Combination of Atorvastatin with Sulindac or Naproxen Profoundly Inhibits Colonic Adenocarcinomas by Suppressing the p65/β-Catenin/Cyclin D1 Signaling Pathway in Rats

Colorectal cancer is one of the major leading causes of death from cancer in the United States as well as in the worldwide. It is predicted to be responsible for the death of almost 50,000 people each year in the United States alone (1, 2). Because the majority of the cause of colon cancer is attributable to lifestyle, diet, and genetic factors, there has been increasing awareness and focus on the prevention of colon cancer (3, 4). In particular, the presence of chronic inflammatory conditions in the colonic environment has been implicated in the development of colorectal cancer, and treatment regimens against inflammatory markers have reduced the risk of colon cancer (5–7).

Increased aberrant expression of inflammatory genes, such as inducible nitric oxide synthase (iNOS) and COX, has been shown in the azoxymethane-induced colon cancer model from the early stage of hyperplastic aberrant crypt foci (ACF) to late-stage adenocarcinoma (8–14). Because many studies report that selective iNOS and COX-2 inhibitors exerted suppressive effects against colon cancer (8, 11, 15–22), there is a rationale for investigating the ability of the combination of low-dose atorvastatin and nonsteroidal anti-inflammatory drugs (NSAIDs) to inhibit iNOS and COX-2 in a colon cancer model where inflammatory genes play a key role in carcinogenesis.

Evidence supports the protective role of NSAIDs and statin (15, 16, 23–27). In our earlier study, we showed that statins such as atorvastatin (Lipitor), as well as NSAIDs as effective agents in suppressing colon cancer in animals (7, 16, 24). Azoxymethane-induced tumors result from mutations in the Wnt/β-catenin pathway (28–30). Aberrant expression of β-catenin can be regarded as a key event during colorectal tumorigenesis (31) and is linked to the increased transcription of a number of genes such as cyclin D1 (32, 33). Cyclin D1 is overexpressed in patients with adenomatous polyps, primary colorectal adenocarcinoma, and familial adenomatous polyposis (32, 34). Cyclin D1 is a target gene of the Wnt signaling pathway (35), and mutations in this pathway are responsible for approximately 90% of colorectal cancer (36). Mutations in genes belonging to the Wnt pathway, such as inactivating mutations in the adenomatous polyposis coli (APC) gene or activating mutations in β-catenin, result in the nuclear accumulation of β-catenin and subsequent complex formation with T-cell factor/lymphoid enhancer factor (TCF/LEF) transcription factors to activate gene transcription (37). TCF/LEF-binding sites on promoters of cell proliferation genes, such as cyclin D1 and c-MYC (35, 38), thus serve to transmit the aberrant mutations to tumorigenic signals within the colonic crypts.

As discussed above, we and others have shown the chemopreventive effects of statins and number of NSAIDs (7, 16, 20, 23, 24, 39). However, many of these studies used higher dose levels and also no efficacy data on the most commonly used NSAID, naproxen, in the colon cancer model is available. Importantly, our aim to establish whether low-dose combinational approaches targeting different pathways would be ideal for human colorectal cancer prevention. Thus, in the present study, experiments were designed to evaluate the efficacies of atorvastatin and NSAIDs, sulindac and naproxen, administered individually and in combination against colon tumorigenesis. We evaluated the chemopreventive potential of a low-dose atorvastatin in combination with NSAIDs with colonic tumor formation as the endpoint and further determined the action of atorvastatin and in combination with NSAIDs in regulating the expression of key protein markers and signaling pathway during colon carcinogenesis.

Atorvastatin, sulindac, and naproxen (Fig. 1) were provided by the DCP Repository at the National Cancer Institute. Conversion products from [³H]-l-arginine to [³H]-l-citrulline were obtained from New England Nuclear Corporation.

Weanling male F344 rats obtained from Charles River Breeding Laboratories were randomly distributed by weight into control and experimental groups. Animals had access to food and water at all times. Food cups were replenished with fresh diet twice weekly. Experimental diets were purchased from Research Diets and stored at 4°C. Beginning at 5 weeks of age, all rats were fed the modified American Institute of Nutrition-76A (AIN-76A) diet. At 7 weeks of age, the animals were given subcutaneous injections of azoxymethane (CAS no. 25843-45-2; Ash Stevens) at a dose rate of 15 mg/kg body weight or saline as solvent control once weekly for 2 weeks. One week after the second azoxymethane treatment, groups of rats were fed AIN-76A diet containing atorvastatin (200 ppm), sulindac (100 ppm), naproxen (150 ppm), or their combinations with low-dose atorvastatin (100 ppm) for 45 weeks. At autopsy, animals were sacrificed by CO₂ asphyxiation, and the colon was removed, rinsed in PBS, opened longitudinally, and flattened on a filter paper. The location and size of each tumor was noted. Mucosal scrapings were collected and stored at −80°C for further analysis. Tumors were removed, fixed in 10% buffered formalin for 24 hours, and transferred to 70% ethanol for histopathologic analysis.

The tumor tissues were dehydrated, embedded in paraffin, and cut into 4 μm thick sections. For histopathology, the sections were hydrated and stained with hematoxylin and eosin according to the standard protocol. The stained sections were analyzed for tumor grades by a pathologist. For immunohistochemical analysis, only noninvasive adenocarcinomas were selected for the evaluation of protein markers. The detailed procedures for immunohistochemical analysis are reported previously (40). The primary antibodies against proliferating cell nuclear antigen (PCNA; 1:1,500 diluted) from BD PharMingen; cyclin D1 (1:500 diluted), β-catenin (1:500 diluted), phospho-p65 (1:250 diluted), and iNOS (1:500 diluted) all from Santa Cruz Biotechnology; and COX-2 (1:200 diluted) from Cayman Chemical were treated on the sections. The images were taken randomly at 400× using Zeiss AxioCam HRc camera fitted to a Zeiss Axioskope 2 Plus microscope. For β-catenin quantification, Image Pro 6.2 Plus (Media Cybernetics, Inc.) was used to obtain the IOD (integrated optical density = average intensity/density of each object) values.

iNOS activities were determined in colonic tumor samples of rats exposed to various experimental diets. Conversion of [³H]-l-arginine to [³H]-l-citrulline was measured by a modification described previously (7). iNOS activity is expressed as nanomoles of [³H]-l-citrulline per milligram of protein per minute.

Colonic mucosa samples were homogenized in a PBS-based buffer solution (PBS, 0.4 mol/L NaCl, 10 mmol/L EDTA, 0.1 mmol/L phenylmethylsulfonylfluoride, 0.1 mol/L benzethonium ion, 0.5% bovine serum albumin, 3.0% aprotinin, and 0.05% Tween 20) on ice using a Tekmar Tissuemiser (Fisher Scientific International, Inc.). The homogenized solution was centrifuged at 10,000 rpm at 4°C for 10 minutes. The supernatant was collected for determination of protein concentration and stored at −20°C. For determination of the levels of interleukin (IL)-1β, IL-4, and TNF-α, tissue homogenates were normalized down to a concentration of 1.0 mg/mL of total protein and then diluted 10-fold in diluent buffer for analysis, following the manufacturer's protocols. Invitrogen Immunoassay kits (BioSource International Inc.) were used to determine the levels of IL-1β (catalogue no. KRC0012), IL-4 (catalogue no. KRC0042), and TNF-α (catalogue no. KRC3012).

Statistical significance was analyzed using Student's t test or ANOVA test followed by Tukey's multiple comparison test. Tumor incidence was analyzed by 2-tailed Fisher's exact probability test.

Body weights of animals fed the experimental diets containing atorvastatin, sulindac, or naproxen individually or in combination were comparable with those fed the control diet throughout the study, indicating that the dose of atorvastatin, sulindac, or naproxen used did not cause any overt toxicity. The maximum tolerated dose (MTD) for each agent was previously determined (sulindac ∼400 ppm, naproxen ∼700 ppm, and atorvastatin >600 ppm). Therefore, the doses were determined on the basis of the information with these agents in AIN-76A diet on the F344 rats (16, 20, 24, 25). In the present study, we used the lower MTD doses of sulindac (∼25% MTD), naproxen (∼20% MTD), and atorvastatin (∼30% and 15%), respectively. Importantly, administration of these dose levels would produce plasma area under the curve (AUC) levels in rats that would somewhat equal the plasma AUC levels of humans given low to mid doses of these agents.

The effects of administration of a low-dose atorvastatin with sulindac or naproxen on azoxymethane-induced colon tumorigenesis were evaluated, and the results are summarized in Table 1. None of the rats in the saline groups (without azoxymethane injection, n = 6 per group) developed tumors when autopsied at week 45 (data not shown). Most of azoxymethane-treated control diet fed rats developed adenocarcinomas at 45 weeks. Histopathologic analysis by hematoxylin and eosin staining revealed more than 90% of the tumors from the control group as adenocarcinomas and the remaining less than 10% were carcinoma in situ. Approximately 95% of the total adenocarcinomas belonged to the noninvasive adenocarcinoma grade whereas the remaining 5% was invasive adenocarcinoma (Table 1). Administration of sulindac and naproxen individually had modest inhibitory (∼25% incidence and ∼33% multiplicity) effect on colon adenocarcinomas. However, atorvastatin (200 ppm) significantly suppressed both colon tumor incidence (52% reduction, P = 0.005) and multiplicity (58% reduction, P = 0.008). Most importantly, total colon tumor incidence was significantly decreased when rats were given low-dose atorvastatin with either sulindac (58% reduction, P = 0.001) or naproxen (63% reduction, P = 0.0005), respectively (Table 1). Colon tumor multiplicities were also profoundly reduced when rats were given low-dose atorvastatin with either sulindac or naproxen (80%–85%, P < 0.0001; Table 2).

As shown in Figure 2A (first row), the histologic evaluation revealed that the majority of tumors were noninvasive adenocarcinomas. The expression of PCNA, a marker for cell proliferation, was determined in the adenocarcinomas from the control and treatment groups. The colon tumors from a low-dose atorvastatin, in combination with sulindac or naproxen, fed group showed significant reduction of PCNA nuclear staining compared with the control group (Fig. 2A, second row). Aberrant expression of β-catenin can be considered as a key event during colorectal tumorigenesis and is linked to the increased transcription of a number of genes such as cyclin D1 (32, 33). β-Catenin was identified along the membrane of the epithelial cells in the control group. Compared with the control, all treatment groups showed marked inhibition of β-catenin membrane staining: sulindac (35.6% inhibition), naproxen (41.6% inhibition), atorvastatin (41.7% inhibition), atorvastatin + sulindac (59.4% inhibition), and atorvastatin + naproxen (54.6% inhibition; Fig. 2A, third row). The colonic crypt cells in the control group showed homogeneous and intense staining for β-catenin in the cytosol as well as in the membrane, with lower and scattered staining in the nucleus. In contrast, the tumors from the treatment groups had no observable nuclear staining. Furthermore, the cytoplasmic expression of β-catenin was also markedly inhibited by the treatment with atorvastatin alone and in combination with sulindac or naproxen (Fig. 2A, third row). Because cyclin D1 is a downstream signaling target of β-catenin, and overexpression of cyclin D1 is reported in patients with colorectal tumors where its lowering has therapeutic significance (32, 33), we determined whether treatment reduces cyclin D1 levels in colon tumors. Positive brownish staining of cyclin D1 in the control group predominantly localized in both the cytoplasm and nucleus. Administration of a low dose of atorvastatin, in combination with sulindac or naproxen, markedly reduced the staining for cyclin D1 in both the cytoplasm and the nucleus (Fig. 2A, fourth row).

Overexpression of inflammatory markers is a hallmark of colorectal tumors (41, 42). This knowledge led us to examine the effects of long-term feeding of treatment with atorvastatin and NSAIDs on the inflammatory markers in the azoxymethane-injected rats. There was significant inhibition of the expression of iNOS and COX-2 proteins within the crypts in the adenocarcinomas from the treatment groups, compared with those from the control group (Fig. 2B). We also determined the effects of each treatment on a key nuclear factor kappaB (NF-κB) signaling molecule, p65, because NF-κB is an upstream factor of both iNOS and COX-2 transcription, and it is critical in the tumorigenesis where ablation of the proteins in this pathway caused the regression of tumors in animal models (43). The activated form of NF-κB subunit p65, that is, phospho-p65, is markedly reduced in the nucleus of the colon tumors from the treatment groups, when compared with those from the control group (Fig. 2B). In addition, significant inhibition of the iNOS enzyme activity was shown in tumors from naproxen, atorvastatin, and combination treatment groups. Importantly, combinational treatment of atorvastatin with sulindac or naproxen showed maximal inhibition on iNOS enzyme activity compared with single agents (Fig. 3). In summary, colon adenocarcinomas from atorvastatin and NSAIDs fed animals showed reduced expression of key inflammatory markers as well as nuclear staining for phospho-p65, a key molecule in the NF-κB pathway.

Inflammatory cytokines are found to be present in human cancers including those of the colorectum, breast, prostate, and bladder (44, 45). The action of cytokines to facilitate carcinogenesis is multifold: DNA damage by reactive oxygen species and reactive nitrogen species; inhibition of DNA repair by reactive oxygen species; functional inactivation of tumor suppressor genes; tissue remodeling via activation of matrix metalloproteinases; and stimulation of angiogenesis and control of cell adhesion molecules (45). ELISA conducted for inflammatory cytokines on mucosal scrapings derived from the azoxymethane-injected rats are shown in Table 2. Sulindac treatment alone strongly inhibited the production of cytokines, TNF-α by 36.2% (P = 0.004), IL-1β by 26.6% (P = 0.04), and IL-4 by 34.0% (P = 0.03). More importantly, administration of a low-dose atorvastatin, in combination with sulindac or naproxen, significantly lowered the levels of cytokines in the colon; TNF-α by 36.8% (P = 0.03) and 33.3% (P = 0.01); IL-1β by 35.7% (P = 0.01) and 31.7% (P = 0.03); IL-4 by 39.9% (P = 0.03) and 40.4% (P = 0.01), respectively.

This is the first report on the combination treatment using low-dose atorvastatin with either low-dose sulindac or naproxen, which greatly suppress the colon adenocarcinoma incidence and multiplicity. Our results suggest that decreased inflammatory cytokines and signaling molecules, particularly inhibition of nuclear p65, β-catenin, and cyclin D1, are responsible for suppression of colonic adenocarcinomas. The present study is an extension of our previous work, which identified atorvastatin, sulindac, and naproxen as effective agents in suppressing colon cancer in animals (7, 16, 24). The results from the current research conducted in colon cancer reveal that administration of a low dose of atorvastatin, in combination with sulindac or naproxen, reduces the colon tumor multiplicity and regulates intermediate signaling pathways of proliferation and inflammation in the colon (Fig. 4).

A comparison of tumor numbers across the different grades of tumor shows an overall reduction by the treatment with a low dose of atorvastatin, in combination with sulindac or naproxen. Importantly, statistical analysis on tumor data revealed a profound inhibitory effect of low-dose atorvastatin with either sulindac or naproxen on adenocarcinomas. Unlike the APC^Min/+ mice, in which most of tumors are localized to the small intestine and those are predominantly adenomas, in the azoxymethane-induced rat, intestinal tumors are mostly localized to distal colon and adenocarcinomas, similar to human etiology. Thus, the results of our present study provide potential significance for human clinical trials. Also, it is important to note that in our previous studies, we have shown that 150 ppm atorvastatin inhibits colon adenocarcinoma incidence and multiplicity (∼34% inhibition, P ≤ 0.05) whereas, in this study, use of 200 ppm of atorvastatin suppressed more than 52% incidence (P = 0.005) and more than 58% multiplicity (P < 0.008) of colon adenocarcinomas. These results suggest that a modest dose increase in atorvastatin (from 150 ppm to 200 ppm) significantly enhances the chemopreventive efficacy.

Azoxymethane-induced tumors result from mutations in the Wnt/β-catenin pathway (28) as does the APC^Min/+ model. However, unlike the APC^Min/+ model, azoxymethane-induced tumors are caused by mutations in the β-catenin gene (29, 30). These mutations result in β-catenin stabilization and aberrant expression of β-catenin, which is considered as a key event during colon tumorigenesis (31). Immunohistochemical analysis revealed abundance of β-catenin mostly in the cytoplasm and relatively low nuclear staining in the adenocarcinomas of rats injected with azoxymethane whereas administration of a low dose of atorvastatin, in combination with sulindac or naproxen, markedly reduced the staining for β-catenin in both the cytoplasm and the nucleus (Fig. 2A).

Cyclin D1 is a very well-known cell-cycle protein targeted by β-catenin (35) and is known to be overexpressed in colonic tumors (32, 34). c-MYC is yet another important protein for cell proliferation regulated by β-catenin and Wnt pathway (38). These observations on cyclin D1 were corroborated by the potency of a low dose of atorvastatin, in combination with sulindac or naproxen, to affect the β-catenin levels in the colon tumors. In our studies, we identified a low dose of atorvastatin, in combination with sulindac or naproxen, to significantly lower the levels of cyclin D1 in the colon tumors induced with azoxymethane (Fig. 2A). More importantly, nuclear levels of β-catenin and cyclin D1 are reported to play more important role in tumorigenesis than the total protein levels (46, 47). In our studies, a low dose of atorvastatin, in combination with sulindac or naproxen, reduced the levels of cyclin D1 and β-catenin in the nucleus (Fig. 2A).

In addition to the effects on β-catenin and cell proliferation, our results indicate the anti-inflammatory property of a low dose of atorvastatin, in combination with sulindac or naproxen. We observed marked reduction in the staining intensities for iNOS, COX-2, and phospho-p65 markers as well as for the iNOS enzyme activity in colon tumors from the combination treatment groups (Figs. 2B and 3). Mucosal levels of inflammatory cytokines, such as TNF-α, IL-1β and IL4, were also significantly downregulated by a low dose of atorvastatin, in combination with sulindac or naproxen (Table 2). Several anti-inflammatory agents that target the nitric oxide or the prostaglandin pathway are reported to present chemopreventive action in the colon (4, 7, 48). A clinical trial on celecoxib, the selective COX-2 inhibitor, at a dose of 400 mg daily reduced advanced adenoma formation in the colon by almost 50% compared with the placebo through a 3-year treatment period (49). In addition to anti-inflammatory and antiproliferative mechanisms, azoxymethane treatment may also generate oxidative stress and significant genotoxicity by inducing methyl–DNA adducts; however, these processes may significantly subside within 12 to 18 hours after the azoxymethane treatment. In this study, we administered chemopreventive agents 1 week after the carcinogen treatment and thus possibility of carcinogenic action of azoxymethane via inhibition of oxidative stress or DNA adducts is very minimal to none by the chemopreventive agents. Promising results with other agents, such as the use of low concentrations of difluoromethylornithine and sulindac as chemopreventive agents in colorectal cancer, highlight the potential role of inflammation in its pathogenesis and the importance of combination strategies (48).

In conclusion, a low dose of atorvastatin, in combination with sulindac or naproxen, inhibits profoundly colon tumorigenesis by regulating the p65/β-catenin/cyclin D1 signaling pathway and the inflammatory responses. Overall, the data indicate that a low dose of atorvastatin, in combination with sulindac or naproxen, holds great promise in the field of colon cancer chemoprevention in humans.

No potential conflicts of interest were disclosed.

The authors thank Maria Hyra and Lamberto R. Navoa of the Animal Facility in the Department of Chemical Biology for their technical assistance in taking care of the animals.

This work was supported by NCI-N01-CN-53300, R01-CA94962, and the Trustees Research Fellowship Program at Rutgers, The State University of New Jersey.

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.