Curdione induces G2/M phage arrest, apoptosis and autophagy via COX2 mediate IDO1 expression through PKCδ/ GSK3β/β-catenin pathway in human uterine leiomyosarcoma

Background Curdione, one of active ingredients of a traditional Chinese herb medicine-Curcuma zedoary, has been reported with benecial therapeutic effects in lots of cancer types. However, the potential anticancer effect in uterine leiomyosarcoma (uLMS) and the underlying mechanisms are still unclear. Methods The aim of this study was to explore the potential effect and mechanisms of curdione on uLMS in vitro and vivo with uLMS cell lines and mouse xenograft tumor model respectively. Results Curdione triggered anti-proliferation effect in uLMS cell lines by inducing cell cycle arrest at G2/M phase, caspase-mediated cell apoptosis, and pro-death cell autophagy. Indoleamine-2,3-dioxygenase-1 (IDO1), which was dependent on cyclooxygenase-2 (COX2), mediated the anti-proliferation effect of curdione. Curdione down-regulated IDO1 expression and promoted the dephosphorylation of protein kinase c (PKC), glycogen synthase kinase-3 beta (GSK3β) and β-catenin in cell lines. PKC/GSK3β/β-catenin pathway responsible for constitutive IDO1 expression in uLMS. COX2 mediate the promotion effect of curdione on the PKC/GSK3β/β catenin pathway activity, which was detected by COX2 inhibitor, and further conrmed by COX2 siRNA and COX2 overexpression. The pathway activity was inhibited by COX2-siRNA and enhanced by COX2 overexpression. In turn, the pathway inhibitor suppressed IDO1 expression. The anti-proliferation effect of curdione on uterine leiomyosarcoma were further conrmed in vivo. Echoing to the results in vitro, the underline mechanism involved in COX2-mediated IDO1 downregulation via PKCδ/GSK3β/β-catenin pathway. Conclusion real-time RT-PCR cDNA RT-PCR relative mRNA COX2 comparative cycle (Ct) repeat times. primer expression. To further clarify the autophagy induced by curdione was pro-survival or pro-death. The cell viability were detected when cells were pretreated with 3-MA (autophagy inhibitor) followed by curdione. 3-MA blocked inhibitory effect of curdione on cell viability. The results indicated that, curdione induced pro-death autophagy in SK-UT-1 and SK-LMS-1 cells. Based on these data, we demonstrated that, curdione inhibited the growth of uLMS by inducing cell cycle arrest, apoptosis and autophagy in SK-UT-1 and SK-LMS-1 cells.

ow cytometry and western blot analysis were performed to evaluate the expression of apoptosis-related proteins.
As an important indicator of cellular fate, autophagy play a double role in the development of tumor [13,14]. The LC3-II level is positively relate to the formation of autophagosomes [15]. Sequestosome1 (SQSTM1/P62), which is highly down-regulated when autophagy occur, participated in autophagic degradation, and often used as a marker of autophagy [16]. Beclin-1 also participated in physiological processes of autophagy, and the protein expression level is positive relate to autophagy [17]. We wondered whether curdione induced autophagy in uLMS cells, the level of autophagic relate marker proteins LC3-II, Beclin-1 and P62 were determined by western blot. IDO1, which plays an important role in promoting the growth of cancers, is over-express in many cancer types and predicts poor outcome [9]. Growing evidence suggest that IDO1 contribute to cancer development [18]. Furthermore, lots of antitumor drugs induced apoptosis and autophagy via declining IDO1 levels in cancers [19,20]. Accordingly, decreasing IDO1 expression is a key to explore new anticancer drugs. COX-2 is highly expressed in many cancer types and is associate with cancer progression [21,22]. Considerable evidences suggest that blocking of COX-2 expression could inhibited tumor progression effectively [23][24][25].
For time immemorial, Chinese herb medicine play a pivotal role in the history of human ghting against diseases. In clinical practice reports of traditional Chinese medicine, herbs or the extractions show widely antibacterial, anti-in ammatory and anticancer e cacy with tolerable toxicity. Therefore, the development of anticancer drugs was refocused on herb medicine. Rhizoma Curcuma has been reported with antitumor effect [26], which mainly dependent on its essential oils. The effective bioactive compounds of the essential oils including: β-elemene, curcumol, curcumin, curdione, furanodiene, furanodienone, and germacrone have been reported with anti-platelet aggregation [27,28], anti-in ammatory [29], antibacterial [30], neuroprotective properties [31], prevent cardiotoxicity [32], and anti-tumor effect [26,30,[33][34][35][36]. Bene cial therapeutic effect of curdione has been reported in breast cancers [34], but the potential anti-cancer effect in uLMS and the underlying mechanism is still unclear.
In this paper, we hypothesized that curdione might be one of the effective active components responsible for the anti-tumor effect. The anti-tumor effect of curdione on uLMS was explored by SK-UT-1 and SK-LMS-1 cells in vitro and SK-UT-1 xenograft mice mode in vivo. Cell Counting Kit-8 (CCK-8) assay, immuno uorescence for 5-ethynyl-2′-deoxyuridine (EdU) and Terminal deoxynucleotidyl transferasemediated dUTP nick-end labeling (TUNEL) assay, ow cytometry, immunohistochemistry, Transient transfection, quantitative real-time reverse transcription-polymerase chain reaction (qRT-PCR), and western blot analysis were conducted to detect the effect of curdione on uLMS, and further elucidate the mechanism.

Curdione suppresses human uterine leiomyosarcoma cells viability
After treatment with grade concentrations (0-500 μM) curdione for 12, 24, 48 and 72 h, the SK-UT-1 and SK-LMS-1 cells viability were detected by CCK8 assay respectively. As shown in Fig. 2, curdione inhibited the cells viability in a concentration and time dependent manner at concentration higher than 10 μM. The 50% inhibited concentration (IC50) were determined to nd the ideal dose for the following experiments.

Curdione inhibited human uterine leiomyosarcoma cells proliferation
To further con rm the anti-proliferation effect of Curdione on uterine leiomyosarcoma cell, SK-UT-1 and SK-LMS-1 cells were treated with curdione for 24 h, EdU assay and the proliferation marker ki67 were detected by immuno uoresent. As Fig.3 shows, compared with control, curdione decreased the EdU and

Curdione induces apoptosis in human uterine leiomyosarcoma cells
Curdione induced both early and late apoptotic of cells in a concentration-dependent manner. After treatment with curdione or not for 24 h, ow cytometry was performed with Annexin V-FITC/ Propidium Iodide (PI) double staining to quantify apoptosis. As shown in Fig.5, compared with control, the early apoptotic ratio elevated by 2.95-fold (1.9% and 1.6%) and 4.25-fold (5.6% and 6.8%) respectively; the late apoptotic ratio increased from 1.9% to 6.4% in SK-UT-1 cells, and from 1.2% to 4.9%in SK-LMS-1 cells; the total apoptosis ratio increased by 8.2% in SK-UT-1, and 8.9% in SK-LMS-1 cells. Which indicated that, both early and late apoptotic ratio of SK-UT-1 and SK-LMS-1 cells increased with curdione treatment in a dosedependent manner.

Curdione induces autophagy in human uterine leiomyosarcoma cells
Autophagy is another important mechanism that regulate cell fate. To detect whether autophagy was induced by curdione, the level of autophagic marker proteins LC3-II, Beclin-1 and P62 in both cells were determined by western blotting. As shown in g 6, curdione upregulated LC3-II and Beclin-1, and downregulated P62 in a dose-dependent manner. To further clarify that autophagy induced by curdione was pro-survival or pro-death, 3-MA (an autophagy inhibitor) was used to detect the in uence of cell viability. Before treated with curdione, SK-UT-1 and SK-LMS-1 cells were incubated with autophagy inhibitor 3-MA for 2 h, the results showed that, 3-MA blocked the inhibitory effect of curdione on cell viability and apoptosis of SK-UT-1 and SK-LMS-1 cells signi cantly.

Curdione induced the crosstalk between apoptosis and autophagy in human uterine leiomyosarcoma cells
Curdione induced apoptosis and autophagy, to explore whether there was crosstalk between apoptosis and autophagy. SK-UT-1 and SK-LMS-1 were pre-treatment with Z-VAD-FMK (apoptosis inhibitor) or 3-MA (autophagy inhibitor) for 2 h, and then incubated with curdione for 24 h, apoptosis ratio were detected by ow cytometry. The apoptosis and autophagy marker proteins were determined by western blot. The results showed that, apoptosis ratio induced by curdione was blocked by Z-VAD-FMK and abolished by 3-MA ( Fig.7 a). Western blot results show that, the decreased expression of LC3II, beclin1 and increased expression of p62 induced by curdione were alleviated by Z-VAD-FMK, the up regulation of cleaved caspase 3, 6, 9 induced by curdione were attenuated by 3-MA ( Fig.7 b).

Curdione inhibited IDO1 expression in human uterine leiomyosarcoma cells
Western blot were performed to quantify the effect of curdione on IDO1 protein expression in SK-UT-1 and SK-LMS-1 cells. As Fig. 8 shows, Curdione down-regulated IDO1 expression of human uterine leiomyosarcoma cells in a concentration -dependent manner. The above data show that, curdione suppressed the proliferation of SK-UT-1 and SK-LMS-1 cells and down-regulated IDO1 expression. To further detected the relationship between regulation effect on IDO1 expression and anti-proliferation of curdione in both cell lines. Cells were pretreatment with IDO1 inhibitor for 2 h, and then incubated with curdione for 24 h. The results showed that, epacadostat reversed the suppression effect on cell viability induced by curdione in both cell lines. To further investigate this effect, IDO1 knockdown and overexpression were performed. The results showed that, compared with curdione alone group, silence of IDO1 overcame the suppression effect on cell viability induced by curdione. While, IDO1 overexpression enhanced the inhibitory effect induced by curdione.
COX2 is necessary for curdione-decreased IDO1 expression in human uterine leiomyosarcoma cells To explore whether IDO1 mediate the suppression effect of curdione is dependent on COX2. NS398, an antagonism of COX2, signi cantly reversed the inhibitory effect of curdione on IDO1 expression in SK-UT-1 and SK-LMS-1 cells. To further con rm that, COX2 mediates curdione down-regulated IDO1 expression, transiently transfection technology approaches to manipulate COX2 expression in both cell lines were designed. Silence or overexpression of COX2 in mRNA and protein levels were con rmed by qRT-PCR and western blot analysis, knockdown COX2 reversed the inhibitory effect of curdione on IDO1 expression, while the inhibitory effect signi cantly enhanced by the over expression of COX2 in both cell lines. These ndings collectively indicated that, dependent on COX2, curdione suppressed IDO1 expression in SK-UT-1 and SK-LMS-1 cells.
PKCδ/GSK3β/β-catenin pathway mediate curdione decrease IDO1 expression in uLMS cells To further illustrate the mechanism involved in COX2-mediated induction of IDO1 expression in uLMS.
The effects of curdione on the activation of PKCδ, GSK3β and β-catenin in human uterine leiomyosarcoma cells were detected. As shown in Fig 10, curdione promoted the PKCδ, GSK3β and βcantein dephosphorylation signi cantly in a concentration-dependent manner. COX2 antagonist NS398 blocked the dephosphorylation of PKCδ, GSK3β and β-catenin pathway. To further con rm this mechanism, SK-UT-1 and SK-LMS-1 cells were transfected with either COX2-siRNA or COX2-pCMV6, and then treated with curdione for 24 h. The results show that, the PKCδ, GSK3β and β-catenin pathway activation were attenuated by COX2-siRNA and enhanced by the overexpression of COX2. The PKCδ, GSK3β and β-catenin pathway activity and the inhibitory effect of curdione on IDO1 expression were abolished by PKCδ inhibitor and GSK3β inhibitor. Which indicated that, the depression effect of curdione on IDO1 expression is mediated by the dephosphorylation of the PKCδ, GSK3β and β-catenin pathways in uLMS cells.
Curdione inhibited cell proliferation by induced cell cycle arrest in G2/M, apoptosis and autophagy in human uterine leiomyosarcoma cells To con rm whether downregulation IDO1 is involved in the anti-proliferation of curdione in uLMS cells and mediated by the PKCδ/GSK3β/βcatenin dependent COX-2 pathway, SK-UT-1 and SK-LMS-1 cells were pretreatment with rottlerin or LY2090314 or NS398 for 2 h, and then incubated with curdione for 24 h. As shown in Fig. 11a, the anti-proliferation effect of curdione was attenuate by selective GSK3β inhibitor Rottlerin and β catenin inhibitor LY2090314, and COX-2 inhibitor NS-398 blocked the anti-proliferation effect of curdione. Fig. 11b shows that, the upregulation of cleaved-caspase 3, LC3, Beclin1, Cdc2, CyclinB1 and down regulation of P62 induced by curdione were abolished by IDO1 inhibitor epacadostat, GSK3β inhibitor Rottlerin, β catenin inhibitor LY2090314 and COX-2 inhibitor NS-398. The above results collectively indicated that IDO1 mediated PKCδ/GSK3β/βcatenin pathway dependent on COX-2 is involved in the anti-proliferative effect of curdione on human uterine leimyosarcoma cells.

Curdione suppress the growth of uterine leiomyosarcoma in vivo
To further explore the anti-uterine leiomyosarcoma effect of curdione in vivo, we established a xenograft tumor model by subcutaneous injection with SK-UT-1 cells into the right anks of BALB/c nude mice.
Tumor xenograft mice were intraperitoneal injection with 100 mg/kg/day or 200 mg/kg/day curdione or not for 21 days. The body weight, histopathological examination, tumor volume and weight were measured to evaluate the effect of curdione on the uterine leiomyosarcoma in vivo. Fig. 12 a-e showed that, compared with control, curdione reduced tumor volume and weight markedly. While, no signi cant body weight loss were detected in curdione treated xenograft mice. The histopathological results showed that, no obvious histopathological lesion were detected in liver and kidney tissue, which meaning curdione induced no liver and kidney injuries in vivo (Fig. 8 f). Above all, Curdione exhibited anti-uterine leiomyosarcoma e cacy with few toxicity in vivo.

Curdione induces apoptosis and autophagy in SK-UT-1 xenografts
To determine the mechanism involved in the inhibitory effect of curdione on human uterine leiomyosarcoma in vivo. The tissues protein expression of IDO1, COX2, the cleaved caspase-3, LC3-II, Becline1, P62, PKCδ, GSK3β, β-catenin were detected by western blot and immunohistochemistry assay. Consistent with the results in vitro, curdione up-regulated the protein expression of IDO1, COX2, the cleaved caspase-3, LC3-II, dephosphorylation of PKCδ, GSK3β, β-catenin and down-regulated ki67, p62 in tumor tissues (Fig. 13 a). The immunohistochemistry assay showed the same changes of protein levels observed in western blot (Fig. 13 b). The above results showed that, echoing to the results in vitro, curdione inhibited the expression of IDO1, COX2 and the phosphorylation of PKCδ/GSK3β/β-catenin pathway in vivo. In conclusion, acting on COX2 mediate IDO1 expression via PKCδ, GSK3β, β-catenin pathway, curdione suppress the growth of human uterine leiomyosarcoma in tumor xenografts.

Discussion
ULMS is a highly malignant tumor with poor outcome (low ve-year survival rate and high recurrence rate) [37]. To date, no safe and effective medications for uterine leiomyosarcoma have been developed [38,39], it is imperative to nd effective medicine for this disease.
With the effect of promoting blood circulation, removing blood stasis and alleviating pain, Rhizoma Curcumae has been widely prescribed for the treatment of gynecological diseases in Traditional Chinese Medicine clinical practice. Mounting studies have shown that, the Rhizoma Curcumae extracts, which have broad-spectrum antitumor effects, play an important role in the effect of herb medicine. Curcumin, which comes from the same family and genus with curdione, has the anti-proliferation effect on uterine sarcoma [36,40]. As one of the extracts of Rhizoma Curcumae, curdione have the same chemical structure with curcumin, so we speculate that curdione has similar anti-tumor effect on uLMS.
In this work, we determined the effect of curdione on the cells viability. The results as Fig. 2 showed that, in a dose and time dependent manner, curdione inhibited the viability of SK-UT-1 and SK-LMS-1 cells. To ensure the delity and reliability, less than a third of IC50 were used as the optimal dose in the following experiment. The anti-proliferation effect of curdione were further con rmed by EdU assay, which often used to re ect the DNA replication activity indirectly, and ki67 expression levels, a marker of proliferation, which was detected by immune orescence. As shown in Fig. 3, curdione suppressed EdU-positive and Ki67-positive orescence images in a dose dependent manner in both cell lines. Cell cycle, which is close related to cell fate, is often used to evaluate the e cacy of new drugs. In this study, curdione increased the cells population of G2/M phase in a concentration dependent manner in uLMS cell lines.
Apoptosis is a form of programmed cell death, also an important mechanism of cellular self-protective.
Here, curdione induced SK-LMS-1 and SK-UT-1 cells pro-death apoptosis in a dose-dependent manner, the apoptosis ratio were detected by ow cytometry analysis with Annexin V-FITC/PI double staining, and the apoptosis morphology changes were determined by TUNEL analysis. Li reported that β-elemene induced caspase-mediated mitochondrial cell apoptotic [41]. Similarly, western blot analysis show that, the proteins expression levels of cleaved caspase3, 6, 9 increased in a dose-dependent manner, while no changes occur in pro-and cleaved-caspase 8 (Fig. 4), which suggest that, curdione induced caspasemediated apoptosis through intrinsic pathway.
Autophagy, just like a double edged sward-pro-survival or pro-death, is also another important mechanism that regulate the behavior of tumor. Here, as shown in Fig. 5, curdione induce autophagy, which were re ect by the increased LC3-II, Beclin-1 expression and down regulated P62 expression. To further clarify the autophagy induced by curdione was pro-survival or pro-death. The cell viability were detected when cells were pretreated with 3-MA (autophagy inhibitor) followed by curdione. 3-MA blocked inhibitory effect of curdione on cell viability. The results indicated that, curdione induced pro-death autophagy in SK-UT-1 and SK-LMS-1 cells. Based on these data, we demonstrated that, curdione inhibited the growth of uLMS by inducing cell cycle arrest, apoptosis and autophagy in SK-UT-1 and SK-LMS-1 cells.
IDO1, which play an important role in the cancer development [42], is highly expressed in many cancer types [43][44][45]. High IDO1 expression is associated with poor overall survival and progression-free survival [46][47][48][49]. According to Liu report, compared with paired normal tissues, IDO1 is highly expressed in uterine carcinosarcoma and uterine corpus endometrial carcinoma [46], but the expression level in uterine leiomyosarcoma is still unclear. In the pre-trail, we found high IDO1 expression in SK-UT-1 and SK-LMS-1 cells. Here, the inhibitory effect of curdione on IDO1 expression were determined, and further, we analyzed the correlation between IDO1 expression and the anti-proliferation effect of curdione, as shown in Fig. 5.
The inhibitor of IDO1 reversed the suppression effect of curdione on uLMS cell lines signi cantly, and the suppression effect were overcame by silence of IDO1, while enhanced by IDO1 overexpression. These data indicated that the suppression effect of curdione on the growth of uLMS is mediated by IDO1.
As one of the immune checkpoints, IDO1 participated in limiting immune surveillance, promoting tumor immune tolerance and immune escape [46,50,51]. As a "brake" or "accelerator" in the tumor immune regulation, IDO1 prepare immunosuppressive tumor microenvironment [52]. However, Ameet I. Thaker reported that, independent on T cell mediated immune regulation, IDO1 induce colon cancer growth directly [42]. In this work, we found that, IDO1 mediated the anti-proliferative effect of curdione in uLMS. Further research is need to analysis whether this effect dependent on IDO1 activity mediate immune response, which was detected by analyzing tryptophan levels, kynurenine levels and immune response.
COX-2 was found to be overexpression in primary or metastatic lesions of many tumor types Previous studies have found that, PKC/GSK3β/βcatenin pathway is a bridge of COX2 triggers IDO1 expression in human cancers [57]. To further explore the signaling pathway responsible for constitutive IDO1 expression in uLMS, the activation of PKC/GSK3β/βcatenin pathway were determined. Curdione promoted the dephosphorylation of the pathway in a concentration dependent manner, this was blocked by COX2 inhibitor NS398, which means that curdione induced the pathway activity is mediated by COX2, this was further con rmed by silence of COX2 and overexpression of COX2. As Fig. 10 shows, the pathway activity was inhibited by COX2-siRNA and enhanced by COX2 overexpression. In turn, the pathway inhibitor suppressed IDO1 expression in both cell lines.

Conclusions
In this paper, for the rst time, we demonstrated the anti-tumor effect of curdione on uLMS and analyzed the underlying mechanisms. Curdione induced cell cycle arrest, apoptosis and autophagy, which jointly promote cell death powerful. The crosstalk between apoptosis and autophagy suppressed the progression of uLMS. The mechanisms, which were involved in COX2 mediated IDO1 expression via PKCδ/GSK3β/βcatenin pathway, were explored further by pharmaceutical inhibitor and transfection technology. In addition, the safety of curdione in vivo was checked by detecting the body weight changes of xenograft tumor models and histological analysis of virtual organs -kidney and liver.
Taken together, curdione, extract from natural herb medicine, is a promising candidate for the treatment or assistant treatment of uLMS. Our research provide preclinical evidence for uterine leiomyosarcoma treatment and open a new perspective in the war against uLMS. Availability of data and materials All data generated or analyzed during this study were included in this published article.

Competing interests
All authors declare no potential competing interests.       expression levels of both cell lines were analyzed by western blot. Histogram represented the statistical analysis of the relative proteins expression level compared with GAPDH. All data were present with means ± SD of three independent experiments. *P < 0.05, **P < 0.01 compared with control. expression levels of both cell lines were analyzed by western blot. Histogram represented the statistical analysis of the relative proteins expression level compared with GAPDH. All data were present with means ± SD of three independent experiments. *P < 0.05, **P < 0.01 compared with control.   The protein expression levels of Beclin1, LC3B II, p62, and cleaved caspase-3, 6, 9 were determined by Western blot assays. The statistical analysis results of relative proteins expression levels compared with to GAPDH were shown in histogram. All date presented as the mean ± SD. *P < 0.05; **P < 0.01 compared with control; # P < 0.05; ## P < 0.01 compared with curdione alone group. The protein expression levels of Beclin1, LC3B II, p62, and cleaved caspase-3, 6, 9 were determined by Western blot assays. The statistical analysis results of relative proteins expression levels compared with to GAPDH were shown in histogram. All date presented as the mean ± SD. *P < 0.05; **P < 0.01 compared with control; # P < 0.05; ## P < 0.01 compared with curdione alone group. Cells were treated with 100 μM curdione for 24, 48, 72 h, relative IDO1 protein expression levels were analyzed by western blot (b). To further explored whether the anti-proliferation effect of curdione in uLMS cells is mediate by IDO1. Cells were transiently transfected with IDO1 speci c siRNA, IDO1 mRNA expression levels were con rmed by RT-PCR (c) and the protein expression levels were con rmed by western blot (d). Cells were transfected with IDO1 PCMV6 and con rmed mRNA expression by RT-PCR (e) and protein expression by western blot (f). Cells were pretreated with IDO1 inhibitor epacadostat (g) or transfected with IDO1 speci c siRNA (h) or IDO1 PCMV6 (i), following with Curdione for 24 h, the cell viability were analyzed with CCK8 assay. Histogram represented the statistical analysis data from three independent trails and present with mean±SD. *P < 0.05; **P < 0.01 compared with control; #P < 0.05; ##P < 0.01 compared with curdione alone group. Cells were treated with 100 μM curdione for 24, 48, 72 h, relative IDO1 protein expression levels were analyzed by western blot (b). To further explored whether the anti-proliferation effect of curdione in uLMS cells is mediate by IDO1. Cells were transiently transfected with IDO1 speci c siRNA, IDO1 mRNA expression levels were con rmed by RT-PCR (c) and the protein expression levels were con rmed by western blot (d). Cells were transfected with IDO1 PCMV6 and con rmed mRNA expression by RT-PCR (e) and protein expression by western blot (f). Cells were pretreated with IDO1 inhibitor epacadostat (g) or transfected with IDO1 speci c siRNA (h) or IDO1 PCMV6 (i), following with Curdione for 24 h, the cell viability were analyzed with CCK8 assay. Histogram represented the statistical analysis data from three independent trails and present with mean±SD. *P < 0.05; **P < 0.01 compared with control; #P < 0.05; ##P < 0.01 compared with curdione alone group. COX2 is essential for curdione-deduced IDO1 expression in human uterine leiomyosarcoma cell lines. SK-UT-1 and SK-LMS-1 cells were pre-treated with NS398 or not for 2 h and then incubated with 100 μM curdione for 24 h. IDO1 expression was determined using western blot and presented in histogram (a).
After transiently transfected with speci c COX2-siRNA or full-length cDNA of human COX2 (COX2-pCMV6). The COX2 mRNA and protein expression was con rmed by qRT-PCR analysis (b, e), and western blot analysis (c, f) respectively. After transfected with COX2-siRNA or COX2-pCMV6, cell were treated with curdione for 24 h, the IDO1 protein expression compared with GAPDH in SK-UT-1 and SK-LMS-1 cells were analyzed by western blot (d, g). All data are represent with mean±SD and present in histogram. All experiments performed triplicate. *P < 0.05; **P < 0.01 compared with control; #P < 0.05, ##P < 0.01 compared with curdione alone group. COX2 is essential for curdione-deduced IDO1 expression in human uterine leiomyosarcoma cell lines. SK-UT-1 and SK-LMS-1 cells were pre-treated with NS398 or not for 2 h and then incubated with 100 μM curdione for 24 h. IDO1 expression was determined using western blot and presented in histogram (a).