Deguelin Suppresses Pancreatic Tumor Growth and Metastasis by Inhibiting Epithelial-To-Mesenchymal Transition in an Orthotopic Model

Home , Liver tumor

ORIGINAL ARTICLE Deguelin suppresses pancreatic tumor growth and metastasis by inhibiting epithelial-to-mesenchymal transition in an orthotopic model

SR Boreddy and SK Srivastava

Deguelin is known to suppress the growth of cancer cells; however, its anti-metastatic effects have not been studied so far in any cancer model. In the present study, we aimed to evaluate the anti-metastatic potential of deguelin in vivo and in tumor growth factor-b1 (TGFb1)-stimulated cells. Our results demonstrate that tumor growth, peritoneal dissemination and liver/lung metastasis of orthotopically implanted PanC-1-luc cells were signiﬁcantly reduced in deguelin-treated mice along with the induction of apoptosis. Furthermore, deguelin-treated tumors showed increased epithelial signature such as increased expression of E-Cadherin and cytokeratin-18 and decreased expression of Snail. Similar observations were made when PanC-1, COLO-357 and L3.6pl cells were treated in vitro with deguelin. Moreover, E-cadherin was transcriptionally upregulated and accumulated in the membrane fraction of deguelin-treated cells, as indicated by increased interaction of E-Cadherin with b-catenin. TGFb1-induced downregulation of E-Cadherin and upregulation of Snail were abrogated by deguelin treatment. In addition, deguelin inhibited TGFb1-induced Smad3 phosphorylation and Smad4 nuclear translocation in PanC-1 cells. Furthermore, when TGFb1-induced nuclear factor kappa B (NFkB) activation was inhibited, TGFb1-induced Snail upregulation or E-Cadherin downregulation was blocked. Deguelin also signiﬁcantly downregulated the constitutive phosphorylation and DNA binding of NFkB in a dose- dependent manner. Interestingly, overexpression of either NFkB or Snail completely abrogated deguelin-mediated epithelial-to- mesenchymal transition (EMT) inhibition, whereas overexpression of NFkB but not Snail rescued cells from deguelin-induced apoptosis. Hence, deguelin targets NFkB to induce reversal of EMT and apoptosis but downstream effectors might be different for both processes. Taken together, our results suggest that deguelin suppresses both pancreatic tumor growth and metastasis by inducing apoptosis and inhibiting EMT.

Oncogene (2013) 32, 3980–3991; doi:10.1038/onc.2012.413; published online 17 September 2012 Keywords: NFkB; TGF-b; RKIP; Snail; metastasis; apoptosis

INTRODUCTION canonical Smad signaling and/or non-canonical TGFb/TAK1/IKK/ Pancreatic cancer is the fourth leading cause of cancer-related NFkB signaling, but both the pathways merges on Snail to repress deaths in the United States.1 Most of the patients with pancreatic E-Cadherin expression.8 The most important suppressors of cancer develop metastasis and die because of the debilitating E-Cadherin are Snail-related zinc-finger transcription factors such metabolic effects of their unrestrained growth.2 Several well- as Snail, Slug and SIP-1/ZEB.9 Interestingly, Snail-knockout animals known factors such as late detection and resistance to chemo- die at gastrulation stages and show defects in EMT, indicating the and radiotherapy contribute to poor survival rate. Pancreatic role of Snail in EMT induction.10 Recently, nuclear factor kappa B cancer commonly metastasize to the abdominal cavity, lymph (NFkB) was identified as a central mediator of EMT by regulating nodes and liver.3 Snail expression.11,12 Huber et al.11 have shown that the activation Emerging evidences suggest that epithelial-to-mesenchymal of NFkB promoted the transition to a mesenchymal phenotype transition (EMT) is a key process in tumor cell invasiveness and even in the absence of TGFb. Moreover, inhibition of NFkBin metastasis. Previous reports have shown a close association metastatic cells resulted in reversal of EMT, suggesting that the between pancreatic cancer progression and EMT.4 In fact several IKK-2/IkBa/NFkB pathway is required for the induction and pancreatic cancer cell lines and surgically resected pancreatic maintenance of EMT in epithelial cancer cells.11 tumors have shown strong EMT characteristics.4–6 Interestingly, Recognition of EMT as a potential mechanism for metastasis increase in fibronectin or vimentin and decrease in E-Cadherin may offer new targets for therapeutic intervention.13 Such expression in tumors correlates with poor survival.5 A recent report therapeutic intervention might prevent tumor invasion and has shown that 43% (13/30) of primary pancreatic tumors and 53% block metastasis, if applied at an early phase of tumor growth. If (8/15) of metastatic tumors have increased levels of N-Cadherin, a primary tumor has already metastasized to distant sites, anti- indicating the potential role of EMT in pancreatic cancer metastasis.6 EMT approach alone may not be sufficient. Tumor growth factor-b1 (TGFb1) can induce and maintain EMT Deguelin, a rotenoid isolated from Mundulea sericea Willd. during embryogenesis and cancer progression.7 It induces EMT by (Leguminosae) has shown to be a potential anti-cancer agent in

Department of Biomedical Sciences and Cancer Biology Center, Texas Tech University Health Sciences Center, Amarillo, TX, USA. Correspondence: Professor SK Srivastava, Department of Biomedical Sciences and Cancer Biology Center, Texas Tech University Health Sciences Center, 1406 Coulter, Suite 1103, ARB, Amarillo, TX 79106, USA. E-mail: [email protected] Received 9 February 2012; revised 11 July 2012; accepted 25 July 2012; published online 17 September 2012 Deguelin suppresses pancreatic tumor metastasis SR Boreddy and SK Srivastava 3981 various cancer models such as breast, colon and lung cancer.14–16 loss or signs of acute or delayed toxicity (Figure 1d). Interestingly, Intravenous injection of deguelin showed a mean residence time deguelin-treated mice developed much lesser metastatic lesions of 6.98 h and terminal half-life (t1/2) of 9.26 h. About 58% and 14% in peritoneum (Figure 2a). As shown in Figure 2b, control tumors of the deguelin was eliminated via feces and urine, respectively, migrated to an average of about 20 mm away from the pancreas, within 5 days of intragastric administration of deguelin.17 Previous whereas deguelin-treated tumors migrated around 5–6 mm, reports have shown that deguelin inhibits the survival of various indicating significant inhibition of tumor cell migration by cancer cells by targeting the key survival pathways such as AKT, deguelin (20.3 vs 6.6 mm; difference ¼ 13.7 mm, 95% CI ¼ 7–20.8, Wnt, NFkB and cell cycle proteins.14–16,18,19 However, the anti- Po0.0006). In fact, 8 out of 10 mice in treated group showed no metastatic potentials of deguelin have not been investigated in metastatic lesions whereas, in the control group, 8 out of 10 mice any cancer model. The present study was designed to elucidate demonstrated metastatic lesions (Figure 2c). Compared with the role of deguelin on EMT and metastasis in pancreatic cancer. control, deguelin-treated mice exhibited reduced migration of PanC-1 cells into various organs such as liver, lung, spleen, intestinal mesentery nodules and abdominal cavity (Figures 2c RESULTS and d, indicating that deguelin inhibits the metastasis of primary Deguelin inhibits primary tumor growth and spontaneous in vivo implanted tumors to liver and lungs. As liver is the preferential site metastasis of pancreatic tumors of metastasis for most of the pancreatic tumors, we imaged the Previous reports have shown that deguelin inhibits the growth of luminescence in liver. Luciferin was injected 10 min before killing cancer cells.16,20,21 In the present study, we evaluated whether the mice, livers were excised and luminescence was measured by deguelin could inhibit the metastasis of pancreatic cancer along the imaging system. As shown in Figure 2e, liver from deguelin- with primary tumor growth. To determine the anti-metastasis treated mice showed no or modest luminescence, whereas control potential of deguelin in vivo, mice that were orthopically mice livers demonstrated strong luminescence indicating that implanted with PanC-1-luc cells were treated with 5 mg/kg deguelin significantly inhibits the metastasis of pancreatic tumors deguelin (1:1 corn oil:dimethylsulphoxide, daily, intraperitoneal to liver. (i.p.)), and tumor growth and metastasis were monitored using IVIS Bio Imaging Station. Our results show that deguelin substantially reduced the primary tumor growth by 73% Deguelin inhibits EMT in vivo (7.6 Â 107 vs 1.8 Â 107 photons/sec; difference ¼ 5.7 Â 107 As the primary objective of the present study was to evaluate the photons/sec, 95% confidence interval (CI) ¼ 6.6 Â 107 to 12 Â 108 anti-metastatic potential of deguelin, we focused further studies photons/sec, P40.087), as compared with control tumors (Figures on metastasis. Inhibition of pancreatic tumor metastasis and 1a and b). Primary tumor weight was about 63% lesser in migration by deguelin in vivo prompted us to examine whether deguelin-treated mice, as compared with control mice (Figure 1c). the anti-metastatic effect of deguelin was due to the inhibition of Deguelin was well tolerated by the mice as depicted by no weight EMT. Interestingly, our immunofluorescence studies show that

1.0×1008 Control 1.5×1008 deguelin treatment P<0.087 07 8.0×10 Treatment started 1.0×1008 6.0×1007

4.0×1007 5.0×1007 Tumor Growth Tumor

2.0×1007 (Photons/Second)

Tumor Growth Photons/Second Growth Tumor 0.0 ***** 0.0 0 5 10 15 20 25 30 35 40 45 50 Control Treated Days after tumor implantation

30 1000 Control p<0.005 Deguelin (5mg/kg) 28 800 Deguelin treatment started (63% Supression) 26 600 24

400 22 Body weight (g) Body 20 200 Pancreas weight (mg) Pancreas 18 0 0 4 8 12 16 20 24 28 32 36 40 44 Control Deguelin treated Days after tumor implantation Figure 1. Deguelin suppresses primary tumor growth of pancreatic tumors. Mice were orthotopically implanted with PanC-1-luc cells and treated with deguelin (5 mg/kg, daily, i.p.) for 37 days. (a) Tumor growth was monitored by measuring luminescence of control and deguelin- treated mice twice a week by IVIS imaging station. (b) Tumor growth curve (photons/sec; at the termination of the experiment) of control and treated mice. (c) Weight of the pancreas with tumor. At the end of the experiment, pancreas along with the primary tumor from the control and deguelin-treated mice were carefully excised, weighted and presented as a bar graph. (d) Weight of the control and deguelin-treated mice presented as bar graph. Values are mean±s.d., *Po0.05.

& 2013 Macmillan Publishers Limited Oncogene (2013) 3980 – 3991 Deguelin suppresses pancreatic tumor metastasis SR Boreddy and SK Srivastava 3982 Control Deguelin Treatment Lateral View Ventral View

P<0.0006

30 10 Control 9 Deguelin treatment 8 20 7 6 5 4 10 3 Number Animals 2 Tumor migration from

the site of injection (mm) 1 0 0 Control Treated Liver Lung pertoneal No- metastasis Metastatic organ

Control Treated Control Treated Lung metastasis Peritoneal dissemination Liver luminescence Liver metastasis

Figure 2. Deguelin inhibits in vivo metastasis of pancreatic tumors. Mice were orthotopically implanted with PanC-1-luc cells and treated with deguelin (5 mg/kg, daily, i.p.). Tumor luminescence was measured twice a week by IVIS in vivo imaging station. (a) A representative images showing peritoneal dissemination of control and deguelin-treated animals were presented. (b) Distance of the secondary tumors that migrated from pancreas was measured by Living Image Software (Caliper Lifesciences, Hopkinton, MA, USA). (c) At the end of the experiment animals were euthanized and pancreatic tumor metastasis to various organs were counted. (d) At the end of the experiment, animals from control and deguelin-treated were imaged at maximum time scale (5 min) to visualize micro metastasis, and representative images were presented. (e) At the end of the experiment, animals were given with luciferin and animals were euthanized. Intact livers were carefully removed from control and deguelin-treated mice and imaged for metastatic lesions. Values are mean±s.d.

E-Cadherin expression was increased around threefold and Deguelin inhibits the migration and invasion of pancreatic cancer vimentin expression decreased by 80% in the tumors of cells deguelin-treated mice, as compared with control tumors Migration and invasion are the critical steps in metastasis of (Figures 3a and b), providing a critical clue that deguelin inhibits primary tumors, as these processes help in escaping the primary EMT in tumors. Similarly, western blot analysis of tumors shows tumor cells from either blood stream or lymph nodes. As deguelin that E-Cadherin expression was signiﬁcantly increased, whereas inhibited in vivo migration of PanC-1 xenografts, we wanted to see N-Cadherin, vimentin and Snail expressions were decreased by whether deguelin could inhibit the migration and invasion of deguelin treatment. In addition, an increase in cleavage of pancreatic cancer cells in vitro. Initially, we evaluated the caspase-3 and PARP was observed in deguelin-treated tumors, cytotoxicity of deguelin in pancreatic cancer cells and found that as compared with control tumors (Figure 3c). Western blots were IC50 (24 h) values for PanC-1, COlO-357 and L3.6pl were 62, 27 and quantitated and presented as bar diagram (Figure 3d). 12 mM, respectively (Figure 4a). Cells were treated with sublethal

Oncogene (2013) 3980 – 3991 & 2013 Macmillan Publishers Limited Deguelin suppresses pancreatic tumor metastasis SR Boreddy and SK Srivastava 3983 DAPI E-Cadherin Vimentin Overlay Control Treated

E-Cadherin Vimentin p<0.0001 p<0.0007 1.0 5 0.8 4 3 0.6 2 0.4 1 0.2 Fold over control Fold over Fold over control Fold over 0 0.0 Control Treated Control Treated

Folds Change (Actin Normalized) Tumors 01234

E-Cadherin E-Cadherin * N-Cadherin

N-Cadherin * Vimentin

Vimentin * p-NFkB Control

NFkB p-NFkB *

Snail NFkB RKIP Treated

Snail * Cl. Caspase 3

Cl. PARP RKIP

Actin Cl. Caspase 3 * Control Deguelin

Cl. PARP *

Figure 3. Deguelin inhibits EMT in vivo.(a) At the end of the experiment, control and deguelin-treated tumors were excised and immunostained for E-Cadherin and vimentin. DAPI was used as an internal reference. (b) Immunofluorescence of tumor section was quantified by Image J software (http://rsbweb.nih.gov/ij/) and presented as bar diagram. Values are mean±s.d. of at least three slides. (c) Tumors from control and treated mice were homogenized and equal amount of protein was subjected to western blot. (d) Tumor western blots were quantified using UN-SCAN-IT (Silk Scientific Inc., Orem, UT, USA) software and presented as bar diagram. Values are mean±s.d. of six separate tumors. *Po0.05.

dose of deguelin and migration and invasion were evaluated. Our Deguelin inhibits EMT in pancreatic cancer cells in vitro results show that control PanC-1 cells migrated into 92% of the As deguelin inhibited the migration and invasion of pancreatic cancer wound area by 36 h; whereas deguelin-treated PanC-1 cells cells from their parental colonies as compared control cells, we next migrated into 49% (Figures 4b and c) of the wound area showing wanted to see whether deguelin suppress mesenchymal (metastatic) 44% of inhibition by deguelin (859 vs 100 mm2, difference ¼ 759 properties of pancreatic cancer cells and induce epithelial cell signature mm2, 95% CI ¼ 625–893, Po0.07). Similar observations were made along with apoptosis induction. Similar to our in vivo observations, our in COLO-357 cells (Figures 4b and c). Similarly, deguelin (10 mM) results show that deguelin signiﬁcantly downregulated vimentin, also inhibited the invasion of PanC-1, COLO-357 and L3.6pl cells upregulated cytokeratin-18 and E-Cadherin expression, and induced (58%, 77% and 59%, respectively), as evaluated by Boyden’s cleavage of caspase-3 and PARP in all the three cell lines in a dose- and chamber (Figure 4d). time-dependent manner (Figures 5a and b).

& 2013 Macmillan Publishers Limited Oncogene (2013) 3980 – 3991 Deguelin suppresses pancreatic tumor metastasis SR Boreddy and SK Srivastava 3984 a PanC-1 COLO-357 L3.6pl 100 100 24 h 24 h 24 h 100 80 48 h 80 48 h 48 h 72 h 72 h 80 72 h 60 60 60 40 40 40 20 20

Percentage Survival Percentage 20 Percentage Survival Percentage Survival 0 0 0 0 25 50 75 100 125 150 175 200 0 10 20 30 40 50 60 70 80 90 100 0 10 20 30 40 50 60 Concentration of Duguelin (μM) Concentration of Deguelin (M) Concentration of Deguelin (μM)

b 0 h 12 h 24 h 36 h 48 h Control PanC-1 Treated

0 h 24 h 48 h Control COLO-357 Treated

c PanC-1 COLO-357 Control Deguelin Treated Control Deguelin Treated 1500 1500 * 1250 1250 * 1000 1000 * * 750 750 * 500 500 Wound Area (µm2) Wound Area (µm2) * 250 250

0 0 0 12243648 0 24 48 Time (h) Time (h)

d PanC-1 COLO-357 L3.6pl P<0.0001 0.25 P<0.0001 0.15 0.30 P<0.0001 0.25 0.20 0.10 0.20 0.15 0.15 0.10 0.05 0.10 0.05 0.05 0.00 0.00 0.00 (OD@570nm) Cells Invaded Invaded (OD@570nm) Cells 0 5 10 20 (OD@570nm) Invaded Cells 0 5 10 20 0 2.5 5 10 Concentration of Duguelin (µM) Concentration of Deguelin (M) Concentration of Deguelin (M) Figure 4. Deguelin inhibits proliferation, migration and invasion of pancreatic cancer cells. (a) Cells were plated in a 96-well plate, treated with various concentrations of deguelin and viability was measured by sulforhodamine B assay. (b) Conﬂuent monolayers of PanC-1 and COLO-357 cells were scratched with 1-ml pipette tip, treated with 5 mM deguelin and photographed using Nikon microscope (Melville, NY, USA). (c) Wound area of control and treated cells was quantiﬁed by Image J software. Values are mean±s.d. *Po0.05. (d) Invasion of pancreatic cells was measured by Boyden’s Transwell assay according to the manufacturer’s instructions. Values are mean±s.d. *Po0.05.

Oncogene (2013) 3980 – 3991 & 2013 Macmillan Publishers Limited Deguelin suppresses pancreatic tumor metastasis SR Boreddy and SK Srivastava 3985 PanC-1 COLO-357 L3.6pl

E-Cadherin

Cytokeratin 18 HPDE-6 E-Cadherin Snail

RKIP Snail

Cl. Caspase 3 Actin Cl. PARP Deguelin (M) 0 5 10 20 Actin

Deguelin (M) 0 5 10 20 0 5 10 20 0 2.5 510

PanC-1COLO-357 PanC-1 COLO-357

E-Cadherin E-Cadherin

Cytokeratin 18 Vimentin

Vimentin Cl. Caspase 3

Snail Cl. PARP

Actin Actin

Time (h) 0124816240 1 2 4 824 0 1225500 5 1020Gemcitabine (nM)

PanC-1

Membrane Cytosolic Nuclear Whole lysate RT-PCR E-cadherin PanC-1 COLO-357 E-Cadherin Snail

GAPDH Loading Control

Deguelin (10μM) - + - + - + - + - + - + Deguelin (10μM)

Control Deguelin (10μM) Control Deguelin (10μM) IP: E-Cadherin E-Cadherin Bright field β-Catenin

Deguelin (10 M) -+ Florescence

E-Cadherin Snail Figure 5. Deguelin inhibits EMT in vitro. Cells were treated with deguelin either in (a) dose- or (b) time-dependent manner, and EMT markers were evaluated western blot. (c) PanC-1 and COLO-357 cells were treated with gemcitabine for 72 h and EMT markers were evaluated by western blot. (d) PanC-1 cells were treated with 10 mM deguelin for 24 h, and membrane, cytosolic, nuclear and whole-cell lysates were subjected to western blot, whereas mRNA was analyzed by RT–PCR for E-Cadherin expression. (e) Equal protein from control and deguelin- treated cells were immunoprecipitated with E-Cadherin antibody and immunoblotted for b-catenin. (d: right panel) PanC-1 cells were treated with deguelin and immunostained for E-Cadherin, Snail (red) and actin (green).

As Snail is considered to be the main transcriptional repressor of expression was signiﬁcantly reduced in all the three cells lines in a E-Cadherin, we monitored the expression of Snail in response to dose- and time-dependent manner (Figures 5a and b). However, deguelin treatment. Interestingly, our results show that Snail Snail and E-Cadherin were not altered by deguelin treatment in

& 2013 Macmillan Publishers Limited Oncogene (2013) 3980 – 3991 Deguelin suppresses pancreatic tumor metastasis SR Boreddy and SK Srivastava 3986 normal human pancreatic ductal epithelial cells (HPDE-6) Deguelin downregulates Snail expression by inhibiting TGFb1 (Figure 5a, right panel). non-canonical NFkB pathway To differentiate between the cytotoxic effects of deguelin with Interestingly, deguelin treatment increased E-Cadherin expression that of reversal of EMT, we used gemcitabine, which is a known in L3.6pl, BxPC-3 and MIA PaCa-2 cells (data not shown), which do chemotherapeutic drug used in clinics against pancreatic cancer. not have intact functional TGFb1 signaling due to mutation in Our results show that gemcitabine treatment decreased the Smad4 (L3.6pl and BxPC-3) or lack of TGFb receptors (MIA PaCa- survival of PanC-1 and COLO-357 cells similar to that of deguelin 2).26 TGFb1 is also known to induce EMT in various cancer types by (Figure 5c and Figure 4a). However, in contrast to deguelin, Smad-independent pathways such as NFkB.27 On the basis of gemcitabine treatment failed to induce E-Cadherin expression or these facts, we hypothesized that deguelin could also target non- reduce vimentin expression in pancreatic cancer cells (Figure 5c). canonical TGFb pathway to inhibit EMT by modulating NFkB. To In fact, previous reports have shown that gemcitabine- and evaluate the role of NFkB in TGFb1-induced EMT, NFkB was tamoxifen-treated cells acquire EMT phenotype.22–24 Hence, silenced using SiRNA in PanC-1 and COLO-357 cells. As expected, deguelin is unique as it not only suppresses pancreatic tumor our results show that NFkB-silenced PanC-1 or COLO-357 cells growth but also suppress metastasis by inhibiting EMT. were completely insensitive to TGFb1-induced EMT (Figure 7a). Furthermore, we observed maximum NFkB phosphorylation (Ser- 536) and nuclear accumulation after 2 h of TGFb1 treatment in Deguelin transcriptionally increases membrane-bound E-Cadherin PanC-1 cells (Figure 7a, right panel). As NFkB was a prerequisite for and downregulates cytosolic E-Cadherin TGFb1-induced EMT, we next wanted to evaluate the effect of Next, we wanted to see whether deguelin induces E-Cadherin deguelin on NFkB signaling pathway. Our results show that expression transcriptionally or by altering the post-translational deguelin significantly inhibit the constitutive IkB (Ser-32/34) and degradation. Our reverse transcription (RT)–PCR analysis shows NFkB (Ser-536) phosphorylation in pancreatic cells without that E-Cadherin was transcriptionally upregulated in deguelin- affecting the protein levels (Figure 7b). Similarly, TGFb1-induced treated PanC-1 (Figure 5d). Surprisingly, the increase in E-Cadherin phosphorylation of IkB (Ser-32/34) and NFkB (Ser-536) were also protein expression by deguelin treatment in the whole-cell lysates drastically reduced by deguelin treatment (Figure 7b, right panel). did not correlated exactly as mRNA levels (Figures 5b and d). Furthermore, deguelin significantly reduced the transcriptional E-Cadherin is known to mis-localize in various portions of the cell activity, DNA-binding activity (Figures 7c and d) and nuclear such as cytosol and nucleus, hence, we wanted to see whether localization (Figure 7d, right panel) of NFkB in PanC-1 and COLO- deguelin alters the localization of E-Cadherin. Interestingly, our 357 cells. results show that membrane-bound E-Cadherin was significantly increased whereas, cytosolic E-Cadherin was downregulated. However, nuclear E-Cadherin was unaltered in deguelin-treated Ectopic expression of Snail or NFkB abrogates EMT inhibitory effects of deguelin PanC-1 cells (Figure 5d, right panel). Furthermore, b-catenin interaction was increased in deguelin-treated cell lysates that was To confirm the role of NFkB and Snail in deguelin-mediated immunoprecipitated with E-Cadherin (Figure 4e), indicating that inhibition of EMT, PanC-1 and COLO-357 cells were transiently E-Cadherin is accumulated in membrane fraction. Immunofluor- transfected with plasmid-DNA encoding either NFkB or Snail gene. escence studies further confirmed the localization of E-Cadherin As expected, overexpression of Snail alone significantly dimin- and its repressor Snail (Figure 5e, right panel). ished the expression of E-Cadherin, cytokeratin and RKIP, whereas upregulated the expression of vimentin in PanC-1 and COLO-357 cells (Figures 8a and b). However, when Snail overexpressing cells Deguelin suppresses TGFb1-induced EMT were treated with deguelin, EMT inhibition mediated by deguelin Family of TGFb growth factors can initiate and maintain EMT in was almost completely blocked in PanC-1 and COLO-357 cells variety of biological systems.25 Hence, we wanted to see whether (Figures 8a and b), indicating the critical role of Snail in deguelin- deguelin would inhibit TGFb1-induced EMT in pancreatic cancer mediated EMT inhibition. Interestingly, NFkB was not altered cells. As shown in Figure 6a, TGFb1-treated (10 ng/ml) PanC-1 cells significantly by Snail overexpression, indicating that NFkB may not showed elongated morphology including protruding lamellipodia be regulated by Snail in our model (Figures 8a and b). (Figure 6a, white arrows) and scattered from parental colonies Similarly, when NFkB was overexpressed in PanC-1 or COLO-357 (Figure 6a, black arrows), as compared with control cells. However, cells, E-Cadherin expression was reduced. On the other hand, deguelin treatment substantially reduced TGFb-induced cell expression of N-Cadherin and vimentin was increased in both the morphology in PanC-1 cells (Figure 6a). Interestingly, TGFb1 cells (Figures 8c and d), indicating the role of NFkB in EMT. enhanced the invasion of PanC-1 and COLO-357 cells in Boyden’s Interestingly, NFkB overexpression also resulted in the increased chamber by 2.12- and 1.7-fold respectively, as compared with expression of Snail, providing direct evidence on the role of NFkB control cells. Nonetheless, deguelin treatment completely blocked in inducing EMT through Snail (Figures 8c and d). Nonetheless, TGFb1-induced cell invasion in PanC-1 cells, as compared with when NFkB overexpressing PanC-1 or COLO-357 cells were treated TGFb1 alone treated cells (Figure 6b). with deguelin, E-Cadherin induction or vimentin and Snail downregulation was abolished (Figures 8c and d). Interestingly, Snail overexpression did not protected the cells from deguelin- Deguelin targets TGFb1 canonical signaling pathway to induced apoptosis, whereas NFkB overexpression significantly inhibit EMT rescued the cells from deguelin-induced apoptosis (Figures 8a–d). Single dose of TGFb (10 ng/ml) for 72 h significantly induced the mesenchymal signature in PanC-1 cells (Figure 6c). Nonetheless, deguelin treatment completely blocked TGFb1-induced Snail DISCUSSION expression and restored TGFb1-reduced E-Cadherin expression In the present study, we investigated the role of deguelin on in both PanC-1 and COLO-357 cells (Figure 6c, right panel). In pancreatic tumor growth and metastasis in vivo and in vitro. Our addition, deguelin reduced the phosphorylation of Smad-3 (Ser- results show that deguelin significantly inhibits the peritoneal 423/425) in PanC-1cells (Figure 6d). Interestingly, Smad-4 expres- dissemination and liver metastasis of orthotopically implanted sion was not altered by deguelin treatment, however; deguelin PanC-1-luc cells in nude mice along with induction of apoptosis. treatment completely blocked TGFb1-induced shuttling of Smad4 Furthermore, deguelin significantly downregulated the mesench- from cytosol to nucleus (Figure 6d, right panel). ymal markers and increased the expression of epithelial markers

Oncogene (2013) 3980 – 3991 & 2013 Macmillan Publishers Limited Deguelin suppresses pancreatic tumor metastasis SR Boreddy and SK Srivastava 3987 Control TGF β1 (10ng/mL) TGF β1 + Deguelin (5μM) PanC-1

PanC-1 COLO-357 P<0.001 P<0.001 P<0.001 P<0.001 0.4 0.5 0.3 0.4 0.3 0.2 0.2

Cell Invasion 0.1 (OD@570nm) Cell Invasion

(OD@570nm) 0.1 0.0 0.0 TGF-β1 (10ng/mL) --++ TGF-β1 (10ng/mL) --++ Deguelin (10μM) -+-+ Deguelin (10μM) -+ -+

PanC-1, TGF-β1 (10 ng/mL) PanC-1 COLO-357 E-Cadherin E-Cadherin

Vimentin Cytokeratin 18

Smad 4 Snail Snail RKIP RKIP Actin Actin TGFβ1 (10 ng/mL) --+ + --++ Time (h) 0244872 Deguelin (10 μM) -+ - + -+ -+

PanC-1 PanC-1 p-Smad3 (Ser-423/425) Cytosol Nuclear

Smad3 Smad4

Smad4 Actin/LaminB

Actin TGF1 (10ng/mL) -++ - ++ Deguelin (10μM) --+- - + Deguelin (μM) 0 5 10 20 Figure 6. Deguelin inhibits TGFb1-induced EMT in pancreatic cancer cells. PanC-1 cells were serum-starved and treated with TGFb1or deguelin or both. (a) After 72 h of treatment, cells were either evaluated for morphological change (white arrows show lamellipodia whereas, black arrows show space between cells) or (b) invasion by Boyden’s Transwell assay or (c) EMT markers by western blot. (d) PanC-1 cells were treated with deguelin for 24 h and whole-cell lysates were evaluated for phosphorylation and expression of Smad3/4 or (d: right panel) nuclear localization of Smad4. in vivo and in vitro, indicating that deguelin targets EMT to inhibit as compared with deguelin-treated tumors. On the basis of these metastasis of pancreatic tumors along with primary tumor growth. observations, we hypothesized that deguelin target tight junction In the present study, we established that deguelin targets TGFb1 proteins, which are necessary to bind epithelial cells to each other, canonical Smad signaling and non-canonical NFkB signaling to prevent tumor diffusion (metastasis). E-Cadherin is a tight pathway to downregulate Snail expression leading to the junction protein that is differentially expressed in compact and induction of E-Cadherin. Deguelin also suppresses primary diffused tumors,28,29 and has a critical role in tumor pancreatic tumor growth by targeting NFkB. metastasis.30,31 E-Cadherin expression was signiﬁcantly increased As pancreatic tumors have high tendency to metastasize to by deguelin treatment; whereas mesenchymal markers were peritoneal organs such as liver, spleen and intestinal mesentery decreased in vitro and in vivo conﬁrming that deguelin inhibits nodules, peritoneal dissemination and liver metastasis were used EMT in pancreatic cancer. It appears that upregulation of as parameters to measure pancreatic tumor metastasis. E-Cadherin by deguelin in our model was due to the inhibition As compared with control tumors, deguelin-treated tumors of Snail, an endogenous repressor of E-Cadherin. reduced peritoneal dissemination and blocked metastasis to liver. Surprisingly, the increase in E-Cadherin mRNA by deguelin did Interestingly, control tumors were more diffused and less compact not exactly correlated with the protein expression of E-Cadherin

& 2013 Macmillan Publishers Limited Oncogene (2013) 3980 – 3991 Deguelin suppresses pancreatic tumor metastasis SR Boreddy and SK Srivastava 3988 PanC-1 COLO-357

NFkB PanC-1, TGF1 (10ng/mL) p-NFkB p-NFkB (Ser-536) (Ser-536) Actin RKIP Time (h) 0 0.5 1 2 8 24 48 72 E-Cadherin Control TEFβ TEFβ+Deg Snail

Cytokeratin 18

Actin

NFkB SiRNA --+ - -+ TGF (10ng/mL) --++ + +

PanC-1COLO-357 L3.6PL PanC-1 p-IkB (Ser-32) p-IkB (Ser-32) p-NFkB (Ser-536) p-NFkB (Ser-536) NFkB NFkB Actin Actin TGF1 (10ng/mL) --+ + Deguelin (M) 0 5 10 20 0 5 10 20 0 2.5 5 10 Deguelin (10M) -++ -

PanC-1 COLO-357 P<0.001 P<0.0001 120 120 100 100 80 80 60 60

Activity (%) 40 40 Activity (%) 20 20 NFkB Transcriptional NFkB Trnascriptional 0 0 0 5 10 20 0 5 10 20 Duguelin concentration (M) Duguelin concentration ( M)

PanC-1 COLO-357 PanC-1 p-NFkB Cytosol Nuclear (Ser-536) p-NFkB (Ser-536)

Actin/Lamin B

TGF1 (10ng/mL) -+ + - + + Deguelin (10M) --+--+ Deguelin (10M) -+ -+

Figure 7. Role of NFkBinTGFb1-induced EMT. (a) PanC-1 and COLO-357 cells were transfected with 100 pM NFkB siRNA, treated with TGFb1for 72 h and whole-cell lysates were analyzed for EMT markers by western blot. (a: right upper panel). PanC-1 cells were treated with TGFb1for various time points, and phosphorylation of NFkB (Ser-536) was monitored by western blotting. (a: right bottom panel) PanC-1 cells were treated with TGFb1 for 2 h and immunostained for NFkB (red). (b) Pancreatic cancer cells were treated with deguelin or (b: right panel) pretreated with TGFb1 followed by deguelin and phosphorylation of IkB (Ser-32), and NFkB (Ser-536) was evaluated by western blot. (c)PanC-1 and COLO-357 cells were treated with deguelin, and NFkB transcriptional activity was measured using commercially available kit from Panomics (Fremont, CA, USA) and (d) DNA-binding activity was measured by electrophoretic mobility shift assay kit (Panomics), according to the manufacturer’s instructions. (d: right panel) PanC-1 cells were pretreated with deguelin, primed with TGFb1, and nuclear and cytosolic fractions were subjected to western blot analysis to evaluate NFkB localization. A full colour version of this ﬁgure is available at the Oncogene journal online.

in PanC-1 cells. Furthermore, the extent of Snail downregulation E-Cadherin expression by deguelin treatment was due to the also did not correlated with E-Cadherin upregulation as expected. increase in membrane-bound E-Cadherin. In agreement with our One plausible reason could be that concomitant with Snail hypothesis, deguelin signiﬁcantly downregulated E-Cadherin downregulation, deguelin might be activating other E-Cadherin expression in cytosolic fraction whereas increased its expression repressors such as Slug and Twist. However, both the repressors in the membrane fraction of the cells. Our hypothesis was also Slug and Twist were downregulated by deguelin treatment (data strengthened by the observation that E-Cadherin and b-catenin not shown). Another possibility that exists is the localization of interactions were increased in deguelin-treated PanC-1 cells, E-Cadherin. Primarily, E-Cadherin is expressed as transmembrane conﬁrming that more E-Cadherin was accumulated in protein that interacts with b-catenin to maintain cell–cell membrane fraction. adhesion.32 Salahshor et al.33 have reported that 24% pancreatic EMT can be induced or regulated by various growth factors tumors show abnormal localization of E-Cadherin in the cytosol such as TGFb1, FGF, HGF, PDGF, Wnt and Notch.7 TGFb is and nucleus. Thus, we hypothesized that overall increase in known to induce EMT by Smad-dependent (canonical) and

Oncogene (2013) 3980 – 3991 & 2013 Macmillan Publishers Limited Deguelin suppresses pancreatic tumor metastasis SR Boreddy and SK Srivastava 3989 PanC-1 COLO-357 Snail Snail NFkB (Ser-536) NFkB (Ser-536) E-Cadherin E-Cadherin Cytokeratin 18 N-Cadherin Vimentin Cytokeratin 18 RKIP Vimentin Cl. Caspase 3 Cl. Caspase 3 Cl. PARP Cl. PARP Actin Actin - - - + + Snail over exp. ---++Snail over exp. - ---+ Empty vector --+ --Empty vector -+--+Deguelin (10μM) -+--+Deguelin (10μM)

PanC-1 COLO-357 NFkB NFkB E-Cadherin E-Cadherin N-Cadherin Vimentin Cytokeratin 18 Snail Vimetin Cl. Caspase 3 Snail PARP Cl. Caspase 3 Actin PARP - - - + + NFkB over exp. Actin --+ -- Empty vector -+- - + Deguelin (10μM) -- - ++NFkB over exp. - - + --Empty vector -+ - - + Deguelin (10μM) Figure 8. Overexpression of Snail or NFkB abrogates deguelin-induced epithelial markers in pancreatic cancer cells. PanC-1 and COLO-357 cells were reverse transfected with (a, b)NFkB plasmid or (c, d) Snail plasmid in PanC-1 and COLO-357 cells, respectively, using FuGene transfection reagent according to the manufacturer’s instructions and treated with deguelin. Equal amount of protein from control and treated cells were analyzed for EMT markers.

Smad-independent (non-canonical) pathways.27 It is possible that TAK1-induced phosphorylation of IkB (Ser-32) in TGFb-treated deguelin specifically targets R-Smads to inhibit TGFb1 signaling, cells. Inhibition of IkB phosphorylation by deguelin lead to as Smad3 activation (Ser-423/425) was significantly suppressed stabilization of NFkB-IkB interaction thus prevent nuclear by deguelin in TGFb1primed cells, without affecting the protein translocation of NFkB and obligatory TGFb1-induced Snail levels of Smad4. TGFb1 modulate NFkB activity in certain induction. As, deguelin treatment significantly induced RKIP epithelial cells,8,34 and NFkB is essential for TGFb1-induced EMT expression, it could lead to the inhibition of NFkB activation. in certain cancers such as breast cancer.11 Interestingly, when To evaluate whether deguelin-induced reversal of EMT and NFkB was silenced in PanC-1 and COLO-357 cells and treated with apoptosis induction were through a common pathway, induction TGFb1, TGFb1-induced Snail induction and E-Cadherin down- of apoptosis (Cl. Caspase-3 and Cl. PARP) was evaluated in regulation was completely prevented. NFkB overexpression deguelin-treated NFkB or Snail overexpressing PanC-1 and COLO- induced Snail expression and downregulated E-Cadherin in 357 cells. Interestingly, deguelin-induced apoptosis was comple- PanC-1 and COLO-357 cells. On the other hand, TGFb1 tely blocked in NFkB overexpressing cells but not in Snail treatment resulted in the phosphorylation of IkB (Ser-32) and overexpressing cells. Nevertheless, EMT reversal by deguelin was NFkB (Ser-536). Taken together, these observations indicate that significantly blocked in both NFkB and Snail overexpressing cells. NFkB-Snail pathway is indispensable for TGFb1-induced EMT in These results suggest that upstream regulator, such as NFkB, has pancreatic cancer cells. Interestingly, deguelin significantly significant role in both deguelin-induced reversal of EMT and inhibited both constitutive and TGFb1-induced activation of IkB apoptosis process, however, the downstream effectors are (Ser-32/34) and NFkB (Ser-536) in PanC-1 cells and orthotopically different. For example, Snail overexpression failed to protect the implanted tumors. Furthermore, overexpression of either NFkBor cells from deguelin-induced apoptosis but significantly prevented Snail completely abrogated EMT inhibition by deguelin, EMT reversal, suggesting that Snail or EMT plays no role in the indicating that deguelin targets NFkB-Snail pathway to inhibit apoptosis induction through deguelin in our model. A schematic EMT in pancreatic cancer. Interestingly, ectopic expression of diagram of deguelin targets is presented in Figure 9. NFkB induced Snail expression but Snail overexpression did not In conclusion, deguelin suppresses both pancreatic tumor induced NFkB expression, indicating that NFkB acts upstream growth and metastasis by inducing apoptosis and inhibiting EMT. and regulates Snail in our model. There could be two possible mechanisms by which deguelin inhibit NFkB activation in pancreatic cancer cells: (a) TGFb1 induce MATERIALS AND METHODS NFkB nuclear localization by TGFb-activated kinase (TAK1)- Chemicals and antibodies 8 dependent IkB phosphorylation, and (b) RKIP-mediated inhibi- Deguelin and TGF-b1 were purchased from Sigma Chemicals (Sigma, tion of NFkB activation by interacting with various upstream St Louis, MO, USA). All the antibodies were procured from Cell Signaling kinases such as NIK and TAK1.35–37 Deguelin significantly inhibited Technology Inc. (Danvers, MA, USA).

& 2013 Macmillan Publishers Limited Oncogene (2013) 3980 – 3991 Deguelin suppresses pancreatic tumor metastasis SR Boreddy and SK Srivastava 3990 days, mice were randomly separated into two groups with 10 mice in each group. Mice in the experimental group received 5 mg/kg deguelin (daily, i.p., 37 days), whereas the control mice received vehicle alone. Dose of the deguelin was selected based on previously published literature.41–43 Tumor luminescence and animal weight was measured twice a week for 6 weeks. At the end of the experiment (day 45), mice were killed; tumors and pancreas were excised from each mouse weighed and snap-frozen for western blot and immunoﬂuorescence analysis. Tumor migration from the site of injection was measured by Living Image software (Caliper LifeSciences).

Cytotoxicity, wound healing, transwell and RT–PCR analysis PanC-1, COLO-357 and L3.6pl cells were treated with various concentrations of deguelin, and cytotoxicity, wound healing and transwell assays were performed as described by us earlier.44,45 Total RNA was extracted from control and deguelin-treated cells, and RT–PCR analysis was performed as described by us earlier45 using E-Cadherin sense and antisense primers 50-CGCCCTATGATTCTCTGCTCG-30 and 50-TCGTCCTCGCC GCCTCCGTA-30, respectively.

TGF-b1 treatment Cells were plated in a six-well plate at a density of 0.3 Â 106 cells/well and left overnight. Next day, cells were serum-starved overnight and media was replaced with fresh media containing 10 ng/ml TGF-b1. After 48 h of incubation, various concentrations of deguelin were added to TGF-b1- Figure 9. Deguelin targets canonical and non-canonical TGFb1 treated cells and coincubated for additional 24 h. Cells were collected and pathway to inhibit EMT and metastasis. TGFb1 is known to induce subjected to either western blot analysis or immunostained and observed EMT by both Smad-dependent and Smad-independent pathways. In under microscope. canonical pathway, upon binding to its receptor TGFb1 leads to the activation of R-Smads, such as Smad2 and Smad3. The activated NFkB transcriptional activity and DNA binding (electrophoretic R-Smads binds to co-Smads such as Smad4 and translocate into the mobility shift assay) nucleus leading to Snail upregulation and eventually EMT. In TGFb1- PanC-1 or COLO-357 cells were treated with deguelin (10 mM) for 24 h ,and non-canonical pathway, TGFb1 receptors activate various down- nuclear and cytosolic fractions were collected. NFkB DNA-binding activity stream molecules such as NFkB and MAPK. TGFb1 is known to and transcriptional activity was measured as described by us earlier.46 activate TAK1, which in turn phosphorylates NFkB inhibitor IkB leading to the proteosomal degradation of IkB. IkB-free NFkB translocate to nucleus and induce Snail expression, which leads to Transient transfections E-Cadherin downregulation. %Indicates deguelin potential targets. PanC-1 or COLO-357 cells were transiently reverse transfected with either NFkB or Snail overexpressing plasmids using FuGene 6 (Roche Diagnostics Corporation, Indianapolis, IN, USA) transfection reagent, according to the Cell culture manufacturer’s instructions. Briefly, either 2 mgofNFkB-plasmid-DNA or 1.5 mg of Snail-plasmid-DNA for PanC-1 and 3 mgofNFkB-plasmid-DNA or Human pancreatic cancer cell lines Panc-1 and L3.6pl cells were kind gift 2 mg of Snail-plasmid-DNA for COLO-357 was diluted separately in Opti- from Dr Thomas L Brown (Wright State University, Dayton, OH, USA) and Dr MEM media to which FuGene transfection reagent was added in a ratio of Ming H Wang (Texas Tech Health Sciences Center, Amarillo, TX, USA), 1:3 (w:v) and incubated at room temperature for 1 h in a six-well plate. respectively. COLO-357 cells were kind gift from Dr Fazlul H Sarkar (Wayne After incubation, Panc-1 or COLO-357 cells were trypsinized and around State University, Detroit, MI, USA) and Dr Paul Chiao (MD Anderson Cancer 6 0.3 Â 10 cells were added separately to DNA–FuGene complexes formed in Center, Houston, TX, USA). PanC-1 cells stably expressing luciferase (PanC- the six-well plates and incubated for another 5 h. Medium was replaced 1-luc) was a kind gift from Dr Frank C Marini (MD Anderson Cancer Center, with fresh media and incubated for additional 48 h. Transfected cells were Houston, TX, USA). Human pancreatic ductal epithelial cells (HPDE-6) were either treated with 0.1% dimethylsulphoxide or deguelin (10 mM) for 24 h. generously provided by Dr Ming-Sound Tsao (University of Toronto, Cells were collected and subjected to western blot analysis. Toronto, Ontario, Canada). All the cell lines were maintained in DMEM medium except HPDE-6, which were cultured in keratinocyte-SFM medium as described by us earlier.38 Western blot analysis Treated and control cells were collected and lysed in lysis buffer containing protease and phosphatase inhibitors by incubating on ice for 20 min The Orthotopic implantation of PanC-1 luciferase cells protein concentration was estimated by Bradford reagent (BioRad, PanC-1-luc cells were orthotopically implanted in the pancreas of nude Hercules, CA, USA). Forty microgram of protein was resolved on 10–12% mice, as described by us earlier.39 All the experiments involving animals SDS–polyacrylamide gel electrophoresis and transferred on to polyvinyli- were approved by the Institutional Animal Care and Use Committee. dene fluoride membrane for western blot analysis as described by us Twenty 4- to 6-week-old female athymic nude mice (Charles River, previously.44,47 Wilmington, MA, USA) were kept on an antioxidant-free AIN-76A diet (TestDiet, Richmond, IN, USA) during the experiment. The mice were used Statistical analysis in each group to get better statistical difference. Sample size was calculated using the software developed by DuPont and Plummer. The All the statistical analysis was performed using Prism 5.0 software (Graph following parameters were used to calculate the sample size: a (0.1), d (0.5), Pad software Inc., San Diego, CA, USA). Results were expressed as ± s (0.5) and power (0.9).40 Animals were anesthetized by Ketamine– mean s.d. of at least three independent experiments. Data were analyzed Xylazine–Acepromazine mixture and a small left abdominal flank incision by Student’s t-test or one way ANOVA followed by Bonferroni’s post-hoc was made. Around 1 Â 106 exponentially growing PanC-1-luc cells analysis for multiple comparisons. Differences were considered statistically suspended in 20 ml PBS were injected into the subcapsular region of the significant at Po0.05. pancreas. The peritoneum and skin incisions were closed sequentially with absorbable suture. Animals were imaged on the next day of surgery for basal luminescence using IVIS Bio Luminescent System equipped with CONFLICT OF INTEREST Living Image software (Caliper LifeSciences, Hopkinton, MA, USA). After 7 The authors declare no conflict of interest.

Oncogene (2013) 3980 – 3991 & 2013 Macmillan Publishers Limited Deguelin suppresses pancreatic tumor metastasis SR Boreddy and SK Srivastava 3991 ACKNOWLEDGEMENTS cancer cells is linked with activation of the notch signaling pathway. Cancer Res We would like to acknowledge National Cancer Institute for supporting the present 2009; 69: 2400–2407. study in part by R01 grants CA106953 and CA129038 (to SKS). The technical assistance 23 Shah AN, Summy JM, Zhang J, Park SI, Parikh NU, Gallick GE. Development and of Kartick C Pramanik, Prabodh K Kandala and Parul Gupta in the in vivo experiment is characterization of gemcitabine-resistant pancreatic tumor cells. Ann Surg Oncol greatly appreciated. We would also like to thank Dr Thomas L Brown (Wright 2007; 14: 3629–3637. State University, Dayton, OH, USA), Dr Ming H Wang (Texas Tech Health Sciences 24 Ali S, Ahmad A, Banerjee S, Padhye S, Dominiak K, Schaffert JM et al. Gemcitabine Center, Amarillo, TX, USA), Dr Fazlul H Sarkar (Wayne State University, Detroit, MI, USA), sensitivity can be induced in pancreatic cancer cells through modulation of miR- Dr Paul Chiao (MD Anderson Cancer Center, Houston, TX, USA), Dr Frank C Marini (MD 200 and miR-21 expression by curcumin or its analogue CDF. Cancer Res 2010; 70: Anderson Cancer Center, Houston, TX, USA) and Dr Ming-Sound Tsao (University of 3606–3617. Toronto, Toronto, Ontario, Canada) for providing the cell lines. 25 Zavadil J, Bottinger EP. TGF-beta and epithelial-to-mesenchymal transitions. Oncogene 2005; 24: 5764–5774. 26 Takekawa M, Tatebayashi K, Itoh F, Adachi M, Imai K, Saito H. Smad-dependent REFERENCES GADD45beta expression mediates delayed activation of p38 MAP kinase by TGF- beta. EMBO J 2002; 21: 6473–6482. 1 Jemal A, Siegel R, Xu J, Ward E. Cancer statistics, 2010. CA Cancer J Clin 2010; 60: 27 Zhang YE. Non-Smad pathways in TGF-beta signaling. Cell Res 2009; 19: 128–139. 277–300. 28 Brinck U, Jacobs S, Neuss M, Tory K, Rath W, Kulle B et al. Diffuse growth pattern 2 Li D, Xie K, Wolff R, Abbruzzese JL. Pancreatic cancer. Lancet 2004; 363: affects E-cadherin expression in invasive breast cancer. Anticancer Res 2004; 24: 1049–1057. 2237–2242. 3 Besmer DM, Curry JM, Roy LD, Tinder TL, Sahraei M, Schettini J et al. Pancreatic 29 Weinel RJ, Neumann K, Kisker O, Rosendahl A. Expression and potential role of ductal adenocarcinoma mice lacking mucin 1 have a profound defect in tumor E-cadherin in pancreatic carcinoma. Int J Pancreatol 1996; 19: 25–30. growth and metastasis. Cancer Res 2011; 71: 4432–4442. 30 Onder TT, Gupta PB, Mani SA, Yang J, Lander ES, Weinberg RA. Loss of E-cadherin 4 Cates JM, Byrd RH, Fohn LE, Tatsas AD, Washington MK, Black CC. Epithelial- promotes metastasis via multiple downstream transcriptional pathways. Cancer mesenchymal transition markers in pancreatic ductal adenocarcinoma. Pancreas 2009; 38: e1–e6. Res 2008; 68: 3645–3654. 5 Javle MM, Gibbs JF, Iwata KK, Pak Y, Rutledge P, Yu J et al. Epithelial-mesenchymal 31 von Burstin J, Eser S, Paul MC, Seidler B, Brandl M, Messer M et al. E-cadherin transition (EMT) and activated extracellular signal-regulated kinase (p-Erk) in regulates metastasis of pancreatic cancer in vivo and is suppressed by a SNAIL/ surgically resected pancreatic cancer. Ann Surg Oncol 2007; 14: 3527–3533. HDAC1/HDAC2 repressor complex. Gastroenterology 2009; 137: 361–371. 6 Nakajima S, Doi R, Toyoda E, Tsuji S, Wada M, Koizumi M et al. N-cadherin 32 Beavon IR. The E-cadherin-catenin complex in tumour metastasis: structure, expression and epithelial-mesenchymal transition in pancreatic carcinoma. function and regulation. Eur J Cancer 2000; 36: 1607–1620. Clin Cancer Res 2004; 10(Pt 1): 4125–4133. 33 Salahshor S, Naidoo R, Serra S, Shih W, Tsao MS, Chetty R et al. Frequent accu- 7 Xu J, Lamouille S, Derynck R. TGFbeta-induced epithelial to mesenchymal mulation of nuclear E-cadherin and alterations in the Wnt signaling pathway in transition. Cell Res 2009; 19: 156–172. esophageal squamous cell carcinomas. Mod Pathol 2008; 21: 271–281. 8 Gingery A, Bradley EW, Pederson L, Ruan M, Horwood NJ, Oursler MJ. TGF-beta 34 Arsura M, Panta GR, Bilyeu JD, Cavin LG, Sovak MA, Oliver AA et al. Transient coordinately activates TAK1/MEK/AKT/NFkB and SMAD pathways to promote activation of NF-kappaB through a TAK1/IKK kinase pathway by TGF-beta1 inhi- osteoclast survival. Exp Cell Res 2008; 314: 2725–2738. bits AP-1/SMAD signaling and apoptosis: implications in liver tumor formation. 9 Nieto MA. The snail superfamily of zinc-ﬁnger transcription factors. Nat Rev Mol Oncogene 2003; 22: 412–425. Cell Biol 2002; 3: 155–166. 35 Wu K, Bonavida B. The activated NF-kappaB-Snail-RKIP circuitry in cancer reg- 10 Carver EA, Jiang R, Lan Y, Oram KF, Gridley T. The mouse snail gene encodes a key ulates both the metastatic cascade and resistance to apoptosis by cytotoxic regulator of the epithelial-mesenchymal transition. Mol Cell Biol 2001; 21: drugs. Crit Rev Immunol 2009; 29: 241–254. 8184–8188. 36 Tang H, Park S, Sun SC, Trumbly R, Ren G, Tsung E et al. RKIP inhibits NF-kappaB in 11 Huber MA, Azoitei N, Baumann B, Grunert S, Sommer A, Pehamberger H et al. cancer cells by regulating upstream signaling components of the IkappaB kinase NF-kappaB is essential for epithelial-mesenchymal transition and metastasis in a complex. FEBS Lett 2010; 584: 662–668. model of breast cancer progression. J Clin Invest 2004; 114: 569–581. 37 Yeung KC, Rose DW, Dhillon AS, Yaros D, Gustafsson M, Chatterjee D et al. Raf 12 Baritaki S, Chapman A, Yeung K, Spandidos DA, Palladino M, Bonavida B. Inhibi- kinase inhibitor protein interacts with NF-kappaB-inducing kinase and TAK1 and tion of epithelial to mesenchymal transition in metastatic prostate cancer cells by inhibits NF-kappaB activation. Mol Cell Biol 2001; 21: 7207–7217. the novel proteasome inhibitor, NPI-0052: pivotal roles of Snail repression and 38 Sahu RP, Srivastava SK. The role of STAT-3 in the induction of apoptosis in pan- RKIP induction. Oncogene 2009; 28: 3573–3585. creatic cancer cells by benzyl isothiocyanate. J Natl Cancer Inst 2009; 101: 13 Huber MA, Kraut N, Beug H. Molecular requirements for epithelial-mesenchymal 176–193. transition during tumor progression. Curr Opin Cell Biol 2005; 17: 548–558. 39 Pramanik KC, Srivastava SK. Apoptosis signal-regulating kinase 1-thioredoxin 14 Lee HY, Oh SH, Woo JK, Kim WY, Van Pelt CS, Price RE et al. Chemopreventive complex dissociation by capsaicin causes pancreatic tumor growth suppression effects of deguelin, a novel Akt inhibitor, on tobacco-induced lung tumorigenesis. by inducing apoptosis. Antioxid Redox Signal (e-pub ahead of print 11 June 2012). J Natl Cancer Inst 2005; 97: 1695–1699. 40 Dupont WD, Plummer Jr WD. Power and sample size calculations. A review and 15 Murillo G, Salti GI, Kosmeder 2nd JW, Pezzuto JM, Mehta RG. Deguelin inhibits the computer program. Control Clin Trials 1990; 11: 116–128. growth of colon cancer cells through the induction of apoptosis and cell cycle 41 Oh SH, Woo JK, Yazici YD, Myers JN, Kim WY, Jin Q et al. Structural basis for arrest. Eur J Cancer 2002; 38: 2446–2454. depletion of heat shock protein 90 client proteins by deguelin. J Natl Cancer Inst 16 Murillo G, Peng X, Torres KE, Mehta RG. Deguelin inhibits growth of breast cancer 2007; 99: 949–961. cells by modulating the expression of key members of the Wnt signaling path- 42 Yan Y, Wang Y, Tan Q, Lubet RA, You M. Efﬁcacy of deguelin and silibinin on way. Cancer Prev Res (Phila) 2009; 2: 942–950. benzo(a)pyrene-induced lung tumorigenesis in A/J mice. Neoplasia 2005; 7: 17 Udeani GO, Zhao GM, Shin YG, Kosmeder 2nd JW, Beecher CW, Kinghorn AD et al. 1053–1057. Pharmacokinetics of deguelin, a cancer chemopreventive agent in rats. Cancer 43 Kang HW, Kim JM, Cha MY, Jung HC, Song IS, JS Kim. Deguelin an Akt inhibitor, Chemother Pharmacol 2001; 47: 263–268. down-regulates NF-kB signaling and induces apoptosis in colon cancer cells and 18 Ito S, Oyake T, Murai K, Ishida Y. Deguelin suppresses cell proliferation via the inhibits tumor growth in mice. Dig Dis Sci (e-pub ahead of print 24 May 2012). inhibition of survivin expression and STAT3 phosphorylation in HTLV-1-trans- 44 Boreddy SR, Pramanik KC, Srivastava SK. Pancreatic tumor suppression by benzyl formed T cells. Leuk Res 2010; 34: 352–357. isothiocyanate is associated with inhibition of PI3K/AKT/FOXO pathway. Clin 19 Dell’Eva R, Ambrosini C, Minghelli S, Noonan DM, Albini A, Ferrari N. The Akt Cancer Res 2011; 17: 1784–1795. inhibitor deguelin, is an angiopreventive agent also acting on the NF-kappaB 45 Boreddy SR, Sahu RP, Srivastava SK. Benzyl Isothiocyanate suppresses pancreatic pathway. Carcinogenesis 2007; 28: 404–413. tumor angiogenesis and invasion by inhibiting HIF-alpha/VEGF/Rho-GTPases: 20 Lee H, Lee JH, Jung KH, Hong SS. Deguelin promotes apoptosis and inhibits pivotal role of STAT-3. PloS ONE 2011; 6: e25799. angiogenesis of gastric cancer. Oncol Rep 2010; 24: 957–963. 46 Batra S, Sahu RP, Kandala PK, Srivastava SK. Benzyl isothiocyanate-mediated 21 Chen Y, Wu Q, Cui GH, Chen YQ, Li R. Deguelin blocks cells survival signal inhibition of histone deacetylase leads to NF-kappaB turnoff in human pancreatic pathways and induces apoptosis of HL-60 cells in vitro. Int J Hematol 2009; 89: carcinoma cells. Mol Cancer Ther 2010; 9: 1596–1608. 618–623. 47 Kandala PK, Srivastava SK. Diindolylmethane mediated Gli1 suppression induces 22 Wang Z, Li Y, Kong D, Banerjee S, Ahmad A, Azmi AS et al. Acquisition of epi- anoikis in ovarian cancer cells in vitro and blocks tumor formation ability in vivo. thelial-mesenchymal transition phenotype of gemcitabine-resistant pancreatic J Biol Chem 2012; 287: 28745–28754.

& 2013 Macmillan Publishers Limited Oncogene (2013) 3980 – 3991