The MUC4 Mucin Mediates Gemcitabine Resistance of Human Pancreatic Cancer Cells Via the Concentrative Nucleoside Transporter Family

Home , Mucin

ORIGINAL ARTICLE The MUC4 mucin mediates gemcitabine resistance of human pancreatic cancer cells via the Concentrative Nucleoside Transporter family

N Skrypek1,2,3, B Ducheˆne1,2,3, M Hebbar1,2,4, E Leteurtre1,2,5, I van Seuningen1,2,3 and N Jonckheere1,2,3

The fluorinated analog of deoxycytidine, Gemcitabine (Gemzar), is the main chemotherapeutic drug in pancreatic cancer, but survival remains weak mainly because of the high resistance of tumors to the drug. Recent works have shown that the mucin MUC4 may confer an advantage to pancreatic tumor cells by modifying their susceptibility to drugs. However, the cellular mechanism(s) responsible for this MUC4-mediated resistance is unknown. The aim of this work was to identify the cellular mechanisms responsible for gemcitabine resistance linked to MUC4 expression. CAPAN-2 and CAPAN-1 adenocarcinomatous pancreatic cancer (PC) cell lines were used to establish stable MUC4-deficient clones (MUC4-KD) by shRNA interference. Measurement of the IC50 index using tetrazolium salt test indicated that MUC4-deficient cells were more sensitive to gemcitabine. This was correlated with increased Bax/BclXL ratio and apoptotic cell number. Expression of Equilibrative/Concentrative Nucleoside Transporter (hENT1, hCNT1/3), deoxycytidine kinase (dCK), ribonucleotide reductase (RRM1/2) and Multidrug-Resistance Protein (MRP3/4/5) was evaluated by quantitative RT–PCR (qRT–PCR) and western blotting. Alteration of MRP3, MRP4, hCNT1 and hCNT3 expression was observed in MUC4-KD cells, but only hCNT1 alteration was correlated to MUC4 expression and sensitivity to gemcitabine. Decreased activation of MAPK, JNK and NF-kB pathways was observed in MUC4-deficient cells, in which the NF-kB pathway was found to have an important role in both sensitivity to gemcitabine and hCNT1 regulation. Finally, and in accordance with our in vitro data, we found that MUC4 expression was conversely correlated to that of hCNT1 in tissues from patients with pancreatic adenocarcinoma. This work describes a new mechanism of PC cell resistance to gemcitabine, in which the MUC4 mucin negatively regulates the hCNT1 transporter expression via the NF-kB pathway. Altogether, these data point out to MUC4 and hCNT1 as potential targets to ameliorate the response of pancreatic tumors to gemcitabine treatment.

Oncogene (2013) 32, 1714–1723; doi:10.1038/onc.2012.179; published online 14 May 2012 Keywords: MUC4; mucin; pancreatic cancer; gemcitabine; hCNT1; NF-kB

INTRODUCTION Nucleoside Transporter 1 and 3 (hCNT1 and hCNT3).4 It has been Pancreatic adenocarcinoma is one of the most deadly cancers in shown that the expression of hENT1 and hCNT3 in pancreatic western countries with an extremely poor prognosis (survival rate tumors is correlated with chemosensitivity and overall survival, 5–7 of 6 months).1 This marked outcome is related to a lack of efficient making them good predictive markers for patient survival. therapeutic tools and early diagnostic markers. At the time of Moreover, hCNT1 was shown to be frequently decreased in diagnosis, more than 80% of patients have metastasis or locally pancreatic tumors and, in most of the PC cell lines, this decrease advanced cancer. Only about 10 to 15% of patients are considered being correlated with gemcitabine cytotoxicity, suggesting eligible for surgical resection. Gemcitabine (Gemzar), a fluorinated the importance of hCNT1 in improving the sensitivity to 8,9 analog of deoxycytidine, is the main chemotherapeutic drug used gemcitabine. in firstline in advanced pancreatic cancer (PC). Despite the On the basis of expression and molecular studies, mucins, improvement of quality of life of patients, the gain in survival especially the MUC4 mucin, have been proposed as factors of 10–12 remains short (6 additional months). This is mainly due to the high chemoresistance. MUC4 is a membrane-bound mucin, which resistance of pancreatic tumor cells to the drug.2,3 Deciphering is not expressed in healthy pancreas but is expressed in the very 13 mechanisms responsible for PC cell resistance to gemcitabine is early steps of pancreatic carcinogenesis. MUC4 involvement in 14–18 thus mandatory if one wants to improve efficacy of the drug and biological properties of PC cells is well described. However, propose more efficient therapies. the mechanisms linking MUC4 expression to gemcitabine One way to explain modifications of cell sensitivity to sensitivity of PC cells remain to be determined. gemcitabine is an alteration of the factors responsible for its The aim of this work was to decipher the molecular mechan- metabolism and more particularly nucleoside transporters. isms linking MUC4 expression and gemcitabine resistance of PC Gemcitabine is taken up by the cell mainly by human Equilibrative cells. For that, we hypothesized that MUC4 could alter expression Nucleoside Transporter 1 (hENT1) and by human Concentrative of factors of gemcitabine metabolism and/or detoxifying channels.

1Inserm, UMR837, Jean-Pierre Aubert Research Center, Lille Cedex, France; 2Universite´ Lille Nord de France, Lille Cedex, France; 3Centre Hospitalier Re´gional et Universitaire de Lille, Lille Cedex, France; 4Centre Hospitalier Re´gional et Universitaire de Lille, Department of Medical Oncology, Lille, France and 5Centre Hospitalier Re´gional et Universitaire de Lille, Centre de Biologie-Pathologie, Department of Pathology, Lille, France. Correspondence: Dr N Jonckheere, Inserm UMR837/JPARC, rue Polonovski, 59045 Lille Cedex, France. E-mail: [email protected] Received 24 December 2011; revised 12 March 2012; accepted 6 April 2012; published online 14 May 2012 Gemcitabine resistance linked to MUC4 expression N Skrypek et al 1715 We show in this paper that the loss of MUC4 in PC cells induces indicate that MUC4 involvement in gemcitabine sensitivity implies sensitivity to gemcitabine and involves MUC4 regulation of hCNT1 alteration of the apoptotic balance in PC cells. expression via the NF-kB pathway.

Alteration of gemcitabine metabolism marker expression in RESULTS MUC4-KD cells MUC4-deficient cells are more sensitive to gemcitabine Gemcitabine efficiency depends on the state of nucleoside Sensitivity of MUC4-deficient cells (MUC4-KD) to gemcitabine was metabolism for its activation and incorporation into DNA. studied in CAPAN-2 and CAPAN-1 PC cells. Measurement of IC50 Modification of nucleoside transporters (hENT1, hCNT1 and 3) shows that the loss of MUC4 leads to an increased sensitivity or deoxycytidine kinase (dCK)19,20 expression contributes to of CAPAN-2 (MUC4-KD ¼ 34.8±1.8 nM vs Mock ¼ 52.9±2.2 nM) gemcitabine efficiency. On the contrary, multidrug resistance- 21 (Figure 1a) and CAPAN-1 (MUC4-KD ¼ 80±2nM vs Mock ¼ associated protein family (MRP) channels and ribonucleotide 22 145±15 nM) (Supplementary Figure 1A) cells to gemcitabine. A reductase (RRM1 and 2) contribute to maintaining a drug- similar effect was observed at a longer time point (6 days) in both resistant phenotype. As shown in Figure 2a, CAPAN-2 MUC4-KD CAPAN-1 and CAPAN-2 cells (data not shown). Sensitivity of cells strongly overexpressed mRNA of MRP3 channel (ninefold), MUC4-KD cells to another cytidine analog, the cytarabine/aracytin hCNT1 (sevenfold) and hCNT3 (twofold) transporters compared ARA-C, was also evaluated after 72 h of treatment. Interestingly, with Mock cells, whereas MRP4 channel mRNA expression was we also observed an increased sensitivity of MUC4-KD cells decreased (twofold). Increased expression of hCNT1, hCNT3 and (IC50 ¼ 1.3 mM±0.3) when compared with Mock cells (3.2 mM±0.6). MRP3, and decreased expression of MRP4, was confirmed at the Cell sensitivity to the alkylating agent oxaliplatin, on the other protein level by western blotting (Figure 2b). Implication of these hand, was not affected by the lack of MUC4 (MUC4- four markers in cell sensitivity to the drug was then tested by KD ¼ 1.35±0.06 mM vs Mock ¼ 1.15±0.05 mM) (data not shown). transient siRNA approach (Supplemental Figure 2A). Decreased The increase of gemcitabine-induced cytotoxicity in MUC4-KD expression of MRP3 did not alter cytotoxicity to gemcitabine cells was accompanied by an increased expression of the pro- (Figure 2c), whereas hCNT1, hCNT3 and MRP4 repression led to a apoptotic marker Bax and a decreased expression of the anti- statistically significant (Po0.0001, P ¼ 0.0372, P ¼ 0.0007, respec- apoptotic marker BclXL, leading to an increase of the Bax/BclXL tively) increase in survival (Figure 2c). These results suggest an ratio, suggesting a higher susceptibility to apoptosis in both cell implication of hCNT1, hCNT3 and MRP4 in gemcitabine sensitivity, lines (Figure 1b and Supplementary Figure 1A). Moreover, the with hCNT1 and hCNT3 effect correlated to both MUC4 expression activation of the apoptotic mediator p53 was increased in MUC4- and MUC4-KD cell sensitivity. In CAPAN-1 MUC4-KD cells, KD cells (Figure 1b). To confirm these observations, the apoptotic increased expression of hCNT1 was also found, whereas hCNT3 index was determined after gemcitabine treatment by annexin-V expression remained the same when compared with Mock cells staining followed by cytometry analysis. MUC4-KD clones, as (Supplementary Figure 1B). Therefore, in the rest of the paper, we expected, have a statistically significant higher apoptotic index focused our studies on hCNT1 transporter, for which biological than Mock cells (P ¼ 0.0025) (Figure 1c). Altogether, these results effects were correlated to MUC4 expression in the two cell lines.

Mock MUC4-KD IC50 100 Gemcitabine -++ - 90 Mock 52.9 ± 2.2nM MUC4 80 MUC4-KD 34.8 ± 1.8nM 70 Bax 60 50 BclXL

40 Phospho- Survival rate (%) 30 p53 20 p53 10 0 β 10-8 10-7.75 10-7.5 10-7.25 10-7 10-6.75 10-6.5 10-6.25 -actin Gemcitabine concentration (M) 3 5 4 XL XL 2 40 n.s ** 3 1 2 Bax/Bcl 30 Bax/Bcl 1 0 0 Mock MUC4 Mock MUC4 20 -KD -KD 3 3 10 2 Apoptotic cells (%) 2 1 1 0 MockMUC4 Mock MUC4 0 0 Phospho p53/p53 KD KD Phospho p53/p53 Mock MUC4 Mock MUC4 -KD -KD - gemcitabine + gemcitabine - gemcitabine + gemcitabine

Figure 1. Sensitivity of the CAPAN-2 MUC4-KD cells to gemcitabine is correlated with apoptosis. (a)IC50 rates were measured after 72 h of gemcitabine treatment. (b) Western blots were performed to analyze the expression of MUC4, Bax, BclXL, Phospho-p53, p53 and b-actin in CAPAN-2 Mock and MUC4-KD cells. The density of each marker was measured and Bax/BclXL ratio was determined and represented as histograms. Expression in Mock cells was arbitrarily set to 1. (c) Apoptosis was studied by annexin-V/propidium iodide staining. All experiments were performed three times independently.

& 2013 Macmillan Publishers Limited Oncogene (2013) 1714 – 1723 Gemcitabine resistance linked to MUC4 expression N Skrypek et al 1716 11 ** 10 MUC4 Mock -KD 9 *** 2.5 8 hCNT1 7 2 6 hCNT3 1.5 5 * 1 3 MRP3 Expression ratio MUC4-KD/Mock 0.5 (MUC4-KD / Mock ) 2 mRNA relative expression *** MRP4 0 1

MRP3 MRP4 0 hCNT1 hCNT3 β-actin

dCK MRP3 MRP4 MRP5 RRM1 RRM2 hENT1 hCNT1 hCNT3

120 *** *** 110

100

90 * n.s 80

60 Survival rate (%) 50

30 siControlsihCNT1sihCNT3 siMRP4 siMRP3 Figure 2. Expression of gemcitabine transporters and channels in CAPAN-2 MUC4-KD cells. (a) mRNA expression of MRP3, MRP4, MRP5, hENT1, hCNT1, hCNT3, RRM1, RRM2 and dCK was analyzed in Mock and MUC4-KD cells by qRT–PCR with speciﬁc primers shown in Table 1. The histogram represents the ratio of their expression in MUC4-KD compared with Mock cells. (b) hCNT1, hCNT3, MRP3, MRP4 and b-actin expression by western blotting. Bands were quantiﬁed by densitometry and shown in the histogram. (c) CAPAN-2 cells were transfected with Control (non-targeting), hCNT1, hCNT3, MRP4 or MRP3 siRNA, and then treated with gemcitabine. Survival rate was measured after 24 h of gemcitabine treatment (100 nM) using the MTT assay. Four independent experiments were performed.

MAPK, JNK and NF-kB pathways are altered in MUC4-KD cells and of MAPK (U0126), JNK (SP600125) and NF-kB (BAY-11-7082) modulate sensitivity to gemcitabine pathways (Figure 5a). The results indicate that inhibition of MAPK To understand the molecular mechanisms that regulate hCNT1 leads to decreased hCNT1 mRNA expression (twofold), whereas expression, we first studied alteration of the main signaling treatment with NF-kB inhibitor induced increase of hCNT1 mRNA pathways in MUC4-KD cells. We show that the loss of MUC4 in level (1.5-fold). No modification of hCNT1 mRNA level was both CAPAN-2 and CAPAN-1 cells induces a decrease of both Erk1/ observed following treatment with JNK pathway inhibitor. 2 and JNK activation, and a decrease of NF-kB p65 expression To confirm the implication of the NF-kB pathway on hCNT1 (Figure 3a and Supplementary Figure 3A). The translocation to the regulation, the inhibitor of kappa B (IkB) was overexpressed in nucleus of NF-kB p50 and p65 subunits was also decreased in both CAPAN-2 and CAPAN-1 cells and hCNT1 mRNA level was MUC4-KD cells, as the nuclear amount of the two subunits was measured (Figure 5b and Supplementary Figure 3C). Decreased diminished in MUC4-KD cells compared with Mock cells activity of NF-kB using the synthetic kB-Luc promoter confirmed (Figure 3b). This was correlated with a decreased activity of the the efficiency of IkB. Similar to pharmacological inhibition, kB-Luc synthetic promoter (Figure 3c and Supplementary overexpression of IkB increased hCNT1 mRNA expression Figure 3B). Moreover, pharmacological inhibition of the MAPK (Figure 5b). Increased hCNT1 expression at the protein level was (U0126) and JNK (SP600125) pathways induced a strong decrease then confirmed after either IkB overexpression or transfection of of NF-kB expression (Figure 3d) and promoter activity (Figure 3e), NF-kB p105 siRNA (Figure 5c). Taken together, these results clearly suggesting a link between these pathways. show that the NF-kB pathway negatively regulates hCNT1 Implication of these three pathways in gemcitabine sensitivity expression. was then investigated using specific siRNAs to target Erk1/2 and JNK kinases, and NF-kB p105 (precursor of NF-kB p50 subunit) (Supplementary Figure 2B). The results indicate that decreased MUC4 expression is conversely correlated to hCNT1 in tissues from NF-kB expression leads to increased sensitivity to gemcitabine patients with pancreatic adenocarcinoma (Figure 4a), correlated to an increased Bax/BclXL ratio (Figure 4b). Having shown in vitro that MUC4 expression is correlated with a On the contrary, decreased Erk1/2 and JNK expression led to a decreased expression of hCNT1 in PC cells, we undertook to study decreased sensitivity to gemcitabine (Figure 4a), correlated to a their expression on a small population of tissues from 16 patients decreased Bax/BclXL ratio (Figure 4b). These results suggest that with pancreatic adenocarcinoma. Immunohistochemical stainings targeting the NF-kB pathway appears to be more appropriate if show that MUC4 was expressed in 10/16 PC samples and hCNT1 one wants to alter the cell sensitivity to gemcitabine. expressed in 9/16 PC tumors (Figure 6a). In MUC4-positive tumors, 7/10 patients did not express hCNT1. Interestingly, 100% of the samples not expressing MUC4 were positive for hCNT1 (6/6), The NF-kB pathway regulates hCNT1 expression suggesting a converse expression of MUC4 and hCNT1. Moreover, The hCNT1 mRNA expression was measured by qRT–PCR after Fisher’s exact test indicated that proportions differed significantly CAPAN-2 cell treatment with specific pharmacological inhibitors (P ¼ 0.0114) (Figure 6b).

Oncogene (2013) 1714 – 1723 & 2013 Macmillan Publishers Limited Gemcitabine resistance linked to MUC4 expression N Skrypek et al 1717 MUC4 Mock -KD

Phospho-Erk1/2

Erk1/2 Cytosolic Nuclear Phospho-JNK MUC4 MUC4 Mock-KD Mock -KD JNK NF-κB p50

NF-κB p65 NF-κB p65 NF-κB p50

Sp1 β-actin

1.5 1.5 Tubulin

1 1 -actin -actin 1.5 1.5 β β 0.5 0.5 p50/ p65/ 1 1 0 0 Mock MUC4 Mock MUC4 0.5 0.5 -KD -KD p65 expression p50 expression (Nuclear/Cytosolic) (Nuclear/Cytosolic) 0 0 1.5 1.5 Mock MUC4 Mock MUC4 -KD -KD 1 1

0.5 0.5 Phospho Erk/Erk

0 Phospho JNK/JNK 0 Mock MUC4 Mock MUC4 -KD -KD

*** *** 100 100

80 DMSO U0126 SP600125 80

Phospho- 60 60 NF-κB p65

40 40 κ NF- B p65 luciferase activity (%) 20 20 Luc B- B-Luc luciferase activity (%) κ κ 0 β-actin 0 Mock MUC4 -KD DMSO U0126 SP600125 Figure 3. Expression and activity of the MAPK, JNK and NF-kB pathways in CAPAN-2 MUC4-KD cells. (a) Expression and activation of MAPK, JNK, NF-kB p65 and p50 subunits were analyzed by western blotting on total extracts from Mock or MUC4-KD cells. (b) Expression of NF-kB p65 and p50 subunits in cytosolic and nuclear extracts was carried out by western blotting. Density of each western blot was quantiﬁed and nuclear to cytosolic ratio represented as histograms where Mock was arbitrarily set to 1. (c) Luciferase activity of the kB-Luc synthetic promoter was measured 48 h after transfection. Luciferase activity in Mock cells was set as 100%. (d) Western blotting of phospho-p65 and constitutive NF-kB p65 subunit of CAPAN-2 cell extracts from cells treated with pharmacological inhibitors of MAPK (U0126, 10 mM) or JNK (SP600125, 10 mM) pathways during 24 h. (e) Relative luciferase activity in cell extracts from CAPAN-2 cells transfected with the kB-Luc promoter followed by treatment with U0126 and SP600125 pharmacological inhibitors. Dimethylsulphoxide (DMSO, control condition) is arbitrarily set to 100%.

DISCUSSION tumors23 and as such has been proposed as a new biomarker. A major and marked characteristic of pancreatic adenocarcinoma Several in vitro and in vivo studies have shown that MUC4 is the high resistance to chemotherapy leading to low efﬁciency of modulates tumor growth, invasion, adherence and therapies, in particular those based on gemcitabine. The under- apoptosis.14,16–18 Very recently, we have also shown that the PC standing of mechanisms underlying chemoresistance is thus cells lacking MUC4 were less proliferative with decreased cyclin D1 mandatory if one wants to identify new predictive and/or expression and G1 cell cycle arrest.15 Altogether, these alterations prognostic markers and develop new therapeutic approaches. In could potentially confer chemoresistance capacity to tumor cells. this study, we show that the MUC4 mucin is involved in PC cell On the basis of these observations, Bafna et al.10 have shown chemoresistance to gemcitabine via two complementary mechan- that MUC4 expression in CD18/HPAF PC cell line decreases Bad isms involving apoptosis repression for one and nucleoside pro-apoptotic activity by its phosphorylation via Erk1/2 kinase, transporter hCNT1 inhibition via the NF-kB pathway for the other. thereby decreasing gemcitabine cytotoxicity. In this report, we Neo-expression of MUC4 in the early steps of pancreatic show that the loss of MUC4 is sufﬁcient to modify Bax carcinogenesis suggests that this membrane-bound mucin is (pro-apoptotic) and BclXL (anti-apoptotic) expression, leading to implicated in the acquisition of an aggressive phenotype by an increase of Bax/BclXL ratio, indicating an increase in susceptibility

& 2013 Macmillan Publishers Limited Oncogene (2013) 1714 – 1723 Gemcitabine resistance linked to MUC4 expression N Skrypek et al 1718

κB siControl siNF- siERK1/2 siJNK

Bax 100 ns **

90 BclXL

80 β–actin 70

60 2

50 *** 1.5 Survival rate (%)

40 XL 1 30 Bax/Bcl 0.5 20 siControlsiNF-κB siERK1/2 siJNK 0 κB siJNK siNF- siControl siERK1/2 Figure 4. Inhibition of the NF-kB pathway sensitizes CAPAN-2 cells to gemcitabine. (a) Cell survival rate was measured following Erk1/2, JNK and NF-kB inhibition in CAPAN-2 cells using speciﬁc siRNA during 48 h before gemcitabine treatment (100 nM). Sensitivity to gemcitabine was compared with a non-targeting siRNA (siControl). (b) Apoptotic markers Bax and BclXL expression was measured by western blotting. The density of each band was measured and Bax/BclXL ratio calculated, where siControl ratio was arbitrarily set to 1.

2.5 2.5

2 100 2 * 80 1.5 ** 1.5 60

1 1 40 luciferase activity (%) 0.5

(Treated/Untreated cells) 0.5 Luc 20 B- κ hCNT1 mRNA relative expression 0 0 hCNT1 mRNA relative expression 0 + IκB + IκB

U0126

SP600125 BAY 11-7082

κ siControl siNF-κB + I B 2 3 1.5 -actin hCNT1 hCNT1 -actin β 2 β 1 β 1 0.5 -actin β-actin hCNT1/ 0 hCNT1/ 0 + IκB κB

siControl siNF- Figure 5. Identiﬁcation of the NF-kB pathway as a regulator of hCNT1 expression. (a) CAPAN-2 cells were treated for 24 h with pharmacological inhibitor of MAPK (U0126, 10 mM), JNK (SP600125, 10 mM) or NF-kB (BAY-11-7082, 10 mM) pathways. hCNT1 mRNA expression level was measured by qRT–PCR. The histogram represents the ratio between treated/untreated cells against the inhibitor. The results are means that represent three separate experiments in triplicate for each inhibitor. (b) Overexpression of IkB in CAPAN-2 cells was performed by co-transfecting pCR3- IkB or pCR3 empty vector for 48 h in the presence of the kB-Luc synthetic promoter. RNA or protein extraction was then carried out. Data are represented as histograms. (c) Western blot of hCNT1 was performed after inhibition of the NF-kB pathway with a speciﬁc siRNA or by IkB overexpression with pCR3-IkB in CAPAN-2 cells.

to apoptosis. The importance of the Bax/BclXL ratio is attested by (Figure 7). Our results are thus in favor of a role for MUC4 as an the fact that it is associated in prostate,24 breast25 and pancreatic apoptosis regulator that would inﬂuence PC cell resistance to cancers26 with chemotherapeutic sensitivity and has been gemcitabine. proposed as a predictive marker. In our cellular models, the However, gemcitabine cytotoxic effect cannot be optimal in the regulation of Bax and BclXL by MUC4 seems in part to be mediated absence of an efﬁcient metabolism pathway to take up the drug. by the NF-kB pathway, and not by Erk1/2 and JNK pathways This is especially true for resistant tumor cells that often exhibit an

Oncogene (2013) 1714 – 1723 & 2013 Macmillan Publishers Limited Gemcitabine resistance linked to MUC4 expression N Skrypek et al 1719

Figure 6. MUC4 and hCNT1 expression is conversely correlated in tissues from patients with pancreatic adenocarcinoma. (a) MUC4 and hCNT1 expression was analyzed by immunohistochemistry. MUC4 þ indicates a tissue sample positive for MUC4 and negative for hCNT1. MUC4 À indicates a tissue sample that does not express MUC4 but is positive for hCNT1. Histological sections were stained with Hematoxylin Eosin (HE). Magnification ( Â 20). (b) MUC4 expression is correlated with a lack of expression of hCNT1. Fisher’s exact test indicates that proportions differ significantly (*P ¼ 0.0114). altered metabolism of gemcitabine. In this study, the absence of are conversely correlated in PC cells, as well as in tumor samples MUC4 in PC cells led to a marked modification in hCNT1 and from patients with pancreatic adenocarcinoma. These data confer hCNT3 transporters, as well as MRP3 and MRP4 drug-detoxifying to MUC4 a potential as a predictive marker and will have to be channels expression levels. We also show that hCNT1, hCNT3 and confirmed in bigger cohorts. MRP4 are involved in gemcitabine sensitivity. MRP channels are Investigation of hCNT1 regulation by MUC4 led us to show that able to export out of the cell a wide range of lipophilic anions such the NF-kB pathway mediates this mechanism. The NF-kB pathway as chemotherapeutic drugs to maintain a drug-resistant pheno- is well known to be constitutively activated in PC compared with type. Overexpression of MRP3 was found in several cancers, normal pancreas.34 NF-kB also promotes tumor progression by such as ovarian,27 lung28 and pancreatic cancers.29 MRP3 over- regulating genes implicated in proliferation, angiogenesis and expression was correlated to cell chemoresistance to several survival. Implication of the NF-kB pathway in chemoresistance to drugs, such as methotrexate, etoposide, doxorubicin and platinum gemcitabine was already described and was found to modulate agents. Nevertheless, in these studies MRP3 was not implicated in apoptosis.35 In our work, in addition to apoptotic regulation, we nucleoside-derived drug resistance. In our work, MRP3 was also highlight a new resistance mechanism mediated by the NF-kB not an actor of drug resistance as opposed to MRP4. Similar to pathway that leads to the downregulation of hCNT1 thereby MRP5 and MRP8, MRP4 is able to transport cyclic nucleotides decreasing gemcitabine uptake and efficiency (Figure 7). Our out of the cell and was previously shown to confer cell results support the promising current effort to improve gemcita- chemoresistance to 6-mercaptopurine and 6-thioguanine in bine efficacy by targeting the NF-kB pathway with natural36,37 or leukemia,30 as well as to methotrexate.31 Although the implica- chemical38 inhibitors or by siRNA delivery.39 tion of MRP5 in 5-fluorouracil and gemcitabine resistance was We have described in this paper two complementary resistance demonstrated,32 the role of MRP4 remains unclear. In this work, mechanisms of PC cells to gemcitabine that are linked to MUC4. we show that MRP4 expression is conversely correlated to They are mediated by the NF-kB pathway and lead to (i) the gemcitabine resistance, making this channel a potential interest- modulation of apoptosis and (ii) the regulation of hCNT1 ing marker that is independent of MUC4 expression (Figure 7). A transporter expression. These data reinforce the value of MUC4 similar result has been previously described in which resistance to as a prognostic marker. Interestingly, cell sensitivity to cytarabine/ several drugs in melanoma cells by Muc4 expression was aracytin (ARA-C), another cytidine analog, mainly used in leukemia conversely correlated to MDR1 and MRP1, which are members or lymphoma, but that is incorporated into human DNA via a of multidrug resistance family.33 We thus propose that MRP4 acts similar metabolic pathway as in gemcitabine, was also altered in independently of MUC4 to convey gemcitabine sensitivity. MUC4-KD cells. This suggests that our findings regarding the role Interestingly, hCNT1 is upregulated when MUC4 is inhibited, of MUC4 in drug resistance could be extended to other nucleoside and this is correlated to drug sensitivity. The same result was analogs and other types of cancer. FOLFIRINOX, a new therapeutic found for hCNT3. This correlation was observed in both CAPAN-1 protocol, which combines therapeutic agents (5-fluorouracil, and CAPAN-2 cells for hCNT1, but only in CAPAN-2 for hCNT3. This irinotecan, oxaliplatin and leucovorin), brings hope as statistical suggests a cell-specific mechanism for hCNT3. Although in vitro and clinical significant benefit compared with gemcitabine alone studies have correlated hCNT1 expression with a drug-resistant was reported in patients with advanced pancreatic cancer.40 phenotype of PC cells,8,9 no significant evidence has proven in vivo Interestingly, 5-fluorouracil was shown to induce MRP3 expression implication with a correlation to a better patient outcome. Our increasing oxaliplatin efficiency,41 whereas MRP4 expression in vitro and ex vivo data indicate that MUC4 and hCNT1 expression increases irinotecan resistance.42 In preliminary experiments, we

& 2013 Macmillan Publishers Limited Oncogene (2013) 1714 – 1723 Gemcitabine resistance linked to MUC4 expression N Skrypek et al 1720

MUC4

Gemcitabine

hCNT1/3 hENT1

P Gemcitabine Erk1/2

NF-κB P JNK P50 P65 dCK

Bax BclXL P MRP4 P P Gemcitabine MRP3

hCNT1 MRP5

DNA Apoptosis dNTP NTP Nucleus

RRM1/2 Decrease of cell sensitivity to gemcitabine

MUC4-dependent MUC4-independent

Figure 7. Schematic representation of the proposed mechanisms of PC cell chemosensitivity to gemcitabine. Left panel: MUC4-dependent mechanisms involving either decrease of hCNT1 nucleoside transporter expression via the NF-kB pathway or modiﬁcation of the Bax/BclXL ratio (upregulation of BclXL and downregulation of Bax expression). Right panel: MUC4-independent mechanisms. MRP4 channel increases cell sensitivity to gemcitabine by an unknown mechanism, whereas MRP3 is not involved. MRP5 is known to decrease cell sensitivity to gemcitabine, as well as modiﬁcation of deoxycytidine kinase (dCK) and ribonucleotide reductase (RRM1 and 2).20,22,32

have also found that the loss of MUC4 leads to an increased Western blotting sensitivity of CAPAN-2 cells to 5-fluorouracil (personal communica- Western blotting on nitrocellulose membrane (0.2 mm, Whatman) was tion). Taking into account these recent data and ours, targeting carried out as described previously.46 Membranes were incubated with MUC4 would certainly improve efficiency of the FOLFIRINOX antibodies against phospho-Erk1/2 (Thr202/Tyr204) (clone 20G11, 1/500), protocol and is at this time under investigation. Erk1/2 (clone I37F5, 1/500), NF-kB Phospho-p65 (Ser536) (clone 93H1, 1/500), NF-kB p65 (clone E498, 1/500), phospho-SAPK/JNK (Thr183/Tyr185) (9251, dilution 1/500), SAPK/JNK (clone 56G8, 1/500), phospho-p53 (Ser 15) (9284, dilution 1/500) and p53 (9282, 1/500), from Cell Signaling MATERIALS AND METHODS Technology (Ozyme, Saint Quentin Yvelines, France); Bax (clone N-20, Cell culture 1/500), BclXL (clone H-5, 1/500), hCNT1 (clone H-70, 1/200), MRP3 (clone CAPAN-1 and CAPAN-2 PC cell lines were cultured as described previously .43 H-16, 1/200), MRP4 (clone M4I-10, 1/200), MUC4 (clone 8G7, 1/500), NF-kB MUC4 knocked-down (MUC4-KD) cells were obtained by stable trans- p50 (clone H-119, 1/500) and Sp1 (PEP2, 1/1000) from Santa Cruz Biotechnology Inc. (Heidelberg, Germany); or hCNT3 (HPA023311; 1/500), fection with pRetroSuper plasmid (SA Biosciences, Qiagen, Courtaboeuf, tubulin (DM1A, 1/1000) and b-actin (A5441, 1/5000) from Sigma-Aldrich France) containing a small-hairpin RNA targeting MUC4 (50-AAGTGG AACGAATCGATTCTGTTCAAGAGACAGAATCGATTCGTTCCACTT-30). Control (St Quentin Fallavier, France). Antibodies were diluted in 5% (w/v) nonfat cells (Mock) were obtained by transfecting the empty-vector pRetro- dry milk in Tris-Buffered Saline Tween-20, except for MUC4 and b-actin 1 Super plasmid. After selection with neomycine (300 mg/ml) and serial limit diluted in Tris-Buffered Saline Tween-20, and incubated overnight at 4 C. dilution, expression of MUC4 was controlled by western blotting. Four Peroxidase-conjugated secondary antibodies (Sigma-Aldrich) were used and immunoreactive bands were visualized using the West Pico selected clones of Mock and MUC4-KD cells were pooled in order to avoid chemoluminescent substrate (Thermo Scientific, Pierce, Brebie`res, clonal variation. All cells were maintained in a 37 1C incubator with 5% CO 2 France). Chemo-luminescence was visualized using the LAS4000 and cultured as the parental cells. Pharmacological inhibitors were added to the cells and incubated for 24 h at the following final concentrations: apparatus (Fujifilm, Courbevoie, France). Densities of bands were integrated using the Gel analyst software (Claravision, Paris, France) and U0126 (10 mM, inhibitor of MAPK, Calbiochem, Merck Chemical Limited, represented as histograms. Three independent experiments were Nottingham, UK), SP600125 (10 mM, inhibitor of JNK, Calbiochem) and BAY- performed. 11-7082 (10 mM, inhibitor of NF-kB, Calbiochem).

Cytotoxicity assay Protein extraction Cells were seeded in growth medium into 96-well plates at a density of 104 Cytosolic, nuclear and total cellular extracts were performed as previously cells per well. After 24 h incubation, the medium was replaced by fresh described in Van Seuningen et al.44 and Jonckheere et al.,45 respectively. medium containing gemcitabine, oxaliplatin or ARA-C at the determined

Oncogene (2013) 1714 – 1723 & 2013 Macmillan Publishers Limited Gemcitabine resistance linked to MUC4 expression N Skrypek et al 1721 concentration (range: 10 nM-20 mM) and incubated for 72 h at 37 1C. extracts (20 ml) was measured on a Mithras Microplate Reader LB 940 The viability of cells was determined using the 3-(4,5-dimethylthiazol-2- (Berthold Technologies, Thoiry, France). Total protein content in the extract yl)-2,5-diphenyltetrazolium bromide assay (MTT, Sigma-Aldrich). Yellow- (4 ml) was measured using the bicinchoninic acid method in 96-well plates colored MTT was dissolved in phosphate-buffered saline (Invitrogen, (Pierce, Thermoscientific). Relative luciferase activity was expressed as fold Life Technologies, Villebon sur Yvette, France) at a final concentration of activation of the kB-Luc vector compared with that of the empty vector 5 mg/ml. The cells were then incubated in a CO2 incubator at 37 1C for 1 h. pGL3 basic (Promega). In co-transfection studies, pCR3-IkB or pCR3 empty Formazan crystal was solubilized in dimethylsulphoxide (Sigma-Aldrich) for vector was transfected with kB-Luc vector. Each experiment was assayed in 5 min at room temperature before being analyzed spectrophotometrically triplicate in at least three independent experiments. at a wavelength of 570 nm with a microplate reader (Bio-Rad, Marnes- la-Coquette, France). Percentage of viability ¼ [(Atreated À Ablank)/(Aneg À qRT–PCR Ablank)] Â 100, where Atreated is the average of absorbance in wells Total RNA from PC cells was prepared using the NucleoSpin RNA II kit containing cells treated with gemcitabine, Aneg is the average of wells (Macherey Nagel, Hoerdt, Germany). cDNA was prepared as described containing cells without gemcitabine treatment and Ablank is the average previously.47 PCR was performed using the SsoFast Evagreen Supermix kit of wells containing medium without cells. following the manufacturer’s protocol using the CFX96 real-time PCR system (Bio-Rad). Primer information is given in Table 1. Each marker was RNA interference assayed in triplicate in three independent experiments. Expression levels of genes of interest were normalized to the mRNA level of GAPDH Transient inhibition of MRP3, MRP4, hCNT1 and hCNT3 was performed housekeeping gene. using a pool of siRNA designed by Santa Cruz Biotechnology with Effectene (Qiagen) following the manufacturer’s instruction. Transient KD for Erk1 (MAPK1), Erk2 (MAPK3), JNK1 (MAPK8), JNK2 (MAPK9) and NF-kB Flow cytometry (p105 NF-kB) was performed using siRNA from Dharmacon (Thermo Cells were plated into 6-well plates at a density of 1.5 105 cells per well 46 Â Scientific) following the protocol described previously. Controls were and incubated for 24 h at 37 1C before gemcitabine treatment for another performed using a non-targeting siRNA (NT) in both protocols. Cells were 72 h. Apoptosis was measured by Annexin V-PE (Sigma-Aldrich). Briefly, 5 seeded at a density of 5 Â 10 cells per well into 6-well plates for RNA and cells were harvested by trypsinization, washed twice with 1 phosphate- 4 Â protein extraction, or at a density of 10 cells per well into 96-well plates buffered saline and 1 Â 105 cells of each condition were incubated with for cytotoxic assay, and left for 48 h at 37 1C before gemcitabine treatment 5 ml of annexin V-PE (50 mg/ml) and 10 ml propidium iodide (100 mg/ml) for (100 nM) for another 24 h at 37 1C. 10 min at room temperature. Staining intensity was determined with Cell Lab QuantaSC MPL (Beckman Coulter, Roissy, France) and analyzed with Quanta-Collection. Luciferase activity Transfection of kB-Luc synthetic promoter containing three kB-binding sites was performed with Effectene (Qiagen). Total cell extracts were Immunohistochemical analyses prepared after 48 h incubation at 37 1C using 1x Reagent Lysis Buffer Pancreatic adenocarcinoma samples were obtained from 16 patients of the (Promega, Charbonnie`res-les-Bains, France). Luciferase activity in the Lille University Hospital. No patient received either chemotherapy or

Table 1. Primer sequences used for quantitative real-time RT–PCR

Gene Forward Primer 50 to 30 Reverse Primer 50 to 30 Amplicon size (bp)

dCK GAGAAACCTGAACGATGGTCTT TCTCTGCATCTTTGAGCTTGC 102 hENT1 CTCTCAGCCCACCAATGAAAG CTCAACAGTCACGGCTGGAA 123 hCNT1 CCTCACCTGTGTGGTCCTCA AGACCCCTCTTAAACCAGAGC 86 hCNT3 CTTTTCTGGAGTACACAGATGCT CGGCAGGACCTTAAATGCAAA 108 RRM1 CTGCAACCTTGACTACTAAGCA CTTCCATCACATCACTGAACACT 108 RRM2 CCACGGAGCCGAAAACTAAAG CTCTGCCTTCTTATACATCTGCC 131 MRP3 GGAGGACATTTGGTGGGCTTT CCCTCTGAGCACTGGAAGTC 90 MRP4 AAGTGAACAACCTCCAGTTCCAG GGCTCTCCAGAGCACCATCT 119 MRP5 AGAACTCGACCGTTGGAATGC TCATCCAGGATTCTGAGCTGAG 104 GAPDH CCACATCGCTCAGACACCAT CCAGGCGCCCAATACG 70

Table 2. Characteristics of patients with pancreatic adenocarcinoma.

Patient Age (years) Gender Metastatic site Location of tumor Tumor TNM stage MUC4 hCNT1 in pancreas differentiation

1 65 F Liver NS M pTxNxM1 þÀ 2 63 F Liver Head M pTxNxM1 þþ 3 46 M Head M pT3N1M0 Àþ 4 51 M Liver, bone NS M pTxNxM1 Àþ 5 61 M Peritoneum Tail M pT3N1M1 þÀ 6 57 F Peritoneum, ovary NS W pT4NxM1 þþ 7 57 F Peritoneum Head M pTxNxM1 þÀ 8 51 M Liver NS W pTxNxM1 þÀ 9 54 M Peritoneum Head W pTxNxM1 Àþ 10 60 M Liver, peritoneum Head M pTxN1M1 þþ 11 78 F Head P pTxNxM0 þÀ 12 61 F NS M pT3NxM0 Àþ 13 47 F NS M pTxN1M0 þÀ 14 66 M NS M pTxN0M0 Àþ 15 53 M Head M pTxN1M0 þÀ 16 73 F Head M pT3N1M0 Àþ Abbreviations: M, moderately differentiated; NS, not speciﬁed; P, poorly-differentiated; W, well-differentiated.

& 2013 Macmillan Publishers Limited Oncogene (2013) 1714 – 1723 Gemcitabine resistance linked to MUC4 expression N Skrypek et al 1722 radiotherapy before the surgical resection. A consent form was obtained in 20,20-difluorodeoxycytidine- induced cytotoxicity. Clin Cancer Res 2003; 9: from each patient, and permission for removal of surgical samples was 5000–5008. obtained from the institutional review board. Patient characteristics are 10 Bafna S, Kaur S, Momi N, Batra SK. Pancreatic cancer cells resistance to gemci- given in Table 2. Samples were processed for paraffin embedding. Tissue tabine: the role of MUC4 mucin. Br J Cancer 2009; 101: 1155–1161. sections (4 mM) were stained with hematoxylin and eosin. Manual hCNT1 11 Kalra AV, Campbell RB. Mucin impedes cytotoxic effect of 5-FU against growth of immunohistochemistry (IHC) was carried out as described in Van der Sluis human pancreatic cancer cells: overcoming cellular barriers for therapeutic gain. et al.48 and automatic MUC4 IHC with an automated immunostainer (ES, Br J Cancer 2007; 97: 910–918. Ventana Medical System, Strasbourg, France) as in Mariette et al.49 hCNT1 12 Kalra AV, Campbell RB. Mucin overexpression limits the effectiveness of 5-FU by (H-70, 1/100) and MUC4 (8G7, 1/100) antibodies were purchased from reducing intracellular drug uptake and antineoplastic drug effects in pancreatic Santa Cruz Biotechnology. Sections were counterstained with hematoxylin, tumours. Eur J Cancer 2009; 45: 164–173. dehydrated and mounted. A positive control for MUC4 and hCNT1 13 Jonckheere N, Skrypek N, van Seuningen I. Mucins and pancreatic cancer. Cancers immunostainings and a negative control in the absence of primary 2010; 2: 1794–1812. antibody were included in each set of experiments. 14 Chaturvedi P, Singh AP, Moniaux N, Senapati S, Chakraborty S, Meza JL et al. MUC4 mucin potentiates pancreatic tumor cell proliferation, survival, and invasive properties and interferes with its interaction to extracellular matrix proteins. Mol Statistical analyses Cancer Res 2007; 5: 309–320. Statistical analyses were performed using the Graphpad Prism 4.0 software 15 Jonckheere N, Skrypek N, Merlin J, Dessein AF, Dumont P, Leteurtre E et al. The (Graphpad softwares Inc., La Jolla, CA, USA). Differences in data of two mucin MUC4 and its membrane partner ErbB2 regulate biological properties of samples were analyzed by Student’s t-test or ANOVA test with selected human CAPAN-2 pancreatic cancer cells via different signalling pathways. PLoS comparison using Tukey’s post-hoc test, and were considered significant One 2012; 7: e32232. for the following P-values: *Po0.05, **Po0.01 or ***Po0.001. IHC results 16 Komatsu M, Tatum L, Altman NH, Carothers Carraway CA, Carraway KL. Poten- were compared using Fisher’s exact test. tiation of metastasis by cell surface sialomucin complex (rat MUC4), a multi- functional anti-adhesive glycoprotein. Int J Cancer 2000; 87: 480–486. 17 Ponnusamy MP, Lakshmanan I, Jain M, Das S, Chakraborty S, Dey P et al. CONFLICT OF INTEREST MUC4 mucin-induced epithelial to mesenchymal transition: a novel mechanism for metastasis of human ovarian cancer cells. Oncogene 2010; 29: The authors declare no conflict of interest. 5741–5754. 18 Workman HC, Sweeney C, Carraway KL 3rd. The membrane mucin Muc4 inhibits apoptosis induced by multiple insults via ErbB2-dependent and ErbB2-indepen- ACKNOWLEDGEMENTS dent mechanisms. Cancer Res 2009; 69: 2845–2852. We thank MH Gevaert and R Siminsky (Department of Histology, Faculty of Medicine, 19 Blackstock AW, Lightfoot H, Case LD, Tepper JE, Mukherji SK, Mitchell BS et al. University of Lille 2) and the technical platform IFR114/IMPRT for flow cytometry (Dr Tumor uptake and elimination of 20,20-difluoro-20-deoxycytidine (gemcitabine) N Jouy) and luciferase (AS Drucbert) analyses. Gemcitabine and oxaliplatin are a kind after deoxycytidine kinase gene transfer: correlation with in vivo tumor response. gift from Dr A Lansiaux (Centre Oscar Lambret, Universite´ Lille Nord de France, Inserm Clin Cancer Res 2001; 7: 3263–3268. UMR837, team 4, Lille, France). Cytarabine/aracytin ARA-C is a kind gift from Pr B 20 Marechal R, Mackey JR, Lai R, Demetter P, Peeters M, Polus M et al. Deoxycitidine Quesnel (Oncohaematology department, Centre Hospitalier Re´gional et Universitaire kinase is associated with prolonged survival after adjuvant gemcitabine for de Lille, Inserm UMR837, team 3, Lille, France). We thank Dr JL Desseyn (Inserm U995) resected pancreatic adenocarcinoma. Cancer 2010; 116: 5200–5206. for his help in designing MUC4 shRNA. Nicolas Skrypek is a recipient of a PhD 21 Kruh GD, Belinsky MG. The MRP family of drug efflux pumps. Oncogene 2003; 22: fellowship of Centre Hospitalier Re´gional et Universitaire (CHRU) de Lille and Re´gion 7537–7552. Nord-Pas de Calais. Dr Nicolas Jonckheere is a recipient of postdoctoral fellowship 22 Jordheim LP, Seve P, Tredan O, Dumontet C. The ribonucleotide reductase large from the Institut National du Cancer (INCa) and Ligue Nationale contre le Cancer subunit (RRM1) as a predictive factor in patients with cancer. Lancet Oncol 2010; (LNCC). This work is supported by a grant from la Ligue Nationale contre le Cancer 12: 693–702. (Equipe Labelliseé Ligue 2010, IVS). Isabelle Van Seuningen is the recipient of a 23 Swartz MJ, Batra SK, Varshney GC, Hollingsworth MA, Yeo CJ, Cameron JL et al. ‘Contrat Hospitalier de Recherche Translationnelle’/CHRT 2010, AVIESAN. The funders MUC4 expression increases progressively in pancreatic intraepithelial neoplasia. had no role in study design, data collection and analysis, decision to publish or in Am J Clin Pathol 2002; 117: 791–796. preparation of the paper. 24 Mackey TJ, Borkowski A, Amin P, Jacobs SC, Kyprianou N. bcl-2/bax ratio as a predictive marker for therapeutic response to radiotherapy in patients with prostate cancer. Urology 1998; 52: 1085–1090. REFERENCES 25 Krajewski S, Krajewska M, Turner BC, Pratt C, Howard B, Zapata JM et al. Prog- nostic significance of apoptosis regulators in breast cancer. Endocr Relat Cancer 1 Jemal A, Bray F, Center MM, Ferlay J, Ward E, Forman D. Global cancer statistics. 1999; 6: 29–40. CA Cancer J Clin 2011; 61: 69–90. 26 Shi X, Liu S, Kleeff J, Friess H, Buchler MW. Acquired resistance of pancreatic 2 Vincent A, Herman J, Schulick R, Hruban RH, Goggins M. Pancreatic cancer. Lancet cancer cells towards 5-Fluorouracil and gemcitabine is associated with altered 2011; 378: 607–620. expression of apoptosis-regulating genes. Oncology 2002; 62: 354–362. 3 O’Reilly EM. Pancreatic adenocarcinoma: new strategies for success. Gastrointest 27 Ohishi Y, Oda Y, Uchiumi T, Kobayashi H, Hirakawa T, Miyamoto S et al. Cancer Res 2009; 3: S11–S15. ATP-binding cassette superfamily transporter gene expression in human primary 4 Damaraju VL, Damaraju S, Young JD, Baldwin SA, Mackey J, Sawyer MB et al. ovarian carcinoma. Clin Cancer Res 2002; 8: 3767–3775. Nucleoside anticancer drugs: the role of nucleoside transporters in resistance to 28 Oguri T, Isobe T, Fujitaka K, Ishikawa N, Kohno N. Association between expression cancer chemotherapy. Oncogene 2003; 22: 7524–7536. of the MRP3 gene and exposure to platinum drugs in lung cancer. Int J Cancer 5 Farrell JJ, Elsaleh H, Garcia M, Lai R, Ammar A, Regine WF et al. Human equili- 2001; 93: 584–589. brative nucleoside transporter 1 levels predict response to gemcitabine in 29 Konig J, Hartel M, Nies AT, Martignoni ME, Guo J, Buchler MW et al. Expression and patients with pancreatic cancer. Gastroenterology 2009; 136: 187–195. localization of human multidrug resistance protein (ABCC) family members in 6 Giovannetti E, Del TM, Mey V, Funel N, Nannizzi S, Ricci S et al. Transcription pancreatic carcinoma. Int J Cancer 2005; 115: 359–367. analysis of human equilibrative nucleoside transporter-1 predicts survival in 30 Chen ZS, Lee K, Kruh GD. Transport of cyclic nucleotides and estradiol 17-beta-D- pancreas cancer patients treated with gemcitabine. Cancer Res 2006; 66: glucuronide by multidrug resistance protein 4. Resistance to 6-mercaptopurine 3928–3935. and 6-thioguanine. J Biol Chem 2001; 276: 33747–33754. 7 Marechal R, Mackey JR, Lai R, Demetter P, Peeters M, Polus M et al. Human 31 Lee K, Klein-Szanto AJ, Kruh GD. Analysis of the MRP4 drug resistance profile in equilibrative nucleoside transporter 1 and human concentrative nucleoside transfected NIH3T3 cells. J Natl Cancer Inst 2000; 92: 1934–1940. transporter 3 predict survival after adjuvant gemcitabine therapy in resected 32 Hagmann W, Jesnowski R, Lohr JM. Interdependence of gemcitabine treatment, pancreatic adenocarcinoma. Clin Cancer Res 2009; 15: 2913–2919. transporter expression, and resistance in human pancreatic carcinoma cells. 8 Bhutia YD, Hung SW, Patel B, Lovin D, Govindarajan R. CNT1 expression influences Neoplasia 2010; 12: 740–747. proliferation and chemosensitivity in drug-resistant pancreatic cancer cells. 33 Hu YP, Haq B, Carraway KL, Savaraj N, Lampidis TJ. Multidrug resistance correlates Cancer Res 2011; 71: 1825–1835. with overexpression of Muc4 but inversely with P-glycoprotein and multidrug 9 Garcia-Manteiga J, Molina-Arcas M, Casado FJ, Mazo A, Pastor-Anglada M. resistance related protein in transfected human melanoma cells. Biochem Nucleoside transporter profiles in human pancreatic cancer cells: role of hCNT1 Pharmacol 2003; 65: 1419–1425.

Oncogene (2013) 1714 – 1723 & 2013 Macmillan Publishers Limited Gemcitabine resistance linked to MUC4 expression N Skrypek et al 1723 34 Wang W, Abbruzzese JL, Evans DB, Larry L, Cleary KR, Chiao PJ. The nuclear factor- 42 Tian Q, Zhang J, Tan TM, Chan E, Duan W, Chan SY et al. Human multidrug kappa B RelA transcription factor is constitutively activated in human pancreatic resistance associated protein 4 confers resistance to camptothecins. Pharm Res adenocarcinoma cells. Clin Cancer Res 1999; 5: 119–127. 2005; 22: 1837–1853. 35 Arlt A, Gehrz A, Muerkoster S, Vorndamm J, Kruse ML, Folsch UR et al. Role of 43 Jonckheere N, Perrais M, Mariette C, Batra SK, Aubert JP, Pigny P et al. A role for NF-kappaB and Akt/PI3K in the resistance of pancreatic carcinoma cell lines human MUC4 mucin gene, the ErbB2 ligand, as a target of TGF-beta in pancreatic against gemcitabine-induced cell death. Oncogene 2003; 22: 3243–3251. carcinogenesis. Oncogene 2004; 23: 5729–5738. 36 Banerjee S, Zhang Y, Ali S, Bhuiyan M, Wang Z, Chiao PJ et al. Molecular evidence 44 Van Seuningen I, Ostrowski J, Bustelo XR, Sleath PR, Bomsztyk K. The K protein for increased antitumor activity of gemcitabine by genistein in vitro and in vivo domain that recruits the interleukin 1-responsive K protein kinase lies adjacent to using an orthotopic model of pancreatic cancer. Cancer Res 2005; 65: 9064–9072. a cluster of c-Src and Vav SH3-binding sites. Implications that K protein acts as a 37 Kunnumakkara AB, Guha S, Krishnan S, Diagaradjane P, Gelovani J, Aggarwal BB. docking platform. J Biol Chem 1995; 270: 26976–26985. Curcumin potentiates antitumor activity of gemcitabine in an orthotopic model of 45 Jonckheere N, Fauquette V, Stechly L, Saint-Laurent N, Aubert S, Susini C et al. pancreatic cancer through suppression of proliferation, angiogenesis, and inhi- Tumour growth and resistance to gemcitabine of pancreatic cancer cells are bition of nuclear factor-kappaB-regulated gene products. Cancer Res 2007; 67: decreased by AP-2alpha overexpression. Br J Cancer 2009; 101: 637–644. 3853–3861. 46 Piessen G, Jonckheere N, Vincent A, Hemon B, Ducourouble MP, Copin MC et al. 38 Wang SJ, Gao Y, Chen H, Kong R, Jiang HC, Pan SH et al. Dihydro- Regulation of the human mucin MUC4 by taurodeoxycholic and taur- artemisinin inactivates NF-kappaB and potentiates the anti-tumor effect of ochenodeoxycholic bile acids in oesophageal cancer cells is mediated by hepa- gemcitabine on pancreatic cancer both in vitro and in vivo. Cancer Lett 2010; 293: tocyte nuclear factor 1alpha. Biochem J 2007; 402: 81–91. 99–108. 47 Van Seuningen I, Perrais M, Pigny P, Porchet N, Aubert JP. Sequence of the 39 Kong R, Sun B, Jiang H, Pan S, Chen H, Wang S et al. Downregulation of 50-ﬂanking region and promoter activity of the human mucin gene MUC5B in nuclear factor-kappaB p65 subunit by small interfering RNA synergizes with different phenotypes of colon cancer cells. Biochem J 2000; 348: 675–686. gemcitabine to inhibit the growth of pancreatic cancer. Cancer Lett 2010; 291: 48 van der Sluis M, Melis MH, Jonckheere N, Ducourouble MP, Buller HA, Renes I et al. 90–98. The murine Muc2 mucin gene is transcriptionally regulated by the zinc-ﬁnger 40 Conroy T, Desseigne F, Ychou M, Bouche O, Guimbaud R, Becouarn Y et al. GATA-4 transcription factor in intestinal cells. Biochem Biophys Res Commun 2004; FOLFIRINOX versus gemcitabine for metastatic pancreatic cancer. N Engl J Med 325: 952–960. 2011; 364: 1817–1825. 49 Mariette C, Perrais M, Leteurtre E, Jonckheere N, Hemon B, Pigny P et al. 41 Theile D, Grebhardt S, Haefeli WE, Weiss J. Involvement of drug transporters in the Transcriptional regulation of human mucin MUC4 by bile acids in oesophageal synergistic action of FOLFOX combination chemotherapy. Biochem Pharmacol cancer cells is promoter-dependent and involves activation of the phosphatidy- 2009; 78: 1366–1373. linositol 3-kinase signalling pathway. Biochem J 2004; 377: 701–708.

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

& 2013 Macmillan Publishers Limited Oncogene (2013) 1714 – 1723