Carcinogenesis vol.25 no.8 pp.1377EE1385, 2004 doI:10.

1093/carcin/bgh122

Suppression of human pancreatic cancer cell proliferation by AGN194204, an RXR-selective retinoid

Sivaprakasam Balasubramanian1, Roshantha A.S.Chandraratna6 and Richard L.Eckert1EE5,7

1

Introduction Pancreatic cancer is the fifth leading cause of cancer-related death in western countries (1). Despite improvements in diagnosis, staging and treatment, the prognosis of advanced pancreatic cancer has not substantially improved during the past three decades. More than 80% of the patients have advanced regional disease or metastasis at the time of diagnosis, and the 5-year survival is 52% with a median survival time of ~5 months (1,2). Therefore, new therapeutic strategies focusing on inhibition of pancreatic cancer cell function are urgently needed. Retinoids are natural or synthetic derivatives of vitamin A. They mediate their biological effects by binding to nuclear, ligand-activated receptors (3,4). Upon ligand binding, these receptors bind to specific DNA elements to regulate transcription (5EE7). These receptors are divided into the retinoic acid receptor (RAR) and retinoid X receptor (RXR) families. Each family includes a, b and g receptor subtypes. RARs display a high affinity for all-trans-retinoic acid (atRA) and related ligands, while RXRs bind 9-cis-retinoic acid (9-cRA) and other RXR ligands (6,8,9). An important anticancer strategy is the design of retinoids that have high efficacy and low toxicity for the treatment of disease. In particular, investigators have designed ligands that target various subclasses of the retinoid receptors (10EE13). Various retinoids have been shown to be effective in pancreatic cancer cell models. These agents act to inhibit cell proliferation, induce differentiation and promote apoptosis (14EE17). However, treatment of human cancers with retinoids has been limited by significant toxicities including mucocutaneous irritation, hypertriglyceridemia and teratogenicity (18,19). Therefore, identification of new retinoids having better therapeutic index with less toxicity is an important goal. It is also appreciated that it would be desirable to have retinoids that work efficiently in conjunction with existing therapeutic agents. RXR-selective retinoids are useful candidate therapeutic agents, as they may display less toxicity than RARselective retinoids. In the present study, we examine the ability of a novel, high affinity, highly selective, RXR-retinoid, AGN194204, to inhibit the function of pancreatic cancer cells. Our results suggest that RXR-selective retinoids are more efficient inhibitors of pancreatic cell function than RAR-selective retinoids, but that not all pancreatic cancer cell lines are retinoid sensitive. Materials and methods
Reagents atRA, 9-cRA, 13-cis-retinoic acid (13-cRA) and cisplatin were purchased from Sigma (St Louis, MO). AGN194204, AGN193109, AGN195393, SR11217 and TTNPB (Table I) were synthesized in the Department of Chemistry, Allergan. Retinoids were prepared as 1000-fold stocks in dimethylsulfoxide and stored in aliquots at 70 C. Recombinant human interferon-gamma (IFNg), prepared in 10 mM acetic acid containing 0.1% human serum albumin

Department of Physiology and Biophysics, 2Department of Dermatology, 3 Department of Biochemistry, 4Department of Reproductive Biology and 5 Department of Oncology, Case Western Reserve University School of Medicine, Cleveland, OH 44106-4970, USA and 6Department of Chemistry, Allergan, Inc., Irvine, CA 92623, USA
7

To whom correspondence should be addressed Email: rle2@po.cwru.edu

Downloaded from http://carcin.oxfordjournals.org/ at Universiti Teknologi Malaysia on March 26, 2012

Retinoids may be useful agents for the treatment of pancreatic cancer. However, retinoic acid receptor (RAR)selective retinoids produce unwanted side effects. In contrast, retinoid X receptor (RXR)-selective retinoids produce fewer side effects; however, it was not known whether RXR-selective retinoids could reduce pancreatic tumor cell proliferation. In the present study, the novel RXR-selective retinoid, AGN194204, was compared with that of other retinoids for the ability to suppress pancreatic cancer cell proliferation. We treated various pancreatic cancer cell lines with receptor-selective ligands and cytotoxic agents and monitored the effects on cell proliferation, markers of apoptosis and cell cycle. Our results indicate that AGN194204, at concentrations 410 nM, inhibits proliferation of MIA PaCa-2 and BxPC-3 cells but not the proliferation of AsPC-1 cells. Moreover, in BxPC-3 and MIA PaCa-2 cells, AGN194204 was 10EE100 times more effective than RAR-selective retinoids. AGN194204dependent suppression of MIA PaCa-2 cell proliferation is associated with reduced cyclin E and cyclin-dependent kinase 6 (cdk6) level, but cyclin D1, cdk2 and cdk4 content is not altered. In addition, p27 level increases 2-fold. The RXR-selective antagonist, AGN195393, reverses the AGN194204-dependent growth inhibition and the decline in cyclin E and cdk6 levels. In contrast, these changes are not reversed by treatment with the RAR antagonist, AGN193109. AGN194204 did not appear to alter cell apoptosis as measured by change in cleavage of procaspase-3, -8 or -9. We also examined the effects AGN194204 co-treatment with cytotoxic agents. Treatment of MIA PaCa-2 cells with AGN194204 cisplatin, gemcitabine, 5-flurouracil, interferon (IFN)a or IFNg resulted in an additive but not synergistic reduction in MIA PaCa-2 cell number. These results indicate that AGN194204, an RXR-selective retinoid, is a more effective inhibitor of pancreatic cell proliferation than the RAR-selective retinoids, and further indicate that AGN194204 produces an additive reduction in cell number when given with other agents. Our results suggest that RXR-selective ligands, which are less toxic than RAR-selective ligands, may be suitable agents for the treatment of pancreatic cancer.
Abbreviations: atRA, All-trans-retinoic acid; cdk, cyclin-dependent kinase; 9-cRA, 9-cis-retinoic acid; 13-cRA, 13-cis-retinoic acid; 5-FU, 5-fluorouracil; RAR, retinoic acid receptors; RXR, retinoid X receptors. Carcinogenesis vol.25 no.8 # Oxford University Press 2004; all rights reserved.

1377

S.Balasubramanian, R.A.S.Chandraratna and R.L.Eckert

Table I. Receptor binding affinity of retinoid ligands Compound Receptor affinity, nM RARa AGN194204 TTNPB SR11217 AGN193109 AGN195393 410 4 20 410 4 1.5 793 RARb 410 4 39 410 4 1.9 950 RARg 410 4 51 410 4 3.3 1498 RXRa 2 9113 388 410 4 1 RXRb 2 4093 121 410 4 1 RXRg 2 2556 915 410 4 2

cells were then analyzed by flow cytometry. Cell cycle analysis was performed three separate times with similar results in each case.

Results Retinoids Table I lists the binding affinity of each of the retinoids used in this study. These ligands were selected to include a potent RAR-selective retinoid (TTNPB) (20) that has a low affinity for RXR receptors, a moderate affinity RXR-selective retinoid (SR11217), and a high affinity RXR-selective retinoid (AGN194204) (11). The table also lists the binding affinities of an RAR-selective antagonist (AGN193109) (21EE23) and an RXR-selective antagonist (AGN195393). AGN194204 regulation of pancreatic cancer cell proliferation To test the effect of AGN194204 on pancreatic cancer cell proliferation, cells were treated with increasing concentrations of AGN194204, TTNPB or SR11217 (0EE10 mM) for 6 days. As shown in Figure 1, AsPC-1 cell number was modestly reduced (20%) by all three retinoids. BxPC-3 cell number was reduced 40% by treatment with 1 mM AGN194204, 20% by treatment with 1 mM SR11217, and 10% by treatment with TTNPB. MIA PaCa-2 cell number was reduced 60% by treatment with 1 mM AGN194204, and 20% by treatment with 1 mM SR11217 or TTNPB. These results indicate that AGN194204-dependent suppression of cell number varies for different pancreatic cell lines, and that the RXR-selective AGN194204 is, overall, a more effective inhibitor of pancreatic cancer cell proliferation than the RAR-selective, TTNPB. The partially RAR-selective ligand, SR11217, has an intermediate efficiency. AGN194204-dependent suppression of cell number is mediated by RXR To confirm that the AGN194204-dependent suppression is mediated by RXR, we monitored the ability of RAR and RXR-selective antagonists to inhibit the AGN194204 response. MIA PaCa-2 cells were incubated with either 100 nM TTNBP (RAR-selective agonist) or 100 nM AGN194204 (RXRselective agonist) in the presence of increasing concentrations of the RAR-selective antagonist, AGN193109. After 6 days of treatment, the cells were harvested and counted. As shown in Figure 2A, addition of TTNBP results in a 20% reduction in cell number and AGN193109 produces a concentration-dependent reversal of this inhibition. In contrast, AGN193109 does not reverse the 40% reduction in cell number produced by AGN194204. The experiment shown in Figure 2B indicates that addition of AGN195393, an RXR-selective antagonist, reverses the AGN194204-dependent inhibition of cell number, but that this agent does not reverse the TTNBP-dependent reduction. These findings suggest that the TTNBP-dependent inhibition is mediated exclusively by RAR subtypes, and that the AGN194204-dependent inhibition of cell number is mediated exclusively by RXR subtypes. AGN194204 modifies the level of some cell cycle regulators We next determined whether AGN194204 influences cell cycle regulatory protein expression in MIA PaCa-2 cells. This analysis shows that cyclin E and cdk6 levels are suppressed, to 40 and 10% of control level, respectively, after 48 h of treatment with 1 or 10 mM AGN194204 (Figure 3A). In contrast, the levels of cyclin D1, cdk2 and cdk4 are not modified. In addition, the level of the cdk inhibitor, p27, is
Downloaded from http://carcin.oxfordjournals.org/ at Universiti Teknologi Malaysia on March 26, 2012

The relative binding affinity of each retinoid was determined for the RAR and RXR subtypes using baculovirus-expressed RAR and RXRa, b and g as described previously (39). was purchased from R&D Systems. Interferon-alpha (IFNa), prepared in phosphate-buffered saline, was obtained from Biochemical. Mouse monoclonal antibodies specific for p27 (sc-1641), p21 (sc-6246), rabbit polyclonal antibodies specific for cyclin E (sc-481), cyclin-dependent kinase (cdk)2 (sc-163), cdk4 (sc-601), cdk6 (sc-7181) caspase-9 (sc-8355) and goat polyclonal anti-caspase-8 (sc-6136) were obtained from Santa Cruz Biotechnology (Santa Cruz, CA). Rabbit polyclonal anti-caspase-3 was purchased from Biosource (AHZ0052). Mouse monoclonal anti-b-actin was obtained from Sigma (A5441) and anti-cyclin D1 (554180) was from BD PharMingen. Cell lines and culture conditions Human pancreatic cancer cell lines MIA PaCa-2, BxPC-3 and AsPC-1 were obtained from American Type Culture Collection (ATCC, Rockville, MD). MIA PaCa-2 cells were maintained in Dulbecco's Modified Eagle's Medium (DMEM) supplemented with 4 mM L-glutamine, 100 U/ml penicillin, 100 U/ml streptomycin, 10% fetal bovine serum (FBS) and 2.5% horse serum. BxPC-3 cells were grown in RPMI supplemented with 4 mM L-glutamine 100 U/ml penicillin, 100 U/ml streptomycin, 1 mM sodium pyruvate, 10 mM HEPES, 10% FBS and 4.5 g/l glucose. AsPC-1 cells were maintained in RPMI supplemented with 2 mM L-glutamine, 100 U/ml penicillin, 100 U/ml streptomycin, 1 mM sodium pyruvate, 20 mM HEPES, 20% FBS and 4.5 g/l glucose. Proliferation assays Pancreatic cancer cells were seeded into 9.6 cm2 dishes at 1000 cells/cm2 for MIA PaCa-2, and 3000 cells/cm2 for BxPc-3 and AsPC-1 in complete medium and allowed to attach for 24 h. The cells were treated by the addition of fresh medium containing an appropriate concentration of retinoid. Treatment was continued for 6 days with addition of fresh retinoid-containing medium on alternate days. After 6 days, the cells were harvested with 0.025% trypsin, 1 mM EDTA and counted using a coulter counter. The cell counts are expressed as the mean standard error of the mean for a minimum of three repeats. Co-treatment with cytotoxic agents MIA PaCa-2 cells were plated at 1000 cells/cm2 in 9.6 cm2 dishes and allowed to attach overnight. The cells were given fresh DMEM medium containing 0EE1 mg cisplatin/ml, 0EE4000 IU IFNa/ml, or 0EE200 IU IFNg/ml, 0EE12.5 nM gemcitabine, or 0EE20 mM 5-fluorouracil delivered in combination with 0EE10 mM AGN194204. The final concentration of the DMSO vehicle did not exceed 0.2% in the culture medium, and all control dishes received the appropriate concentration of vehicle. After 6 days, with fresh treatment on alternate days, the cells were harvested with trypsinEEEDTA and counted. Western blot analysis For immunoblot, an equivalent amount of protein was electrophoresed on denaturing 8EE10% polyacrylamide gels, transferred to nitrocellulose membrane, and the membrane was incubated in 10 mM TrisEEHCl pH 7.2 containing 100 mM NaCl, 0.1% Tween-20 and 5% non-fat dry milk to block non-specific binding. After incubation with primary antibody, the blot was washed and exposed to the appropriate horseradish peroxidase-conjugated secondary antibody. Specific antibody binding was visualized using chemiluminescence detection reagents (Amersham Biosciences). Cell cycle analysis MIA PaCa-2 cells were grown in DMEM with supplements as above. When 30% confluent, the cells were treated with DMSO or 1 or 10 mM AGN194204 for 24 or 48 h. The cultures were trypsinized, washed with phosphate-buffered saline, and collected by centrifugation. The cells were re-suspended in phosphate-buffered saline and fixed in cold methanol. After an additional wash in phosphate-buffered saline, the cells were incubated with RNase (10 mg/ml) at 37 C for 30 min, and stained with 50 mg/ml propidium iodide for 60 min. The

1378

Pancreatic cancer regulation by retinoids

increased 2-fold following treatment with 10 mM AGN194204. p21 was not detected. Thus, AGN194204 changes the expression level of only selected cell cycle-associated regulators. RXR antagonist reversal of AGN194204-dependent changes in cyclin E, cdk6 and p27 level As noted above, the RXR antagonist, AGN195393, can reverse the AGN194204-dependent reduction in cell number. To determine whether the AGN194204-dependent modification of cell cycle regulatory protein level can be reversed by this ligand, MIA PaCa-2 cells were treated with 10 mM AGN194204 in the presence of RAR-selective antagonist AGN193109, or RXRselective antagonist AGN195393. As shown in Figure 3B, AGN193109 and AGN195393 do not alter cyclin E, cdk6 or p27 level. In addition, the AGN194204-dependent suppression of cyclin E and cdk6 expression and the increase in p27 expression is not reversed by treatment with RAR antagonist. However, the RXR-selective antagonist, AGN195393, reverses these AGN194204-dependent changes. Simultaneous treatment with AGN194204 and other growth inhibitors Simultaneous treatment with retinoids and other growth inhibitory agents may be useful for the treatment of pancreatic cancer. To assess whether co-treatment results in enhanced inhibition of cell proliferation, we examined the effect of treatment with selected growth inhibitors in conjunction with AGN194204 treatment. In Figure 4A MIA PaCa-2 cells were treated with increasing concentrations of IFNa in the absence (closed circles) or presence (open circles) of 100 nM AGN194204. IFNa produced a concentration-dependent reduction in cell number with a 50% suppression at an IFNa concentration of 4000 IU/ml. We next tested the response of IFNa in the presence of 100 nM AGN194204. As shown in Figure 4A, the inclusion of AGN194204 shifted the IFNa doseEEresponse downward by 30%, but did not change the slope. Figure 4B shows a similar response for the inverse experiment in which cells were treated with increasing concentrations of AGN194204 in the absence or presence of 500 IU/ml IFNa. The addition of IFNa shifts the AGN194204 response curve downward, but does not change the slope. We also examined the effects of co-treatment of MIA PaCa-2 cells with AGN194204 and IFNg. Figure 5 shows that, similar to IFNa, IFNg produces an additive inhibition when coadministered with AGN194204. This holds whether the cells are treated at a fixed concentration of AGN194204 with increasing IFNg (Figure 5A), or treated at fixed IFNg with increasing AGN194204 (Figure 5B). Cisplatin is a cytotoxic agent that is widely used as a chemotherapeutic agent. Figure 6A shows that cisplatin treatment produces a concentration-dependent inhibition of MIA PaCa-2 cell proliferation. When combined with 100 nM AGN194204, the cisplatin suppression curve is moved lower, but the slope of the curve does not changethus the agents produce additive inhibition. The inverse experiment, using a fixed cisplatin concentration, with increasing AGN194204, reveals a similar additive inhibition (Figure 6B). Gemcitabine and 5-fluorouracil (5-FU) are also used presently to treat pancreatic cancer. Gemcitabine is at present the most effective chemotherapeutic drug for the treatment of this disease (24). As shown in Figure 7A, 5-FU treatment reduces MIA PaCa-2 cell number, while co-treatment with 1379

Downloaded from http://carcin.oxfordjournals.org/ at Universiti Teknologi Malaysia on March 26, 2012

Fig. 1. Retinoid regulation of pancreatic cancer cell number. Pancreatic cancer cells were plated as outlined in Materials and methods. After 24 h, treatment was initiated with the indicated retinoid. Treatment was continued for 6 days with the addition of fresh medium and retinoid on alternating days. At 7 days, the cells were harvested and counted. The cells were harvested with 0.025% trypsin, 1 mM EDTA and counted using a coulter counter. The error bars represent the mean SEM for three experiments. The cell density at the beginning of the experiment corresponds to 5% in each plot (i.e. ~25 000 cells/35 mm dish).

S.Balasubramanian, R.A.S.Chandraratna and R.L.Eckert

Downloaded from http://carcin.oxfordjournals.org/ at Universiti Teknologi Malaysia on March 26, 2012

Fig. 2. RXR receptors mediate the AGN194204-dependent reduction in cell number. MIA PaCa-2 cells were plated at 1000 cells/cm2 in 9.6 cm2 dishes, and the cells were permitted to attach for 24 h before treatment. (A) Cells were treated with increasing concentrations of AGN193109 without addition of a second retinoid, or with addition of 100 nM TTNPB or 100 nM AGN194204. Fresh medium and retinoid was added on alternate days, and the cells were harvested and counted on day 6. (B) Cells were treated with increasing concentrations of AGN195393 without addition of a second retinoid, or with addition of 100 nM TTNPB or 100 nM AGN194204. After 6 days, the cells were harvested and counted. Vehicle concentration (DMSO) was maintained constant in all treatment groups.

100 M AGN194204 shifts the 5-FU response curve but does not change the slope. Figure 7B indicates a similar relationship for the inverse experiment where a fixed concentration of 5-FU (2.5 mM) is used with increasing concentrations of AGN194204. Figure 8A shows a concentration-dependent inhibition of MIA PaCa-2 cell proliferation by gemcitabine. Treatment with 7.5 nM gemcitabine results in a 52% reduction in cell number. Addition of 100 nM AGN194204 shifts the gemcitabine doseEEresponse downward, but did not change the slope. Treatment with 5 nM gemcitabine and increasing concentration of AGN194204 results in a similar additive inhibition (Figure 8B). Effect of AGN194204 compared with natural retinoids Figure 9 compares the ability of AGN194204 and the other natural retinoids to suppress MIA PaCa-2 cell number. AGN194204 (10 mM) reduces cell number by 72%. atRA, 9-cRA and 13-cRA, at 10 mM, inhibit by 67, 66 and 55%, respectively. AGN194204 does not enhance MIA PaCa-2 cell apoptosis The above findings suggest that AGN194204 suppresses MIA PaCa-2 cell proliferation, in conjunction with a reduction in cyclin E and cdk6 expression, and increase in p27 level. To determine whether apoptotic processes are also involved, we treated cells with the agents as indicated in Figure 10 and then assayed for the ability of each agent to reduce procaspase-8, -9 and -3 levels. As shown in Figure 10, none of the agents tested promote enhanced cleavage of procaspase-8. With respect to procaspase-9 and -3, only 5-FU promotes a reduction in level.
Fig. 3. AGN194204 treatment of MIA PaCa-2 cells reduces cyclin E and cdk6 level. MIA PaCa-2 cells were plated at 1000 cells/cm2 in 9.6 cm2 dishes, and permitted to attach. The cells were treated with 0, 1 or 10 mM AGN194204. (A) After 48 h of treatment, the cells were harvested and cell extracts were prepared for detection of cell cycle regulatory protein and retinoid receptor expression by immunoblot. b-Actin level was monitored to normalize protein loading. Similar results were observed in each of three experiments. (B) RXR-antagonist inhibition of AGN194204-dependent changes in cell cycle protein expression. Cells were treated for 48 h with each indicated ligand or in combination (each at 10 mM). The cells were then assayed for cyclin E, cdk6, p27 and b-actin level as described above.

Discussion Retinoids as cancer therapeutics Retinoids are important regulators of cell proliferation and differentiation. As such they are important candidates for the treatment of disease conditions, including cancer (12). The retinoid receptor family includes six distinct members RARa, b and g, and RXRa, b and g (3,4,6,9,25). These

1380

Pancreatic cancer regulation by retinoids

Downloaded from http://carcin.oxfordjournals.org/ at Universiti Teknologi Malaysia on March 26, 2012

Fig. 4. Co-treatment of MIA PaCa-2 cells with AGN194204 and IFNa. MIA PaCa-2 cells were plated as 1000 cells/cm2 in 9.6 cm2 dishes and permitted to attach overnight. (A) Cells were treated with increasing concentrations of IFNa in the absence (closed circles) or presence (open circles) of 100 nM AGN194204. After 6 days of treatment, the cells were harvested and counted. (B) Cells were treated with increasing concentrations of AGN194204 in the absence (closed circles) or presence (open circles) of 500 IU/ml IFNa. After 6 days of treatment, the cells were harvested and counted.

Fig. 5. Co-treatment of MIA PaCa-2 cells with AGN194204 and IFNg. MIA PaCa-2 cells were plated as 1000 cells/cm2 in 9.6 cm2 dishes and permitted to attach overnight. (A) Cells were treated with increasing concentrations of IFNg in the absence (closed circles) or presence (open circles) of 100 nM AGN194204. After 6 days of treatment, the cells were harvested and counted. (B) Cells were treated with increasing concentrations of AGN194204 in the absence (closed circles) or presence (open circles) of 12.5 IU/ml IFNg. After 6 days of treatment, the cells were harvested and counted.

receptors cooperate to regulate cell function by forming homoand heterodimers that act to regulate cell transcription in a cell type- and receptor subtype-specific manner. Because of the tissue-selective pattern of expression and the selective role of the RAR and RXR subtypes in generating specific biological responses, retinoid development has focused on identifying retinoid analogs that selectively interact with specific retinoid receptor subtypes (10,26). The ultimate goal is the design of a specific cancer-targeted therapy with minimal side effects. RXR-selective ligands are frequently less toxic than RARselective ligands (11,27). For example, Wu et al. (28), working with C3(1)-SV40 Tag mice, a mammary tumorigenesis model, showed that a RAR-selective ligand was modestly chemopreventive and highly toxic. In contrast, LGD1069, a RXR ligand, was an effective chemopreventive agent without toxicity.

Based on this and other studies, it is hoped that RXR-selective ligands may be efficacious inhibitors of pancreatic tumor cell growth, without producing significant unwanted side effects (28,29). The major goal of the present study is to evaluate the ability of an RXR-selective retinoid, AGN194204, to inhibit the growth of several pancreatic cancer cell lines. Inhibition of pancreatic cancer cell proliferation by AGN194204 The antitumor effects of retinoids have been noted in a variety of tissues and model systems, including pancreatic cancer cells (15,30EE32). Previous reports indicate that a range of natural retinoids, including atRA, 9-cRA and 13-cRA, can inhibit pancreatic cancer cell proliferation (14EE16,31,33,34). 1381

S.Balasubramanian, R.A.S.Chandraratna and R.L.Eckert

Downloaded from http://carcin.oxfordjournals.org/ at Universiti Teknologi Malaysia on March 26, 2012

Fig. 6. Co-treatment of MIA PaCa-2 cells with AGN194204 and cisplatin. MIA PaCa-2 cells were plated as 1000 cells/cm2 in 9.6 cm2 dishes and permitted to attach overnight. (A) Cells were treated with increasing concentrations of cisplatin in the absence (closed circles) or presence (open circles) of 100 nM AGN194204. (B) Cells were treated with increasing concentrations of AGN194204 in the absence (closed circles) or presence (open circles) of 0.2 mg cisplatin/ml. After 6 days of treatment, the cells were harvested and counted.

Additional studies suggest that the response requires specific RAR receptor subtypes. Some investigators report that the retinoid-associated inhibition of pancreatic cancer cell proliferation requires the RARg receptor (14,35), while others attribute the suppression to the RARa receptor (36). For example, treating DSL-6A/C1 pancreatic adenocarcinoma cells with an RARa-selective synthetic retinoid, Ro 40-6055, suppresses cell proliferation (36)co-treatment with a synthetic RARaantagonist, Ro 41-5253, inhibits this response. Another RARa-selective synthetic retinoid, TAC-101, inhibits proliferation of BxPC-3, MIA PaCa-2 and AsPC-1 cells (37). In addition, mofarotene (Ro 40-8757), a synthetic retinoid that does not bind to or activate any retinoid receptor, also suppresses pancreatic cancer cell proliferation (38). All of these studies focus on the potential utility of RAR-selective retinoids. Thus, a potential role for RXR-targeting ligands has not been investigated. In the present study we investigate the anti-proliferative effects of two RXR receptor-selective retinoids, SR11217 and AGN194204 (39). SR11217 selectively binds, with moderate affinity, to RXR receptors, while AGN194204 binds to RXR receptors with high selectivity and affinity (39). Our studies indicate that MIA PaCa-2 cells are markedly growth suppressed by retinoid treatment. BxPC-3 cells, in contrast, are moderately retinoid-sensitive and AsPC-1 cells are the least sensitive. Among the various compounds tested, AGN194204 is the most effectivemore effective than SR11217 or the high-affinity RAR-selective retinoid, TTNPB (20). These results suggest that RXR-selective ligands, acting via RXR receptors, can suppress pancreatic cancer cell proliferation. This idea that the suppression requires RXR receptor function is supported by the observation that the AGN194204-dependent suppression of proliferation is reversed by treatment with the RXR-selective antagonist, AGN195393, but is not reversed by treatment AGN193109, a RAR-selective antagonist (11). These findings indicate that high-affinity RXR-selective ligands can efficiently suppress pancreatic cancer cell proliferation. The relative difference in sensitivity of the cell lines is not likely to be due to differences in retinoid receptor subtype 1382

expression, as most pancreatic cell lines express RXRa and RXRb (32). The response difference could be due to differing expression of co-activators and co-repressors (7,40EE42), however, this has not been examined. AGN194204 regulation of cell cycle protein expression Pancreatic cancer cells have defective G1 cell cycle control (43). To determine whether the AGN194204-dependent reduction in cell number is associated with specific changes in the level of cell cycle regulatory proteins, we measured the effect of AGN194204 on expression of these regulators. No change in cyclin D1, cdk2 or cdk4 level is observed following AGN194204 treatment. However, a reduction in cyclin E and cdk6 level is observed at 48 h of treatment. In addition, a moderate increase of p27 is observed. p21 is not detected. Treatment with AGN195393, the RXR-antagonist (44), reverses the AGN194024-dependent changes in cell cycle protein expression. Previous studies describe retinoid effects on cell cycle regulatory proteins in pancreatic cancer cells. In MIA PaCa-2 cells, TAC-101, a RARa-selective synthetic retinoid, inhibits cell proliferation leading to arrest in G1 phase (37). This arrest is associated with reduced pRB and cyclin A level, and increased p21 and p27 level (37). Another study reports that the synthetic retinoid, Ro 40-8757, inhibits pancreatic cancer cell proliferation by increasing p21 and p27 levels (16). Thus, although the protein targets are subtly different, both RAR- and RXR-selective ligands appear to target cell cycle regulatory proteins that control events in the G1 cell cycle phase. A consistent observation for both RARand RXR-selective ligands is that growth suppression correlates with increased cdk inhibitor expression (i.e. p21 and/or p27). These results anticipate arrest in the G1 phase of the cell cycle. Indeed, cell cycle analysis of untreated cells reveals 46% in G1 , 24% in G2 /M and 30% in S phase. In contrast, the distribution was 64% G1 , 18% G2 /M and 18% S phase in cells treated for 24 or 48 h with 1 or 10 mM AGN194204 (not shown). In addition to regulation of cell cycle progression, retinoids have also been shown to influence cancer cell apoptosis

Pancreatic cancer regulation by retinoids

Downloaded from http://carcin.oxfordjournals.org/ at Universiti Teknologi Malaysia on March 26, 2012

Fig. 7. Co-treatment of MIA PaCa-2 cells with AGN194204 and 5-FU. MIA PaCa-2 cells were plated as 1000 cells/cm2 in 9.6 cm2 dishes and allowed to attach overnight. (A) Cells were treated with increasing concentrations of 5-FU in the absence (closed circles) or presence (open circles) of 100 nM AGN194204. (B) Cells were treated with increasing concentrations of AGN194204 in the absence (closed circles) or presence (open circles) of 2.5 mM 5-FU. After 6 days of treatment, the cells were harvested and counted.

Fig. 8. Co-treatment of MIA PaCa-2 cells with AGN194204 and gemcitabine. MIA PaCa-2 cells were plated as 1000 cells/cm2 in 9.6 cm2 dishes and allowed to attach overnight. (A) Cells were treated with increasing concentrations of gemcitabine in the absence (closed circles) or presence (open circles) of 100 nM AGN194204. After 6 days of treatment, the cells were harvested and counted. (B) Cells were treated with increasing concentrations of AGN194204 in the absence (closed circles) or presence (open circles) of 5 nm gemcitabine. After 6 days of treatment, the cells were harvested and counted.

(45EE47). Procaspase cleavage is a commonly used marker of apoptosis. Thus, we measured the ability of the RAR- and RXR-selective retinoids and the various cytotoxic agents to promote cleavage of procaspase 3, 8 and 9. Our studies indicate that neither RAR- nor RXR-selective retinoid enhance apoptosis, as measured by enhanced procaspase cleavage. Thus, at least in MIA PaCa-2 cells, the main regulation appears to be regulation of cell cycle-associated events.

Regulation of MIA PaCa-2 cell proliferation by AGN194204 in combination with cytotoxic agents Clinical experience indicates that pancreatic cancer is not controlled by chemotherapy or other conventional treatments (30). Because it is possible that co-therapy with cytotoxic agents and retinoids may improve the activity of chemotherapeutic anticancer agents, we evaluated the ability of AGN194204 to enhance the action of various cytotoxic 1383

S.Balasubramanian, R.A.S.Chandraratna and R.L.Eckert

Our results indicate that IFNa inhibits MIA PaCa-2 cell proliferation, but not as efficiently as IFNg. Gemcitabine, cisplatin and 5-FU treatment also produced a concentrationdependent reduction in cell number. Moreover, the combination of AGN194204 with IFNa, IFNg, gemcitabine, cisplatin or 5-FU enhances inhibition in an additive, but not synergistic, manner. This is in contrast to a report that suggests that IFNa can act synergistically with retinoids (19). On balance, our present studies suggest that RXR-selective retinoids may provide a less-toxic alternative to RAR-selective ligands for the treatment of pancreatic cancer. Acknowledgements
This work utilized the facilities of the Skin Diseases Research Center of Northeast Ohio (NIH, AR39750) and was supported by Allergan Inc., and American Institute for Cancer Research (R.L.E.).

Downloaded from http://carcin.oxfordjournals.org/ at Universiti Teknologi Malaysia on March 26, 2012

References
Fig. 9. Regulation of MIA PaCa-2 cell number by natural retinoids. MIA PaCa-2 cells were plated as 1000 cells/cm2 in 9.6 cm2 dishes, allowed to attach overnight, and then treated with the indicated concentrations of atRA, 9-cRA, 13-cRA and AGN194204. After 6 days of treatment, the cells were harvested and counted. 1. Arbuck,S.G. (1990) Chemotherapy for pancreatic cancer. Baillieres Clin. Gastroenterol., 4, 953EE968. 2. Rosenberg,L. (2000) Pancreatic cancer: a review of emerging therapies. Drugs, 59, 1071EE1089. 3. Petkovich,M., Brand,N.J., Krust,A. and Chambon,P. (1987) A human retinoic acid receptor which belongs to the family of nuclear receptors. Nature, 330, 444EE450. 4. Brand,N., Petkovich,M., Krust,A., Chambon,P., de The,H., Marchio,A., Tiollais,P. and Dejean,A. (1988) Identification of a second human retinoic acid receptor. Nature, 332, 850EE853. 5. Mangelsdorf,D.J., Umesono,K., Kliewer,S.A., Borgmeyer,U., Ong,E.S. and Evans,R.M. (1991) A direct repeat in the cellular retinol-binding protein type II gene confers differential regulation by RXR and RAR. J. Cell Biochem., 66, 555EE561. 6. Mangelsdorf,D.J., Borgmeyer,U., Heyman,R.A., Zhou,J.Y., Ong,E.S., Oro,A.E., Kakizuka,A. and Evans,R.M. (1992) Characterization of three RXR genes that mediate the action of 9-cis retinoic acid. Genes Dev., 6, 329EE344. 7. Westin,S., Rosenfeld,M.G. and Glass,C.K. (2000) Nuclear receptor coactivators. Adv. Pharmacol., 47, 89EE112. 8. Heyman,R.A., Mangelsdorf,D.J., Dyck,J.A., Stein,R.B., Eichele,G., Evans,R.M. and Thaller,C. (1992) 9-Cis retinoic acid is a high affinity ligand for the retinoid X receptor. J. Cell Biochem., 68, 397EE406. 9. Zelent,A., Krust,A., Petkovich,M., Kastner,P. and Chambon,P. (1989) Cloning of murine alpha and beta retinoic acid receptors and a novel receptor gamma predominantly expressed in skin. Nature, 339, 714EE717. 10. Chandraratna,R.A. (1996) Tazarotenefirst of a new generation of receptor-selective retinoids. Br. J. Dermatol., 135 (suppl. 49), 18EE25. 11. Standeven,A.M., Thacher,S.M., Yuan,Y.D., Escobar,M., Vuligonda,V., Beard,R.L. and Chandraratna,R.A. (2001) Retinoid X receptor agonist elevation of serum triglycerides in rats by potentiation of retinoic acid receptor agonist induction or by action as single agents. Biochem. Pharmacol., 62, 1501EE1509. 12. Nagpal,S. and Chandraratna,R.A. (2000) Recent developments in receptor-selective retinoids. Curr. Pharm. Des, 6, 919EE931. 13. Thacher,S.M., Nagpal,S., Klein,E.S., Arefieg,T., Krasinski,G., DiSepio,D., Agarwal,C., Johnson,A., Eckert,R.L. and Chandraratna,R.A. (1999) Cell type and gene-specific activity of the retinoid inverse agonist AGN 193109: divergent effects from agonist at retinoic acid receptor gamma in human keratinocytes [In Process Citation]. Cell Growth Differ., 10, 255EE262. 14. Pettersson,F., Dalgleish,A.G., Bissonnette,R.P. and Colston,K.W. (2002) Retinoids cause apoptosis in pancreatic cancer cells via activation of RAR-gamma and altered expression of Bcl-2/Bax. Br. J. Cancer, 87, 555EE561. 15. Pettersson,F., Colston,K.W. and Dalgleish,A.G. (2001) Retinoic acid enhances the cytotoxic effects of gemcitabine and cisplatin in pancreatic adenocarcinoma cells. Pancreas, 23, 273EE279. 16. Kawa,S., Nikaido,T., Aoki,Y., Zhai,Y., Kumagaya,T., Furihata,K., Fujii,S. and Kiyosawa,K. (1997) Arotinoid mofarotene (RO40-8757) up-regulates p21 and p27 during growth inhibition of pancreatic cancer cell lines. Int. J. Cancer, 72, 906EE911.

Fig. 10. AGN194204 does not promote caspase cleavage. MIA PaCa-2 cells were treated for 48 h with the indicated concentration of each agent. Total extracts were then prepared and equivalent quantities of protein were electrophoresed, transferred to nitrocellulose and immunoblotted with antibodies specific for procaspase-3, -8, -9 and b-actin.

agents and growth suppressors. We selected INFa, INFg, cisplatin, gemcitabine and 5-FU for testing, as these agents are used to treat pancreatic cancer patients (30). Treatment with 13-cRA and IFNa produces partial remission in some pancreatic cancer patients and prompts disease stabilization in others (48), and co-treatment with IFNb and retinol palmitate has been reported to be effective in treating metastatic pancreatic carcinoma patients (18). However, an important complicating issue is that these treatments produce toxic side effects, some of which are due to the retinoid therapy (18,48). In this respect, RXR-selective ligands have an important advantage, as in vivo side effects are minimal compared with RAR-selective ligands (28,29). 1384

Pancreatic cancer regulation by retinoids

17. Rosewicz,S., Stier,U., Brembeck,F., Kaiser,A., Papadimitriou,C.A., Berdel,W.E., Wiedenmann,B. and Riecken,E.O. (1995) Retinoids: effects on growth, differentiation, and nuclear receptor expression in human pancreatic carcinoma cell lines. Gastroenterology, 109, 1646EE1660. 18. Recchia,F., Sica,G., Casucci,D., Rea,S., Gulino,A. and Frati,L. (1998) Advanced carcinoma of the pancreas: phase II study of combined chemotherapy, beta-interferon, and retinoids. Am. J. Clin. Oncol., 21, 275EE278. 19. Riecken,E.O. and Rosewicz,S. (1999) Retinoids in pancreatic cancer. Ann. Oncol., 10 (suppl. 4), 197EE200. 20. Strickland,S., Breitman,T.R., Frickel,F., Nurrenbach,A., Hadicke,E. and Sporn,M.B. (1983) Structure-activity relationships of a new series of retinoidal benzoic acid derivatives as measured by induction of differentiation of murine F9 teratocarcinoma cells and human HL-60 promyelocytic leukemia cells. Cancer Res., 43, 5268EE5272. 21. Johnson,A.T., Klein,E.S., Gillett,S.J., Wang,L., Song,T.K., Pino,M.E. and Chandraratna,R.A. (1995) Synthesis and characterization of a highly potent and effective antagonist of retinoic acid receptors. J. Med. Chem., 38, 4764EE4767. 22. Teng,M., Duong,T.T., Klein,E.S., Pino,M.E. and Chandraratna,R.A. (1996) Identification of a retinoic acid receptor alpha subtype specific agonist. J. Med. Chem., 39, 3035EE3038. 23. Agarwal,C., Chandraratna,R.A., Johnson,A.T., Rorke,E.A. and Eckert,R.L. (1996) AGN193109 is a highly effective antagonist of retinoid action in human ectocervical epithelial cells. J. Biol. Chem., 271, 12209EE12212. 24. Haller,D.G. (2003) Chemotherapy for advanced pancreatic cancer. Int. J. Radiat. Oncol. Biol. Phys., 56, 16EE23. 25. Giguere,V., Ong,E.S., Segui,P. and Evans,R.M. (1987) Identification of a receptor for the morphogen retinoic acid. Nature, 330, 624EE629. 26. Johnson,A. and Chandraratna,R.A. (1999) Novel retinoids with receptor selectivity and functional selectivity. Br. J. Dermatol., 140 (suppl. 54), 12EE17. 27. Thacher,S.M., Standeven,A.M., Athanikar,J., Kopper,S., Castilleja,O., Escobar,M., Beard,R.L. and Chandraratna,R.A. (1997) Receptor specificity of retinoid-induced epidermal hyperplasia: effect of RXR-selective agonists and correlation with topical irritation. J. Pharmacol. Exp. Ther., 282, 528EE534. 28. Wu,K., Zhang,Y., Xu,X.C. et al. (2002) The retinoid X receptor-selective retinoid, LGD1069, prevents the development of estrogen receptornegative mammary tumors in transgenic mice. Cancer Res., 62, 6376EE6380. 29. Wu,K., Kim,H.T., Rodriquez,J.L., Hilsenbeck,S.G., Mohsin,S.K., Xu,X.C., Lamph,W.W., Kuhn,J.G., Green,J.E. and Brown,P.H. (2002) Suppression of mammary tumorigenesis in transgenic mice by the RXRselective retinoid, LGD1069. Cancer Epidemiol. Biomarkers Prev., 11, 467EE474. 30. Rosenberg,L. and Lipsett,M. (2003) Biotherapeutic approaches to pancreatic cancer. Expert. Opin. Biol. Ther., 3, 319EE337. 31. Albrechtsson,E., Ohlsson,B. and Axelson,J. (2002) The expression of retinoic acid receptors and the effects in vitro by retinoids in human pancreatic cancer cell lines. Pancreas, 25, 49EE56. 32. Vickers,S.M., Sampson,L.K., Ying,W. and Phillips,J.O. (1997) Receptordependent growth inhibition of human pancreatic cancer by 9-cis retinoic acid. J. Gastrointest. Surg., 1, 174EE181. 33. Egawa,N., Maillet,B., VanDamme,B., De Greve,J. and Kloppel,G. (1996) Differentiation of pancreatic carcinoma induced by retinoic acid or sodium

butyrate: a morphological and molecular analysis of four cell lines. Virch. Arch., 429, 59EE68. 34. Rosewicz,S., Brembeck,F., Kaiser,A., Marschall,Z.V. and Riecken,E.O. (1996) Differential growth regulation by all-trans retinoic acid is determined by protein kinase C alpha in human pancreatic carcinoma cells. Endocrinology, 137, 3340EE3347. 35. Kaiser,A., Wolf-Breitinger,M., Albers,A., Dorbic,T., Wittig,B., Riecken,E.O. and Rosewicz,S. (1998) Retinoic acid receptor gamma1 expression determines retinoid sensitivity in pancreatic carcinoma cells. Gastroenterology, 115, 967EE977. 36. Brembeck,F.H., Kaiser,A., Detjen,K., Hotz,H., Foitzik,T., Buhr,H.J., Riecken,E.O. and Rosewicz,S. (1998) Retinoic acid receptor alpha mediates growth inhibition by retinoids in rat pancreatic carcinoma DSL-6A/C1 cells. Br. J. Cancer, 78, 1288EE1295. 37. Fujimoto,K., Hosotani,R., Doi,R., Wada,M., Lee,J.U., Koshiba,T., Miyamoto,Y., Tsuji,S., Nakajima,S. and Imamura,M. (1999) Induction of cell-cycle arrest and apoptosis by a novel retinobenzoic-acid derivative, TAC-101, in human pancreatic-cancer cells. Int. J. Cancer, 81, 637EE644. 38. Teelmann,K., Tsukaguchi,T., Klaus,M. and Eliason,J.F. (1993) Comparison of the therapeutic effects of a new arotinoid, Ro 40-8757, and all-trans- and 13-cis-retinoic acids on rat breast cancer. Cancer Res., 53, 2319EE2325. 39. Vuligonda,V., Thacher,S.M. and Chandraratna,R.A. (2001) Enantioselective syntheses of potent retinoid X receptor ligands: differential biological activities of individual antipodes. J. Med. Chem., 44, 2298EE2303. 40. Kawasaki,H., Eckner,R., Yao,T.P., Taira,K., Chiu,R., Livingston,D.M. and Yokoyama,K.K. (1998) Distinct roles of the co-activators p300 and CBP in retinoic-acid-induced F9-cell differentiation. Nature, 393, 284EE289. 41. Leo,C. and Chen,J.D. (2000) The SRC family of nuclear receptor coactivators. Gene, 245, 1EE11. 42. Xu,L., Glass,C.K. and Rosenfeld,M.G. (1999) Coactivator and corepressor complexes in nuclear receptor function. Curr. Opin. Genet. Dev., 9, 140EE147. 43. Perugini,R.A., McDade,T.P., Vittimberga,F.J.,Jr and Callery,M.P. (1998) The molecular and cellular biology of pancreatic cancer. Crit. Rev. Eukaryot. Gene Expr., 8, 377EE393. 44. Xiao,J.H., Ghosn,C., Hinchman,C., Forbes,C., Wang,J., Snider,N., Cordrey,A., Zhao,Y. and Chandraratna,R.A. (2003) APC-independent regulation of beta-catenin degradation via a retinoid X receptor-mediated pathway. J. Biol. Chem., 278, 29954EE29962. 45. Ralhan,R. and Kaur,J. (2003) Retinoids as chemopreventive agents. J. Biol. Regul. Homeost. Agents, 17, 66EE91. 46. Lotan,R. (2003) Receptor-independent induction of apoptosis by synthetic retinoids. J. Biol. Regul. Homeost. Agents, 17, 13EE28. 47. Rishi,A.K., Zhang,L., Boyanapalli,M. et al. (2003) Identification and characterization of a cell cycle and apoptosis regulatory protein-1 as a novel mediator of apoptosis signaling by retinoid CD437. J. Biol. Chem., 278, 33422EE33435. 48. Brembeck,F.H., Schoppmeyer,K., Leupold,U., Gornistu,C., Keim,V., Mossner,J., Riecken,E.O. and Rosewicz,S. (1998) A phase II pilot trial of 13-cis retinoic acid and interferon-alpha in patients with advanced pancreatic carcinoma. Cancer, 83, 2317EE2323. Received October 26, 2003; revised January 19, 2004; accepted February 10, 2004

Downloaded from http://carcin.oxfordjournals.org/ at Universiti Teknologi Malaysia on March 26, 2012

1385