhibition not only induced apoptosis in STS cells, but also enhanced the antitumoral effect of doxorubicin in sarcomas. Clinical trials with a similar compound (PRI-724) are currently in preparation for patients with metastatic colorectal cancer. Further characterization of the results of our study with in vivo models and in clinical settings, could benefit sarcoma patients with molecular alterations in this signaling pathway.

J. Martín-Broto has received speakers bureau honoraria from PharmaMar, Novartis, and Lilly and is a consultant/advisory board member of PharmaMar, Lilly, and Novartis. No potential conflicts of interest were disclosed by the other authors.

Conception and design: E. Martinez-Font, J. Martín-Broto, A. Obrador-Hevia

Development of methodology: E. Martinez-Font, I. Felipe-Abrio, R. Ramos

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): S. Calabuig-Fariñas, R. Ramos, A. Obrador-Hevia

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): E. Martinez-Font, I. Felipe-Abrio, J. Terrasa, O. Vögler, R. Alemany, A. Obrador-Hevia

Writing, review, and/or revision of the manuscript: E. Martinez-Font, I. Felipe-Abrio, S. Calabuig-Fariñas, J. Terrasa, O. Vögler, R. Alemany, J. Martín-Broto, A. Obrador-Hevia

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): R. Ramos, A. Obrador-Hevia

Study supervision: A. Obrador-Hevia

We thank the Department of Oncology at the Hospital Universitari Son Espases and patients for providing samples, Drs. Catalina Crespí and Dr. Javier Pierola for technical support, and Andrea Ochoa for help with viability experiments and Dr. Aina Yañez for statistical support.

This study was partially supported by the Institute of Health Carlos III (Ministerio de Economía y Competitividad) and the EC [European Regional Development Fund (ERDF); PI12/01748; to A. Obrador-Hevia, S. Calabuig-Fariñas, R. Ramos]. This study was also supported by PharmaMar, S.A_JM.

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.