(2013) 32, 4825–4835 & 2013 Macmillan Publishers Limited All rights reserved 0950-9232/13 www.nature.com/onc

ORIGINAL ARTICLE Inhibition of the Nrf2 by the alkaloid trigonelline renders pancreatic cells more susceptible to through decreased proteasomal expression and activity

A Arlt1,4, S Sebens2,4, S Krebs1, C Geismann1, M Grossmann1, M-L Kruse1, S Schreiber1,3 and H Scha¨fer1

Evidence accumulates that the transcription factor nuclear factor E2-related factor 2 (Nrf2) has an essential role in cancer development and chemoresistance, thus pointing to its potential as an anticancer target and undermining its suitability in chemoprevention. Through the induction of cytoprotective and proteasomal , Nrf2 confers apoptosis protection in tumor cells, and inhibiting Nrf2 would therefore be an efficient strategy in anticancer therapy. In the present study, pancreatic carcinoma cell lines (Panc1, Colo357 and MiaPaca2) and H6c7 pancreatic duct cells were analyzed for the Nrf2-inhibitory effect of the coffee alkaloid trigonelline (trig), as well as for its impact on Nrf2-dependent proteasome activity and resistance to -related apoptosis-inducing ligand (TRAIL) and anticancer drug-induced apoptosis. Chemoresistant Panc1 and Colo357 cells exhibit high constitutive Nrf2 activity, whereas chemosensitive MiaPaca2 and H6c7 cells display little basal but strong tert-butylhydroquinone (tBHQ)-inducible Nrf2 activity and drug resistance. Trig efficiently decreased basal and tBHQ-induced Nrf2 activity in all cell lines, an effect relying on a reduced nuclear accumulation of the Nrf2 . Along with Nrf2 inhibition, trig blocked the Nrf2-dependent expression of proteasomal genes (for example, s5a/ and a5/) and reduced proteasome activity in all cell lines tested. These blocking effects were absent after treatment with Nrf2 siRNA, a condition in which proteasomal and proteasome activity were already decreased, whereas siRNA against the related transcription factor Nrf1 did not affect proteasome activity and the inhibitory effect of trig. Depending on both Nrf2 and proteasomal gene expression, the sensitivity of all cell lines to anticancer drugs and TRAIL-induced apoptosis was enhanced by trig. Moreover, greater antitumor responses toward anticancer drug treatment were observed in tumor-bearing mice when receiving trig. In conclusion, representing an efficient Nrf2 inhibitor capable of blocking Nrf2-dependent proteasome activity and thereby apoptosis protection in pancreatic cancer cells, trig might be beneficial in improving anticancer therapy.

Oncogene (2013) 32, 4825–4835; doi:10.1038/onc.2012.493; published online 29 October 2012 Keywords: chemoresistance; oxidative stress; tumorigenesis; pancreas

INTRODUCTION both conditions favoring tumorigenesis on one hand and making A great number of malignant tumors,1–11 for example, colonic, tumor cells more refractory to chemo- and radiotherapy on the 2,5,23 thyroid, endomethrial, lung, ovarian, breast and pancreatic cancer, other hand. By inducing a battery of phase II and exhibit an increased activity of the antioxidative transcription detoxification genes that protect cells from anticancer drug factor nuclear Factor E2-related factor 2 (Nrf2). This enhanced Nrf2 toxicity, Nrf2 directly confers chemoresistance, as quite recently 24,25 activation has been shown to originate from rare gain-of-function reported for several types of tumors, including pancreatic 3,4,9 mutations of Nrf2 itself12,13 and from loss-of-function mutations, adenocarcinoma (PDAC). A recent study also identified a role promoter hypermethylation or micro RNA targeting of the of NRF2 in promoting tumor angiogenesis through the HIF-1a/ Nrf2 inhibitory protein Keap1/INRF.14–16 Besides these genetic VEGF pathways, not only underscoring the potential of Nrf2 to 26 and epigenetic alterations, an exaggerated Nrf2 activity may also sustain tumor growth and survival, but also providing an result from cellular adaptation to metabolic stress, for example, interesting insight into how hypoxia-related signals and oxidative fumarate accumulation leading to Keap1 succination17,18 or to stress adaptation may be linked to each other. oxidative stress,19–21 for example, emerging along with persistent Another mode of action by which Nrf2 activation favors inflammation during chronic colitis or pancreatitis. As a conse- tumorigenesis relates to the induction of proteasomal genes quence of the enhanced Nrf2 activity, tumor cells acquire protec- having an impact on the ubiquitine–proteasome signaling path- tion from apoptosis1,20–22 and are more capable of proliferation, way. It is well known that alterations in proteasome activity, which

1Department of Internal Medicine I, Laboratory of Molecular Gastroenterology & Hepatology, UKSH, Kiel, Germany; 2Institute of Experimental Medicine, Inflammatory Carcinogenesis Research Group, Kiel, Germany and 3Institute of Clinical Molecular Biology, UKSH, Kiel, Germany. Correspondence: Professor H Scha¨fer, Department of Internal Medicine I, Laboratory of Molecular Gastroenterology & Hepatology, Bldg. 6, UKSH-Campus Kiel, Schittenhelmstr. 12, Kiel, Schleswig-Holstein, D-24105, Germany. E-mail: [email protected] 4These authors contributed equally to this work. This work is part of a doctoral thesis (SK). Received 13 December 2011; revised 30 August 2012; accepted 13 September 2012; published online 29 October 2012 Nrf2 inhibition and pancreatic cancer A Arlt et al 4826 is crucial for cellular homeostasis and regular , are Effect of trigonelline on Nrf2 activity in PDAC and H6c7 cells involved in tumor development. Upregulation of proteasomal ARE-luciferase assays conducted with Panc1 and Colo357 cells 1,27–30 gene expression in tumors has been previously reported, subject to trig treatment at various concentrations (0.01–10 mM) for and in this context Nrf2 has an important role owing to its 16 h revealed a dose-dependent inhibition of ARE-driven lucifer- inducing effect on proteasomal gene expression and thereby on ase expression by trig. The greatest inhibition was seen at 1,31–34 26S/20S proteasome activity. Recent studies identified a concentrations between 0.1 and 1 mM (Figure 1c). Similarly, number of proteasomal genes (for example, a5/PSMA5, PSMB5, ARE-driven luciferase expression could be dose-dependently b1/PSMB6 and s5a/PSMD4) transcriptionally regulated by Nrf2 inhibited by trig in all cell lines treated with 50 mM tBHQ for 8 h through one or multiple antioxidant response elements (ARE) and preincubated (1 and 16 h) with trig before tBHQ administra- 35,36 within their gene promoters. In colon cancer, greater Nrf2 tion. The 16-h preincubation with trig was less efficient than the activity and increased proteasomal gene expression correlate with 1-h preincubation. At the latter time point and at a dose of 0.1 mM, 1 an elevated proteasome activity, and human colonocytes and most significant inhibitory effects were observed in MiaPaca2 cells colon cancer cells acquire an increased apoptosis protection exhibiting a 54% decrease of ARE-driven luciferase expression in from elevated proteasomal gene expression and proteasome comparison with vehicle-treated cells. In the other cell lines, the 1,28,30 activation. inhibitory effect of trig at 0.1 mM was somewhat lower, ranging Thus, evidence has accumulated that Nrf2 exhibits profound between 35 and 50% inhibition. At higher doses (1 and 10 mM), the 37,38 protumorigenic activity and its targeting may therefore have inhibition of ARE-driven luciferase by trig was not stronger or even 23,39 great potential in antitumor therapy, particularly in over- less pronounced.45 coming chemoresistance, for example, in PDAC. However, Nrf2 has also gained attention in terms of its use as a chemopreventive Trig affects nuclear localization of Nrf2 in PDAC and H6c7 cells target, because the activation of Nrf2 by certain antioxidants such as sulforaphane and oltipraz leads to protection from toxic To elucidate the mechanisms by which trig interferes with Nrf2 DNA damage and thereby from tumorigenesis.40,41 Obviously, activation, its effect on the subcellular distribution of Nrf2 was in normal cells exhibiting tightly controlled Nrf2 activation, the investigated. The inhibitory impact of trig on Nrf2 could be impact of Nrf2 is rather preventive against cancer, but in verified by a decline in the nuclear level of Nrf2 protein that was permanently stressed or even transformed cells exhibiting seen in Panc1 and Colo357 cells treated with the compound for deregulated Nrf2 activity its impact is rather protumorigenic.37,42 8 h (Figure 2a). Similarly, H6c7 and MiaPaca2 cells stimulated with Besides the induction of phase II expression through Nrf2, tBHQ exhibited a decreased accumulation of Nrf2 protein in the accounting for direct detoxification of anticancer drugs, the nucleus when subjected to pretreatment with trig (Figure 2b). In impact of Nrf2 on the ubiquitine–proteasome signaling pathway contrast, no differences in the amount of Nrf2 protein were confers general apoptosis resistance that manifests in protection detected in total cellular extracts (Figures 2a and b) Thus, the from various apoptotic stimuli, for example, death ligands, and in higher basal and induced level of nuclear Nrf2 protein in Panc1 or enhanced proliferation. One goal therefore would be to establish Colo357 cells and H6c7 or MiaPaca2 cells, respectively, were novel compounds that selectively block Nrf2 in tumor cells. markedly affected by trig, indicating that the effect of the drug Recently, certain signaling pathways—for example, through on the decline in Nrf2-driven gene transcription was related retinoic acid receptor alpha or estrogen-related receptor beta— to a decrease in its nuclear localization. Immunofluorescence have been described that interfere with Nrf2 activation,43,44 but in microscopy verified the reduced nuclear accumulation of Nrf2 PDAC cells these nuclear receptors are not expressed. Thus, other protein in the PDAC cell lines and H6c7 cells when subjected to natural compounds that have been identified to inhibit Nrf2, for trig treatment (Supplementary Figure S1). example, the alkaloid trigonelline (trig),45 would be attractive tools When treated with leptomycin-B (LMB), an inhibitor of the for the sensitization of tumor cells to apoptosis. crm1-dependent nuclear export of , tBHQ-stimulated In the present study, we elucidated the effects of trig on Nrf2- Panc1, Colo357, H6c7 and MiaPaca2 cells still exhibited lower dependent proteasome activity and antiapoptotic protection in levels of nuclear Nrf2 protein after administration of trig, whereas PDAC cell lines (MiaPaca2, Panc1 and Colo357), as well as in the protein levels of Keap1 were not affected (Figure 2c, lower panel). Therefore, the effect of trig did not depend on the nuclear export human pancreatic duct cell line H6c7. It could be shown that, in all 46 cell lines investigated, trig efficiently suppressed Nrf2 activity of Nrf2 and/or Keap1, but rather on the import of Nrf2 into the along with a decreased proteasome activity, leading to an nucleus. This is underlined by the increased level of Nrf2 protein in elevated sensitivity to apoptosis induction in vitro and in vivo, the cytoplasm if trig was administered to LMB- and tBHQ-treated indicating its suitability as an Nrf2 inhibitor adding to anticancer cells (Figure 2c, upper panel). Similarly, ARE-luciferase assays therapy. revealed a decreased ARE-driven reporter gene expression in all four cell lines subjected to LMB and tBHQ treatment if trig had been administered (Figure 2d).

RESULTS Effect of trig on proteasome activity in PDAC and H6c7 cells Basal and inducible Nrf2 activity in PDAC and H6c7 cells Next, we investigated the effect of trig on basal and tBHQ-induced MiaPaca2, Panc1 and Colo357 PDAC cells, as well as the human proteasome activity. For this purpose, PDAC and H6c7 cells pancreatic duct cell line H6c7, were analyzed for basal and subjected to tBHQ treatment or not and in the absence or inducible Nrf2 activity. analysis revealed greater presence of trig were submitted to the Suc-LLVY-AMC proteasome nuclear protein level of Nrf2 (visible as 100 kDa band) in Panc1 and assay. As shown in Figure 3a, proteasome activity was induced by Colo357 cells as compared with MiaPaca2 and H6c7 cells tBHQ in all the tested cell lines. In the presence of 100 nM trig, (Figure 1a). Treatment with 50 mM tBHQ for 8 h increased the basal proteasome was significantly suppressed in Colo357 and amount of Nrf2 in nuclear extracts from all cell lines (Figure 1a), Panc1 cells (39% and 57%, respectively), and tBHQ-induced exerting the greatest effect in MiaPaca2 and H6c7 cells. Similarly, proteasome activity was markedly reduced by trig in all the cell ARE-luciferase assays demonstrated (Figure 1b) that untreated lines. This effect was most prominent in H6c7, MiaPaca2 Panc1 and Colo357 cells possess greater Nrf2 activity than and Colo357 cells (58%, 64% and 72% decrease, respectively) MiaPaca2 and H6c7 cells. Upon treatment with tBHQ for 8 h, all and somewhat lower in Panc1 cells (36% decrease). four cell lines exhibited an increase of ARE-driven luciferase To confirm the Nrf2 dependency of the blocking effect of trig activity that was strongest in MiaPaca2 and H6c7 cells. on proteasome activity, all cell lines were transfected with control

Oncogene (2013) 4825 – 4835 & 2013 Macmillan Publishers Limited Nrf2 inhibition and pancreatic cancer A Arlt et al 4827

Figure 1. Nrf2 activity in PDAC and H6c7 cells and its inhibition by trig. (a) Nuclear extracts from the indicated cell lines either left untreated or treated with 50 mM tBHQ for 8 h were submitted to western blotting for the detection of Nrf2 and Hsp90 as loading control. A representative out of four independent experiments is shown. (b) The indicated cell lines were transfected with either an empty or an ARE-driven firefly luciferase (ff) construct together with a constitutively renilla luciferase (rl)-expressing construct. Cells were either left untreated or were treated with 50 mM tBHQ for 8 h. Then, luciferase activity was determined and relative luciferase units (RLU) were calculated from the ff/rl ratios. Data are expressed as n-fold activity of ARE-driven RLU normalized to the empty vector RLU. Mean values±s.d. from four independent experiments are shown; *Po0.05. (c) The indicated cell lines were transfected with either an empty or an ARE-driven firefly luciferase (ff) construct together with a constitutively renilla luciferase (rl) expressing construct. Then, the cells were treated with trig at the indicated doses for 16 h or not (left panel), or cells were treated with 50 mM tBHQ for 8 h either preincubated ( À 1 h and À 16 h, respectively) with trig at the indicated doses, or not (center & right panels). Luciferase activity was determined and relative luciferase units (RLU) were calculated from the ff/rl ratios. Data are expressed as the percentage of specific ARE-driven RLU normalized to the empty vector RLU. Mean values±s.d. from six independent experiments are shown; *Po0.05, **Po0.01. and Nrf2 siRNA and the proteasome activity was determined. As basal expression level of S5a and a5 were significantly lower in shown in Figure 3b, tBHQ-induced proteasome activity was Colo357 and Panc1 cells subjected to trig treatment as compared significantly lower in Nrf2 siRNA-treated H6c7, MiaPaca2, Colo357 with untreated cells. Similarly, the increase of s5a and a5 mRNA and Panc1 cells (28%, 41%, 29% and 32% decrease, respectively) levels in response to tBHQ treatment was suppressed by trig in when compared with control siRNA-treated cells. Notably, the MiaPaca2 and H6c7 cells. A similar alteration in response to trig addition of 100 nM trig only marginally affected tBHQ-induced treatment was seen when analyzing the expression of glutamate– proteasome activity in all the cell lines subjected to Nrf2 siRNA cysteine ligase catalytic subunit as a member of phase II enzymes treatment. Moreover, transfection of H6c7, MiaPaca2, Colo357 and among the Nrf2 target genes. Panc1 cells with siRNA targeting the related transcription factor Western blot analysis confirmed the trig-dependent decline Nrf147 did not significantly affect tBHQ-induced proteasome activity of s5a and a5 expression in Colo357 and Panc1 cells, as well as in in Colo357 and Panc1 cells (Figure 3b) or even slightly increased tBHQ-treated MiaPaca2 and H6c7 cells (Figure 4b). proteasome activation in H6c7 and MiaPaca2 cells. In contrast to Nrf2 knockdown, after silencing of Nrf1 expression, trig treatment was still effective in the inhibition of tBHQ-induced proteasome The impact of trig on TRAIL and etoposide-induced apoptosis in activity in H6c7, MiaPaca2, Colo357 and Panc1 cells (44%, 45%, 31% PDAC and H6c7 cells and 27% decrease, respectively). To study the impact of the suppressed Nrf2 activation on death ligand or anticancer drug-induced apoptosis, Colo357 and Panc1 cells were treated with 100 nM trig for 1 h followed by the Altered Nrf2-dependent expression of proteasomal genes in PDAC administration of tumor necrosis factor-related apoptosis-inducing and H6c7 cells subjected to treatment with trig ligand (TRAIL) or etoposide, and apoptosis was determined by To elucidate whether the effect of trig on Nrf2-dependent assaying caspase 3/7 activation. Etoposide was used because proteasome activation relates to alterations of proteasomal gene this anticancer drug exerts the greatest sensitization in many expression, real-time PCR was conducted for the analysis of the chemoresistant PDAC cell lines.48,49 Whereas both cell lines per se expression of the proteasomal 19S and 20S subunit proteins S5a were only moderately sensitive to TRAIL-induced apoptosis (psmd4) and a5 (psma5), respectively. As shown in Figure 4a, the (Figure 5a), exhibiting an increase of caspase 3/7 activation by

& 2013 Macmillan Publishers Limited Oncogene (2013) 4825 – 4835 Nrf2 inhibition and pancreatic cancer A Arlt et al 4828

Figure 2. Trig treatment decreases nuclear protein level of Nrf2 independent of its nuclear export. (a) Panc1 and Colo357 cells were treated with 0.5 mM trig (trig) for 24 h, or (b) H6c7 and MiaPaca2 cells were treated with 50 mM tBHQ for 8 h after preincubation with trig (0.5 mM)for1h or not. Then, total cell lysates or nuclear extracts were prepared and submitted to western blots for the detection of Nrf2 and Hsp90 or lamin- A/C. Representative results of two independent experiments are shown. (c) Panc1, Colo357, MiaPaca2 and H6c7 cells were treated with 50 mM tBHQ for 8 h in the presence of 20 ng/ml of the crm1 inhibitor LMB, either preincubated with trig (0.5 mM) or not. Then, cytosolic (c) or nuclear (n) extracts were prepared and submitted to western blots for the detection of Nrf2, Keap1 and tubulin or lamin-A/C. Representative results of two independent experiments are shown. (d) Panc1, Colo357, MiaPaca2 and H6c7 cells either transfected with an empty or an ARE-driven firefly luciferase vector were treated with 50 mM tBHQ for 8 h in the presence of 20 ng/ml LMB, either preincubated with trig (0.5 mM) or not. Then, luciferase activity was determined and relative luciferase units (RLU) were calculated from the ff/rl ratios. Data are expressed as the percentage of specific ARE-driven RLU normalized to the empty vector RLU. Mean values±s.d. from three independent experiments are shown, *Po0.05.

1.8 and 2.7-fold, respectively, pretreatment with trig signifcantly cleavage in MiaPaca2 and H6c7 cells after tBHQ treatment was enhanced apotosis by TRAIL (2.8 and 4.3-fold increase of caspase blocked by pretreatment with trig (Figure 5b, lower panels). 3/7 activity, respectively). Treatment with etoposide elicited a 1.4 and 1.8-fold increase in caspase3/7 activity in Colo357 and Panc1 cells (Figure 5a), respectively, and preincubation with trig led to an The apoptosis-sensitizing effect of trig depends on Nrf2 and increased apoptosis response to etoposide (2.3 and 3.1-fold higher proteasomal gene expression caspase3/7 activity, respectively). Moreover, tBHQ pretreatment of To elucidate whether the apoptosis-sensitizing effect of trig MiaPaca2 and H6c7 cells significantly reduced apoptotic depends on Nrf2 and proteasomal gene expression, Panc1 responses to TRAIL (from 5.5 to 3.6 and 4.4 to 3.5-fold caspase and MiaPaca2 cells were chosen for knockdown experiments. 3/7 activity, respectively) and to etoposide (from 6.7 to 4.5 and 4.8 After transfection with Nrf2 siRNA, the sensitivity of Panc1 cells to 3.6-fold caspase 3/7 activity, respectively). In the presence of toward TRAIL or etoposide-induced apoptosis was enhanced, as trig, tBHQ-induced protection from TRAIL and etoposide-induced shown by a greater caspase 3/7 activation (Figure 6a). The apoptosis were diminished (Figure 5b). addition of trig was not capable of substantially increasing TRAIL- To confirm these apoptosis effects, poly(ADP-ribose)–polymerase 1 or etoposide-induced apoptosis in Nrf2 siRNA-transfected cells (PARP1) cleavage was analyzed by western blotting. As shown in any further, as shown by similar rates of caspase 3/7 activity in Figure 5a (lower panels), PARP1 cleavage after etoposide and TRAIL trig-treated and untreated cells subjected to Nrf2 knockdown treatment was enhanced by trig in Colo357 and Panc1 cells. (Figure 6a). In addition, the diminishing effect of tBHQ treatment Moreover, the decrease of etoposide- and TRAIL-induced PARP1 on TRAIL- and etoposide-induced apoptosis in MiaPaca2 cells was

Oncogene (2013) 4825 – 4835 & 2013 Macmillan Publishers Limited Nrf2 inhibition and pancreatic cancer A Arlt et al 4829 less pronounced after knocking down Nrf2 expression, as shown by greater caspase3/7 activation in response to both apoptosis stimuli (Figure 6b). Moreover, the addition of trig under Nrf2 knockdown was not capable of substantially increasing TRAIL- or

Figure 3. Effect of trig on proteasome activity in PDAC and H6c7 cells. (a) The indicated cell lines were either left untreated or were treated with 50 mM tBHQ for 24 h, in the absence or presence of 0.1 mM trig, added 1 h before. Then, the cells were submitted to the Suc-LLVY-AMC proteasome assay. Fluoresence units were normal- ized to the protein content determined in parallel. Data are expressed as n-fold AMC activity/mg protein and represent mean values±from six indepedent experiments; *Po0.05. (b) The indicated cell lines were treated with control, Nrf2 and Nrf1 siRNAs. After 48 h, cells were either left untreated or were treated with 50 mM tBHQ in the absence or presence of 0.1 mM trig. The efficacy of the siRNAs was validated by qPCR analysis (Supplementary Figure S5). Data are expressed as n-fold AMC-activity/mg protein and represent mean values±s.d. from four indepedent experiments, *Po0.05.

Figure 4. Altered Nrf2-dependent expression of proteasomal genes in PDAC and H6c7 cells subjected to treatment with trig. (a, b) Colo357 and Panc1 cells were treated with 0.1 mM trig alone for 16 h, or MiaPaca2 and H6c7 cells were treated with 50 mM tBHQ for 16 h, or not, either in the absence or presence of 0.1 mM trig. (a) Total RNA was submitted to reverse transcription and real-time PCR for the detection of the proteasomal genes s5a and a5, as well as the phase II enzyme glutamate–cysteine ligase catalytic subunit (GCLC). For normalization, beta actin was analyzed as control. Data are expressed as normalized mRNA level and represent the mean values±s.d from four independent experiments performed in duplicate; *Po0.05. (b) Total cell lysates were submitted to western blotting for the detection of s5a and a5 proteins, as well as Hsp90 as loading control. A representative result out of four is shown.

& 2013 Macmillan Publishers Limited Oncogene (2013) 4825 – 4835 Nrf2 inhibition and pancreatic cancer A Arlt et al 4830

Figure 5. Impact of trig on TRAIL- and etoposide-induced apoptosis in PDAC and H6c7 cells. (a) Colo357 and Panc1 cells were treated with 0.1 mM trig alone for 16 h, or (b) MiaPaca2 and H6c7 cells were treated with 50 mM tBHQ for 16 h, or not, either in the absence or presence of 0.1 mM trig. Then, the cells were either left untreated or were treated with 20 mg/ml etoposide or 10 ng/ml TRAIL for 24 h and 6 h, respectively. Apoptosis was determined by analyzing caspase 3/7 activity (upper panel) or by western blot analysis of PARP1 cleavage (lower panel) using tubulin as loading control. Data are expressed as n-fold caspase 3,7 activity normalized to the cellular protein content and represent the mean value±s.d. from six independent experiments performed in duplicate; *Po0.05. The western blots are representative of three independent experiments.

etoposide-induced apoptosis in tBHQ-treated MiaPaca2 expression was knocked down. Moreover, tBHQ-treated MiaPaca2 cells. Similar findings were obtained when analyzing Panc1 cells were almost refractory to trig (Figure 7b) when treated and MiaPaca2 cells by PARP1 western blotting (Supplementary with s5a or a5 siRNA. Similar findings were obtained when Figure S2). analyzing Panc1 and MiaPaca2 cells by PARP1 western blotting When transfecting Panc1 cells with s5a or a5 siRNA, the (Supplementary Figure S3). apoptosis in response to TRAIL and etoposide was significantly increased as compared with control siRNA-transfected cells, and the addition of trig only weakly enhanced TRAIL- or Greater antitumor response through combined trig/etoposide etoposide-dependent caspase 3/7 activation (Figure 7a). Similarly, treatment in Colo357 and Panc1 tumor bearing-SCID–beige mice MiaPaca2 cells subjected to tBHQ treatment were more sensitive To verify the chemotherapy sensitizing activity of trig, a to TRAIL- and etoposide-induced apoptosis if s5a and a5 subcutaneous xenograft tumor model in female severe combined

Oncogene (2013) 4825 – 4835 & 2013 Macmillan Publishers Limited Nrf2 inhibition and pancreatic cancer A Arlt et al 4831

Figure 7. The apoptosis-sensitizing effect of trig depends on proteasomal gene expression. The indicated cell lines were treated with control, s5a or a5 siRNA siRNA for 48 h. Then, (a) Panc1 cells were treated with 0.1 mM trig for 16 h followed by the administration of 20 mg/ml etoposide and 10 ng/ml TRAIL for 24 h and 6 h, respectively, or (b) MiaPaca2 cells were treated with 50 mM tBHQ or not for 16 h either in the presence or absence of 0.1 mM trig (added 1 h before), followed by etopside and TRAIL treatment. Apoptosis was determined by analyzing caspase 3/7 activity. Data are expressed as n-fold caspase 3/7 activity normalized to the cellular protein content and represent the mean value±s.d. from four independent experiments performed in duplicate; *Po0.05.

Figure 6. The apoptosis-sensitizing effect of trig depends on Nrf2. 113±45 mm3) and weight (Colo357: 336±136 mg; Panc1: The indicated cell lines were treated with control or Nrf2 siRNA for 84±38 mg) were found to be significantly lower (Figure 8b) in 48 h. Then, (a) Panc1 cells were treated with 0.1 mM trig for 16 h the combination group as compared with the NaCl group followed by adminstration of 20 mg/ml etoposide and 10 ng/ml (Colo357: 671±318 mm3 and 549±293 mg, respectively; Panc1: TRAIL for 24 h and 6 h, respectively, or (b) MiaPaca2 cells were 328±203 mm3 and 220±118 mg, respectively) and the etoposide treated with 50 mM tBHQ or not for 16 h either in the presence or ± 3 ± M group (Colo357: 661 461 mm and 536 264 mg, respectively; absence of 0.1 m trig (added 1 h before), followed by etopside and 3 TRAIL treatment. Apoptosis was determined by analyzing caspase 3/ Panc1: 216±126 mm and 124±74 mg, respectively). 7 activity. Data are expressed as n-fold caspase 3/7 activity In line with these findings and the in vitro data, immunohis- normalized to the cellular protein content and represent the mean tochemical analysis detected a lower nuclear level of activated value±s.d. from four independent experiments performed in Nrf2 protein in tumors from the combination group when duplicate; *Po0.05. compared with tumors from the other two groups (Figure 8c).

DISCUSSION immune deficiency (SCID)-beige mice was employed using Pancreatic ductal adenocarcinoma (PDAC) is still one of the most Colo357 and Panc1 cells. Three weeks after tumor cell inoculation, aggressive and deadly tumors characterized by rapid growth and when the tumors had reached a size of 5 Â 5mm2 or more, tumor- infiltration of surrounding tissues. At the time of diagnosis, most bearing animals were treated (i.p) either with 0.9% NaCl only (as patients suffer from advanced disease and are already in an control), etoposide alone (10 mg/kg body weight) or with a incurable state. Besides a surgical intervention accessible for only combination of etoposide and trig (1 mg/kg body weight). During 10–20% of PDAC patients, no other therapeutical measure is and after treatment, the etoposide/trig combination groups currently available offering cure from the disease.50 On one hand, (III) exhibited a reduced gain of tumor sizes (Figure 8a) this limitation of therapeutic options is due to the rapid and when compared with the NaCl (I)- and etoposide-treated (II) extensive tumor spread of PDAC giving rise to metastases. On the groups, which exhibited continuing tumor growth. After scari- other hand, this is due to the profound resistance of PDAC cells fication of mice, tumor volume (Colo357: 391±155 mm3; Panc1: toward anticancer therapy. This therapy resistance relies on an

& 2013 Macmillan Publishers Limited Oncogene (2013) 4825 – 4835 Nrf2 inhibition and pancreatic cancer A Arlt et al 4832

Figure 8. Greater antitumor response through combined trig/etoposide treatment in Colo357 and Panc1 tumor-bearing SCID–beige mice. SCID mice bearing subcutaneously inoculated Colo357 or Panc1 tumors were treated daily (i.p. injection) for 4 days with 0.9% NaCl (group I) or with 10 mg/kg body weight etoposide either alone (group II) or in combination with 1 mg/kg body weight trig (group III). One day before, animals already received 0.9% NaCl (group I & II) or trig (group III). After 3 days, the same treatment protocol was applied. (a) Tumor sizes (mm2) were analyzed weekly until the scarification of mice 3 weeks after the start of treatment (the treatment periods are indicated by the arrows). Data are expressed as mean±s.d. (n ¼ 6); *Po0.05 compared with group I. At the end of experiment, (b) the volume and weight of freshly excised tumors were determined. Data are expressed as mean±s.d. (n ¼ 6); *Po0.05 compared with group I. (c) Tumor cryosections were stained with a Phoshpo-Nrf2 (Ser40) antibody detecting activated and mainly nuclear Nrf2. Representative images are shown in the lower panel. Expression scores were calculated and the mean±s.d. are shown (upper panel); *Po0.05 compared with group I.

efficient apoptosis protection, reversible by, for example, inhibition of Nrf2 might be an attractive tool to overcome this proteasome inhibitors,48 as well as on the capability of PDAC resistance and to sensitize PDAC cells to anticancer therapy. cells to get rid of the anticancer drugs. Among the mechanisms by Thus, we tested the natural compound trig45 for its capacity to which tumor cells—including PDAC cells—gain apoptotic protec- inhibit Nrf2 and to sensitize PDAC cells to apoptosis and PDAC tion and exclusion to anticancer drugs, the activation of the tumors to chemotherapy. Among all four cell lines tested, trig Keap1/Nrf2 signaling pathway has an important role that is not yet exerted the greatest inhibitory effect on Nrf2 activity at fully defined. Nrf2, either being induced by the cellular alterations submaximal doses (0.1–1 mM); higher doses were less efficient or described above or even by anticancer drugs themselves,24 can even ineffective. This Nrf2 inhibitory potential of trig following a lead to protection from anticancer drugs by inducing a battery of biphasic dose dependency with the highest efficiency at detoxifying enzymes and proteins, as well as by increasing the submicromolar doses has been reported recently.45 Our proteasome activity and thereby apoptosis protection.1,48 investigations further revealed that trig decreases the nuclear As shown in the present study, a line of PDAC cells, as well as level of Nrf2 protein (Figure 2), whereas the overall amount of Nrf2 the human pancreatic duct cell line H6c7, is efficiently protected expression was not altered. As the effect of trig on nuclear Nrf2 from apoptosis by basal and/or tBHQ-induced Nrf2 activity. In protein level, as well as on ARE-dependent luciferase expression, accordance with recent data,4 Colo357 and Panc1 cells are was still present in all the investigated cell lines if stimulated with characterized by higher basal Nrf2 activity, rendering these cells tBHQ and treated with the nuclear export inhibitor LMB, one can less sensitive to chemotherapy-induced apoptosis. This Nrf2- conclude that trig impairs the nuclear import of Nrf2. Under dependent protection includes an increase in proteasomal gene conditions of Nrf2 activation through its release from Keap1 expression, as well as greater proteasome activity, providing the inhibition—here by tBHQ treatment—the inhibitory effect of trig cells with an accelerated turnover of regulatory proteins involved on nuclear accumulation of Nrf2 was still observed even if its in growth regulation and survival.1,22,51 Given this significant crm1-dependent nuclear export was blocked by LMB.46 Moreover, impact of Nrf2 on the resistance of PDAC cells against apoptosis, the addition of the proteasome inhibitor Mg132 (Supplementary

Oncogene (2013) 4825 – 4835 & 2013 Macmillan Publishers Limited Nrf2 inhibition and pancreatic cancer A Arlt et al 4833 Figure S4), which prevents Keap1-dependent intranuclear degra- Fluorometric proteasome assay 52,53, dation of Nrf2 did not affect this function of trig. This indicates For the determination of proteasomal activity in living cells, a fluorometric that trig affects the nuclear import of Nrf2, but not its export or assay using the proteasome substrate Suc-LLVY-AMC in the absence or intranuclear stability. Along with the strong inhibitory effect of trig presence of the proteasome inhibitor Mg132 was performed as described 1 on Nrf2 activation, a suppression of proteasome activity was also recently. All measurements were recorded in duplicates. noted with this compound. Notably, siRNA-mediated knockdown of Nrf2 but not of Nrf1 expression abrogated the inhibitory effect Western blotting of trig on proteasome activity, indicating an exclusive Nrf2 Nuclear and cytosolic extracts or total cell lysates were prepared as dependency under these conditions. This allows to speculate described before.60 After electrophoresis and semidry electroblotting onto that the reported role of Nrf1 in proteasome activity47,54 is PVDF membranes, the following primary antibodies were used for restricted to a feedback effect upon proteasome inhibition. In fact, immunodetection at 1:1000-fold dilution in 5% (w/v) nonfat milk powder the Nrf1 protein is much more susceptible to proteasomal and 0.05% Tween20 in TBS (Tris-buffered saline: 50 mM Tris/HCl, pH 7.6, and degradation than Nrf2.20,47 The latter, however, is particularly 150 mM NaCl): Nrf2, Keap1, lamin-A/C and Hsp90 (Santa Cruz involved in the upregulation of the proteasome upon antioxidant Biotechnology, Heidelberg, Germany), PARP1 (Cell Signaling Technology, 20,55,56 Frankfurt, Germany) and tubulin (Sigma). After incubation overnight at stimulation and during cellular stress response. 4 1C, the blots were exposed to the appropriate horse-radish peroxidase- Moreover, proteasomal gene expression was decreased in the conjugated secondary antibody (Santa Cruz) diluted (1:1000) in blocking investigated cell lines subjected to trig treatment. As proteasome buffer and developed using the Dura detection kit (Perbio Sciences, Bonn, activation contributes to antiapoptotic protection by Nrf2 in Germany). Data acquisition was done with the Chemidoc-XRS gel tumor cells,1,22,51 not only the cellular response to anticancer documentation system (BioRad, Munich, Germany) using the Quantity drugs depending on phase II enzyme and detoxification gene One software (Bio-Rad). Hsp90, lamin-A/C and tubulin served as loading control. expression, but also the responsiveness to death ligands would be affected by Nrf2 inhibition. Accordingly, Nrf2-dependent protection RNA preparation and real-time PCR to anticancer drugs and TRAIL-induced apoptosis was significantly Isolation of total RNA and reverse transcription into single-stranded cDNA abrogated in all cell lines when pretreated with trig. Besides the was carried out as described.20 cDNA was subjected to real-time PCR siRNA-mediated knockdown of Nrf2 expression, the knockdown of (iCycler; BioRad) using the SYBR-Green assay20 with gene-specific primers proteasomal genes such as a5 and s5a similarly increased at a final concentration of 0.2 mM. The primer sequences and PCR apoptosis and abrogated the sensitizing effect of trig. Thus, conditions can be appreciated from Supplementary Table S1. the antiapoptotic action of Nrf2 manifests not only through the reported detoxification of anticancer drugs42 but also through siRNA treatment proteasome activation, thereby broadly protecting cells from For siRNA (Qiagen, Hilden, Germany) treatment, cells grown in 12-well apoptosis induction. In this manner, the blockade of Nrf2 by plates were submitted to lipofection using 6 ml of the HiPerfect reagent phytochemicals—such as trig—would greatly enhance the tumor (Qiagen) and 150 ng/well of either negative control siRNA, Nrf2 killing effects by TRAIL and anticancer drugs. By using a (no. SI03246614), Nrf1 (no. SI00657909), S5a (no. SI03019331) or a-5 siRNA subcutaneous xenograft tumor model in SCID-beige mice, it could (no. SI100043316). be further demonstrated that the chemosensitizing effect of trig also operates in vivo. Accordingly, Colo357 and Panc1 tumor- Dual Luciferase assay bearing mice that had received the combination therapy with ARE-driven reporter gene expression in cells was determined using a etoposide and trig exhibited a significantly greater reduction in commercial pathway detection kit (SABioscience/Qiagen) and as described tumor sizes (Figure 8) than tumor-bearing mice subjected to recently.20 treatment with the chemotherapeutic drug etoposide alone. This finding underscores the suitability of trig to efficiently enhance Caspase-3/-7 assay the antitumor response of anticancer drugs. Apoptosis induced by Killer-TRAIL or etoposide was determined by the Hence, trig has the potential to be used in combination therapy measurement of caspase3/7 activity (Promega, Mannheim, Germany) of highly resistant tumors such as pancreatic cancer. As no adverse according to the manufacturer’s instructions and as described.20 All assays effects of this natural coffee constituent have been reported so far, were performed in duplicates. Caspase3/7 activity was normalized to the and as its suitability for application in humans has been already protein content of the analyzed cell lysates. shown in diabetes and metabolic syndrome patients,57,58 further testing of trig in clinical trials seems to be promising. Subcutaneous xenograft tumor model in SCID mice Colo357 and Panc1 cells (2 Â 106/200 ml saline) were inoculated into the shaved flank of 8-week-old female SCID–beige mice (body weight around 20 g; Charles River, Sulzfeld, Germany). When the tumors had reached a MATERIALS AND METHODS size of 5 Â 5mm2 or more, the animals were randomized into three groups Materials (n ¼ 6) receiving one of the following treatments, which were given by Trig and tert-butylhydroquinone (tBHQ) were purchased from Sigma intraperitoneal injections (200 ml): group I, 0.9% NaCl (day 1–5 & (Deisenhofen, Germany). LMB and Killer-TRAIL were from Enzo Life- 9–13); group II, 10 mg/kg body weight etoposide in 0.9% NaCl (day 2–5 Science/Alexis (Lo¨rrach, Germany). N-succinyl-L-leucyl-L-leucyl-L-valyl-L-tyr- & 10–13); and group III, 0.02 mg/kg body weight trig in 0.9% NaCl (day 1–5 osyl-7-amido-4-methylcumarin (Suc-LLVY-AMC) and Mg132 were from & 9–13) and 0.2 mg/kg body weight etoposide in 0.9% NaCl (day 2–5 & Biomol (Taunusstein, Germany). Etoposide (Vepesid) was from Bristol- 9–13). Before, during and after treatment, tumor sizes were measured Myers/Sqibb (Munich, Germany). weekly until the scarification of mice (day 21 after start of treatment). Next, the tumors were excised, weighed and snap-frozen in liquid nitrogen. The experiments were approved by the local animal research authority. Cell culture The human pancreatic ductal epithelial cell line H6c7 (kindly provided by Immunohistochemistry MS Tsao, Ontario Cancer Center, Toronto, Canada) was cultured as Six-micrometer cryostat sections were mounted on uncovered glass slides, described.59 The human PDAC cell lines MiaPaca2, Panc1 (both from ATCC/ air-dried overnight at room temperature, fixed in chilled acetone LSC) and Colo357 (donated by H Kalthoff, Experimental Cancer Research (Merck, Darmstadt, Germany) for 10 min and air-dried again for 10 min. Institute, UKSH-Campus Kiel) were cultured in RPMI 1640 containing 10% Next, the slides were washed in PBS. To avoid nonspecific binding, sections FCS, 1% L-glutamine and 1% sodium pyruvate (all from PAA-Laboratories, were treated with 4% bovine serum albumin (BSA) (Serva, Heidelberg, Co¨lbe, Germany). Cells were cultured at 37 1C, 5% CO2 and 85% humidity. Germany) for 20 min followed by incubation with the monoclonal rabbit

& 2013 Macmillan Publishers Limited Oncogene (2013) 4825 – 4835 Nrf2 inhibition and pancreatic cancer A Arlt et al 4834 Ser40-Phospho-Nrf2 antibody (Epitomics, Berlin, Germany) at 1:200 dilution 10 Singh A, Misra V, Thimmulappa RK, Lee H, Ames S, Hoque MO et al. in 1% BSA/PBS. After primary antibody incubation (overnight, 4 1C), the Dysfunctional KEAP1-NRF2 interaction in non-small-cell . PLoS Med sections were washed three times in PBS and then treated with EnVision 2006; 3: e420. peroxidase conjugates (DakoCytomation, Hamburg, Germany) for 30 min. 11 Solis LM, Behrens C, Dong W, Suraokar M, Ozburn NC, Moran CA et al. Nrf2 and Thereafter, the sections were washed three times in PBS. Next, peroxidase Keap1 abnormalities in non-small cell lung carcinoma and association with substrate reaction was performed with the AEC peroxidase substrate kit clinicopathologic features. Clin Cancer Res 2010; 16: 3743–3753. (DakoCytomation) according to the manufacturer’s instructions. Subsequently, 12 Hu Y, Ju Y, Lin D, Wang Z, Huang Y, Zhang S et al. Mutation of the Nrf2 gene in sections were washed in water, counterstained in 50% haemalaun (Merck) non-small cell lung cancer. Mol Biol Rep 2012; 39: 4743–4747. and mounted with glycerol-gelatin. The same protocol was performed for 13 Kim YR, Oh JE, Kim MS, Kang MR, Park SW, Han JY et al. Oncogenic NRF2 negative controls, either omitting the first antibody or using an isotype- mutations in squamous cell carcinomas of oesophagus and skin. J Pathol 2010; matched control antibody. Evaluation of Nrf2 expression was done 220: 446–451. by scoring the percental distribution (0 ¼ 0%, 1 ¼ 1–10%; 2 ¼ 10–50%; 14 Wang R, An J, Ji F, Jiao H, Sun H, Zhou D. Hypermethylation of the Keap1 gene in 3 ¼ 50–90% and 4 ¼ 490%) and expression intensity (0 ¼ none, 1 ¼ low; human lung cancer cell lines and lung cancer tissues. Biochem Biophys Res 2 ¼ medium and 3 ¼ high) in each section. The expression score was Commun 2008; 373: 151–154. calculated by multiplication of the intensity and distribution scores. 15 Eades G, Yang M, Yao Y, Zhang Y, Zhou Q. miR-200a regulates Nrf2 activation by targeting Keap1 mRNA in breast cancer cells. J Biol Chem 2011; 286: Statistics 40725–40733. The data represent the mean±standard deviation and were 16 Nioi P, Nguyen T. A mutation of Keap1 found in breast cancer impairs its ability to repress Nrf2 activity. Biochem Biophys Res Commun 2007; 362: 816–821. analyzed by Student’s t-test; P-values o0.05 were considered statistically significant. 17 Kinch L, Grishin NV, Brugarolas J. Succination of Keap1 and activation of Nrf2- dependent antioxidant pathways in FH-deficient papillary renal cell carcinoma type 2. Cancer Cell 2011; 20: 418–420. ABBREVIATIONS 18 Adam J, Hatipoglu E, O’Flaherty L, Ternette N, Sahgal N, Lockstone H et al. Renal ARE, antioxidant response element; LMB, leptomycin-B; Nrf2 & -1, cyst formation in Fh1-deficient mice is independent of the Hif/Phd pathway: nuclear factor E2-related factor 2 & -1; PARP1, poly(ADP-ribose)– roles for fumarate in KEAP1 succination and Nrf2 signaling. Cancer Cell 2011; 20: polymerase 1; PDAC, pancreatic ductal adenocarcinoma; SCID, 524–537. 19 Kim J, Cha YN, Surh YJ. A protective role of nuclear factor-erythroid 2-related severe combined immunodeficiency; Suc-LLVY-AMC, N-succinyl-L- factor-2 (Nrf2) in inflammatory disorders. Mutat Res 2010; 690: 12–23. leucyl-L-leucyl-L-valyl-L-tyrosyl-7-amido-4-methylcumarin; tBHQ, 20 Sebens S, Bauer I, Geismann C, Grage-Griebenow E, Ehlers S, Kruse ML et al. tertbutylhydroxyquinone; TRAIL, tumor necrosis factor-related Inflammatory macrophages induce NRF2 dependent proteasome activity in apoptosis-inducing ligand; trig, trigonelline colonic NCM460 cells and thereby confer anti-apoptotic protection. J Biol Chem 2011; 286: 40911–40921. 21 Singh S, Vrishni S, Singh BK, Rahman I, Kakkar P. Nrf2-ARE stress response CONFLICT OF INTEREST mechanism: a control point in oxidative stress-mediated dysfunctions and chronic The authors declare no conflict of interest. inflammatory diseases. Free Radic Res 2010; 44: 1267–1288. 22 Du ZX, Yan Y, Zhang HY, Liu BQ, Gao YY, Niu XF et al. Proteasome inhibition induces a p38 MAPK pathway-dependent antiapoptotic program via Nrf2 in ACKNOWLEDGEMENTS thyroid cancer cells. J Clin Endocrinol Metab 2011; 96: E763–E771. 23 Singh A, Bodas M, Wakabayashi N, Bunz F, Biswal S. Gain of Nrf2 function in non- Technical assistance by Frauke Grohmann, Dagmar Leisner and Iris Bauer is small-cell lung cancer cells confers radioresistance. Antioxid Redox Signal 2010; 13: acknowledged. The project was funded by the German Research Society DFG (SCHA 1627–1637. 677/9-1), the Sander Foundation (2010.076.1) and the German Cluster of Excellence 24 Shim GS, Manandhar S, Shin DH, Kim TH, Kwak MK. Acquisition of doxorubicin ‘Inflammation-at-Interfaces’. resistance in ovarian carcinoma cells accompanies activation of the NRF2 path- way. Free Radic Biol Med 2009; 47: 1619–1631. 25 Zhang DD. The Nrf2-Keap1-ARE signaling pathway: the regulation and dual REFERENCES function of Nrf2 in cancer. Antioxid Redox Signal 2010; 13: 1623–1626. 1 Arlt A, Bauer I, Schafmayer C, Tepel J, Muerkoster SS, Brosch M et al. Increased 26 Kim TH, Hur EG, Kang SJ, Kim JA, Thapa D, Lee YM et al. NRF2 blockade suppresses proteasome subunit protein expression and proteasome activity in colon cancer colon tumor angiogenesis by inhibiting hypoxia-induced activation of HIF-1a. relate to an enhanced activation of nuclear factor E2-related factor 2 (Nrf2). Cancer Res 2011; 71: 2260–2275. Oncogene 2009; 28: 3983–3996. 27 Chen L, Madura K. Increased proteasome activity, -conjugating enzymes, 2 Akhdar H, Loyer P, Rauch C, Corlu A, Guillouzo A, Morel F. Involvement of Nrf2 and eEF1A translation factor detected in breast cancer tissue. Cancer Res 2005; 65: activation in resistance to 5-fluorouracil in human colon cancer HT-29 cells. 5599–5606. Eur J Cancer 2009; 45: 2219–2227. 28 Hu XT, Chen W, Zhang FB, Shi QL, Hu JB, Geng SM et al. Depletion of the 3 DeNicola GM, Karreth FA, Humpton TJ, Gopinathan A, Wei C, Frese K et al. proteasome subunit PSMA7 inhibits colorectal cancer cell tumorigenicity and Oncogene-induced Nrf2 transcription promotes ROS detoxification and tumor- migration. Oncol Rep 2009; 22: 1247–1252. igenesis. Nature 2011; 475: 106–109. 29 Ren S, Smith MJ, Louro ID, McKie-Bell P, Bani MR, Wagner M et al. The p44S10 4 Hong YB, Kang HJ, Kwon SY, Kim HJ, Kwon KY, Cho CH et al. Nuclear factor locus, encoding a subunit of the proteasome regulatory particle, is amplified (erythroid-derived 2)-like 2 regulates drug resistance in pancreatic cancer cells. during progression of cutaneous malignant melanoma. Oncogene 2000; 19: Pancreas 2010; 39: 463–472. 1419–1427. 30 Rho JH, Qin S, Wang JY, Roehrl MH. Proteomic expression analysis of surgical 5 Jiang T, Chen N, Zhao F, Wang XJ, Kong B, Zheng W et al. High levels of Nrf2 human colorectal cancer tissues: up-regulation of PSB7, PRDX1, and SRP9 and determine chemoresistance in type II endometrial cancer. Cancer Res 2010; 70: hypoxic adaptation in cancer. J Proteome Res 2008; 7: 2959–2972. 5486–5496. 31 Kwak MK, Wakabayashi N, Itoh K, Motohashi H, Yamamoto M, Kensler TW. 6 Kim TH, Hur EG, Kang SJ, Kim JA, Thapa D, Lee YM et al. NRF2 blockade suppresses Modulation of gene expression by cancer chemopreventive dithiolethiones colon tumor angiogenesis by inhibiting hypoxia-induced activation of HIF-1alpha. through the Keap1-Nrf2 pathway. Identification of novel gene clusters for cell Cancer Res 2011; 71: 2260–2275. survival. J Biol Chem 2003; 278: 8135–8145. 7 Kim SK, Yang JW, Kim MR, Roh SH, Kim HG, Lee KY et al. Increased expression of 32 Kwak MK, Wakabayashi N, Greenlaw JL, Yamamoto M, Kensler TW. Antioxidants Nrf2/ARE-dependent anti-oxidant proteins in tamoxifen-resistant breast cancer enhance mammalian proteasome expression through the Keap1-Nrf2 signaling cells. Free Radic Biol Med 2008; 45: 537–546. pathway. Mol Cell Biol 2003; 23: 8786–8794. 8 Konstantinopoulos PA, Spentzos D, Fountzilas E, Francoeur N, Sanisetty S, 33 Kapeta S, Chondrogianni N, Gonos ES. Nuclear erythroid factor 2-mediated pro- Grammatikos AP et al. Keap1 mutations and Nrf2 pathway activation in epithelial teasome activation delays senescence in human fibroblasts. J Biol Chem 2010; ovarian cancer. Cancer Res 2011; 71: 5081–5089. 285: 8171–8184. 9 Lister A, Nedjadi T, Kitteringham NR, Campbell F, Costello E, Lloyd B et al. Nrf2 is 34 Schaedler S, Krause J, Himmelsbach K, Carvajal-Yepes M, Lieder F, Klingel K et al. overexpressed in pancreatic cancer: implications for cell proliferation and therapy. Hepatitis B virus induces expression of antioxidant response element-regulated Mol Cancer 2011; 10:37. genes by activation of Nrf2. J Biol Chem 2010; 285: 41074–41086.

Oncogene (2013) 4825 – 4835 & 2013 Macmillan Publishers Limited Nrf2 inhibition and pancreatic cancer A Arlt et al 4835 35 Kwak MK, Kensler TW. Induction of 26S proteasome subunit PSMB5 by the 49 Fritsche P, Seidler B, Schuler S, Schnieke A, Gottlicher M, Schmid RM et al. HDAC2 bifunctional inducer 3-methylcholanthrene through the Nrf2-ARE, but not the mediates therapeutic resistance of pancreatic cancer cells via the BH3-only pro- AhR/Arnt-XRE, pathway. Biochem Biophys Res Commun 2006; 345: 1350–1357. tein NOXA. Gut 2009; 58: 1399–1409. 36 Kwak MK, Itoh K, Yamamoto M, Kensler TW. Enhanced expression of the 50 Stathis A, Moore MJ. Advanced pancreatic carcinoma: current treatment and transcription factor Nrf2 by cancer chemopreventive agents: role of antioxidant future challenges. Nat Rev 2010; 7: 163–172. response element-like sequences in the nrf2 promoter. Mol Cell Biol 2002; 22: 51 Elangovan M, Oh C, Sukumaran L, Wojcik C, Yoo YJ. Functional differences 2883–2892. between two major ubiquitin receptors in the proteasome; S5a and hRpn13. 37 Lau A, Villeneuve NF, Sun Z, Wong PK, Zhang DD. Dual roles of Nrf2 in cancer. Biochem Biophys Res Commun 2010; 396: 425–428. Pharmacol Res 2008; 58: 262–270. 52 Nguyen T, Sherratt PJ, Nioi P, Yang CS, Pickett CB. Nrf2 controls constitutive 38 Reuter S, Gupta SC, Chaturvedi MM, Aggarwal BB. Oxidative stress, inflammation, and inducible expression of ARE-driven genes through a dynamic pathway and cancer: how are they linked? Free Radic Biol Med 2010; 49: 1603–1616. involving nucleocytoplasmic shuttling by Keap1. J Biol Chem 2005; 280: 39 Taguchi A, Politi K, Pitteri SJ, Lockwood WW, Faca VM, Kelly-Spratt K et al. Lung 32485–32492. cancer signatures in plasma based on proteome profiling of mouse tumor 53 Niture SK, Jaiswal AK. Prothymosin-a mediates nuclear import of the INrf2/ models. Cancer Cell 2011; 20: 289–299. Cul3_Rbx1 complex to degrade nuclear Nrf2. J Biol Chem 2009; 284: 40 Hayes JD, McMahon M. NRF2 and KEAP1 mutations: permanent activation of an 13856–13868. adaptive response in cancer. Trends Biochem Sci 2009; 34: 176–188. 54 Radhakrishnan SK, Lee CS, Young P, Beskow A, Chan JY, Deshaies RJ. Transcription 41 Osburn WO, Kensler TW. Nrf2 signaling: an adaptive response pathway for factor Nrf1 mediates the proteasome recovery pathway after proteasome inhi- protection against environmental toxic insults. Mutat Res 2008; 659: 31–39. bition in mammalian cells. Mol Cell 2010; 38: 17–28. 42 Wang XJ, Sun Z, Villeneuve NF, Zhang S, Zhao F, Li Y et al. Nrf2 enhances 55 Xie Y. Feedback regulation of proteasome gene expression and its implications in resistance of cancer cells to chemotherapeutic drugs, the dark side of Nrf2. cancer therapy. Cancer Metastasis Rev 2010; 29: 687–693. Carcinogenesis 2008; 29: 1235–1243. 56 Nguyen T, Nioi P, Pickett CB. The Nrf2-antioxidant response element 43 Wang XJ, Hayes JD, Henderson CJ, Wolf CR. Identification of retinoic acid as an signaling pathway and its activation by oxidative stress. J Biol Chem 2009; 284: inhibitor of transcription factor Nrf2 through activation of retinoic acid receptor 13291–13295. alpha. Proc Natl Acad Sci USA 2007; 104: 19589–19594. 57 van Dijk AE, Olthof MR, Meeuse JC, Seebus E, Heine RJ, van Dam RM. Acute effects 44 Zhou W, Lo SC, Liu JH, Hannink M, Lubahn DB. ERRbeta: a potent inhibitor of Nrf2 of decaffeinated coffee and the major coffee components chlorogenic acid and transcriptional activity. Mol Cell Endocrinol 2007; 278: 52–62. trigonelline on glucose tolerance. Diabetes Care 2009; 32: 1023–1025. 45 Boettler U, Sommerfeld K, Volz N, Pahlke G, Teller N, Somoza V et al. Coffee 58 Olthof MR, van Dijk AE, Deacon CF, Heine RJ, van Dam RM. Acute effects of constituents as modulators of Nrf2 nuclear translocation and ARE (EpRE)- decaffeinated coffee and the major coffee components chlorogenic acid and dependent gene expression. J Nutr Biochem 2011; 22: 426–440. trigonelline on incretin hormones. Nutr Metab (Lond) 2011; 8:10. 46 Kannan S, Jaiswal AK. Low and high dose UVB regulation of transcription factor 59 Geismann C, Arlt A, Bauer I, Pfeifer M, Schirmer U, Altevogt P et al. Binding of the NF-E2-related factor 2. Cancer Res 2006; 66: 8421–8429. transcription factor slug to the L1CAM promoter is essential for transforming 47 Steffen J, Seeger M, Koch A, Kruger E. Proteasomal degradation is transcriptionally growth factor-beta1 (TGF-beta)-induced L1CAM expression in human pancreatic controlled by TCF11 via an ERAD-dependent feedback loop. Mol Cell 2010; 40: ductal adenocarcinoma cells. Int J Oncol 2011; 38: 257–266. 147–158. 60 Mu¨ erko¨ster S, Arlt A, Sipos B, Witt M, Gromann M, Klo¨ppel G et al. Increased 48 Arlt A, Vorndamm J, Breitenbroich M, Folsch UR, Kalthoff H, Schmidt WE et al. expression of the E3-ubiquitin ligase receptor subunit bTRCP1 relates to con- Inhibition of NF-kappaB sensitizes human pancreatic carcinoma cells to apoptosis stitutive NF-kB activation and chemoresistance in pancreatic carcinoma cells. induced by etoposide (VP16) or doxorubicin. Oncogene 2001; 20: 859–868. Cancer Res 2005; 65: 1316–1324.

Supplementary Information accompanies the paper on the Oncogene website (http://www.nature.com/onc)

& 2013 Macmillan Publishers Limited Oncogene (2013) 4825 – 4835