Oncogene (2004) 23, 1539–1548
& 2004 Nature Publishing Group All rights reserved 0950-9232/04 $25.00 www.nature.com/onc
RNA interference targeting the M2 subunit of ribonucleotide reductase enhances pancreatic adenocarcinoma chemosensitivity to gemcitabine Mark S Duxbury1, Hiromichi Ito1, Michael J Zinner1, Stanley W Ashley1 and Edward E Whang*,1 1
Department of Surgery, Brigham and Women’s Hospital, Harvard Medical School, 75 Francis Street, Boston, Massachusetts, USA
Ribonucleotide reductase is emerging as an important determinant of gemcitabine chemoresistance in human cancers. Activity of this enzyme, which catalyses conversion of ribonucleotide 50 -diphosphates to their 20 -deoxynucleotides, is modulated by levels of its M2 subunit (RRM2). Here we show that RRM2 overexpression is associated with gemcitabine chemoresistance in pancreatic adenocarcinoma cells, and that suppression of RRM2 expression using RNA interference mediated by small interfering RNA (siRNA) enhances gemcitabine-induced cytotoxicity in vitro. We demonstrate the ability of systemically administered RRM2 siRNA to suppress tumoral RRM2 expression in an orthotopic xenograft model of pancreatic adenocarcinoma. Synergism between RRM2 siRNA and gemcitabine results in markedly suppressed tumor growth, increased tumor apoptosis and inhibition of metastasis. Our findings confirm the importance of RRM2 in pancreatic adenocarcinoma gemcitabine chemoresistance. This is the first demonstration that systemic delivery of siRNA-based therapy can enhance the efficacy of an anticancer nucleoside analog. Oncogene (2004) 23, 1539–1548. doi:10.1038/sj.onc.1207272 Published online 8 December 2003 Keywords: pancreatic adenocarcinoma; cancer; gemcitabine; ribonucleotide reductase; siRNA; RNA interference
Introduction Pancreatic adenocarcinoma has recently been recognized to be the fourth leading cause of cancer death in the United States (Jemal et al., 2003). Chemoresistance is a major cause of pancreatic adenocarcinoma treatment failure, such that even the principal chemotherapeutic agent used in this cancer, gemcitabine (20 ,20 -difluorodeoxycytidine), has only marginal effects on clinical outcome (Burris III et al., 1997). Expression and activity of ribonucleotide reductase have been reported to be determinants of gemcitabine chemoresistance in human tumor cells (Goan et al., *Correspondence: EE Whang; E-mail:
[email protected] Received 4 September 2003; revised 3 October 2003; accepted 6 October 2003
1999; Jung et al., 2001). Ribonucleotide reductase catalyses conversion of ribonucleotide 50 -diphosphates to their 20 -deoxynucleotide form, a rate-limiting step in the production of 20 -deoxyribonucleoside 50 -triphosphates (dNTP) required for DNA synthesis (Reichard, 1978, 1985; Cory, 1997; Goan et al., 1999). Ribonucelotide reductase promotes gemcitabine chemoresistance in a number of ways, primarily through expansion of the dNTP pool. Expansion of the dNTP pool has the following consequences: (1) Excess dNTPs inhibit the activity of deoxycytidine kinase, an enzyme required for gemcitabine phosphorylation, thereby inhibiting gemcitabine incorporation into DNA (Heinemann et al., 1988; Plunkett et al., 1995; Goan et al., 1999). (2) dNTPs directly compete with gemcitabine for incorporation into DNA. (3) dNTPs are reported to increase the activity of deoxycytidine deaminase, an enzyme that degrades gemcitabine (Plunkett et al., 1996). Other additional mechanisms may contribute to gemcitabine chemoresistance. Altered expression and activity of the enzyme deoxycytidine kinase (Bergman et al., 2002) and of transmembrane transporters such as P-glycoprotein, the product of the mdr-1 gene, and multidrug resistance-associated protein (Endicott and Ling, 1989; Grant et al., 1994) can lead to a gemcitabineresistant phenotype. Ribonucleotide reductase consists of two subunits: M1 and M2. Ribonucleotide reductase enzymatic activity is modulated by levels of its M2 subunit (RRM2) (Eriksson and Martin Jr, 1981). In addition, a recently discovered p53-dependent gene, p53R2, has been found to encode a ribonucleotide reductase subunit (Tanaka et al., 2000). The p53R2 protein is similar to the M2 subunit and, through an association with the M1 subunit, forms an active ribonucleotide reductase heterodimer (Guittet et al., 2001). RRM2 itself is a dimer of two 44 kDa moieties, each containing a tyrosine free-radical and non-heme iron (Thelander et al., 1985). Overexpression of RRM2 is associated with resistance to a variety of chemotherapeutic agents, including gemcitabine (Zhou et al., 1995; Goan et al., 1999). The synthetic flavone flavopiridol has been reported to promote gemcitabine cytotoxicity in gastrointestinal cancer cells by suppressing RRM2 expression (Jung et al., 2001). In this study, we hypothesized that directly targeting RRM2 would enhance the chemosensitivity of pancreatic adenocarcinoma to gemcitabine.
RNA interference enhances chemosensitivity MS Duxbury et al
1540
RNA interference using small interfering RNA (siRNA) is emerging as a strategy for the highly specific post-transcriptional suppression of gene expression, both in vitro and, more recently, in vivo (Filleur et al., 2003; Li et al., 2003). siRNA is reported to have greater in vitro and in vivo potency in suppressing gene expression than antisense oligonucleotides (Bertrand et al., 2002; Aoki et al., 2003). Further, the feasibility of reversing chemoresistance using siRNA-based strategies has been demonstrated in vitro (Nieth et al., 2003). We sought to determine whether RRM2-specific siRNAbased treatment would be an effective strategy to increase the sensitivity of pancreatic adenocarcinoma to gemcitabine. Here we report for the first time that siRNA-targeting RRM2 enhances gemcitabine-induced cytotoxicity both in vitro and in vivo. Our findings confirm RRM2 as a rational therapeutic target in pancreatic adenocarcinoma, and indicate that siRNA-mediated gene silencing is a promising in vivo antitumor strategy.
Specific suppression of RRM2 expression by RRM2 siRNA in vitro Based on these observations, we sought to determine whether siRNA directed against RRM2 would be an
Results RRM2 overexpression is associated with pancreatic adenocarcinoma gemcitabine chemoresistance RRM2 expression was quantified in the pancreatic ductal adenocarcinoma cell lines PANC1, MIAPaCa2, BxPC3 and Capan2 by Western blot. Differential expression of RRM2 was observed. The highest levels of RRM2 expression occurred in PANC1 and MIAPaCa2, cell lines displaying the greatest gemcitabine resistance. The lowest RRM2 expression was found in Capan2, the cell line with the least resistance to gemcitabine (Figure 1a; Table 1). We assayed ribonucleotide reductase activity in the four cell lines, using a polymerase-based assay (Jong et al., 1998). The highest levels of ribonucleotide reductase activity were confirmed in PANC1 and MIAPaCa2 (Figure 1b). Owing to its inherently low level of RRM2 expression, we chose Capan2 to determine the effects of induced RRM2 overexpression on the gemcitabine 50% inhibitory concentration (IC50). RRM2 overexpression was induced by transient transfection using the pcDNA-DEST RRM2 expression construct. Western blot analysis confirmed RRM2 overexpression. Overexpression of RRM2 markedly increased the gemcitabine IC50 of Capan2 (Figure 2). Transfection with the empty vector had no effect on the gemcitabine IC50, compared to that of untransfected cells.
Figure 1 (a) Representative Western blot for RRM2 (44 kDa) in untreated exponentially growing PANC1, MIAPaCa2, BxPC3 and Capan2 pancreatic ductal adenocarcinoma cells. Densitometry values are means (7s.d.) of three independent determinations. (b) Ribonucleotide reductase activity was quantified using a DNA polymerase-based assay. [14C]CDP is converted to [14C]dCDP by ribonucleotide reductase, and the resulting [14C]dCTP incorporation into nascent DNA is quantified. The highest ribonucleotide reductase activity was observed in PANC1 and MIAPaCa2 cells: cell lines with the most pronounced gemcitabine chemoresistance. Mean values from triplicate experiments using four determinations per cell line are shown. *Po0.05 versus PANC1
Table 1 Gemcitabine IC50 of pancreatic adenocarcinoma cell lines
Gemcitabine IC50 Relative RRM2 expression (arbitrary units)
PANC1
MIAPaCa2
BxPC3
Capan2
50 nM 0.71
40 nM* 0.42*
18 nM* 0.16*
12 nM* 0.13*
*Po0.05 versus PANC1 by Kruskal–Wallis test. R2 ¼ 0.94 between gemcitabine IC50 and relative RRM2 expression by Spearman correlation Oncogene
RNA interference enhances chemosensitivity MS Duxbury et al
1541
Figure 2 RRM2 overexpression increases the gemcitabine chemoresistance of Capan2. Capan2 was chosen as it inherently expresses low levels of RRM2. Capan2 cells were transfected with a RRM2 expression construct. RRM2 overexpression was confirmed by Western blot (a). The gemcitabine IC50 was determined by MTT assay (b). *Po0.05 versus empty vector
effective strategy for the suppression of gemcitabine resistance in pancreatic adenocarcinoma. Three siRNAs were designed to target the coding region of RRM2. Of the three RRM2-specific sequences evaluated, the sequence most effective at suppressing RRM2 expression was used in subsequent experiments (RRM2 sense: 50 -UGCUGUUCGGAUAGAACAGd(TT)-30 ; RRM2 antisense: 50 -CUGUUCUAUCCGAACAGCAd(TT)30 . A single-base mismatched siRNA lacking sequence homology with any human gene was used as a control (control sense: 50 -UGCUUUUCGGAUAGAACAGd (TT)-30 ; control antisense: 50 -CUGUUCUAUCCGAA AAGCAd(TT)-30 ). Pancreatic adenocarcinoma cells were transfected with double-stranded siRNA oligonucleotides. At 48 h post-transfection, cells were harvested and RRM2 expression was analysed by Western (Figure 3) and Northern blot (Figure 4). RRM2 transcript was visualized as two bands of 3.4 and 1.6 kb, the presence of which is reported to be the result of alternative polyadenylation (Thelander and Berg, 1986; Wright et al., 1987; Pavloff et al., 1992). Up to 90% suppression of RRM2 protein expression occurred within 24 h of RRM2 siRNA transfection and persisted at this level at 6 days post-transfection. Control siRNA had no effect on RRM2 expression, compared to that of
Figure 3 Representative Western blot for RRM2 48 h following treatment with control siRNA () or RRM2 siRNA ( þ ). Singlebase mismatch control siRNA did not significantly affect RRM2 expression compared to untreated cells, whereas RRM2-specific siRNA induced marked suppression of RRM2 expression. Expression of p53R2, RRM1 and actin was unaffected by either siRNA treatment. RRM2 densitometry values are means (7s.d.) of three independent determinations. *Po0.05 versus control siRNA
untransfected cells. Actin, RRM1 and p53R2 expression was unaffected by treatment with either RRM2 or control siRNA (Figures 3,4), indicating that nonspecific downregulation of protein expression was not induced by RRM2 siRNA treatment. RRM2 siRNA treatment suppresses pancreatic adenocarcinoma cell gemcitabine chemoresistance in vitro Next, we determined the effect of RRM2 siRNA on cellular proliferation and gemcitabine chemoresistance in vitro. At 48 h following transfection with RRM2 siRNA, cellular proliferation was determined by MTT assay in the absence of gemcitabine and in the presence of gemcitabine at concentrations ranging from 1 nM to 10 mM. From these data, the gemcitabine IC50 was calculated. RRM2 siRNA induced a marked decrease in the gemcitabine IC50 of all the four cell lines (Figure 5a). The decrease in gemcitabine IC50 was associated with large increases in the apoptotic fraction, determined by in situ TUNEL staining (Figure 5b). This increase in apoptosis was accompanied by increased caspase 3 activity (Figure 5c). Treatment with RRM2 siRNA induced a minor reduction in cellular proliferation in the absence of gemcitabine (Figure 5d). The decrease was only significant in the case of PANC1. Control siRNA Oncogene
RNA interference enhances chemosensitivity MS Duxbury et al
1542
had no effect on cellular proliferation, compared to that of untreated cells, indicating a low level of transfectioninduced cytotoxicity. Comparable results were obtained following exposure to gemcitabine for 48 h. Chemosensitization by RRM2 siRNA is associated with an obtunded NF-kB response to gemcitabine
Figure 4 Representative Northern blot for RRM2, 48 h following treatment with control siRNA () or RRM2 siRNA ( þ ). RRM2 transcript was detected as two bands at 1.6 and 3.4 kb. Suppression of RRM2 transcript was demonstrated following administration of RRM2 siRNA, but not control siRNA. Actin transcripts were unaffected by either siRNA treatment. *Po0.05 versus control siRNA
Recent reports have identified nuclear factor kB (NFkB) as an important mediator of pancreatic adenocarcinoma gemcitabine chemoresistance, with inhibition of NF-kB enhancing gemcitabine-induced cytotoxicity in pancreatic adenocarcinoma cells (Arlt et al., 2003; Muerkoster et al., 2003). We therefore determined the effect of RRM2 siRNA on basal NF-kB activity and its effect on NF-kB activity following exposure to 1 mM gemcitabine for 6 h. NF-kB activity was quantified in two ways: luciferase activity was measured in lysates of cells transfected with a NF-kB luciferase reporter; NFkB DNA-binding activity in nuclear extracts was quantified by ELISA at identical time points. The absence of binding to the negative control oligonucleotide containing a mutant NF-kB-binding sequence was confirmed. At 24 h following transfection with the NFkB reporter, the four cell lines were exposed to gemcitabine for 6 h, following which luciferase activity was quantified. Treatment with 1 mM gemcitabine of each cell line for 6 h induced significant increases in luciferase activity in each case. This experiment was also performed on cells transfected 48 h previously with RRM2-specific or
Figure 5 RRM2 siRNA enhances gemcitabine-induced cytotoxicity in vitro. Cells were transfected with RRM2 siRNA or control siRNA. At 48 h post-transfection, the gemcitabine cytotoxicity assay was commenced for each cell line (a), *Po0.05 versus control siRNA. Following exposure for 4 days to 1 mM gemcitabine, apoptotic cells were identified by TUNEL staining and expressed as an apoptotic fraction (%, b). *Po0.05 versus control siRNA þ gemcitabine. Caspase 3 activity was determined by fluorometric assay (c). Cellular proliferation was quantified following siRNA treatment and expressed relative to untreated cells. RRM2 siRNA caused only a minor decrease in cellular proliferation relative to control siRNA (d). *Po0.05 versus control siRNA Oncogene
RNA interference enhances chemosensitivity MS Duxbury et al
1543
control siRNA. Following RRM2 siRNA treatment, basal NF-kB activity was decreased. Luciferase activity was not affected by control siRNA treatment. When RRM2 siRNA-treated cells were exposed to 1 mM gemcitabine for 6 h, the increase in NF-kB activity that was induced in untreated control siRNA-treated cells was attenuated (Figure 6). NF-kB reporter luciferase assay and nuclear extract NF-kB ELISA results corresponded closely. RRM2 siRNA suppresses RRM2 expression and enhances gemcitabine-induced cytotoxicity in a nude mouse xenograft model In view of these in vitro findings, we sought to determine whether RRM2 siRNA would potentiate the therapeutic
Figure 6 NF-kB activity in MIAPaCa2 following control and RRM2 siRNA alone, and in combination with gemcitabine. NFkB activity was quantified by luciferase reporter assay (a). Nuclear NF-kB/NF-kB consensus sequence DNA-binding activity was quantified by TransFactor ELISA performed on nuclear extracts under identical conditions (b). Nuclear NF-kB activity correlates linearly with the OD560. Gemcitabine exposure (1 mM, 6 h) increased NF-kB reporter activity and NF-kB activity in nuclear extracts. This effect was attenuated by prior transfection with RRM2 siRNA, but not control siRNA. Luciferase reporter assay and nuclear extract ELISA produced a similar pattern of results. All experiments were performed in triplicate with three determinations per condition. *Po0.05 versus PBS-treated and control siRNA-treated cells. wPo0.05 versus control siRNA þ gemcitabine
efficacy of gemcitabine in a nude mouse xenograft model of pancreatic adenocarcinoma. Nude mice were subcutaneously implanted with 106 MIAPaCa2 cells. Once the tumors reached 100 mm3 (20 days postimplantation), treatment was initiated according to the groups summarized in Table 2. Following 6 weeks of treatment, necropsy was performed. Primary tumors were excised, weighed, sectioned and homogenized. Tumor protein lysates were analysed by Western blot and caspase 3 assay. Mice were examined for liver metastases, which were counted, excised and confirmed histologically. No adverse events were observed in any treatment group. The ability of systemically administered RRM2 siRNA to inhibit RRM2 expression in vivo was confirmed by Western blot analysis of tumor homogenates. Systemic administration of RRM2 siRNA suppressed tumoral RRM2 expression. Control siRNA had no effect on RRM2 expression (Figure 7a). Treatment with RRM2 siRNA alone caused a small but significant decrease in the rate of tumor growth. Control siRNA had no significant effect on tumor growth. We observed the synergism between RRM2 siRNA and gemcitabine treatment, resulting in marked suppression of tumor growth. RRM2 siRNA in combination with gemcitabine induced an approximately eight-fold suppression in the mean tumor mass, compared to tumors from mice treated with gemcitabine alone (Po0.05; Figure 7b, c). Control siRNA did not affect the response to gemcitabine. Treatment with gemcitabine in combination with RRM2 siRNA, but not control siRNA, resulted in complete inhibition of liver metastasis (Po0.05; Figure 7d). In order to quantify tumor apoptosis, tumor sections were TUNEL stained and tumor homogenates were assayed for caspase 3 activity. Tumor sections from mice treated with RRM2 siRNA in combination with gemcitabine showed a higher percentage of apoptotic cells than those from mice treated with gemcitabine alone or in combination with control siRNA (Figure 7e). There was no significant difference in tumor apoptosis between mice treated with gemcitabine alone and those treated with gemcitabine and control siRNA. Similarly, caspase 3 activity was the highest in the group treated with RRM2 siRNA and gemcitabine (Figure 7f). These observations demonstrate that systemically administered RRM2 siRNA inhibits RRM2 expression and enhances gemcitabineinduced apoptosis in vivo.
Table 2 Twice weekly in vivo treatment regimen in the nude mouse xenograft model Group
Treatment
1 2 3 4 5 6
PBS Gemcitabine (150 mg/kg) Control siRNA (150 mg/kg) RRM2 siRNA (150 mg/kg) Gemcitabine (150 mg/kg)+control siRNA (150 mg/kg) Gemcitabine (150 mg/kg)+RRM2 siRNA (150 mg/kg) Oncogene
RNA interference enhances chemosensitivity MS Duxbury et al
1544
Figure 7 Following subcutaneous implantation of 106 MIAPaCa2 cells, mice were treated with PBS, control siRNA or RRM2 siRNA alone or in combination with gemcitabine for 6 weeks. Necropsy was performed, tumors were sectioned and TUNEL stained. Tumor homogenates were assayed for RRM2 and caspase 3 activity, as described. Suppression of RRM2 expression by in vivo RRM2 siRNA treatment was confirmed in tumors from mice receiving RRM2 siRNA, but not those receiving control siRNA (a). Tumor growth was markedly retarded, and the final mean tumor mass was reduced by 87%, in mice treated with RRM2 siRNA in combination with gemcitabine, compared to those treated with gemcitabine alone or in combination with control siRNA. *Po0.05 versus both gemcitabine alone and control siRNA þ gemcitabine (b, c). Hepatic metastasis occurred in 80% of PBS-treated mice and 40% of gemcitabine-treated mice. There were no metastases in mice treated with RRM2 siRNA and gemcitabine (d). The mean tumor apoptotic fraction (e) and caspase 3 activity (f) were significantly increased in tumors from mice treated with RRM2 siRNA in combination with gemcitabine, compared to those treated with control siRNA and gemcitabine and gemcitabine alone. *Po0.05 versus both gemcitabine alone and control siRNA þ gemcitabine
Discussion This study was designed to determine to role of RRM2 in pancreatic adenocarcinoma gemcitabine chemoresistance, and to assess the efficacy of siRNA directed against RRM2 in suppressing resistance to gemcitabine. We have demonstrated that RRM2 overexpression represents a chemoresistance mechanism in pancreatic adenocarcinoma cells. We have also shown that siRNA can specifically suppress RRM2 expression and enhance gemcitabine-induced cytotoxicity, both in vitro and in vivo. RRM2 siRNA exhibits therapeutic synergism with gemcitabine, resulting in strong suppression of tumor progression and inhibition of metastasis in the nude mouse xenograft model. Together, these findings confirm RRM2 as both a chemoresistance mechanism and as a potential therapeutic target in pancreatic adenocarcinoma. Oncogene
RRM2 is a determinant of malignant cellular behavior in a variety of human cancers (Zhou et al., 1995, 1998, 2001; Fan et al., 1996; Goan et al., 1999). In the pancreatic adenocarcinoma cells tested, higher inherent levels of RRM2 expression occur in cell lines, demonstrating the greatest gemcitabine chemoresistance. In addition, artificial overexpression of RRM2 results in a further increase in gemcitabine chemoresistance. The antiproliferative and proapoptotic effects of RRM2 siRNA treatment alone are relatively minor in comparison to the synergistic effects observed following combined administration of RRM2 siRNA and gemcitabine, suggesting that RRM2 siRNA targets an important mechanism contributing to gemcitabine chemoresistance in these pancreatic adenocarcinoma cells. Decreases in malignant cellular proliferation have recently been reported by Lee et al. (2003), following
RNA interference enhances chemosensitivity MS Duxbury et al
1545
treatment with RRM2-directed phosphorothioate antisense oligonucleotides. Administration of 10 mg/kg RRM2-specific antisense oligonucleotide to severe combined immunodeficient (SCID) mice, on alternate days, induced an approximately 50% inhibition of pancreatic adenocarcinoma growth. However, phosphorothioate oligonucleotides exhibit a high degree of nonspecific serum protein binding, and are reported to induce immune stimulation in rodents (Monteith et al., 1997). Although SCID mice lack NK-, B- and T-cell functions, it remains possible that immunomodulation contributes to the antitumor effects of nucleotide-based treatments. Despite a lack of evidence regarding in vivo immunomodulation by siRNA, we minimized the likelihood of nonspecific effects by using a low concentration of siRNA (Editorial, 2003). The concentration of siRNA effective in vitro was less than 1 nM, indicating the high potency of siRNA-mediated gene silencing. Mice were treated with 150 mg/kg siRNA twice weekly, a dosage commensurate with that used in a recent study reported by Filleur et al. (2003). We observed that cellular proliferation and apoptosis in monolayer culture remain virtually unaffected by RRM2 siRNA, while tumor growth is suppressed by this treatment. RRM2 may protect xenografted tumor cells from hostderived modifiers encountered in vivo that are not present in monolayer culture. Other groups have reported therapeutic synergism between target-specific treatment and gemcitabine. Target-specific therapy offers the theoretical benefit of fewer side effects with consequently improved quality of life for cancer patients undergoing such therapy. Kelley et al. (2001) combined gemcitabine with a cancerassociated Sm-like antisense RNA strategy, and found extension of survival time beyond either therapy alone in a murine model of pancreatic cancer. A recent phase I clinical trial has reported H-ras antisense treatment combined with gemcitabine to be a well-tolerated and clinically active combination (Adjei et al., 2003). In our study, combining RRM2-specific siRNA with gemcitabine resulted in highly effective synergism. Our observation that in vivo systemic siRNA treatment promotes gemcitabine chemosensitivity is important and, given the absence of side effects in the murine model, systemic administration of siRNA requires further evaluation as a therapeutic approach. In vivo RNA interference has previously been reported in a mouse tumor model (Filleur et al., 2003). We generally observe significant gene silencing within 1 week of in vivo siRNA administration (unpublished data). RNA interference mediated by siRNA is generally transient, largely due to siRNA dilution during cell division. While, in our study, repeated administration of siRNA overcame this limitation of siRNA-mediated gene silencing, future strategies employing plasmid and virus-based siRNA delivery systems (Brummelkamp et al., 2002a, b) may further improve the efficacy of this approach. Our observation that RRM2 siRNA suppresses NFkB activity is interesting for a number of reasons. NFkB is now recognized as an important mediator of chemoresistance to agents including gemcitabine (Arlt
et al., 2003), and its activity is associated with pancreatic adenocarcinoma malignant cellular behavior (Sclabas et al., 2003). Gemcitabine-resistant clones displaying RRM2 overexpression have also been reported to demonstrate altered NF-kB-binding activities (Zhou et al., 2001). Furthermore, inhibition of ribonucleotide reductase by the novel inhibitor Didox, and by hydroxyurea, have both been shown to suppress NFkB activity (Calzado et al., 2000; Inayat et al., 2002). Consistent with these observations, we observed suppression of NF-kB activity following RRM2 siRNA treatment alone. NF-kB activity increases following exposure of pancreatic adenocarcinoma cells to gemcitabine (Arlt et al., 2003). We observed almost complete abolition of the gemcitabine-induced increase in NF-kB activity following RRM2 siRNA treatment. This finding may, in part, account for the chemosensitizing effect of RRM2 siRNA treatment. The mechanisms by which RRM2 affects NF-kB activity are the subject of ongoing investigation. We observed complete inhibition of metastasis in mice treated with RRM2 siRNA in combination with gemcitabine. This observation may reflect the smaller tumor burden resulting from this treatment. Suppression of RRM2 expression may also have antimetastatic effects through its influence on NFkB activity, which is reported to be an important determinant of pancreatic adenocarcinoma metastasis (Fujioka et al., 2003; Sclabas et al., 2003). In conclusion, this study identifies RRM2 as an important determinant of pancreatic adenocarcinoma gemcitabine chemoresistance, and demonstrates that RRM2 gene silencing by siRNA is an effective therapeutic adjunct to gemcitabine treatment both in vitro and in vivo. The gene specificity of siRNA, as well as the ability to generate novel siRNAs rapidly, suggest that treatment could be modified in response to dynamic tumor characteristics. Monitoring of tumor gene-expression profiles may allow rational targetdirected therapy to be developed and tailored to tumor behavior.
Methods Cell lines and cell culture PANC1, MIAPaCa2, BxPC3 and Capan2 pancreatic ductal adenocarcinoma cells were obtained from ATCC (Rockville, MD, USA). Cells were maintained in Dulbecco’s Modified Eagle Medium (DMEM) containing 10% fetal bovine serum (FBS, Gibco BRL, Gaithersburg, MD, USA). Cells were incubated in a humidified (371C, 5% CO2) incubator and passaged upon reaching 80% confluence. siRNA The siRNA sense and antisense strands were purchased from Qiagen (Qiagen-Xeragon, Germantown, MD, USA). The target sequences assessed were: RRM2, 50 -AATGCTGTTC GGATAGAACAG-30 ; 50 -AAGAAGAAGGCAGACTGGG CC-30 ; 50 -AAATGAAGTGAAGATGTGCCC-30 . Single-base mismatch siRNA was used as a control for the active siRNA selected for use in all further experiments (control sense: 50 Oncogene
RNA interference enhances chemosensitivity MS Duxbury et al
1546 UGCUUUUCGGAUAGAACAGd(TT)-30 ; control antisense: 50 -CUGUUCUAUCCGAAAAGCAd(TT)-30 ). The siRNAs were dissolved in buffer (100 mM potassium acetate, 30 mM HEPES-potassium hydroxide, 2 mM magnesium acetate, pH 7.4) to a final concentration of 20 mM, heated to 901C for 60 s and incubated at 371C for 60 min prior to use, to disrupt any higher aggregates formed during synthesis. In vitro transfection was performed using siPORTAmine transfection reagent (Ambion, Austin, TX, USA), which was added to serum-free OptiMEM for a final complexing volume of 200 ml, vortexed and incubated at room temperature for 20 min. Following addition of siRNA solution and gentle mixing, this mixture was incubated at room temperature for 15 min. A measure of 1 mg siRNA was used for 106 cells. Following transfection, cells were incubated in DMEM containing 10% FBS. RRM2 transient transfection RRM2 cDNA inserts were produced by PCR (primer 1: 50 CACCATGCTCTCCCTCCGTGT-30 ; primer 2: 50 -TTAGAAGTCAGCATCCAAG-30 ). The PCR products were inserted into the pENTR TOPO vector, expanded in OneShot TOP10 cells and transferred to the pcDNA3.2-DEST expression vector, in accordance with the manufacturer’s instructions (Invitrogen, Carlsbad, CA, USA). Transfection was performed using Lipofectamine 2000 Reagent (Invitrogen). Assays were performed 48 h post-transfection.
film (Eastman Kodak Co., Rochester, NY, USA). RRM2 mRNA was detected as two bands of 3.4 and 1.6 kb, as previously reported (Pavloff et al., 1992; Lee et al., 2003). After detection, the filter was stripped and re-probed for b-actin. The 3.4 kb RRM2 signal was quantified using ImagePro Plus software version 4.0 (Media Cybernetic, Silver Spring, MD, USA), and normalized to that of b-actin. Western blotting At 48 h post-transfection, whole-cell extracts were prepared using cell lysis buffer (20 mM Tris, pH 7.5, 0.1% Triton X, 0.5% deoxycholate, 1 mM PMSF, 10 mg/ml aprotinin, 10 mg/ml leupeptin) and cleared by centrifugation at 12 000 g, 41C. Tissues were homogenized and lysed in the same lysis buffer. Total protein concentration was measured using the BCA assay kit (Sigma, St Louis, MO, USA) with bovine serum albumin as a standard. Cell lysates containing 30 mg total protein were analysed by immunoblotting. Anti-RRM2, p53R2 and RRM1 antibodies were obtained from Santa Cruz (CA, USA). Antiactin antibodies were obtained from LabVision (Freemont, CA, USA). Chemoluminescent detection (Upstate, Lake Placid, NY, USA) was performed in accordance with the manufacturer’s instructions. The RRM2 signal was quantified using ImagePro Plus software version 4.0, and normalized to that of b-actin. Ribonucleotide reductase assay
Cell-proliferation assays Cell proliferation was quantified by 3-(4,5-dimethylthiazol-2yl)-2,5-diphenyltetrazolium bromide (MTT) assay (Trevigen, Gaithersburg, MD, USA), in accordance with the manufacturer’s instructions, and confirmed by cell counting. Results of the MTT assay have been shown to correlate well with [3H]thymidine incorporation in pancreatic adenocarcinoma cell lines (Raitano et al., 1990). At 48 h following siRNA treatment, cells were seeded into 96-well plates at 104 cells/ well and allowed to adhere overnight. Cells were cultured in medium containing 10% FBS and 0–10 mM gemcitabine. The medium was replaced every 48 h and cellular proliferation was determined after 48 and 96 h. Plates were read using a Vmax microplate spectrophotometer (Molecular Devices, Sunnyvale, CA, USA) at a wavelength of 570 nm, referenced to 650 nm. In all, 10 samples were used for each experimental condition and experiments were performed in triplicate. The 50% inhibitory concentration of gemcitabine (IC50) was calculated from these data. At identical time points, cell counting was performed. Cells were trypsinized to form a single cell suspension. Viable cells, determined by Trypan blue exclusion, were counted using a Neubauer hemocytometer (Hausser scientific, Horsham, PA, USA). Cell counts were used to confirm MTT results. Northern blot analysis Total RNA samples were prepared using Trizol reagent (Gibco BRL, Rockville, MD, USA), according to the manufacturer’s instructions. A measure of 20 mg total RNA, quantified by UV spectrophotometry, was electrophoresed through a denaturing gel (1.2% agarose, 2.2 M formaldehyde) and blotted onto a positively charged nylon membrane. After crosslinking with 120 mJ/cm2 ultraviolet irradiation, the filter was prehybridized for 4 h at 421C in 5 SSPE, 5 Denhardt’s solution, 0.1% SDS and 100 mg/ml denatured salmon sperm DNA. Blots were hybridized with [a-32P]ATP probe synthesized from the RRM2 cDNA (McClarty et al., 1987). Hybridization was carried out with 50 ng probe at 421C for 16 h. The filter was washed four times in 5 SSPE, 0.1% SDS for 10 min and exposed to XAR Oncogene
Cells (107) were collected, washed and extracts were prepared on ice in the presence of RNAse A, as previously described (Jong et al., 1998). Protein concentration was measured by BCA assay and samples containing equal total protein were analysed. Endogenous nucleotides were removed using a Sephadex G-25 spin column. Ribonucleotide reductase activity was quantified using a [14C]CDP and DNA polymerase-based assay, as described by Jong et al. (1998). Ribonucleotide reductase catalyses the conversion of [14C]CDP to [14C]dCDP, which is converted to [14C]dCTP by endogenous nucleotide diphosphate kinase and incorporated into nascent DNA by DNA polymerase. The duration of the ribonucleotide reductase reaction was 30 min and the duration of the coupling reaction was 50 min. Nascent DNA radioactivity was quantified using a liquid scintillation counter (Beckman Coulter Inc., Fullerton, CA, USA). Negative controls included the omission of polymerase and DNA primers. All experiments were performed three times with four determinations per cell line. NF-kB luciferase assay NF-kB activity was quantified using a commercially available NF-kB reporter construct (Stratagene, La Jolla, CA, USA). At 24 h following transfection with siRNA, cells were transfected with the reporter construct in accordance with the manufacturer’s instructions. After 24 h, cells were exposed to 1 mM gemcitabine for 6 h. Luciferase activity was assessed after 6 h exposure using a proprietary luciferase assay kit (BD BioSciences, San Jose, CA, USA). Luminescence was quantified using a TD-20/20 luminometer (Turner BioSystems, Inc., Sunnyvale, CA, USA). Experiments were performed in triplicate with three determinations per condition. NF-kB enzyme-linked immunoassay Levels of NF-kB consensus sequence DNA binding in nuclear extracts were detected using the BD Mercury Transfactort kit (BD Biosciences). This nonradioactive technique has 10-fold higher sensitivity than electrophoretic mobility shift assays
RNA interference enhances chemosensitivity MS Duxbury et al
1547 (Shen et al., 2002). DNA oligonucleotides containing the NF-kB consensus sequence (50 -GGGGACTTTCCC-30 ) (Pessara and Koch, 1990; Kunsch et al., 1992) immobilized on a 96-well plate were used to capture NF-kB from pancreatic adenocarcinoma nuclear extracts containing 20 mg of total protein. An oligonucleotide containing a mutated consensus sequence motif (50 -GGCGACTTTCCC-30 ) was used as a negative control. The DNA-bound transcription factor was detected and quantified by ELISA using a NF-kB-specific monoclonal antibody and TMB-based detection, in accordance with the manufacturer’s instructions. Plates were read on a Vmax microplate spectrophotometer at 650 nm. Nude mouse xenograft model Male athymic nu/nu mice, 5 weeks of age, weighing 20–22 g and specific pathogen-free, were obtained from Charles River Laboratories (Wilmington, MA, USA). Mice were housed in microisolator cages with autoclaved bedding in a specific pathogen-free facility with 12 h light–dark cycles. Animals received water and food ad libitum, and were observed for signs of tumor growth, activity, feeding and pain, in accordance with the guidelines of the Harvard Medical Area Standing Committee on Animals. Mice were anesthetized with intraperitoneal ketamine (200 mg/kg), xylazine (10 mg/kg), and inoculated with 106 gemcitabine-resistant PANC1GemRes cells in 20 ml PBS by subcutaneous implantation. After inoculation, mice were allocated to one of three treatment groups. Treatment commenced 20 days following implantation: Group 1 (control, n ¼ 10) received 100 ml PBS by intraperitoneal injection and 50 ml PBS by tail vein injection twice weekly. Group 2 (n ¼ 10) received gemcitabine (150 mg/kg) in 100 ml PBS vehicle by intraperitoneal injection, and 50 ml PBS by tail vein injection twice weekly. Group 3 (n ¼ 10) received 100 ml PBS by intraperitoneal injection and 150 mg/kg control siRNA in 50 ml PBS by tail vein injection twice weekly. Group 4 received 100 ml PBS by intraperitoneal injection and 150 mg/kg RRM2 siRNA in 50 ml PBS by tail vein injection twice weekly. Group 5 received 150 mg/kg gemcitabine in 100 ml PBS by intraperitoneal injection and 150 mg/kg control siRNA in 50 ml PBS by tail vein injection twice weekly. Group 6 received 150 mg/kg gemcitabine in 100 ml PBS by intraperitoneal
injection and 150 mg/kg RRM2 siRNA in 50 ml PBS by tail vein injection twice weekly (summarized in Table 2). Tumor dimensions were measured weekly using micrometer calipers. Tumor volumes were calculated using the formula: Volume ¼ 12ab2 where a and b represent the larger and smaller tumor diameters, respectively. After 6 weeks of treatment, necropsy was performed, primary tumors were excised, immediately weighed, and the mice were examined for liver metastases. The liver of each animal was harvested, metastatic foci were counted under a dissecting microscope (Gorelik et al., 1993) and confirmed by hematoxylin and eosin staining. Tumor growth inhibition (TGI) was calculated using the formula TGI (%) ¼ (1MT/ MC) 100, where MT and MC are the mean tumor masses in the treatment and control groups, respectively (Sun et al., 2003). TUNEL staining was performed on 5 mm sections of the excised tumors. The number of apoptotic cells in five random high-power fields were counted and expressed as a percentage of total cells (apoptotic fraction). Tumor homogenates were assayed for RRM2 and actin by Western blot and for caspase 3 activity, as previously described. Statistical analysis Data are expressed as means7s.d. Analysis was performed using ANOVA, two-sided t-test, and Mann–Whitney U-test or Kruskal–Wallis test for nonparametric data, as appropriate. Correlations between nonparametric data were analysed by Spearman correlation test. Analysis was performed using Statisticat 5.5 (Stat Soft, Tulsa, OK, USA), with Po0.05 being considered statistically significant. Acknowledgements We gratefully acknowledge the technical assistance of Jan Rounds. This work was supported by The National Pancreas Foundation and departmental funds from the Department of Surgery, Brigham and Women’s Hospital.
References Adjei AA, Dy GK, Erlichman C, Reid JM, Sloan JA, Pitot HC, Alberts SR, Goldberg RM, Hanson LJ, Atherton PJ, Watanabe T, Geary RS, Holmlund J and Dorr FA. (2003). Clin. Cancer Res., 9, 115–123. Aoki Y, Cioca DP, Oidaira H, Kamiya J and Kiyosawa K. (2003). Clin. Exp. Pharmacol. Physiol., 30, 96–102. Arlt A, Gehrz A, Muerkoster S, Vorndamm J, Kruse ML, Folsch UR and Schafer H. (2003). Oncogene, 22, 3243–3251. Bergman AM, Pinedo HM and Peters GJ. (2002). Drug Resist. Updat., 5, 19–33. Bertrand JR, Pottier M, Vekris A, Opolon P, Maksimenko A and Malvy C. (2002). Biochem. Biophys. Res. Commun., 296, 1000–1004. Brummelkamp TR, Bernards R and Agami R. (2002a). Science, 296, 550–553. Brummelkamp TR, Bernards R and Agami R. (2002b). Cancer Cell, 2, 243–247. Burris III HA, Moore MJ, Andersen J, Green MR, Rothenberg ML, Modiano MR, Cripps MC, Portenoy RK, Storniolo AM, Tarassoff P, Nelson R, Dorr FA, Stephens
CD and Von Hoff DD. (1997). J. Clin. Oncol., 15, 2403–2413. Calzado MA, Macho A, Lucena C and Munoz E. (2000). Clin. Exp. Immunol., 120, 317–323. Cory JG. (1997). Purine and pyrimidine nucleotide metabolism. Devlin T (ed). Biochemistry with Clinical Correlations. Wiley-Liss: New York, pp. 489–523. Editorial (2003). Nat. Cell Biol., 5, 489–490. Endicott JA and Ling V.. (1989). Annu. Rev. Biochem., 58, 137–171. Eriksson S and Martin Jr DW. (1981). J. Biol. Chem., 256, 9436–9440. Fan H, Villegas C and Wright JA. (1996). Proc. Natl. Acad. Sci. USA, 93, 14036–14040. Filleur S, Courtin A, Ait-Si-Ali S, Guglielmi J, Merle C, HarelBellan A, Clezardin P and Cabon F. (2003). Cancer Res., 63, 3919–3922. Fujioka S, Sclabas GM, Schmidt C, Frederick WA, Dong QG, Abbruzzese JL, Evans DB, Baker C and Chiao PJ. (2003). Clin. Cancer Res., 9, 346–354. Oncogene
RNA interference enhances chemosensitivity MS Duxbury et al
1548 Goan YG, Zhou B, Hu E, Mi S and Yen Y. (1999). Cancer Res., 59, 4204–4207. Gorelik E, Kim M, Duty L, Henion T and Galili U. (1993). Clin. Exp. Metastasis, 11, 439–452. Grant CE, Valdimarsson G, Hipfner DR, Almquist KC, Cole SP and Deeley RG. (1994). Cancer Res., 54, 357–361. Guittet O, Hakansson P, Voevodskaya N, Fridd S, Graslund A, Arakawa H, Nakamura Y and Thelander L. (2001). J. Biol. Chem., 276, 40647–40651. Heinemann V, Hertel LW, Grindey GB and Plunkett W. (1988). Cancer Res., 48, 4024–4031. Inayat MS, Chendil D, Mohiuddin M, Elford HL, Gallicchio VS and Ahmed MM. (2002). Cancer Biol. Ther., 1, 539–545. Jemal A, Murray T, Samuels A, Ghafoor A, Ward E and Thun MJ. (2003). CA Cancer J. Clin., 53, 5–26. Jong AY, Yu K, Zhou B, Frgala T, Reynolds CP and Yen Y. (1998). J. Biomed. Sci., 5, 62–68. Jung CP, Motwani MV and Schwartz GK. (2001). Clin. Cancer Res., 7, 2527–2536. Kelley JR, Fraser MM, Schweinfest CW, Vournakis JN, Watson DK and Cole DJ. (2001). Surgery, 130, 280–288. Kunsch C, Ruben SM and Rosen CA. (1992). Mol. Cell. Biol., 12, 4412–4421. Lee Y, Vassilakos A, Feng N, Lam V, Xie H, Wang M, Jin H, Xiong K, Liu C, Wright J and Young A. (2003). Cancer Res., 63, 2802–2811. Li K, Lin SY, Brunicardi FC and Seu P. (2003). Cancer Res., 63, 3593–3597. McClarty GA, Chan AK, Engstrom Y, Wright JA and Thelander L. (1987). Biochemistry, 26, 8004–8011. Monteith DK, Henry SP, Howard RB, Flournoy S, Levin AA, Bennett CF and Crooke ST. (1997). Anticancer Drug Des., 12, 421–432. Muerkoster S, Arlt A, Witt M, Gehrz A, Haye S, March C, Grohmann F, Wegehenkel K, Kalthoff H, Folsch UR and Schafer H. (2003). Int. J. Cancer, 104, 469–476.
Oncogene
Nieth C, Priebsch A, Stege A and Lage H. (2003). FEBS Lett., 545, 144–150. Pavloff N, Rivard D, Masson S, Shen SH and Mes-Masson AM. (1992). DNA Seq., 2, 227–234. Pessara U and Koch N. (1990). Mol. Cell. Biol., 10, 4146–4154. Plunkett W, Huang P, Searcy CE and Gandhi V. (1996). Semin. Oncol., 23, 3–15. Plunkett W, Huang P, Xu YZ, Heinemann V, Grunewald R and Gandhi V. (1995). Semin. Oncol., 22, 3–10. Raitano AB, Scuderi P and Korc M. (1990). Pancreas, 5, 267–277. Reichard P. (1978). Fed. Proc., 37, 9–14. Reichard P. (1985). Basic Life Sci., 31, 33–45. Sclabas GM, Fujioka S, Schmidt C, Evans DB and Chiao PJ. (2003). Int. J. Gastrointest. Cancer, 33, 15–26. Shen Z, Peedikayil J, Olson GK, Siebert PD and Fang Y. (2002). Biotechniques, 32(5), 1168, 1170–2, 1174 passim. Sun FX, Tohgo A, Bouvet M, Yagi S, Nassirpour R, Moossa AR and Hoffman RM. (2003). Cancer Res., 63, 80–85. Tanaka H, Arakawa H, Yamaguchi T, Shiraishi K, Fukuda S, Matsui K, Takei Y and Nakamura Y. (2000). Nature, 404, 42–49. Thelander L and Berg P. (1986). Mol. Cell. Biol., 6, 3433–3442. Thelander M, Graslund A and Thelander L. (1985). J. Biol. Chem., 260, 2737–2741. Wright JA, Alam TG, McClarty GA, Tagger AY and Thelander L. (1987). Somat. Cell Mol. Genet., 13, 155–165. Zhou B, Mo X, Liu X, Qiu W and Yen Y. (2001). Cytogenet. Cell Genet., 95, 34–42. Zhou BS, Hsu NY, Pan BC, Doroshow JH and Yen Y. (1995). Cancer Res., 55, 1328–1333. Zhou BS, Tsai P, Ker R, Tsai J, Ho R, Yu J, Shih J and Yen Y. (1998). Clin. Exp. Metastas., 16, 43–49.
