Pancreatic ductal adenocarcinomas are thought to arise from noninvasive, intraductal precursor lesions called pancreatic intraepithelial neoplasias (PanIN). The study of PanINs holds great promise for the identification of early detection markers and effective cancer-preventing strategies. Cyclooxygenase-2 (COX-2) represents an intriguing target for therapeutic and preventive approaches in various human malignancies. The aim of the present study was to evaluate the efficacy of a selective COX-2 inhibitor to prevent the progression of PanINs in a conditional KrasG12D mouse model. Offspring of LSL-KRASG12D x PDX-1-Cre intercrosses were randomly allocated to a diet supplemented with the selective COX-2 inhibitor nimesulide (400 ppm) or a control diet. After 10 months, animals were sacrificed. Successful recombination in the pancreas was evaluated by PCR. The pancreas of KRASG12D;PDX-1-Cre mice was analyzed for the presence of murine PanINs. Animals fed the COX-2 inhibitor had significantly fewer PanIN-2 and PanIN-3 lesions than control animals (P < 0.05). Ten percent of all pancreatic ducts in the nimesulide-fed animals showed PanIN-2 or PanIN-3 lesions, whereas 40% of the pancreatic ducts in the control animals had PanIN-2 or PanIN-3 lesions. Intrapancreatic prostaglandin E2 levels were reduced in nimesulide-fed animals. Immunohistochemistry confirmed COX-2 expression in early and late PanINs. In summary, we found that the selective COX-2 inhibitor nimesulide delays the progression of pancreatic cancer precursor lesions in a preclinical animal model. These data highlight the importance of COX-2 in the development of pancreatic cancer. Inhibition of COX-2 may represent an intriguing strategy to prevent pancreatic cancer in high-risk patients. [Cancer Res 2007;67(15):7068–71]

Despite advances in the field of molecular genetics in human pancreatic cancers, the identification of various putative molecular targets with the development of targeted therapies has not yet translated to an improved overall patients' survival (1). The precise mechanistic contributions of activating mutations in proto-oncogenes and of inactivating mutations in tumor suppressor genes to the malignant phenotype of pancreatic cancers are essentially elusive (2). Only a minority of patients are diagnosed with this disease at a stage that renders them eligible for surgical resection. In this small cohort, complete tumor resection has yielded 5-year survival rates exceeding 30% in the most recent series of high-volume centers in the United States (3). In an effort to reduce the incidence of pancreatic cancers or to detect this tumor at an early stage eligible to curative surgical resection, recent research interest in this field has adopted models and approaches to identify risk factors for the development of pancreatic cancers to understand the biology of pancreatic cancer precursor lesions and to discover early detection markers. Today, it is generally accepted that pancreatic ductal adenocarcinomas arise from the stepwise progression of precursor lesions called pancreatic intraepithelial neoplasias (PanIN; ref. 4). The preclinical study of PanINs has recently been made possible by the generation of genetically modified animal models, which recapitulate human PanINs on a genetic and histomorphologic level (5). The conditional KrasG12D model, first described by Hingorani et al. (6), is considered a very valuable tool to study PanIN biology. In this model, the KrasG12D mutation, commonly found in human pancreatic cancers, is specifically expressed from its endogenous gene locus in pancreatic progenitor cells during embryologic development. To direct the mutation to pancreatic progenitor cells, LSL-KRASG12D mice, in which the KrasG12D mutation is silenced by a floxed STOP cassette, are crossed with PDX-1-Cre mice, which express Cre recombinase under the pancreas-specific promoter Pdx-1. Cre recombinase-mediated excision of the STOP cassette in pancreas progenitor cells leads to expression of the KrasG12D mutation in virtually all mature pancreatic cell lineages. LSL-KRASG12D;PDX-1-Cre mice will develop early murine PanIN (mPanIN) lesions, which progress to advanced mPanINs and eventually to frank pancreatic cancers (6).

Data from other carcinogenesis models strongly support the notion that cyclooxygenase-2 (COX-2), the inducible isoform of COXs, plays an important role in the development of various human tumors (712). Inhibiting COX-2 activity therefore represents an intriguing approach to attenuate PanIN progression and ultimately to lower the incidence of pancreatic cancers. In the present study, we describe the potential benefit of a selective COX-2 inhibitor in inhibiting mPanIN progression in the conditional KrasG12D model.

Reagents and animals. The selective COX-2 inhibitor nimesulide was purchased from Cayman Chemical. The experimental diets were prepared by Dyets, Inc. LSL-KRASG12D mice were obtained from the National Cancer Institute Mouse Models of Human Cancers Consortium (Frederick, MD), and PDX-1-Cre mice were kindly provided by Andrew Lowy (Division of Surgical Oncology, University of Cincinnati, Cincinnati, OH).

Conditional KrasG12D mouse model. To study the effect of a selective COX-2 inhibitor on pancreatic cancer precursor lesions, the conditional KrasG12D model from Hingorani et al. (6) was used. LSL-KRASG12D and PDX-1-Cre mice were maintained as heterozygous lines. Offspring of LSL-KRASG12D and PDX-1-Cre mice were fed an AIN-93G–based purified diet supplemented with the selective COX-2 inhibitor nimesulide (400 ppm; n = 30). Mice fed the AIN-93G–based diet without the inhibitor served as controls (n = 30). Mice were individually housed. Food intake and body weight of each animal were measured twice weekly. After 10 months, animals were euthanized and the entire pancreas was harvested for further analyses.

Genotyping analysis. Successful excision-recombination events were confirmed by the presence of a single LoxP site in the pancreas as described elsewhere (6). Briefly, genomic DNA was extracted from snap-frozen pancreas tissue samples using the DNeasy Tissue kit (Qiagen, Inc.). PCR primers for detecting the single LoxP site were as follows: 5′-GGGTAGGTGTTGGGATAGCTG and 3′-TCCGAATTCAGTGACTACAGATGTACAGAG. PCR was carried out using the Advantage GC 2 PCR Kit and Polymerase Mix (Clontech Laboratories, Inc.). PCR products were separated on a 2% agarose gel. Successful recombination (single LoxP site) yields a 325-bp product (285-bp product in wild-types).

Histologic evaluation. Formalin-fixed, paraffin-embedded tissues were sectioned (4 μm) and stained with H&E. Six to eight sections (100 μm apart) of pancreatic tissues were histologically evaluated by a gastrointestinal pathologist blinded to the experimental groups. mPanIN lesions were classified according to histopathologic criteria as recommended elsewhere (5, 13). To quantify the progression of mPanIN lesions, the total number of ductal lesions and their grade were determined. Only the highest grade lesion per pancreatic lobule was evaluated. About 100 to 130 pancreatic ducts of the entire fixed specimen (head, body, and tail of the pancreas) were analyzed for each animal. The relative proportion of each mPanIN lesion to the overall number of analyzed ducts was recorded for each animal.

COX-2 immunohistochemistry. Tissue sections (4 μm) were deparaffinized in xylene and rehydrated in a graded ethanol series. Sections were subjected to heat-induced epitope retrieval using boiling citrate buffer (pH 6.0). Endogenous peroxidase activity was quenched by 0.3% H2O2 and unspecific binding sites were blocked with 5% bovine serum albumin. Immunostaining was done with a rabbit monoclonal COX-2 antibody (clone SP21, Lab Vision Corp.). Immunoreactivity was detected by the UltraVision Detection System using 3-amino-9-ethylcarbazol as the chromogen according to the manufacturer's recommendation (Lab Vision).

Intrapancreatic prostaglandin E2 levels. Total tissue extracts were prepared as described previously (14). Briefly, tissue specimens were homogenized and sonicated in antiprotease lysis buffer (Roche Applied Science) and prostaglandin E2 (PGE2) levels were determined by a PGE2 Immunoassay (R&D Systems) and normalized to total protein content.

Statistical analysis. Data are presented as mean ± SD. Differences in the mean of two samples were analyzed by an unpaired t test. Comparisons of more than two groups were made by a one-way ANOVA with post hoc Holm-Sidak analysis for pairwise comparisons and comparisons versus control. An α value of 0.05 was used to determine significant differences. All statistics were done in SigmaStat 3.1 (Systat Software, Inc.).

The study of pancreatic carcinogenesis and thus the evaluation of dietary and chemoprevention regimens have recently been made possible by the generation of genetically modified mouse models, which recapitulate the main characteristics of pancreatic cancer development on the histopathologic and genetic level (5, 13). The conditional KrasG12D model displays intraductal lesions that closely resemble human PanINs. The mPanINs thereby progress from early to advanced lesions and eventually to invasive pancreatic adenocarcinomas (6). To study the efficacy of a selective COX-2 inhibitor on mPanIN development and progression, we fed LSL-KRASG12D x PDX-1-Cre mice with a diet supplemented with nimesulide (400 ppm; n = 30) or a control diet (n = 30) for 10 months. At 10 months, >80% of pancreatic ducts usually display mPanINs with >30% mPanIN-2 and mPanIN-3 lesions (6). The dose of nimesulide was chosen based on previously published reports showing antitumor efficacy in mice (7, 9). There was no difference in food intake and weight gain between both experimental groups during the entire study period (data not shown). Because the successful recombination event (i.e., Cre-mediated excision of the STOP cassette) in this model can only be detected in the pancreas but not in other tissues (i.e., tails or ears), all animals (n = 60) were euthanized after 10 months and the entire pancreas was harvested for further studies. PCR analyses of pancreatic genomic DNA of all 60 mice revealed that the successful recombination event was present in six and five mice fed the nimesulide-supplemented and the control diet, respectively. Histopathologic examination by a blinded pathologist showed that mice fed the selective COX-2 inhibitors had a significant delay in the progression of mPanIN lesions. About 100 to 130 pancreatic ducts were analyzed per mouse. Control diet–fed animals had mPanIN-1, mPanIN-2, and mPanIN-3 lesions in about 50%, 30%, and 10% of all pancreatic ducts, respectively, whereas ∼70% and 10% of pancreatic ducts in nimesulide-fed animals showed mPanIN-1 and mPanIN-2, respectively (Fig. 1). No mPanIN-3 lesions were detected in nimesulide-fed animals. Conversely, only 10% of pancreatic ducts in the control animals appeared normal, whereas ∼20% of pancreatic ducts were classified normal in the nimesulide-fed animals (Fig. 1). Because human PanIN-3 is generally considered “carcinoma in situ” with a high likelihood to develop into frank cancer, whereas the biological significance of PanIN-1 and PanIN-2 is currently debatable, we specifically analyzed the effect of nimesulide on mPanIN-3 lesions compared with all other lesions combined. The proportion of mPanIN-3 was significantly decreased, whereas the relative numbers of all other lesions combined (normal ducts, mPanIN-1, and mPanIN-2) were significantly increased in the nimesulide-fed animals (Fig. 1). No invasive cancers were present in either group at 10 months. Although invasive pancreatic cancer can be detected in very few mice within 10 months, the majority of LSL-KRASG12D x PDX-1-Cre mice will develop pancreatic cancers at 12 to 15 months of age (5, 6). In mice without the recombination event, no pancreatic pathology was observed in both experimental groups (data not shown). To our knowledge, this is the first report showing that the genetically predetermined progression of mPanIN lesions in the conditional KrasG12D mouse model can be attenuated by interventional strategies. Immunohistochemistry showed COX-2 protein expression in early and advanced mPanINs of LSL-KRASG12D;PDX-1-Cre mice (Fig. 2) with no obvious difference in COX-2 staining between control and nimesulide-fed animals. COX-2 protein was not detected in normal ducts of LSL-KRASG12D;PDX-1-Cre mice and wild-type littermates. It is noteworthy that COX-2 was generally expressed in only a subset of epithelial cells within an individual mPanIN lesion. This staining pattern was confirmed by two additional anti-COX-2 antibodies (data not shown). The pathways leading to COX-2 gene transcription and protein expression in mPanIN lesions in this model are unclear at the moment. Because we observed only punctuated COX-2 expression in transformed epithelial cells, the KrasG12D mutation alone, which is present in all mature exocrine and endocrine pancreatic cells in this model, seems necessary but not sufficient to induce COX-2 expression. The pancreata of nimesulide-fed LSL-KRASG12D;PDX-1-Cre mice had significantly lower levels of PGE2, a surrogate marker for COX-2 activity, indicating that oral intake of nimesulide at a concentration of 400 ppm in the diet achieved tissue levels that were sufficient to reduce COX-2 enzymatic activity (Fig. 2). Our data clearly suggest that COX-2 and COX-2–derived prostanoids are critical mediators in pancreatic carcinogenesis. This notion is strongly supported by a recent report, which described the development of intraductal papillary mucinous neoplasm–like and PanIN-like structures in transgenic mice in which COX-2 expression in pancreatic ductal cells was driven by a keratin 5 promoter (15). The corollary may be that strategies aimed at inhibiting the prostaglandin production pathway have a tumor-preventive potential in pancreatic cancer development. A chemopreventive effect of selective COX-2 inhibitors has been documented in other genetically modified mouse models. Mammary tumors in HER-2/neu transgenic mice expressed the COX-2 protein, and dietary intake of the selective COX-2 inhibitor celecoxib significantly reduced the incidence of these tumors (10, 16). Selective COX-2 inhibitors also decreased the rate of intestinal polyp formation in mouse models of familial adenomatous polyposis (12, 17). Furthermore, celecoxib reduced the incidence of prostatic intraepithelial neoplasia and prostate cancer in the transgenic adenocarcinoma of the mouse prostate model (8, 18). However, the efficacy of a selective COX-2 inhibitor in a genetically modified mouse model of pancreatic cancer development has not been reported. In conclusion, our studies showed that oral intake of a selective COX-2 inhibitor delayed the progression of mPanINs in a conditional KrasG12D mouse model of pancreatic carcinogenesis, thus highlighting the potential benefit of inhibiting the COX-2/prostanoid pathway in preventing pancreatic cancer development. Further studies using LSL-KRASG12D;PDX-1-Cre mice treated for a longer time period or crossed on a Trp53R172H- or Ink4a/Arf-deficient background, which accelerates pancreatic cancer development (19, 20), will elucidate whether the actual reduction of mPanIN-3 lesions ultimately decreases the incidence of frank invasive pancreatic cancers.

Figure 1.

A, representative pancreatic histologies of LSL-KRASG12D;PDX-1-Cre mice at 10 months of age. Top left, normal duct; top center, mPanIN-1B; top right, mPanIN-1B high magnification; middle left, mPanIN-2; middle center, mPanIN-2; middle right, mPanIN-2 high magnification; bottom left, mPanIN-3; bottom center, mPanIN-3 high magnification; bottom right, mPanIN-3. Black box, area that was magnified in the following image. B, quantitative analysis of mPanINs in experimental groups. Top, percentage of pancreatic ducts with no pathology (nl), mPanIN-1A (1A), mPanIN-1B (1B), mPanIN-2 (2), and mPanIN-3 (3) lesions in control mice (n = 5) and nimesulide-fed mice (n = 6); bottom, percentage of combined normal ducts/mPanIN-1/mPanIN-2 and mPanIN-3 in control mice and nimesulide-fed animals. *, P < 0.05, control versus nimesulide.

Figure 1.

A, representative pancreatic histologies of LSL-KRASG12D;PDX-1-Cre mice at 10 months of age. Top left, normal duct; top center, mPanIN-1B; top right, mPanIN-1B high magnification; middle left, mPanIN-2; middle center, mPanIN-2; middle right, mPanIN-2 high magnification; bottom left, mPanIN-3; bottom center, mPanIN-3 high magnification; bottom right, mPanIN-3. Black box, area that was magnified in the following image. B, quantitative analysis of mPanINs in experimental groups. Top, percentage of pancreatic ducts with no pathology (nl), mPanIN-1A (1A), mPanIN-1B (1B), mPanIN-2 (2), and mPanIN-3 (3) lesions in control mice (n = 5) and nimesulide-fed mice (n = 6); bottom, percentage of combined normal ducts/mPanIN-1/mPanIN-2 and mPanIN-3 in control mice and nimesulide-fed animals. *, P < 0.05, control versus nimesulide.

Close modal
Figure 2.

A, COX-2 immunohistochemistry in mPanINs of LSL-KRASG12D;PDX-1-Cre mice at 10 months of age. B, intrapancreatic PGE2 levels in nimesulide-fed animals as measured by ELISA relative to control animals. *, P < 0.05, control versus nimesulide.

Figure 2.

A, COX-2 immunohistochemistry in mPanINs of LSL-KRASG12D;PDX-1-Cre mice at 10 months of age. B, intrapancreatic PGE2 levels in nimesulide-fed animals as measured by ELISA relative to control animals. *, P < 0.05, control versus nimesulide.

Close modal

Grant support: NIH grant R01 CA104027 (G. Eibl), Jonsson Cancer Center Foundation (G. Eibl), and Hirshberg Foundation for Pancreatic Cancer Research (H.A. Reber).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1
Eckel F, Schneider G, Schmid RM. Pancreatic cancer: a review of recent advances.
Expert Opin Investig Drugs
2006
;
15
:
1395
–410.
2
Su GH, Kern SE. Molecular genetics of ductal pancreatic neoplasia.
Curr Opin Gastroenterol
2000
;
16
:
419
–25.
3
Kazanjian KK, Hines OJ, Duffy JP, et al. Improved survival for adenocarcinoma of the pancreas after pancreaticoduodenectomy.
Gastroenterology
2005
;
128
:
A813
–4.
4
Maitra A, Fukushima N, Takaori K, Hruban RH. Precursors to invasive pancreatic cancer.
Adv Anat Pathol
2005
;
12
:
81
–91.
5
Hruban RH, Adsay NV, bores-Saavedra J, et al. Pathology of genetically engineered mouse models of pancreatic exocrine cancer: consensus report and recommendations.
Cancer Res
2006
;
66
:
95
–106.
6
Hingorani SR, Petricoin EF, Maitra A, et al. Preinvasive and invasive ductal pancreatic cancer and its early detection in the mouse.
Cancer Cell
2003
;
4
:
437
–50.
7
Fukutake M, Nakatsugi S, Isoi T, et al. Suppressive effects of nimesulide, a selective inhibitor of cyclooxygenase-2, on azoxymethane-induced colon carcinogenesis in mice.
Carcinogenesis
1998
;
19
:
1939
–42.
8
Gupta S, Adhami VM, Subbarayan M, et al. Suppression of prostate carcinogenesis by dietary supplementation of celecoxib in transgenic adenocarcinoma of the mouse prostate model.
Cancer Res
2004
;
64
:
3334
–43.
9
Kitamura T, Itoh M, Noda T, Matsuura M, Wakabayashi K. Combined effects of cyclooxygenase-1 and cyclooxygenase-2 selective inhibitors on intestinal tumorigenesis in adenomatous polyposis coli gene knockout mice.
Int J Cancer
2004
;
109
:
576
–80.
10
Lanza-Jacoby S, Miller S, Flynn J, et al. The cyclooxygenase-2 inhibitor, celecoxib, prevents the development of mammary tumors in Her-2/neu mice.
Cancer Epidemiol Biomarkers Prev
2003
;
12
:
1486
–91.
11
Liu W, Nakamura H, Tsujimura T, et al. Chemoprevention of spontaneous development of hepatocellular carcinomas in fatty liver Shionogi mice by a cyclooxygenase-2 inhibitor.
Cancer Sci
2006
;
97
:
768
–73.
12
Oshima M, Murai N, Kargman S, et al. Chemoprevention of intestinal polyposis in the ApcΔ716 mouse by rofecoxib, a specific cyclooxygenase-2 inhibitor.
Cancer Res
2001
;
61
:
1733
–40.
13
Hruban RH, Rustgi AK, Brentnall TA, et al. Pancreatic cancer in mice and man: the Penn Workshop 2004.
Cancer Res
2006
;
66
:
14
–7.
14
Eibl G, Takata Y, Boros LG, et al. Growth stimulation of COX-2-negative pancreatic cancer by a selective COX-2 inhibitor.
Cancer Res
2005
;
65
:
982
–90.
15
Muller-Decker K, Furstenberger G, Annan N, et al. Preinvasive duct-derived neoplasms in pancreas of keratin 5-promoter cyclooxygenase-2 transgenic mice.
Gastroenterology
2006
;
130
:
2165
–78.
16
Howe LR, Subbaramaiah K, Patel J, et al. Celecoxib, a selective cyclooxygenase 2 inhibitor, protects against human epidermal growth factor receptor 2 (HER-2)/neu-induced breast cancer.
Cancer Res
2002
;
62
:
5405
–7.
17
Swamy MV, Patlolla JM, Steele VE, et al. Chemoprevention of familial adenomatous polyposis by low doses of atorvastatin and celecoxib given individually and in combination to APCMin mice.
Cancer Res
2006
;
66
:
7370
–7.
18
Narayanan BA, Narayanan NK, Pittman B, Reddy BS. Regression of mouse prostatic intraepithelial neoplasia by nonsteroidal anti-inflammatory drugs in the transgenic adenocarcinoma mouse prostate model.
Clin Cancer Res
2004
;
10
:
7727
–37.
19
Aguirre AJ, Bardeesy N, Sinha M, et al. Activated Kras and Ink4a/Arf deficiency cooperate to produce metastatic pancreatic ductal adenocarcinoma.
Genes Dev
2003
;
17
:
3112
–26.
20
Hingorani SR, Wang L, Multani AS, et al. Trp53R172H and KrasG12D cooperate to promote chromosomal instability and widely metastatic pancreatic ductal adenocarcinoma in mice.
Cancer Cell
2005
;
7
:
469
–83.