Acquired chemoresistance in pancreatic carcinoma cells - Nature

Oncogene (2006) 25, 3973–3981

& 2006 Nature Publishing Group All rights reserved 0950-9232/06 $30.00 www.nature.com/onc

ORIGINAL ARTICLE

Acquired chemoresistance in pancreatic carcinoma cells: induced secretion of IL-1b and NO lead to inactivation of caspases S Sebens Muërko¨ster1, J Lust1, A Arlt1, R Ha¨sler2, M Witt1, T Sebens1, S Schreiber2, UR Fo¨lsch1 and H Scha¨fer1 1

Laboratory of Molecular Gastroenterology & Hepatology, 1st Department of Medicine, UKSH, Campus Kiel, Kiel, Germany and Mucosal Immunology Research Group, 1st Department of Medicine, UKSH, Campus Kiel, Kiel, Germany

2

Pancreatic cancer exhibits profound chemoresistance resulting either from pre-existing (intrinsic) mechanisms, or from anticancer drug treatment itself (acquired chemoresistance). To identify molecular alterations leading to acquired chemoresistance, the chemosensitive pancreatic carcinoma cell line PT45-P1 was exposed to low-dose treatment with etoposide for 6 weeks. Afterwards, these cells (PT45-P1res) were much more resistant to high-dose treatment with anticancer drugs than parental cells. Among several differentially expressed genes in PT45-P1res cells, IL-1b was most significantly upregulated, a finding in line with our previous observation that IL-1b accounts for intrinsic chemoresistance of pancreatic carcinoma cells. Elevated IL-1b expression in PT45-P1res cells was confirmed by real-time PCR and ELISA, and treatment with the IL-1 receptor antagonist restored drug-induced apoptosis. The increased IL-1b secretion was accompanied by an elevated formation of nitric oxide (NO) and a NO-dependent inhibition of the etoposide-induced caspase-3/-7/-8/-9 activity. Caspase activation was restored either by the iNOS inhibitor 1400W, the reducing agent dithiothreitol or the IL-1 receptor antagonist, resulting in greater sensitivity towards anticancer drug treatment. Conversely, IL-1b or the NO-donor SNAP decreased caspase activation and apoptosis in etoposide-treated PT45-P1 cells. These data confirm IL-1b and NO as determinants of chemoresistance in pancreatic cancer, and indicate that the intrinsic and acquired chemoresistance rely to some extent on common molecular targets beneficial for improved therapeutical strategies. Oncogene (2006) 25, 3973–3981. doi:10.1038/sj.onc.1209423; published online 13 February 2006 Keywords: chemoresistance; interleukin-1b; pancreatic cancer; nitric oxide

Correspondence: Dr H Scha¨fer, Laboratory of Molecular Gastroenterology & Hepatology, 1st Department of Medicine, UKSHCampus Kiel, Schittenhelmstrae 12, 24105 Kiel, Germany. E-mail: [email protected] Received 4 August 2005; revised 1 January 2006; accepted 3 January 2006; published online 13 February 2006

Introduction Pancreatic ductal adenocarcinoma (PDAC) is 4–5th in the rank order of fatal malignant diseases in Western Countries, exhibiting a still increasing prevalence (Bramhal et al., 1995; Parker et al., 1997). Owing to a highly malignant and progressive phenotype and a lack of specific symptoms, PDAC is diagnosed mostly at an already advanced state (Brand and Tempero, 1998; Lillemore, 1998) precluding a surgical intervention as the only curative option. For more than 85% of the patients, solely palliative options rely on radioor chemotherapy, both still with moderate success (Blaszkowsky, 1998; Neoptolemos et al., 2001). This strong limitation of conventional treatment is mainly owing to the profound resistance of PDAC cells towards anti cancer drugs emerging from two distinct modes: (1) the intrinsic or innate chemoresistance, pre-existing before chemotherapeutic treatment, that is often induced by autocrine or paracrine mechanisms given by tumor–tumor or tumor–stroma interactions (Arlt et al., 2002; Muërko¨ster et al., 2004); and (2) the acquired chemoresistance emerging during the course of chemotherapy (Banerjee et al., 2002; Kang et al., 2004). An important way by which tumor cells gain protection from chemotherapeutic drugs is given by an altered balance of pro- and antiapoptotic proteins, along with a reduced apoptotic response (Morisaki and Katano, 2003; Soengas and Lowe, 2003). In this context, the functional loss of the tumor-suppressor p53 (Brown and Attardi, 2005), the upregulation of antiapoptotic molecules, for example, bcl-2, c-FLIP, XIAP (Medema et al., 1999; Kim et al., 2004; Yan et al., 2004), or the downregulation of proapoptotic proteins, for example Smac/Diabolo and Bid (Erler et al., 2004; Yan et al., 2004), have been extensively reported. Another important determinant of antiapoptotic protection in cancer cells, including PDAC, is the constitutive activation of the transcription factor NF-kB (Wang et al., 1999; Schmid and Adler, 2000). Its inhibition significantly increases the sensitivity of PDAC cells towards chemotherapy in vitro and in vivo (Arlt et al., 2001; Muërko¨ster et al., 2003). Furthermore, we recently demonstrated that in various chemoresistant PDAC cells, the constitutive NF-kB activity relates to the autocrine action of IL-1b (Arlt et al., 2002). This

Acquired chemoresistance in pancreatic cancer S Sebens Muërko¨ster et al

3974

Results Long-term cytotoxic drug treatment reduces chemosensitivity of PT45-P1 cells After treatment with a daily dose of 0. 2 mg/ml etoposide over 6 weeks (reflecting a course of chemotherapy), these long-term drug-treated PT45-P1 cells and their parental counterpart were both cultured under normal cell culture conditions for up to 10 weeks in order to perform further analyses. For detecting the induction of apoptosis after cytostatic drug treatment, both cell lines were either left untreated or exposed to 20 mg/ml etoposide or 5 mg/ml gemcitabine. After 24 h, cells were analysed for annexinV staining by flow cytometry. As shown in Figure 1a, treatment with etoposide induced significantly less apoptosis in long-term drug-treated PT45-P1 cells (therefore named PT45-P1res) than that in the parental cell line PT45-P1. Whereas 45% of PT45-P1 cells were apoptotic after cytostatic drug treatment, only 27% of the PT45-P1res cells showed an apoptotic phenotype indicated by annexinV staining. Similar results were obtained after treatment with gemcitabine (Figure 1a). In addition, the resistance towards treatment with TRAIL (2.5 mg/ml) and anti-CD95 (CH-11, 100 ng/ml), already existing in PT45-P1 cells was not significantly affected by the long-term etoposide exposure of PT45-P1res cells (Figure 1a). To exclude that a reduced proliferation rate accounts for the reduced chemosensitivity in PT45-P1res cells, the cell cycle profile was analysed in PT45-P1 and PT45-P1res cells. Staining with the DNA intercalating dye propidium iodide (PI) revealed similar cell cycle profiles in Oncogene

a % AnnexinV positive cells (over basal)

60

* *

PT45-P1

50

PT45-P1res 40 30 20 10 0

b 60

gemcitabine

PT45-P1 PT45-P1res

50 40 30 20 10 0 G1

S

G2/M

c cell number (n x 105)

etoposide

% of cells

autocrine loop obviously evolves from tumor–stroma interactions, as shown for the chemosensitive PDAC cell lines PT45-P1 and T3M4 that, upon coculture with stromal fibroblasts, gained NF-kB-dependent chemoresistance preceeded by elevated IL-1b-secretion (Muërko¨ster et al., 2004). The importance of these factors in the development of innate chemoresistance was further supported by immunohistochemical stainings of human PDAC specimens, which revealed increased expression of the activated NF-kB subunit p65 and IL-1b in a large number of the tested tumor samples but not in normal pancreatic tissue (Muërko¨ster et al., 2004; Muërko¨ster et al., 2005). As we have already described a model for the gain of innate chemoresistance in PT45-P1 cells emerging from tumor–stroma interactions, the aim of this study was to identify mechanisms underlying the acquired chemoresistance of this anti cancer-drug-sensitive cell line. For this purpose, PT45P1 cells were treated over 6 weeks with a low dose of etoposide. Afterwards, these cells (PT45-P1res) together with the chemosensitive parental cell line were used for comparative analysis. The identification of molecular mechanisms accounting for both the innate and the acquired chemoresistance might offer common targets for the development of more efficient therapeutical strategies in the treatment of pancreatic carcinoma.

CH11 20 18 16 14 12 10 8 6 4 2 0

Trail

PT45-P1 PT45-P1res

0h

24h 48h time of culture

72h

Figure 1 Long-term drug treatment leads to chemoresistance in PT45-P1 cells. (a) PT45-P1 and PT45-P1res cells were either left untreated or treated with 20 mg/ml etoposide, 5 mg/ml gemcitabine, 100 ng/ml CH11 antibody, 2.5 mg/ml recombinant human Trail. After 24 h, apoptosis was determined by annexinV/PI staining followed by fluorescence flow cytometry. Data are presented as % annexinV-positive cells over basal level. Basal apoptosis was 10– 15% in PT45-P1 cells and 12–20% in PT45-P1res cells. (b) PT45-P1 and PT45-P1res cells were cultured for 24 h. Measurement of propidium iodide (PI) uptake and cell cycle analysis was performed by fluorescence flow cytometry. Means7s.d. from three independent experiments are shown. *Indicates Po0.05. (c) PT45-P1 and PT45-P1res cells were seeded into six-well plates. Cell numbers were determined 24, 48 and 72 h after seeding. One representative experiment out of three is shown.

chemosensitive PT45-P1 and in chemoresistant PT45P1res cells (Figure 1b). Correspondingly, the doubling times of both cell lines were nearly indistinguishable (Figure 1c). These results demonstrate that long-term drug exposure led to a stable chemoresistant phenotype of PT45-P1 cells. Furthermore, cellular mechanisms other than a reduced proliferation must account for this acquired chemoresistance. Long-term drug treatment induces NF-kB and IL-1b secretion in PT45-P1res cells, accounting for chemoresistance As we have previously demonstrated that many PDAC cells show an enhanced NF-kB activation and elevated secretion of IL-1b, accounting for chemoresistance in vitro and in vivo (Arlt et al., 2002; Muërko¨ster et al., 2004), we paid particular attention here to these two factors. Interestingly, among several other differentially expressed genes, IL-1b was most significantly (more than fourfold) upregulated in PT45-P1res cells in comparison with PT45-P1 cells, as revealed by comparative cDNA-micorarray analysis of 409 different apoptosis-associated genes (data not shown). Real-time PCR (Figure 2a) and RT–PCR analysis (data not


3975

mRNA amount (arb. units)

a

60

PT45-P1

50

PT45-P1res

40 30 20 10 0

c

5

*

4 3 2 1 PT45-P1

PT45-P1res

*

0.6 NF-κB activation (OD 490nm)

6

0

β-actin

IL1β

b pg IL1 β/105 Zellen

shown) confirmed the strongly increased levels of IL-1b mRNA in PT45-P1res cells. To determine the amount of secreted mature IL-1b, cell culture supernatants of both

PT45-P1 PT45-P1res

0.5 0.4 0.3 0.2 0.1 0

d

untreated

sulfasalazine PT45-P1 PT45-P1res

% AnnexinV positive cells (over basal)

80 70

*

IL1-RA *

60 50 40 30 20 10

0 etoposide IL1-RA IL1β

+ -

+ + -

+ +

+ -

+ + -

+ +

Figure 2 IL-1b is induced by long-term drug treatment in PT45P1res cells and mediates chemoresistance. (a) RNA from PT45-P1 and PT45-P1res cells was subjected to reverse transcription and subsequent real-time PCR using primers specific for IL-1b. In parallel, a real-time PCR was conducted for b-actin, which was used as a control. Results from one representative selection procedure are shown. Data are expressed as amount of mRNA in arbitrary units. Each sample was measured in duplicates. (b) Supernatants of PT45-P1 and PT45-P1res cells were subjected to an IL-1b immunoassay. The amount of IL-1b was normalized to equal cell number determined in parallel (expressed as pg human IL-1b/ 105 cells). Data represent the means7s.d. from three independent experiments. (c) NF-kB activation was measured by an ELISAbased NF-kB-binding assay detecting the p65/p50 subunits of NFkB in nuclear extracts of PT45-P1 and PT45-P1res cells, either left untreated or treated with 0.5 mmol/l sulfasalazine or 250 ng/ml IL1-RA for 24 h. Data represent the means7s.d. from three independent experiments. (d) PT45-P1 and PT45-P1res cells were either left untreated or treated with 20 mg/ml etoposide for 24 h. In the indicated samples, either 250 ng/ml of the IL1-RA were added 1 h before etoposide or 20 ng/ml IL-1b were added 8 h after etoposide. Then, cells were annexinV/PI stained and analysed by fluorescence flow cytometry. Data are presented as % annexinVpositive cells over basal level. Basal apoptosis was 10–15% in PT45-P1 cells and 12–20% in PT45-P1res cells. Treatment with IL1b and IL1-RA alone, respectively, did not effect cell viability. Means7s.d. from three independent experiments are shown. *Indicates Po0.05.

cell lines were subjected to an IL-1b-ELISA. Whereas PT45-P1 cells hardly secreted any IL-1b, PT45-P1res cells showed significant IL-1b secretion of 3–4 pg/105 cells (Figure 2b). Furthermore, an ELISA detecting the activated NF-kB subunit p65 in nuclear extracts (Figure 2c) as well as gelshift assays (data not shown) revealed that PT45-P1res cells showed significantly higher levels of activated NF-kB than PT45-P1 cells. This elevated basal NF-kB activity can be clearly reduced by the NF-kB inhibitor sulfasalazine (0.5 mmol/l) or the IL-1 receptor antagonist (IL1-RA, 250 ng/ml), which efficiently inhibits binding and signaling of IL-1b by binding to the IL-1 receptor. These results indicate an autocrine loop of constitutive NF-kB activation via the release of IL-1b as we already described for various chemoresistant PDAC cell lines (Arlt et al., 2002) as well as for T3M4 and PT45-P1 cells upon coculture with stromal fibroblasts (Muërko¨ster et al., 2004). To verify IL-1b as a mediator of chemoresistance in PT45-P1res cells, the IL1-RA was used before treatment with cytotoxic drugs and apoptosis was determined by annexinV staining. Treatment with IL-1b and IL1-RA alone, respectively, did not affect the viability of either cell line (data not shown), whereas pretreatment with the IL1-RA increased the number of apoptotic PT45P1res cells upon etoposide treatment from 27 to 66% (Figure 2d), indicating that the elevated IL-1b expression contributes to the chemoresistant phenotype of PT45-P1res cells. Accordingly, administration of 20 ng/ml IL-1b significantly reduced the number of apoptotic PT45-P1 cells upon anticancer drug treatment from 46 to 28% (Figure 2d), supporting the role of IL-1b in the mediation of chemoresistance in PDAC cells. Decreased caspase activation in PT45-P1res cells is IL-1b dependent Next, we analysed the mechanisms of IL-1b-mediated chemoresistance in PT45-P1res cells in more detail. As it is known that the NF-kB target genes inhibitor of apoptosis proteins (IAPs) inhibit caspases and thereby account for apoptosis resistance (Friesen et al., 1999), the expression levels of the most relevant caspases involved in apoptosis (the initiator caspases-8 and -9 and the effector caspases-3 and -7), as well as the expression of their endogenous inhibitors cIAP1/2 and XIAP were analysed by Western blotting. As shown in Figure 3a, PT45-P1 and PT45-P1res cells showed similar basal levels of all analysed procaspases (full-length form). In contrast, the appearance of cleaved caspases upon stimulation with etoposide, indicating activation of procaspases, was strongly attenuated in PT45-P1res cells compared to PT45-P1 cells. In addition, a homogeneous chemoluminiscent assay detecting the activity of the two downstream effector caspases-3 and -7 revealed a threefold reduced caspase-3/7 activity after etoposide treatment in PT45-P1res cells compared to PT45-P1 cells (Figure 3b), thus confirming the Western blot data. To verify whether increased expression of IAPs might account for the decreased caspase activation and Oncogene


3976

a

PT45-P1 etoposide: + IL1-RA: procaspase 3

etoposide-induced caspase-3/7 activity increased in PT45-P1res cells when treated with the IL1-RA (Figure 3b). As inhibition of IL-1b signaling increased also the number of annexinV-positive PT45-P1res cells after cytotoxic drug treatment (Figure 2d), it can be concluded that the elevated IL-1b production and secretion results in a decreased caspase activation, thereby accounting for chemoresistance in PT45-P1res cells.

PT45-P1res + + +

(35 kDa) >

cleaved caspase 3 (17/19 kDa) > procaspase 7

(35 kDa) >

cleaved caspase 7 (20 kDa) > procaspase 8 (57 kDa) > cleaved caspase 8 (41/43 kDa) > procaspase 9

(47 kDa) >

cleaved caspase 9 (35/37 kDa) > α-tubulin (55 kDa) >

10 9 8 7 6 5 4 3 2 1 0

*

c

re s

etoposide etoposide+ IL1-RA

PT 45 -P 1 PT 45 -P 1

Induction of caspase 3/7 activity (n-fold of basal)

b

cIAP1 (62 kDa) > cIAP2 (70 kDa) > XIAP (53 kDa) >

PT45-P1 PT45-P1res

α-tubulin (55 kDa) >

Figure 3 Decreased caspase activation in PT45-P1res cells is IL-1b-dependent. (a) PT45-P1 and PT45-P1res cells were either left untreated or treated with 20 mg/ml etoposide for 48 h. Additionally, PT45-P1res cells were coincubated with 250 ng/ml IL1-RA. Cellular lysates were submitted to Western blotting using antibodies for the detection of full length and cleaved caspase-3, 7, -8 and -9. An a-tubulin antibody was used as a control for equal protein load. (b) Cells were treated as above and analysed for caspase-3/7 activity expressed as n-fold induction of caspase-3/7 activity over basal. Means7s.d. from three independent experiments are shown. *Indicates Po0.05. (c) Detection of cIAP1, cIAP2 and XIAP in untreated cellular lysates of PT45-P1 and PT45-P1res cells by Western blotting. An a-tubulin antibody was used as a control for equal protein load.

chemoresistance in PT45-P1res cells, expression analysis of cIAP1, cIAP2 and XIAP were performed by Western blotting. However, both cell lines showed similar expression levels of all three endogenous caspase inhibitors (Figure 3c), as already indicated by the lack of distinct mRNA expression levels found with the DNA-microarray analysis. Thus, other mechanisms might be involved in the broad inhibition of caspase activation in PT45-P1res cells. Next, we investigated whether the elevated IL-1b secretion accounts for the decreased caspase activation in PT45-P1res cells. Western blot analysis of PT45-P1res cell lysates revealed that administration of the IL1-RA enhanced etoposide-induced cleavage of the procaspases-3, -7, -8 and -9 (Figure 3a) as compared to cells treated with etoposide alone, indicating that IL-1b attenuates anticancer-drug-dependent caspase activation in PT45-P1res cells. Moreover, IL-1b diminished cleavage of all four caspases after etoposide treatment in PT45-P1 cells (Figure 5a). In support of this finding, the Oncogene

Drug-induced IL-1b enhances secretion of NO in PT45-P1res cells IL-1b is a known inducer of the inducible nitric oxide synthetase (iNOS) and production of nitric oxide (NO). Other reports demonstrated that inhibition of caspase activity might occur by S-nitrosylation of cysteine residues in the active site of the caspases (Li et al., 1997; Maejima et al., 2005). In order to elucidate the mechanism of IL-1b-mediated caspase inhibition, we first analysed iNOS expression and NO secretion in PT45-P1 and PT45-P1res cells. In PT45-P1 cells, low levels of iNOS mRNA, strongly inducible by 20 ng/ml IL-1b within 3–6 h were detected by real-time PCR (Figure 4a) and RT–PCR (data not shown). In contrast, PT45-P1res cells exhibited much higher iNOS mRNA levels that were reduced by the IL1-RA (Figure 4a). Compared to PT45-P1 cells (2.1 mmol/105 cells), an elevated NO-release was found in untreated PT45-P1res cells (3.7 mmol/105 cells) and this secretion could be reduced by the IL1-RA (2.2 mmol/105 cells) (Figure 4b). These data indicate that drug-induced IL-1b secretion leads to an increased NO release. IL-1b-dependent NO secretion leads to decreased activation of caspases and chemoresistance of PT45-P1res cells To confirm whether the IL-1b-dependent NO secretion accounts for caspase inhibition, as seen in PT45-P1res cells, PT45-P1 cells were treated with IL-1b or without in the absence or presence of the specific iNOS inhibitor 1400W before etoposide administration, and PT45P1res cells were treated with a combination of 1400W and etoposide for 48 h. Finally, caspase activation as well as apoptosis were analysed. Treatment with 10 mmol/l 1400W led to an efficient iNOS inhibition and significantly suppressed NO secretion: either that one in PT45-P1res cells endogenously induced in an IL-1b-dependent fashion or in PT45-P1 cells induced by exogenously administered IL-1b (data not shown). Along with iNOS inhibition, 1400W increased the etoposide-induced cleavage of the caspases-3, -7, -8 and -9 (Figure 5a) as well as caspase-3/7 activity (Figure 5b) in IL-1b-treated PT45-P1 cells and in PT45-P1res cells, as shown by Western blot analysis and caspase-3/7 assay, respectively. In PT45-P1 cells, IL-1b-diminished etoposide-induced caspase-3/7 activity from 4.4- to 2.8-fold, an effect that was completely abolished by 10 mmol/l 1400W (Figure 5b), and in PT45-P1res cells, treatment with 1400W significantly increased etoposide-induced caspase-3/7 activity (from 2.1- to 4-fold). Furthermore,


1

mRNA amount (arb. units)

3977

a

0.8

PT45-P1 PT45-P1res

PT45-P1: actin PT45-P1: iNOS

PT45-P1res: actin PT45-P1res: iNOS

0.6 0.4

a 0.2 0

actin

iNOS

-

3h

6h

24h

-

6h

1400W :

-

-

PT45-P1res + +

- - - - +

IL1RA

4 3.5

µmol NO/105 cells

24h

PT45-P1 + + + + + +

-

Eto : IL1β : IL1β

b

during etoposide treatment, 1400W significantly elevated the number of annexinV-positive PT45-P1 cells preincubated with IL-1b, as well as of annexinV-positive PT45-P1res cells (Figure 5c). These data provide clear evidence that IL-1b-dependent NO secretion accounts

procaspase 3

(35 kDa) >

cleaved caspase 3

(29 kDa) >

PT45-P1 PT45-P1res

3

cleaved caspase 3 (17/19 kDa) >

2.5

procaspase 7

2

(35 kDa) >

1.5

cleaved caspase 7 (20 kDa) >

1 0.5

procaspase 8 3h 6h 24h -

-

3h 6h 24h

Figure 4 Drug-induced IL-1b level induce secretion of NO by PT45-P1res cells. (a) RNA samples from PT45-P1 and PT45-P1res cells were subjected to reverse transcription and subsequent realtime PCR using primers specific for iNOS. PT45-P1 cells were either treated with 20 ng/ml IL-1b for the indicated periods or not, and PT45-P1res cells were either treated with 250 ng/ml IL1-RA for the indicated periods or not. In parallel, real-time PCR was conducted for b-actin, which was used as a control. Results from one representative selection procedure are shown. Data are expressed as amount of mRNA in arbitrary units. Each sample was measured in duplicates. (b) Cells were treated as above, supernatants were cleared and subjected to a NO immunoassay. The amount of NO was normalized to equal cell number, which was determined in parallel (expressed as mmol NO/105 cells). Data from one representative experiment of three experiments are shown.

Figure 5 IL-1b-dependent NO secretion leads to decreased activation of caspases and chemoresistance of PT45-P1res cells. (a) PT45-P1 and PT45-P1res cells were either left untreated or treated with 20 mg/ml etoposide for 48 h. As indicated, PT45-P1 and PT45-P1res cells were additionally treated with 20 ng/ml IL-1b and with 10 mmol/l 1400W, respectively, 1 h before and 24 h after etoposide. Cellular lysates were submitted to Western blotting using antibodies for the detection of full length and cleaved caspase-3, -7, -8 and -9. An a-tubulin antibody was used as a control for equal protein load. (b) Cells were treated as above and analysed for caspase-3/7 activity expressed as n-fold induction over basal. Means7s.d. from three independent experiments are shown. (c) PT45-P1 and PT45-P1res cells were either left untreated or treated with 20 mg/ml etoposide for 24 h. As indicated, PT45-P1 cells were additionally treated with 20 ng/ml IL-1b, 10 mmol/l 1400W alone or 1 h before etoposide treatment. PT45-P1res cells were additionally treated with 10 mmol/l 1400W alone or 1 h before etoposide. Then, cells were annexinV/PI stained and analysed by fluorescence flow cytometry. Data are presented as % annexinVpositive cells over basal level. Basal apoptosis was 10–15% in PT45-P1 cells and 12–20% in PT45-P1res cells. Means7s.d. from three independent experiments are shown. *Indicates Po0.05.

(57 kDa) >


procaspase 9

(47 kDa) >

cleaved caspase 9 (35 kDa) > cleaved caspase 9 (17/19 kDa) > α-tubulin (55 kDa) >

b

Induction of caspase 3/7 activity (n-fold of basal)

-

PT45-P1res

PT45-P1

7

*

*

6 5 4 3 2 1 0

etoposide : β: IL1β 1400W :

+

-

c

+ +

+ + +

-

+

+

-

-

+

PT45-P1res

PT45-P1

* % AnnexinV-positive cells (over basal)

0 + IL1β + IL1RA

*

50 40 30 20 10

0 etoposide : β: IL1β 1400W :

+

-

+ +

-

+ + +

+

-

+

-

+ Oncogene


3978

a

procaspase 3

procaspase 7

+

+

-

-

+

PT45-P1res

-

- + - + - - - - - + +

+ + +

+ +

(35 kDa) >

(35 kDa) >

cleaved caspase 7 (20 kDa) >

procaspase 8

(57 kDa) >


procaspase 9

(47 kDa) >


α-tubulin (55 kDa) >

b

One major cause of the poor outcome of advanced pancreatic adenocarcinoma is the notoriously chemo-

Oncogene

-


Discussion

Figure 6 NO-induced caspase inhibition is abrogated under reducing conditions. (a) PT45-P1 and PT45-P1res cells were either left untreated or treated with 20 mg/ml etoposide for 48 h. As indicated, PT45-P1 and PT45-P1res cells were additionally treated with 200 mmol/l SNAP and with 200 mmol/l DTT, respectively, 1 h before and 24 h after etoposide. Cellular lysates were submitted to Western blotting using antibodies for the detection of full length and cleaved caspase-3, -7, -8 and -9. An a-tubulin antibody was used as a control for equal protein load. (b) Cells were treated as above and analysed for caspase-3/7 activity expressed as n-fold induction over basal. Means7s.d. from three independent experiments are shown. (c) PT45-P1 and PT45-P1res cells were either left untreated or treated with 20 mg/ml etoposide for 24 h. As indicated, PT45-P1 cells were additionally treated with 200 mmol/l SNAP, 200 mmol/l DTT alone or 1 h before etoposide treatment. PT45P1res cells were additionally treated with 200 mmol DTT alone or 1 h before etoposide. Then, cells were annexinV/PI stained and analysed by fluorescence flow cytometry. Data are presented as% annexinV-positive cells over basal level. Basal apoptosis was 10– 15% in PT45-P1 cells and 12–20% in PT45-P1res cells. Means7s.d. from three independent experiments are shown. *Indicates Po0.05.

PT45-P1

-

DTT 48h : SNAP 48h : etoposide 48h :

Induction of caspase 3/7 activity (n-fold over basal)

NO-induced caspase inhibition is abrogated under reducing conditions To further elucidate the mechanism by which NO inhibits caspase activation, PT45-P1res cells were pretreated with 200 mmol/l of the reducing agent dithiothreitol (DTT), an established inhibitor of S-nitrosylation (Li et al., 1997; Kim et al., 2002), to check whether etoposide-induced caspase activation and apoptosis could be increased. Furthermore, the NO-donor SNAP (200 mmol/l) was added in the absence or presence of DTT to PT45-P1 cells before etoposide treatment, and caspase activation as well as apoptosis were analysed. As shown by Western blot analysis and caspase-3/7 assay, SNAP reduced the amount of cleaved (activated) caspases-3, -7, -8 and -9 (Figure 6a) as well as the activity of caspase-3/7 (Figure 6b) in PT45-P1 cells subjected to etoposide treatment. This caspase-suppressive effect by SNAP was abolished by DTT. Accordingly, DTT treatment increased etoposide-induced cleavage of these four caspases in PT45-P1res cells (Figure 6a), as well as the level of caspase-3/7 activity (Figure 6b). AnnexinV staining revealed a reduced number of apoptotic PT45-P1 cells (27 versus 47%) in response to etoposide if SNAP was added before (Figure 6c), indicating a NO-dependent gain of chemoresistance. Consequently, DTT abolished chemoresistance induced either by SNAP in PT45-P1 cells or by endogenous NO in PT45-P1res cells (Figure 6c). These data demonstrate that NO, either derived from the NO-donor SNAP or from IL-1b-induced iNOS, suppresses caspase activation and activity in PT45-P1 cells most likely by S-nitrosylation, thereby confering chemoresistance.

resistant phenotype of PDAC cells precluding any curative option of conventional treatment. In order to improve current therapy, a better understanding is needed of how pancreatic carcinoma cells become chemoresistant, either by intrinsic mechanisms (innate chemoresistance), or by cytotoxic drug treatment itself (acquired chemoresistance). Our previous work using an in vitro coculture model demonstrated that the tumor microenvironment essentially contributes to the gain of innate chemoresistance in pancreatic cancer cells. We

7

PT45-P1

PT45-P1res

*

6

*

*

5 4 3 2 1 0

etoposide SNAP DTT

+ -

+

c

+ -

+ -

PT45-P1 % AnnexinV-positive cells (over basal)

for decreased caspase activation and chemoresistance in PT45-P1res cells.

60

+ +

+

+ -

+ +

PT45-P1res

*

*

50 40

*

30 20 10 0

etoposide SNAP DTT

+ -

+

+ -

+ -

+ +

+

+ -

+ +


3979

could show that intense interactions of the tumor cells with surrounding stromal fibroblasts provide a mutual amplification loop of IL-1b and NO that confers chemoresistance (Muërko¨ster et al., 2004). The aim of this study was to elucidate the mechanisms leading to acquired chemoresistance of pancreatic carcinoma cells and to identify molecular targets in common with innate chemoresistance that might be of substantial benefit for improved therapeutical strategies. For this purpose, the PDAC cell line PT45-P1, already used in the in vitro coculture model, was exposed to a low dose of etoposide over 6 weeks, reflecting the exposure of tumor cells to a chemotherapy regimen. Although gemcitabine is currently the standard agent in the treatment of pancreatic cancer, the impact of this drug in terms of tumor remission and survival time is rather marginal. On the basis of our previous in vitro and in vivo studies demonstrating etoposide as a more effective drug in overcoming chemoresistance in various pancreatic cancer cells than gemcitabine (Arlt et al., 2001, 2003; Muërko¨ster et al., 2003), the former drug was mainly used throughout our experiments. PT45-P1 cells that passed through this treatment showed a stable chemoresistant phenotype and were termed PT45-P1res cells. Among several differentially expressed genes in these cells, IL-1b was most significantly upregulated as compared to the parental cell line, as identified by cDNA microarray analysis (not shown) and confirmed by real-time PCR and ELISA. We and others have previously demonstrated that IL-1b is an important determinant of chemoresistance in PDAC cells (Arlt et al., 2002; Muërko¨ster et al., 2004) and in leukemia (Turzanski et al., 2004) or glioma cells (Poppenborg et al., 1999) as well. Consequently, inhibition of IL-1b signaling in these cells leads to the sensitization towards several cytostatic drugs. Reminiscent of the coculture model, after long-term exposure with etoposide, an IL-1b-dependent loop emerged also in PT45-P1res cells secreting elevated amounts of IL-1b on the one hand, and exhibiting upregulated iNOS expression and an elevated release of NO on the other hand. Thus, the innate as well as the acquired chemoresistance in PDAC cells are both characterized by these two cellular mediators. In addition, another common mediator closely related to IL-1b and NO is NF-kB, which is constitutively activated in the course of innate as well as acquired chemoresistance (Muërko¨ster et al., 2004, 2005). Furthermore, we provide evidence for a mechanism by which constitutive IL-1b secretion contributes to chemoresistance in pancreatic carcinoma cells. IL-1binduced NO leads to broad inhibition of caspases in PT45-P1res cells and hence, to a decreased sensitivity towards cytotoxic drug treatment. Apparently, this inhibition is caused by cystein S-nitrosylation of caspases as DTT (Li et al., 1997; Kim et al., 2002) reversed the NO-dependent caspase inhibition. As already shown for cardiomyocytes (Maejima et al., 2005) or hepatocytes (Li et al., 1997), NO and the NO-derived species NO, ONOO and NO þ can lead to S-nitrosylation of the cysteine residue in the active site of several caspases,

such as caspase-1,-2, -3, -4, -6, -7 and -8. This redox modification of the caspases results in the inhibition of their proteolytic activation and a direct suppression of their activity (Jaiswal et al., 2001). Indeed, while exhibiting similar expression levels of procaspase-3, -7, -8 and -9 compared to PT45-P1 cells, PT45-P1res cells displayed a strongly reduced etoposide-dependent cleavage and activity of these four caspases. The NOdependence of this effect was confirmed by our observation that the iNOS inhibitor 1400W abolished caspase inhibition and chemoresistance in PT45-P1res cells as well as in IL-1b treated PT45-P1 cells. Furthermore, treatment of PT45-P1 cells with the NO-donor SNAP (Stamler, 1994) resulted also in a blockade of druginduced caspase activation. In support of the idea that direct S-nitrosylation of the caspases might account for the reduced caspase activation in PT45-P1res cells as well as in SNAPtreated PT45-P1 cells, the strong reducing agent DTT effectively reversed caspase inhibition and chemoresistance induced by SNAP in PT45-P1 cells or by endogenous NO in PT45-P1res cells. Of course, the mode of action by which NO causes chemoresistance in PDAC cells may not only rely on the direct modification of caspases but also on indirect alterations of the balance between cell survival and cell death (Kolb, 2000). NO might induce the expression of antiapoptotic proteins such as heat-shock proteins via NF-kB activation, thereby providing antiapoptotic protection (Sreedhar and Csermely, 2004). Furthermore, increased iNOS expression and NO levels might contribute to the stimulation of angiogenesis by inducing VEGF. This growth-stimulating mechanism together with the apoptosis-inhibiting effect thereby enhance the survival of genetically altered cells and promote carcinogenesis (Jaiswal et al., 2001). Interestingly, the induction and manifestation of innate and acquired chemoresistance similarly involve a constitutive activation of NF-kB, an elevated release of IL-1b and an enhanced secretion of NO, thus underlining the important and common role of these signaling molecules in the gain of both modes of chemoresitance in pancreatic cancer. As NF-kB and IL-1b are both involved in the induction of iNOS and NO, these factors seem to be appropriate molecular structures that can be targeted to impair tumor progression and to increase chemosensitivity.

Materials and methods Cell lines and culture The human PDAC cell line PT45-P1 and its handling were described previously (Kalthoff et al., 1993). PT45-P1 and PT45-P1res cells were cultured (371C, 5% CO2, 85% humidity) in RPMI40-medium (PAA-Laboratories, Co¨lbe, Germany) supplemented with 1% glutamine (Life Technologies, Eggenstein, Germany) and 10% FCS (Biochrom KG, Berlin, Germany). PT45-P1res cells were derived from PT45-P1 cells cultured in the presence of 0.2 mg/ml etoposide for 6 weeks, daily subjected to medium exchange and etoposide treatment. After etoposide treatment, cells were either stored in liquid Oncogene


3980 nitrogen or were cultured under standard conditions and used for further experiments for 10 weeks. Three independent drug selection experiments were performed, each time yielding cells with identical phenotypical changes. Results shown in the paper were performed with PT45-P1res cells obtained from one of these selection procedures. Reagents Recombinant human IL-1b, the IL-1 receptor antagonist (IL1RA) and recombinant human Trail were obtained from R&D Systems (Wiesbaden, Germany). S-Nitroso-N-acetyl-D,L-penicillamine (SNAP) was purchased from Alexis (Gru¨nberg, Germany) and dithiothreitol (DTT) from Active Motif (Rixensart, Belgium). The iNOS inhibitor 1400W was obtained from Calbiochem (via Merck Biosciences, Schwalbach/Ts, Germany). Etoposide was obtained from Bristol Myers Squibb (Mu¨nchen, Germany), and sulfasalazine from Sigma-Aldrich Chemie (Taufkirchen, Germany). The anti-Fas (CH11) antibody was purchased from Upstate via Biomol (Hamburg, Germany). ELISA-based NF-kB assay NF-kB activation was analysed by the TransAM NF-kB-kit (Active Motif) using nuclear extracts as described previously (Muërko¨ster et al., 2005). For each sample, 20 ml of nuclear extracts containing 5 mg protein (determined by the Dc-Protein assay, BioRad; Mu¨nchen, Germany) were used according to the manufacturer’s instructions. For the detection of activated NF-kB, antibodies against the p65/RelA- or p50-subunit were used, followed by a secondary antibody conjugated to horseradish peroxidase. For the colorimetric readout (450 nm), an ELISA plate-reader was used. ELISAs Cells/well, 1 105, were grown in a six-well culture plate for 24 h. Then, medium was replaced and supernatants were taken 48 h later and precleared by centrifugation (5000 r.p.m., 10 min) before analysis. Human IL-1b and NO secreted into cell culture supernatants were quantified using the QuantikineHS human IL-1b immunoassay and the Total nitric oxide (NO) colorimetric assay, respectively (both from R&D Systems). The assays were performed following the manufacturer’s instructions. Concentrations of measured IL-1b and NO were normalized to the cell numbers determined in parallel. Measurement of apoptosis and caspase activity Cells/well, 1 105, were grown in a six-well culture plate for 24 h. Then, medium was replaced and cells were left untreated or treated as indicated. Apoptosis was determined by staining with annexinV/PI (Biocarta, Hamburg, Germany) according to the manufacturer’s instructions. Analysis was performed by fluorescence flow cytometry (GalaxyArgon Plus; DAKO Cytomation, Hamburg, Germany), using the FLOMAX software. Cells exhibiting high specific annexinV staining were regarded as apoptotic. For the detection of caspase-3/7 activity, a homogeneous luminiscent assay was performed according to the manufacturer’s instructions (Promega, Mannheim, Germany). All samples were measured in duplicates. Cell cycle analysis Cells were seeded (3–5 105 cells/well) into a six-well culture plate. After 24 h, medium was replaced and cells were cultured for 24 h. Then, cells were detached by trypsinization and Oncogene

washed twice in ice-cold PBS þ 5 mmol/l EDTA. After fixation in 100% ethanol (20 min at RT) and subsequent centrifugation, cells were resuspended in 500 ml PBS-EDTA containing 37.5 mg of RNAse-A (Sigma-Aldrich Chemie) and incubated for 30 min at RT. Finally, 500 ml of 250 mg/ml PI (SigmaAldrich Chemie) were added and PI-uptake was determined by fluorescence flow cytometry. Cell cycle analysis was performed using the FLOMAX software. Determination of viable cell number PT45-P1 and PT45-P1res cells (2 105 cells/well) were grown in a six-well culture plate for 72 h. The number of viable cells was determined 24, 48 and 72 h after seeding by cell counting. Before counting, cells were stained with trypane blue. Data are expressed as cell number (n 105). Western blotting Cells/well, 5–10 106, were seeded into a six-well culture plate and further treated as indicated. Then, cells were washed once with PBS and lysed with 1 volume of 2 SDS sample buffer (128 mmol/l Tris-Base, 4.6% SDS, 10% glycerol). Samples were heated for 5 min at 951C and put on ice for 2 min. Protein concentrations were determined using the Dc Protein assay (BioRad). Ten microgram of protein adjusted to an appropriate volume of SDS sample buffer containing 0.2 mg/ml bromphenolblue (Serva, Heidelberg, Germany) and 2.5% b-mercaptoethanol (Biomol, Hamburg, Germany) were submitted to electrophoresis on a 4–20% ProGel-Tris-glycin-gel (Anamed, Darmstadt, Germany) and immunoblotting was performed as described previously (Arlt et al., 2001). For detection of caspases, the following antibodies (Cell Signaling via New England Biolabs, Frankfurt aM, Germany) were used at the indicated dilutions: anticaspase-3 (proenzyme), 1:1000 in 5% nonfat milk powder in 0.05% TBS-Tween (blotto-TBST); anticleaved caspase-3, 1:250 in blotto-TBST; anticaspase-7 (pro- and cleaved enzyme), 1:1000 in blotto-TBST; anticaspase-8 (pro- and cleaved enzyme), 1:1000 in 5% BSA (Serva) in TBS-Tween; anticaspase-9 (proenzyme), 1:1000 in blotto-TBST; anticaspase-9 (cleaved), 1:250 in blotto-TBST. Detection of the IAP was performed using a goat anti-XIAP and a goat anti-cIAP1 antibody (both from R&D Systems) and a rabbit anti-cIAP2 antibody (Chemicon), each diluted at 1:1000 in 5% blotto-TBST. As control of equal protein load, a monoclonal antibody for a-tubulin (Sigma-Aldrich Chemie) was diluted at 1:5000 in 5% BSA in TBST. Incubation with the primary antibodies was performed overnight at 41C. For detection of the primary antibodies, anti-rabbit, anti-goat and anti-mouse HRP-linked antibodies (Cell Signaling), respectively, were used at a dilution of 1:2000 in blotto-TBST at room temperature for 1 h. After washing in TBST, blots were developed using the LumiGlo peroxidase detection kit (Cell Signaling). Real-time PCR Two microgram of total RNA were reverse-transcribed into single-stranded cDNA, as described previously (Scha¨fer et al., 1999). Of cDNA, 1.3 ml, and 0.2 mmol/l gene-specific primers were adjusted with RNAse-free water to a volume of 10 ml. To this mixture, 10 ml of iQ SYBR-Green Supermix (BioRad) were added. Primers for the detection of iNOS were from Biosource (Ratingen, Germany) and used under the following PCR conditions: 951C/5 min; 951C/45 s, 601C/45 s, 721C/45 s for 32 cycles; 721C/10 min. Primers and PCR conditions for the detection of IL-1b were used as described previously (Arlt et al., 2002). For control, b-actin was amplified in parallel using primers from BD Biosciences Clontech. The real-time


3981 PCR was performed with a MyiQ Single Color real-time PCR Detection System (BioRad). Data were collected during annealing steps and were further analysed using the i-Cycler iQ Optical system software (BioRad). All samples were analysed in duplicates and data are expressed as amount of mRNA in arbitrary units. Statistics Data are presented as mean7s.d. and analysed by Student’s t-test. A P-value o0.05 (indicated as *in the figures) was considered as statistically significant.

Abbreviations DTT, Dithiothreitol; IL-1b, Interleukin 1 beta; IL1-RA, IL-1 receptor antagonist; NF-kB, Nuclear factor kappa B; NO, Nitric oxide; PDAC, Pancreatic ductal adenocarcinoma; SNAP, S-Nitroso-N-acetyl-D,L-penicillamine. Acknowledgements This work was supported by the German Research Society DFG Scha 677/7-2 (HS).

References Arlt A, Gehrz A, Muërko¨ster S, Vorndamm J, Kruse M-L, Fo¨lsch UR et al. (2003). Oncogene 22: 3243–3251. Arlt A, Vorndamm J, Muërko¨ster S, Yu H, Schmidt WE, Fo¨lsch UR et al. (2002). Cancer Res 62: 910–916. Arlt A, Vorndamm J, Breitenbroich M, Fo¨lsch UR, Kalthoff H, Schmidt WE et al. (2001). Oncogene 20: 859–868. Banerjee D, Mayer-Kuckuk P, Capiaux G, Budak-Alpdogan T, Gorlick R, Bertino JR. (2002). Biochim. Biophys. Acta 1587: 164–173. Blaszkowsky L. (1998). Front. Biosci. 3: 214–225. Bramhal SR, Allum WH, Jones AG, Allwood A, Cummins C, Neoptolemos JP. (1995). Br J Surg 82: 111–115. Brand RE, Tempero MA. (1998). Curr Opin Oncol 10: 362–366. Brown JM, Attardi LD. (2005). Nat Rev Cancer 5: 231–237. Erler JT, Cawthorne CJ, Williams KJ, Koritzinsky M, Wouters BG, Wilson C et al. (2004). Mol Cell Biol 24: 2875–2889. Friesen C, Fulda S, Debatin KM. (1999). Leukemia 13: 1854–1858. Jaiswal M, LaRusso NF, Gores GJ. (2001). Am J Physiol Gastrointest Liver Physiol 281: G626–634. Kalthoff H, Schmiegel W, Roeder C, Kasche D, Schmidt A, Lauer G et al. (1993). Oncogene 8: 289–298. Kang HC, Kim IJ, Park JH, Shin Y, Ku JL, Jung MS et al. (2004). Clin Cancer Res 10: 272–284. Kim K-M, Kim PKM, Kwon YG, Bai S-K, Nam W-D, Kim Y-M. (2002). J Biochem Mol Biol 35: 127–133. Kim R, Emi M, Tanabe K, Toge T. (2004). Cancer 101: 2491–2502. Cancer Res 59: 3505–351. Kolb JP. (2000). Leukemia 14: 1685–1694. Li JL, Billiar TR, Talanian RV, Kim YM. (1997). Biochem Biophys Res Commun 240: 419–424. Lillemore KD. (1998). Surg Oncol Clin North Am 7: 199–216.

Maejima Y, Adachi S, Morikawa K, Ito H, Isobe M. (2005). J Mol Cell Cardiol 38: 163–174. Medema JP, de Jong J, van Hall T, Melief CJM, Offringa R. (1999). J Exp Med 7: 1033–1038. Morisaki T, Katano M. (2003). Curr Med Chem 10: 2517–2521. Muërko¨ster S, Arlt A, Sipos B, Witt M, Gromann M, Klo¨ppel G et al. (2005). Cancer Res F 65: 1316–1324. Muërko¨ster S, Arlt A, Witt M, Gehrz A, Haye S, March C et al. (2003). Int J Cancer 104: 469–476, 2003. Muërko¨ster S, Wegehenkel K, Arlt A, Witt M, Sipos B, Kruse ML et al. (2004). Cancer Res 64: 1331–1337. Neoptolemos JP, Dunn JA, Stocken DD, Almond J, Link K, Beger H et al. (2001). Lancet 358: 1576–1585. Parker SL, Tong T, Bolden S, Wingo PA. (1997). CA Cancer J Clin 47: 5–27. Poppenborg H, Knupfer MM, Galla HJ, Ernst J, Wolff A. (1999). Cytokine 11: 689–695. Scha¨fer H, Arlt A, Trauzold A, Hu¨nermann-Jansen A, Schmidt WE. (1999). Biochem Biophys Res Commun 262: 139–145. Schmid RM, Adler G. (2000). Gastroenterology 118: 1208–1228. Soengas MS, Lowe SW. (2003). Oncogene 22: 3138–3151. Sreedhar AS, Csermely P. (2004). Pharmacol Ther 101: 227–257. Stamler JS. (1994). Cell 78: 931–936. Turzanski J, Grundy M, Russell NH, Pallis M. (2004). Leukemia 18: 1662–1670. Wang W, Abbruzzese JL, Evans DB, Larry L, Cleary KR, Chiao PJ. (1999). Clin Cancer Res 5: 119–127. Yan Y, Mahotka C, Heikaus S, Shibita T, Wethkamp N, Liebmann J et al. (2004). Br J Cancer 91: 1349–1357.

Oncogene