Targeting Hedgehog pathway and DNA methyltransferases suppress the phenotype of uterine leiomyosarcoma cells


 Background: Uterine leiomyosarcoma (LMS) is an aggressive tumor with poor prognosis, high rates of recurrence and metastasis. Several studies have shown an essential role of Hedgehog (HH) signaling in tumor development. The role of HH pathway in LMS is still unclear. Methods: Uterine smooth muscle (USTM) and LMS cells lines were used to evaluate the mRNA and protein expression of HH components. MTT assay was performed to evaluate the inhibitory effect of SMO, GLI1 and DNMT inhibitors for proliferation. Nuclear and cytoplasm fractions were prepared. Protein and RNA expression levels were determined by Western blot and q-PCR analysis. Migration, invasion and apoptosis assays were performed to determine the LMS phenotype after the treatments. The statistical analysis was performed using GraphPad Prism 5, and the statistical significance was accepted for p<0.05.Results: LMS analyses showed upregulation of SMO and GLIs (1, 2 and 3) in LMS compared to the myometrium and uterine leiomyoma. Increased nuclear translocation of GLIs proteins was found in LMS cells as well. After that, we hypothesized that HH pathway could be activated in uterine LMS and its blocking would lead to LMS malignant phenotype suppression. Inhibition of SMO and GLI after treatment with LDE225 and Gant61 respectively, induced both protein downregulation as well as decreased GLIs nuclear translocation. Also, abnormal increased expression of DNMT (1, 3a, and 3b) was observed in LMS, and treatment with a DNMT inhibitor decreased both DNMTs expression and GL1 nuclear translocation. Conclusion: Our analyses showed that deactivation of SMO, GLI and DNMTs was able to inhibit LMS phenotype. Importantly, the combination of those treatments exhibited a potentiated effect on LMS malignant phenotype suppression, leading to a deactivating HH pathway. In conclusion, our studies may provide novel options for uterine LMS patients´ target therapy.

methylation status of PTCH1 gene were reported in different types of cancer, suggesting the epigenetic role of PTCH1 in tumor development (22,(24)(25)(26). In this study, we aimed to assess the role and molecular mechanisms of HH pathway activation in LMS and to determine its value as a potential target for anti-LMS therapy.

Cells and Reagents
The immortalized human leiomyoma cell line (HuLM) and immortalized human uterine smooth muscle Proliferation assay Cell proliferation was measured using dimethythiazoldiphenyltetra-zoliumbromide (MTT Sigma Aldrich, St. Louis, MO, USA) assay. 2x10³ cells per well were seeded into 96-well tissue culture plates, treated as described in the gure legends and MTT assay was performed at different time points (24, 48, 72 and 96h). Absorbance was measured in a synergy HT multi-detection microplate reader (BioTek, Broadview, Illinois, USA) at 570 nm. This assay was performed three times in triplicate Cell treatment using Hedgehog pathway and DNA methyltransferase inhibitors LMS cells were seeded at 8x10 4 per well in a six-well plate and cultured overnight, then LMS cells were treated with SMO-LDE225 (10 µM), GLI-Gant61 (30 µM) or DNMT-5'-Aza-dc (2 µM) inhibitors for 72 hours, with daily replacement/change. After the treatment, the cells were harvested for protein/RNA expression measurement,and other studies. The experiments were performed three times in triplicate.

RNA extraction and gene expression
Total RNA was isolated using Trizol reagent (Invitrogen, California USA). The concentration of total RNA was determined using NanoDrop (Thermo Scienti c, Waltham, MA). One microgram of total RNA from each sample was reverse-transcribed to complementary DNA (cDNA) using the High Capacity cDNA Transcription Kit (Thermo Scienti c, Waltham, MA). Quantitative real-time polymerase chain reaction (qRT-PCR) was performed to determine the messenger RNA (mRNA) expression of several genes listed with their primer sequences in Table S1, all primers were selected from literature and the sequences were con rmed using Primer-Blast (https://www.ncbi.nlm.nih.gov/tools/primer-blast/). The real-time PCR reactions were performed using CFX96 PCR instrument using SYBR Green Supermix (Bio-Rad, Hercules, California, USA). GAPDH, B2M, 18S and β-ACTIN were tested as housekeeping genes and B2M was used as an internal control. The results are presented as relative gene expression using CFX Maestro TM . This assay was performed three times in triplicate.

Protein extraction and Western Blot
Cells were collected and lysed in RIPA lysis buffer with protease and phosphatase inhibitor cocktail (Thermo Scienti c, Waltham, MA), protein was quanti ed using the Bradford method (Bio-Rad protein Assay kit). The cytoplasmic and nuclear fractionation was performed using Ner-Per nuclear and cytoplasmic Kit (Thermo Fisher) following the manufacturer's instructions. The information about primary antibodies, including antibody dilution and source of antibodies in this study is listed in Table S2. The antigen-antibody complex was detected with Trident Femto western HRP substrate (GeneTex, Irvine, CA). Speci c protein bands were visualized using ChemiDoc XRS þ molecular imager (Bio-Rad).
Migration assay LMS cells were seeded at 5×10 5 per well in a 24-well plate and cultured overnight. When the cell density reached 100%, a straight scratch was created with a 200 µl pipette tip held perpendicular to the bottom of the 24-well plate (27). The cells were then washed three times with PBS and cultured in serum-free medium with varied treatments. Results were expressed as the space between the edges of individual wounds every 24 hours for 72 hours in comparison with initial (start) time using Image J Software. This assay was performed three-time points in triplicate Apoptosis Assay Apoptosis was determined by Annexin V staining (APC Annexin V Apoptosis Detection Kit with 7-AAD, BioLegend, San Diego, California USA) following the manufacturer's instructions. 8×10 3 LMS cells were grown in 6 well plates. The cells were treated with SMO, GLI and DNMT inhibitors for 72 hours. Staurosporine was used as a positive control. The Annexin V stain was evaluated using ow cytometry (Beckman Coulter Gallios). SMO gene silencing 5x10 5 LMS cells were seeded in six-well plates for transfection. Lipofectamine RNAiMAx reagent (Invitrogen, California USA) was used to transfect small interfering RNAs of SMO (esiRNA SMO, Sigma Aldrich St. Louis, MO, USA). RNA and protein expression levels were determined at 72 hours posttransfection following the manufacturer's instruction. PTCH1 Methylation 8x10 4 LMS cells were seeded in six-well plates and treated with DNMT inhibitor. The methylation status of PCTH1 promoter region was determined using the Human cancer EpiTect Methyl II PCR Array® (QIAGEN Sciences, Maryland, USA) following the manufacturer's instructions. The ABI 7500 system for real-time PCR was used to read the plates. The relative amount of methylated and unmethylated DNA was calculated using the standard ΔCt method using an Excel spreadsheet provided by the manufacturer.

Statistical analysis
Comparison of 2 groups was carried out using student t-test for parametric distribution and Mann Whitney test for nonparametric distribution. Comparison of multiple groups was carried out by analysis of variance (ANOVA) followed by a post-test using Tukey for parametric distribution and Kruskal-Wallis test followed by a post-test Dunns for nonparametric distribution, using GraphPad Prism 5 Software. Data were presented as mean ± standard error (SE). The signi cant difference was de ned as p<0.05.

Hedgehog is activated in LMS with increased GLI nuclear translocation
The constitutive (basal) expression levels of HH signaling components in UTSM, HuLM, and LMS cells were evaluated by qRT-PCR. Higher expressionof SMO, SUFU, and GLI1 were detected in LMS compared to UTSM (P<0.05) and HuLM. The expression of PTCH1 was down-regulated in LMS. GLI2 did not show difference among the cells, while GLI3 was up-regulated in HuLM ( Figure 1A). The protein expression levels of SMO and GLI1 were highest in LMS among three detected cell lines. The protein levels of GLI2 and GLI 3 were also upregulated in HuLM. However, SUFU did not show difference among the cell lines ( Figure 1B). The expression of HH ligands (IHH, DHH, and SHH) were not detected in these cells.
We further evaluated the level of GLIs nuclear translocation in UTSM, HuLM and LMS cells ( Figure 1C). The expression levels of GLI1 were highest in the nucleus of the LMS cells, low in UTSM and undetected in HuLM cells. GLI2 and GLI3 were mainly expressed in the nucleus of all cell lines, with the highest expression levels in LMS. ( Figure 1C). These results showed that HH pathway was activated in LMS due to higher expression of SMO and GLI1 with increased GLI nuclear translocation.
Inhibition of Hedgehog pathway using SMO and GLI inhibitors in LMS cells SMO inhibitors (LDE225 and GDC0449) and GLI inhibitors (Gant58 and Gant61) were selected to determine their effect on LMS cells. MTT assay was performed using three-time points (24,48  LMS cells were treated with SMO inhibitor, LDE225, for 24, 48 and 72 hours, with drug replacement every 24 hours. RNA and protein expression of HH components were evaluated for all time points. qRT-PCR showed that the expression levels of SMO, GLI1, GLI2, and GLI3 were signi cantly downregulated after 72 hours LDE225 treatment (p<0.05), while alteration of expression levels after 24 and 48 hours treatment was not observed (Figure 2A). The protein expression levels of these key HH members were decreased in a time-dependent manner in response to the treatment of SMO inhibitor (LDE225) ( Figure 2B). For GLI inhibitor, qRT-PCR analysis demonstrated that RNA expression of SMO, GLI1, GLI2, and GLI3 was not altered after Gant61 treatment (Figure 2 C). However, protein expression of GLI1, GLI2, and GLI3 were decreased in response to Gant61 treatment ( Figure 2D).
The GLIs nuclear translocation was evaluated in LMS cells treated with LDE225 or Gant61 inhibitor for 24, 48, 72 hours, respectively. The results showed that translocation of GLIs into the nucleus was markedly decreased in three-time point treatments with SMO or GLI inhibitor as compared to the untreated control (Figure 2 E and F).
To determine the effects of treatment with LDE225 or Gant61 inhibitor on phenotype of LMS, MTT assay was performed to determine the effect of the treatment on cell proliferation. Our data demonstrated that the proliferation was decreased after treatment with SMO or GLI inhibitor ( Figure 2 G and H).
Cell migration is a multi-step process that plays an important role in tumorigenesis. To evaluate the effect of SMO and GLI inhibitors on the migration of LMS cells, would-healing assay was performed. The results showed that the SMO and GLI inhibitor signi cantly decreased the migration capacity in LMS cells ( Figure 3A) (p<0.05). Apoptosis assay was performed and demonstrated that SMO and GLI inhibitors induced apoptosis in LMS cells ( Figure 3B), the GLI inhibitor showed a more potent effect compared to SMO inhibitor.
To explore the possible synergistic or additive effect of combination treatment with SMO and GLI inhibitors. We evaluated their inhibitory effect on LMS cells using MTT assay. The results showed no synergism or additive effect using a combination of the treatment ( Figure S5) Treatment with SMO or GLI inhibitors individually showed an inhibitory effect on proliferation, migration, and invasion while induced apoptosis in LMS cells. The inhibition of GLI nuclear translocation was more potent using SMO inhibitor. Thus, the SMO gene was selected for knockdown. The knockdown of the SMO gene was performed using interference RNA and the HH components were evaluated in LMS cells.
SMO protein expression was evaluated to verify the e ciency of the knockdown and was highly decreased after the knockdown ( Figure 3C). Figure 3D shows that knockdown of SMO decreased the expression of SMO, GLI1, GLI2, and GLI3 (p<0.05).

Inhibition of DNA methyltransferase regulated HH signaling in LMS cells
To understand the molecular mechanism of activation of the HH signaling pathway in LMS in the context of epigenetic regulation, we determined the expression of DNA methyltransferases (DNMT) in LMS and UTSM cells. Our studies showed that RNA expression of DNMT1, DNMT3a, and DNMT3b was upregulated in LMS compared to UTSM cells (p<0.05) ( Figure 4A). Accordingly, the protein expression of DNMT1 and DNMT3a were also increased in the LMS compared to UTSM cells ( Figure 4B). Next, we determined whether inhibition of DNA methyltransferases with 5'-Aza-2'-Deoxycytidine (5'-Aza-dc) suppressed LMS phenotype. We performed a MTT assay using a different dose at three-times point (24,48 and 72 hours). After 72 hours of treatment, 5'-Aza-dc at the concentration of 2 µM showed 50% of inhibition in proliferation (Figure S6 A). PTCH1 DNA methylation was evaluated to verify if the treatment with DNA methyltransferase inhibitor was able to reverse the methylation pro le in LMS cells. The basal level of the percentage of PTCH1 DNA methylation in LMS was 2.3%. The percentage of PTCH1 DNA methylation after 72 hours of 5'-Aza-dc treatment was decreased to 1%.
RNA and protein expression levels of DNMTs were evaluated in LMS cells in response to 5'-Aza-dc treatment. The RNA expression of DNMT1, DNMT3a, and DNMT3b was decreased in 5'-Aza-dc-treated LMS cells compared to the control ( Figure 4C). The protein expression of DNMT1 and DNMT3a in LMS cells was also decreased after treatment with 5'-Aza-dc ( Figure 4D).
To evaluate the impact of DNA methyltransferase inhibition on the HH signaling pathway, the components of the HH signaling were evaluated in the presence or absence of 5'-Aza-dc in LMS cells. Although the RNA expression of PTCH1, SMO, SUFU, GLI2, and GLI3 was not altered after5'-Aza-dc treatment, the decreased RNA expression of GLI1 was observed in response to 5'-Aza-dc treatment ( Figure  5A). WB analysis exhibited decreased expression levels of SMO and GLI1 ( Figure 5B). Moreover, GLI1 and GLI2 nuclear translocation were decreased in response to 5'-Aza-dc treatment. On the other hand, 5'-Azadc treatment increased the nuclear translocation of GLI3 ( Figure 5C).
Next, determine the effect of DNMT inhibition on proliferation, migration, and apoptosis in LMS cells. The results showed that 5'-Aza-dc decreased proliferation ( Figure 5D), concomitantly with decreased expression of PCNA proliferation marker in LMS cells (Figure S 6B). Migration capacity was decreased after 5'-Aza-dc treatment (p<0.05) ( Figure 5E). Moreover, 5'-Aza-dc treatment was capable of inducing apoptosis in LMS cells ( Figure 5F).

Inhibition of both DNA Methylation and Hedgehog Signaling in Human LMS cell Lines
The treatment with DNA methyltransferase and HH inhibitors showed an inhibitory effect on HH signaling via decreasing GLI1 transcription and protein expression, as evidenced by decreasing proliferation, migration and invasion while inducing apoptosis. Next, we performed experiments to explore whether the DNA methyltransferase inhibitor, in combination with HH inhibitors, can exhibit an additive or synergistic effect in LMS cells.
MTT assay was performed to evaluate the combined effect of DNA methylation and HH inhibitors on LMS proliferation. The results showed that the combination treatment with DNA methylation and SMO inhibitors did not show synergism or additive effect. However, the combination of 5'-Aza-dc with GLI inhibitor showed a synergistic effect ( Figure S7A). Since the treatment of 1 µM 5'-Aza-dc with 30 µM GLI inhibitor exhibited a most potently inhibitory effect on LMS proliferation, this combination treatment was used for further studies ( Figure S7B).
The RNA expression of key HH members including of SMO, GLI1, GLI2, and GLI3 was measured with and without combination treatment, our data demonstrated that combination treatment decreased the RNA expression of SMO, GLI1, GLI2, and GLI3 (p<0.05) compared to the control ( Figure 6A). The combination treatment also resulted in decreased protein levels of GLI1 ( Figure 6B). Moreover, the combination treatment decreased GLI1 nuclear translocation in a time-dependent manner compared to the control ( Figure 6C).
To explore this combination effect on LMS cells phenotype, the cells were treated with both inhibitors for 72 hours, and proliferation was evaluated every 24 hours. Figure 6D showed that combination treatment decreased the proliferation of LMS cells. In addition, the wound-healing assay demonstrated that combination treatment decreased the migration capacity in the LMS cells (p<0.05) ( Figure 6E). The combination showed a more potent effect comparing to the single treatment, with decrease expression of HH signaling components, proliferation, decreasing GLI1 nuclear translocation and migration capacity in LMS.

Discussion
Uterine leiomyosarcoma is a rare but extremely aggressive tumor that represents a treatment challenge due to its unresponsiveness to available therapies. As a consequence, the patients commonly present high rates of tumor recurrence, progression and metastasis (5).Previously, it was demonstrated thatthe protein levels of SMO and GLI1, the key members of HH signaling, were highly expressed in human uterine LMS (6). However, how HH signaling is activated and its contribution to LMS malignant phenotype are still unknown. In this study, we demonstrated for the rst time that HH is activated in LMS cells via upregulation of key HH members concomitantly with increased nuclear translocation of GL1. Importantly, epigenetic mechanisms are involved in the HH pathway activity in LMS cells. Moreover, targeting DNA methylation and HH pathways exhibited a potent inhibitory effect on LMS phenotype. SMO, GLI, and PTCH1 were selected for evaluation as triggers of the HH signaling in LMS cells. LDE225 and Gant61 were used in this study because they are already established as potent inhibitors of SMO and GLI, respectively. Although both drugs had suppressed the LMS cells proliferation, LDE225 showed a more potent effect in decreasing of GLI nuclear expression than Gant61 treatment.
The anti-tumor effect of LDE225 observed in our model was consistent with literature for other types of cancer. For instance, in renal carcinoma cells, treatment with LDE225 showed reduced cell proliferation concomitant with lower GLI1 and GLI2 expression (28). In melanoma cells, treatment with LDE225 showed a decrease in apoptosis rate and cell proliferation, as well as an increase in cell cycle arrest. In the melanoma animal model, treatment with LDE225 inhibited GLI1 expression (29). In Hepatoma cells, the treatment suppressed cell proliferation and decreased the protein levels of GLI2 and ABCC1 transporter (30). In lung cancer, the combined treatment with LDE225 and Erlotinib (an EGFR inhibitor) showed a reduction in cell invasion, migration, colony formation, proliferation, and induced apoptosis (31). In chronic myeloid leukemia, LDE225 treatment inhibited the cell growth with GLI1 downregulation (32). Studies in LMS and other types of cancers using LDE225 demonstrated that targeting SMO is su cient to suppress the phenotype and pathogenesis of these aggressive tumors.
Targeting GLI1 with Gant61 has been studied in other types of cancers showing bene cial effects. In breast cancer, Gant61 treatment decreased cell proliferation by reducing GLI1 and PTCH1 gene expression and inhibited GLI1 nuclear translocation (33). In vivo studies showed that Gant61 treatment reduced both tumor growth and GLI1 expression in rhabdomyosarcoma (34). In prostate cancer, treatment with Gant61 induced suppression of tumor growth with decreased GLI1 and PTCH1 expression (35). Our results showed that Gant61 is able to impair GLI protein expression, reduce GL1 nuclear translocation as well as exhibited an inhibitory effect on LMS cells proliferation. All those studies demonstrated the important role of GLIs in cancer development, and targeting GLIs suppressed the phenotypes of several types of neoplasms.
LMS has high recurrence and metastasis rates (36). Therefore, evaluation of the effects of SMO and GLI inhibitors in the migration process was extremely important, since this mechanism is directly involved with metastasis and progression of the disease. Regarding this, our results showed that LDE225 and Gant61 have inhibitory activity in cell migration. This inhibitory effect by HH inhibitors was consistent with studies in other types of tumors (28,(37)(38)(39). In prostate cancer, Gant61 and LDE225 treatment showed a decrease in the migration capacity of the cells (40), (41). Gant61 treatment in ovarian cancer cells showed a reduction in migration with downregulation of GLI1 activity (42). In glioblastoma cells, treatment with Gant61 suppressed migration with impairment of the expression of HH components (43).
In addition to migration, Gant61 and LDE225 have been reported to induce apoptosis. In pancreatic cancer, Gant61 induces apoptosis with downregulation of GLI expression (44). In prostate cancer, Gant61 induced apoptosis (45). In melanoma cells, LDE225 increased the percentage of apoptotic cells, inhibited cell proliferation and reduced the expression of HH pathway components (46).
Inhibition of SMO or GLI using LDE225 and Gant61 respectively blocked the HH pathway in LMS, decreasing proliferation, migration and inducing apoptosis. In addition, our results showed that the downregulation of PTCH1 in LMS indicates a possible aberrant CpG island hypermethylation of the PTCH1 gene. Activation of the HH signaling pathway can also occur due to the methylation mechanism. PTCH1 promoter region has been shown to be hypermethylated in rhabdomyosarcoma and medulloblastoma (19) and breast cancer (20,21). The treatment with 5'-Aza-dc in rhabdomyosarcoma and medulloblastoma showed a decrease in PTCH1 gene methylation (19). However, we did not nd effects on PTCH1 expression after 5'-Aza-dc treatment.
Interestingly, the treatment with 5'-Aza-dc in LMS regulated GLI1 expression, decrease proliferation, migration and inducing apoptosis. The mechanism in which 5'-Aza-dc regulates GLI1 expression in LMS is not understood. It may occur due to indirect regulation of GLI1 expression by cross-talk with other signalings (AKT1, TGFβ, WIP1 and HDAC), therefore activate GLI1 expression.
We explored the effect of combination treatment using 5'-Aza-dc and Gant61. The results showed a synergistic effect with those drugs. Combination treatment exhibited a more potent effect compared to single treatment in the context of inhibiting the HH pathway and affecting LMS cells proliferative and migratory behavior. The combination treatment with HH and epigenetic inhibitors has been shown to exhibit a more potent effect compared to a single treatment in other types of tumors. In liver cancer, SMO and HDAC inhibitors showed decrease cell viability, colony formation and increase apoptosis (47). In aerodigestive cancer, the combined treatment with SMO and HDAC inhibitors promote cell cycle arrest, suppress SMO and PTCH1 expression and delay tumor growth in animal model prolonging survival more than single agent alone (48). Our study using a combination treatment strategy to produce synergistic activity is one possible advance towards achieving higher cure rates in LMS.
We proposed a mechanism model of HH pathway activation in LMS based on our novel ndings that 1) SMO and GLI expression are deregulated in LMS cells, 2) GLI nuclear translocation is increased in LMS cells, 3) deregulation of the epigenetic mechanism is involved in GLI1 expression, and 4) targeting HH pathway via pharmacological inhibition is able to inhibit the LMS phenotype (Figure 6 F).

Conclusion
In Summary, our studies demonstrated for the rst time that HH signaling is activated in LMS cells with evidence of increased expression and nuclear translocation of GLI. We also demonstrated that impairment of HH signaling with SMO and GLI inhibitors was capable of suppressing LMS phenotype along with decreased GLI nuclear translocation. Besides, the DNMTi treatment alone or combination with GLI inhibitor had a more potent effect on the LMS malignant phenotype. All these data open new perspectives for uterine management, focusing on the development of novel non-invasive speci c therapeutics for this aggressive tumor.

Competing interests
The authors declare that they have no con ict of interest.   Staurosporine treatment for 24 h was used as a positive control. * p<0.05, ** p<0.01, *** p<0.001