Colorectal cancer chemoprevention by 2 β-cyclodextrin inclusion compounds of auraptene and 4′-geranyloxyferulic acid

Authors

  • Takuji Tanaka,

    Corresponding author
    1. Department of Oncologic Pathology, Kanazawa Medical University, Uchinada, Ishikawa, Japan
    2. Tohkai Cytopathology Institute: Cancer Research and Prevention (TCI-CaRP), Gifu City, Gifu, Japan
    • Department of Oncologic Pathology, Kanazawa Medical University, 1-1 Daigaku, Uchinada, Ishikawa 920-0293, Japan
    Search for more papers by this author
    • Fax: +81-76-286-6926

  • Mariangela B.M. de Azevedo,

    1. Cnen/Ipen Instituto De Pesquisa Energéticas E Nucleares, Ed. Centro De Biotecnologia, Cidade Universitária, SP, Brazil
    Search for more papers by this author
  • Nelson Durán,

    1. Instituto de Química, Universidade Estadual de Campinas, UNICAMP, Campinas, CEP, SP, Brazil
    Search for more papers by this author
  • Joel B. Alderete,

    1. Organic Chemistry Department, Universidad de Concepción, Chile
    Search for more papers by this author
  • Francesco Epifano,

    1. Dipartimento di Scienze del Farmaco, Università “G. D'Annunzio” di Chieti-Pescara, Chieti Scalo (CH), Italy
    Search for more papers by this author
  • Salvatore Genovese,

    1. Dipartimento di Scienze del Farmaco, Università “G. D'Annunzio” di Chieti-Pescara, Chieti Scalo (CH), Italy
    Search for more papers by this author
  • Mayu Tanaka,

    1. Department of Pharmacy, Kinjo Gakuin University of Pharmacy, Moriyama-Ku, Nagoya, Aichi, Japan
    Search for more papers by this author
    • Mayu Tanaka and Takahiro Tanaka, who contributed equally to this work, were the summer students of the Department of Oncologic Pathology at Kanazawa Medical University.

  • Takahiro Tanaka,

    1. Department of Physical Therapy, Kansai University of Health Sciences, Kumatori-Machi, Sennan-Gun, Osaka, Japan
    Search for more papers by this author
    • Mayu Tanaka and Takahiro Tanaka, who contributed equally to this work, were the summer students of the Department of Oncologic Pathology at Kanazawa Medical University.

  • Massimo Curini

    1. Dipartimento di Chimica e Tecnologia del Farmaco, Sezione di Chimica Organica, Università degli Studi di Perugia, Perugia, Italy
    Search for more papers by this author

Abstract

The inhibitory effects of novel prodrugs, inclusion complexes of 3-(4′-geranyloxy-3′-methoxyphenyl)-2-trans propenoic acid (GOFA) and auraptene (AUR) with β-cyclodextrin (CD), on colon carcinogenesis were investigated using an azoxymethane (AOM)/dextran sodium sulfate (DSS) model. Male CD-1 (ICR) mice initiated with a single intraperitoneal injection of AOM (10 mg/kg body weight) were promoted by the addition of 1.5% (w/v) DSS to their drinking water for 7 days. They were then given a basal diet containing 2 dose levels (100 and 500 ppm) of GOFA/β-CD or AUR/β-CD for 15 weeks. At Week 18, the development of colonic adenocarcinoma was significantly inhibited by feeding with GOFA/β-CD at dose levels of 100 ppm (63% reduction in multiplicity, p < 0.05) and 500 ppm (83% reduction in the multiplicity, p < 0.001), when compared with the AOM/DSS group (multiplicity: 3.36 ± 3.34). In addition, feeding with 100 and 500 ppm (p < 0.01) of AUR/β-CD suppressed the development of colonic adenocarcinomas. The dietary administration with GOFA/β-CD and AUR/β-CD inhibited colonic inflammation and also modulated proliferation, apoptosis and the expression of several proinflammatory cytokines, such as nuclear factor-kappaB, tumor necrosis factor-α, Stat3, NF-E2-related factor 2, interleukin (IL)-6 and IL-1β, which were induced in the adenocarcinomas. Our findings indicate that GOFA/β-CD and AUR/β-CD, especially GOFA/β-CD, are therefore able to inhibit colitis-related colon carcinogenesis by modulating inflammation, proliferation and the expression of proinflammatory cytokines in mice.

There were ∼1 million new cases of colorectal cancer (CRC) in 2002 (9.4% of the total cancers).1 Globally, the mortality of CRC was reported to be 655,000 deaths per year in 2005.2 There is at least a 25-fold variation in the occurrence of CRC worldwide.1 The highest rates of incidence are in North America, Australia/New Zealand, Western Europe and Japan, especially in Japanese men.1 These large geographic differences for CRC are probably explained by differences in environmental exposures and lifestyles.

There are several types of pathogenesis of CRC.3 Among them, inflammation is linked with CRC development.4 The risk of CRC in patients with inflammatory bowel disease (IBD), including ulcerative colitis, increases with the increasing extent and duration of the disease.3, 5, 6 A mouse model was recently established for colitis-related colon carcinogenesis7 to facilitate the investigation of pathogenesis8–10 and the chemoprevention11, 12 of inflammation-related CRC. In this mouse model of inflammation-related two-stage colon carcinogenesis, different types of colonic carcinogens can be used in combination with a colitis-inducing agent, such as dextran sodium sulfate (DSS), and many colonic tumors develop within a short-term period.7, 13–15 The powerful tumor-promoting effect of DSS may be due to the oxidative/nitrosative stress that is caused by DSS-induced colitis.8–10 This suggests that the oxidative/nitrosative DNA damage associated with inflammation is involved in carcinogenesis, and, therefore, it is important to control the events that result in inflammation-related carcinogenesis.16 In humans, the inflammatory cytokines and oxidative stress also play a key role in the pathogenesis of IBD-related intestinal damage.17, 18 As our understanding of the pathogenesis of IBD is currently inadequate, drug therapy of IBD and INB-related CRC has been empirical, i.e., it is not based on a sound understanding of the etiology of the disease: drug therapy for IBD initially appears successful in the majority of IBD patients, and it comes with the risk of significant side effects. Therefore, we need new strategies, including chemoprevention, for IBD19 and IBD-associated CRC.3, 20–23

Abbreviations

AOM: azoxymethane; AUR: auraptene; CDs: cyclodextrins; COX: cyclooxygenase; CRC: colorectal cancer; DSS: dextran sodium sulfate; dUTP: deoxyuridine triphosphate; GOFA: 3-(4′-geranyloxy-3′-methoxyphenyl)-2-trans propenoic acid (4′-geranyloxy-ferulic acid); IBD: inflammatory bowel disease; IL: interleukin; iNOS: inducible nitric oxide synthase; NF-κB: nuclear factor-kappaB; Nrf2: NF-E2-related factor 2; TdT: terminal deoxynucleotidyl transferase; Tnf: tumor necrosis factor; TUNEL: TdT-mediated dUTP-biotin nick end labeling

The natural and semisynthetic cyclodextrins (CDs) have been extensively studied to improve certain properties of the drugs, such as solubility, stability and bioavailability.24 The CDs are suitable drug delivery systems because of their ability to greatly modify the physicochemical and biological properties of guest molecules through labile interactions by the formation of inclusion complexes. We have recently shown that modification of the physicochemical properties of violacein was achieved by the preparation of inclusion complexes with β-CD, thus leading to growth-inhibitory effects of its β-CD inclusion complexes against HL60 cells.25–27 Many drugs currently used in the therapeutic management of colon diseases have been used as inclusion complexes with CDs.28 The inclusion of the active principles in the cage represented by CD-protected drugs from absorption in the stomach and the upper portion of the lower intestine led to degradation of the saccharide portion in the large bowel by intestinal microflora, thereby ensuring a specific colon delivery with the maximum of bioavailability. This is the scope of the drugs that are being used in the therapy of malignant forms of colon cancer and IBD.

The 3-(4′-geranyloxy-3′-methoxyphenyl)-2-trans propenoic acid (4′-geranyloxy-ferulic acid, GOFA) (Fig. 1a) is a prenyloxycinnamic acid that was extracted from the Australian small plant Acronychia baueri Schott (Family, Rutaceae) in 1966, and, in the last decade, was seen to exert valuable anticancer effects, particularly against tumors affecting the gastrointestinal apparatus.23, 29 Auraptene (AUR) (Fig. 1b) is a geranyloxycoumarin that is widespread in the natural kingdom and was extracted from plants belonging to several families (mainly Rutaceae and Apiaceae), comprising many edible fruits and vegetables, such as lemons, grapefruits and oranges. Like GOFA, AUR was seen in recent years to exert valuable pharmacological properties,30 including dietary feeding colon cancer chemopreventive properties.22

Figure 1.

Chemical structures of (a) 3-(4′-geranyloxy-3′-methoxyphenyl)-2-trans propenoic acid (GOFA) and (b) auraptene (AUR).

As a continuation of our studies, we aimed to acquire further insights into the anticancer properties of selected prenyloxyphenylpropanoids. In our study, we wish to report the colon cancer chemopreventive activity of 2 novel prodrugs of GOFA/β-CD and AUR/β-CD that were obtained as their inclusion complexes with β-CD, using an inflammation-associated mouse colon carcinogenesis initiated with azoxymethane (AOM) and promoted by DSS.7 For the mechanistic investigation of the effects of the 2 prodrugs on AOM/DSS-induced tumorigenesis, we determined the immunohistochemical expression of the proinflammatory cytokines, including nuclear factor-kappaB (NF-κB),31–33 NF-E2-related factor 2 (Nrf2),31, 34 tumor necrosis factor (Tnf)-α35–37 and STAT336 in adenocarcinomas that developed in the colon. In addition, the expression of interleukin (IL)-638, 39 and IL-1β40 was evaluated in the colonic epithelial malignancies. The effects of GOFA/β-CD and AUR/β-CD in the diet on cell proliferation and apoptosis of colonic adenocarcinomas were evaluated using proliferating cell nuclear antigen (PCNA)21, 41 for proliferative activity, apoptosis indices by terminal deoxynucleotidyl transferase (TdT)-mediated deoxyuridine triphosphate (dUTP)-biotin nick end labeling (TUNEL) method21 and positive rate of survivin42 for apoptosis-inhibiting activity.

Material and Methods

Preparation of the inclusion complexes

GOFA and AUR were prepared according to previously reported methods.43, 44 β-CD was purchased from Aldrich Chemical. The inclusion complexes with a 1:1 molar ratio of GOFA to β-CD and AUR to β-CD (113.5 mg, 0.1 mmol) were obtained by dissolving the geranyloxy derivative (0.1 mmol) in 100 mL of acetone, and soon thereafter, slowly evaporating the solution to dryness under vacuum in a rotatory evaporator at 45°C.45 The structure of both inclusion compounds was determined by thermal analysis, X-ray diffraction, IR and NMR analysis, as already described.46 The thermogravimetric analysis (TGA) data were obtained using a Polymer Laboratories (STA-625) thermal analyzer. The samples (2–6 mg) were heated in sealed aluminum pans under nitrogen flow (50 cm3 min−1) at a heating rate of 10°C min−1 from 50 to 500°C.25, 27, 45, 46 The powder X-ray diffraction patterns were recorded using a 6000-XRD (Shimadzu X-ray diffractometer) under the following conditions: Ni-filtered CuK radiation, voltage 40 kV, current 30 mA, at a scanning speed of 2° min−1 and count range 1,000 CPS. The detector was a proportional counter with a 1.7-kV detector voltage.47–49 The samples of the solid dispersions and the physical mixtures of the complex and free drug and free CD were mixed with KBr and was pressed into a small tablet, which was mounted in the infrared beam. The spectra were recorded on the Perkin Elmer Model 1760X FTIR spectrometer from the KBr discs in the 500–4,000 cm−1 region.

Preclinical chemopreventive experiment

Animals, chemicals and diets

Male Crj: CD-1 (ICR) mice (Charles River Japan, Tokyo, Japan), 5 weeks of age, were used in our study. The animals were maintained in the Kanazawa Medical University Animal Facility according to the Institutional Animal Care Guidelines. All the animals were housed in plastic cages (5 mice/cage) and had free access to tap water and a pelleted Charles River Formula (CRF)-1 basal diet (Oriental Yeast, Tokyo, Japan) during quarantine under controlled conditions of humidity (50 ± 10%), lightning (12-hr light/dark cycle) and temperature [(23 ± 2)°C]. They were quarantined for 7 days after arrival and randomized by body weight into the experimental and the control groups. A colonic carcinogen AOM was purchased from Sigma-Aldrich Chemical (St. Louis, MO). DSS with a molecular weight of 36,000–50,000 Da (Lot no. 6046H) was purchased from MP Biomedicals, LLC (Aurora, OH). DSS for the induction of colitis was dissolved in water at 1.5% (w/v). β-CD inclusion complexes of GOFA (GOFA/β-CD) and AUR (AUR/β-CD) were synthesized, as described earlier. The experimental diets containing 0, 100 and 500 ppm of GOFA/β-CD (MW 1465.43) or AUR/β-CD (MW 1433.39) in a powdered basal diet CRF-1 were prepared weekly in our laboratory and stored in a cold room. The doses were selected based on our previous studies.22, 23 The animals had access to food and water at all times. The food cups were replenished daily with a fresh diet. All the handling and procedures were carried out in accordance with the Institutional Animal Care Guidelines.

Experimental procedures

The Institutional Animal Care and Use Committee evaluated all the animal procedures that were associated with our study and assured that all the proposed methods were appropriate.

A total of 150 male ICR mice were divided into 5 experimental and control groups (Supporting Information Fig.). The mice in Groups 1–5 were initiated with AOM by a single intraperitoneal injection (10 mg/kg body weight). One week after the injection, 1.5% DSS (w/v) in drinking water was administrated to mice of Groups 1–5 for 7 days, followed by no further treatment for 18 weeks. The mice of Group 1 were maintained on the CRF-1 diet throughout the study. The mice of Groups 2 and 5 were fed CRF-1 diets containing 100 ppm GOFA/β-CD (Group 2), 500 ppm GOFA/β-CD (Group 3), 100 ppm AUR/β-CD (Group 4) and 500 ppm AUR/β-CD (Group 5) for 15 weeks, respectively, starting 1 week after the cessation of DSS exposure. Group 6 received AOM injection alone. Group 7 was treated with DSS alone. Groups 8 and 9 did not receive AOM and DSS and were fed CRF-1 diets containing 500 ppm GOFA/β-CD and AUR/β-CD, respectively. Group 10 did not receive any treatments and served as an untreated control. At the end of study (Week 18), all the mice were killed by CO2 asphyxiation for careful necropsy, with emphasis on colon, liver, kidney, lung and heart.

At necropsy, the colons were flushed with saline, excised, their length measured (from ileocecal junction to the anal verge), cut open longitudinally along the main axis and then washed with saline. They were cut and fixed in 10% buffered formalin for at least 24 hr. A histological examination was performed on the paraffin-embedded sections after hematoxylin and eosin (H&E) staining by one (T.T.) of the investigators. Colonic tumors were diagnosed according to Ward's description.50 In brief, if the tumor cells with tubular formation invaded into the depth of the submucosa, the tumor was diagnosed as adenocarcinoma. When the tumor cells with glandular structure did not invade the submucosa and compressed the surrounding crypts, the tumor was diagnosed as adenoma.

Scoring of inflammation in the large bowel

Inflammation in the large bowel was scored on the H & E-stained sections made from all the mice. For scoring, the large intestinal inflammation was graded according to the following morphological criteria51: Grade 0, normal appearance; Grade 1, shortening and loss of the basal 1/3 of the actual crypts with mild inflammation in the mucosa; Grade 2, loss of the basal 2/3 of the crypts with moderate inflammation in the mucosa; Grade 3, loss of the entire crypts with severe inflammation in the mucosa and submucosa, but with retention of the surface epithelium; Grade 4, presence of mucosal ulcer with severe inflammation (infiltration of neutrophils, lymphocytes and plasma cells) in the mucosa, submucosa, muscularis propria and/or subserosa. The scoring was made on the entire colon with or without proliferative lesions and expressed as a mean average score/mouse.

Immunohistochemistry of NF-κB, Nrf2, Tnf-α, Stat3, IL-6, IL-1β, PCNA, TUNEL and survivin

The immunohistochemical analysis of the colon adenocarcinomas for the antibodies of NF-κB, Nrf2, Tnf-α, Stat3, IL-6, IL-1β, PCNA, TUNEL and survivin was performed on 4-μm-thick paraffin-embedded sections by applying the labeled streptavidin biotin method using a LSAB KIT (DAKO Japan, Kyoto, Japan), with microwave accentuation. The paraffin-embedded sections from the colonic neoplasms of the mice in each group (n = 18 in Group 1, n = 11 in Group 2, n = 6 in Group 3, n = 6 in Group 4 and n = 3 in Group 5) were heated for 30 min at 65°C, deparaffinized in xylene and rehydrated through graded ethanol at room temperature. Tris-HCl buffer (0.05 M, pH 7.6) was used to prepare the solutions and was used for the washes between the various steps. The incubations were performed in a humidified chamber.

The sections were treated for 40 min at room temperature with 2% bovine serum albumin and incubated overnight at 4°C with primary antibodies. The primary antibodies included anti-NF-κB p50 (H-119) rabbit polyclonal antibody (#sc-7178, 1:500 dilution; Santa Cruz Biotechnology, Santa Cruz, CA), anti-rabbit Nrf2 polyclonal antibody (#ab31163, 1:500 dilution; Abcam, Cambridge, MA), anti-human Tnf-α rabbit polyclonal antibody (#ab6671, 1:500 dilution; Abcam), anti-mouse Stat3 rabbit polyclonal antibody (#ab31370, 1:250 dilution; Abcam), anti-rabbit IL-6 polyclonal antibody (#ab6672, 1:400 dilution; Abcam), anti-mouse IL-1β rabbit polyclonal antibody (#LS-B40, 1:250 dilution; LifeSpan BioSciences, Seattle, WA), anti-rabbit survivin (71G4B7E) monoclonal antibody (#2808, 1:2,000 dilution; Cell Signaling Technology, Danvers, MA) and anti-human PCNA mouse monoclonal antibody (DAKO #U 7032, 1:1,000 dilution; DakoCtytomation, Kyoto, Japan). These antibodies were applied to the sections according to the manufacturer's protocol. The horseradish peroxidase activity was visualized by the treatment with H2O2 and 3,3′-diaminobenzidine for 5 min. At the last step, the sections were weakly counterstained with Mayer's hematoxylin (Merck, Tokyo, Japan). For each case, the negative controls were performed on the serial sections without the first antibodies.

The levels of apoptosis in tumor tissues determined by the TUNEL method were done on 4-μm formalin-fixed, paraffin-embedded tissue sections of the colonic adenocarcinomas, according to the manufacturer's instructions using the Apoptosis in situ Detection Kit Wako (Cat. No. 298-60201, Wako Pure Chemical Industries, Osaka, Japan). The kit is based on the TUNEL procedure. The appropriate positive and negative controls for determining the specificity of staining were generated. The negative controls were processed in the absence of the TdT enzyme in the reaction buffer. Sections of tissue digested with nuclease enzyme and colon lymphoid nodules, which are known to exhibit high rates of apoptosis, were used as the positive controls. The color was developed with the peroxidase substrate 3,3′-diaminobenzidine and the sections were counterstained with Mayer's hematoxylin (Merck).

Immunohistochemical evaluation and scoring

The immunoreactivity against the antibodies, except PCNA, TUNEL and survivin, was assessed in the large colonic adenocarcinomas (more than 3 mm in diameter) developed in Groups 1–5 using a microscope (Olympus BX41, Olympus Optical, Tokyo, Japan). The intensity and localization of the immunoreactivity against the primary antibodies were determined by a pathologist (T.T.) who was unaware of the treatment group to which the slide belonged. The immunoreactivity was evaluated against the NF-κB, Nrf2, Tnf-α, Stat3, IL-6 and IL-1β antibodies with grading between 0 and 5: 0 (∼15% of the colonic cancer cells showing positive reactivity), 1 (16–30% of the colonic cancer cells showing positive reactivity), 2 (31–45% ofthe colonic cancer cells showing positive reactivity), 3 (46–60% of the colonic cancer cells presenting positive reactivity), 4 (61–75% of the colonic cancer cells showing positive reactivity) and 5 (∼75% of the colonic cancer cells showing positive reactivity).

The number of nuclei with positive reactivity for PCNA-, TUNEL- and survivin-immunohistochemistry was counted in a total of 3 × 100 cells in 3 different areas of the colonic cancer and expressed as a percentage (mean ± SD).

Statistical evaluation

Where applicable, the data were analyzed using 1-way ANOVA with Tukey-Kramer Multiple Comparisons Test (GraphPad Instat version 3.05, GraphPad Softwear, San Diego, CA) with p < 0.05 as the criterion of significance. The Fisher's exact probability test was used for comparison of the incidence of lesions between the 2 groups.

Results

General observation

During the experiment, a few animals of Groups 1–5 and 7 (DSS alone) had bloody stool, but the symptom disappeared soon after stopping the DSS treatment. At Week 18, some of the mice of Groups 1–5 had bloody stool again and anal prolapse because of rectal tumor. The mice belonging to Groups 6 (AOM alone), 8 (GOFA/β-CD alone), 9 (AUR/β-CD alone) and 10 (untreated) did not have any symptoms related to the treatments during the experimental period. As summarized in the Supporting Information Table, there was no significant change between the experimental groups with respect to the parameters tested (body and spleen weights). The liver and relative liver weights of Groups 6 and 8 were significantly smaller in comparison to Group 10. With respect to colon length, the value of Group 1 was significantly lower in comparison to Groups 6 (p < 0.05) and 7 (p < 0.05). The colon length of Group 3 was significantly larger in comparison to Group 1 (p < 0.001).

Pathological findings

Macroscopically, nodular and/or polypoid colonic tumors developed in the middle and distal colon of the mice in Groups 1–5. These tumors were histopathologically tubule adenoma (Fig. 2a) or adenocarcinoma (well and moderately differentiated) (Fig. 2b) with a few adenocarcinomas that invaded into the serosa (Fig. 2c). A mucosal ulcer (Fig. 3a) was also observed surrounding the neoplasms. The enlarged lymph nodes with inflammation were present around the large bowel with tumors. The mice of Groups 6–10 had no tumors in all the organs examined, including the colon. A mucosal ulcer was observed in the colon of some of the mice of Group 7.

Figure 2.

Representative colonic neoplasms induced by azoxymethane (AOM)/dextran sodium sulfate (DSS) in a mouse (Group 1). (a) A tubular adenoma, (b) a tubular adenocarcinoma with moderately differentiated and (c) a tubular adenocarcinoma invaded into the submucosa. Note: The severe inflammation around the tumors; Hematoxylin and eosin stain; the inserted bars indicate magnification (μm).

Figure 3.

Representative colonic lesions induced by azoxymethane (AOM)/dextran sodium sulfate (DSS) in a mouse (Group 1). (a) Mucosal ulcer and (b) dysplastic crypts (circled). Hematoxylin and eosin stain, the inserted bars indicate magnification (μm).

The incidences and multiplicities of the colon neoplasms are summarized in Table 1. Group 1 (AOM/DSS) had 64% incidence of colonic adenocarcinoma with a multiplicity of 1.96 ± 2.24. The incidences of colonic adenocarcinoma of Groups 2 (AOM/DSS → 100 ppm GOFA/β-CD, 24%), 3 (AOM/DSS → 500 ppm GOFA/β-CD, 13%) and 5 (AOM/DSS → 500 ppm AUR/β-CD, 25%) were significantly smaller in comparison to Group 1 (p < 0.005, p = 0.0001 and p < 0.005, respectively). Also, the multiplicities of colonic adenocarcinoma of Groups 2 (0.52 ± 1.16, p < 0.01), 3 (0.25 ± 0.74, p < 0.001) and 5 (0.42 ± 0.83, p < 0.05) were significantly smaller in comparison to Group 1. The incidence (46%) and multiplicity (1.21 ± 1.61) of Group 4 (AOM/DSS → 100 ppm AUR/β-CD) were lower in comparison to Group 1, but the differences between the groups were insignificant. The incidences and multiplicities of colonic adenomas and total colonic tumors in Groups 2–5 were also lower in comparison to Group 1 (Table 1).

Table 1. Effects of compounds A and B on the development of colonic adenoma and adenocarcinoma
inline image

Other colonic lesions, including colitis with or without mucosal ulcer (Fig. 3a) and cryptal dysplastia (Fig. 3b), were also observed in the colon of mice in Groups 1–5 and/or 7 (Table 2). With respect to the inflammation score (Table 2) determined on H&E-stained sections at Week 18, the value of Group 1 was the highest among the groups and the scores of Groups 2–5 were significantly smaller in comparison to Group 1 (p < 0.001 for each comparison). Similarly, as shown in Table 2, the number of colonic mucosal ulcer per colon of Group 1 was the greatest, and the values of Groups 2–5 were significantly smaller in comparison to Group 1 (p < 0.01 or p < 0.001). The inflammation score and number of mucosal ulcer of Group 7 were the second among the group. Colonic inflammation in the mice of Groups 6, 8, 9 and 10 was slight, if present, and there were mucosal ulcers in the colon of the mice belonging to these groups.

Table 2. Effects of compounds A and B on colonic inflammation and development of mucosal ulcer and high-grade dysplasia
inline image

PCNA- and survivin-labeling index in the colonic adenocarcinomas

The data for the PCNA-, TUNEL- and survivin-positive rates of adenocarcinomas are illustrated in Figure 4. As shown in Figure 4a, the mean labeling indices of PCNA of Groups 2 (56.3 ± 11.2, p < 0.001), 3 (57.3 ± 9.6, p < 0.001), 4 (71.5 ± 9.4, p < 0.01) and 5 (61.7 ± 9.4, p < 0.001) were significantly lower in comparison to Group 1 (84.5 ± 9.4) (Fig. 4d). The mean TUNEL-positive rates of Groups 2 (12.00 ± 4.56, p < 0.05), 3 (13.70 ± 4.04, p < 0.05), 4 (10.55 ± 3.62) and 5 (13.17 ± 2.79, p < 0.01) were greater in comparison to Group 1 (7.67 ± 1.28) (Fig. 4d). With respect to the positive rates of survivin, the values of Groups 2 (38.0 ± 8.5, p < 0.01), 3 (23.0 ± 7.6, p < 0.001), 4 (53.4 ± 11.3) and 5 (41.3 ± 6.6, p < 0.05) were smaller in comparison to Group 1 (58.1 ± 12.6) (Fig. 4d).

Figure 4.

(a) The PCNA-labeling index (%, mean ± SD), (b) the TUNEL-positive rate (%, mean ± SD) and (c) the survivin-positive rate (%, mean ± SD) of colonic adenocarcinomas developed in mice from Groups 1–5. (d) The photos show representative PCNA-, TUNEL- and survivin-immunohistochemistry from Group 1.

Scores of NF-κB, Nrf2, Tnf-α, Stat3, IL-6 and IL-1β immunohistochemistry

The data for the scores of the immunohistochemical expression of these proinflammatory cytokines in colonic adenocarcinomas are illustrated in Figures 5a5c and 6a6c. Adenocarcinomas and the inflammatory mononuclear cells in the colon positively reacted with the antibodies of the proinflammatory cytokines, such as NF-κB, Nrf2, Tnf-α, Stat3, IL-6 and IL-1β (Figs. 5d and 6d). The scores of NF-κB (Fig. 5a), Stat3 (Fig. 6a), IL-6 (Fig. 6b) and IL-1β (Fig. 6c) of Groups 2–5 were significantly lower in comparison to Group 1. Similarly, the mean scores of Nrf2 (Fig. 5b) and Tnf-α (Fig. 5c) of Groups 2, 3 and 5 were significantly smaller in comparison to Group 1. Both values of Group 3 were lower in comparison to Group 1, but the differences were insignificant.

Figure 5.

The scores (mean ± SD) of (a) NF-κB-, (b) Nrf2- and (c) Tnf-α-immunoreactivity of colonic adenocarcinomas developed in mice from Groups 1–5. (d) The photos show representative NF-κB-, Nrf2- and Tnf-α-immunohistochemistry from Group 1. Note: The adenocarcinoma cells strongly expressed NF-κB, Nrf2 and Tnf-α.

Figure 6.

The scores (mean ± SD) of (a) Stat3-, (b) IL-1β- and (c) IL-6-immunoreactivity of colonic adenocarcinomas developed in the mice from Groups 1–5. (d) The photos show representative Stat3-, (b) IL-1b- and (c) IL-6-immunohistochemistry from Group 1. Note: The adenocarcinoma cells strongly expressed Stat3, IL-1β and IL-6.

Discussion

The results of our study clearly indicated that the novel prodrugs, GOFA/β-CD and AUR/β-CD, effectively inhibited AOM/DSS-induced colitis-related colonic carcinogenesis without any adverse effects in mice. The effect of GOFA/β-CD was superior in comparison to AUR/β-CD. Dietary feeding with both prodrugs exerted their cancer chemopreventive ability by modulating cell proliferation, inducing apoptosis and suppressing the proinflammatory cytokines (NF-κB, Nrf2, Tnf-α, Stat3, IL-6 and IL-1β) in adenocarcinomas that developed in the inflamed colon. In turn, the expression of these cytokines may be involved in AOM/DSS-induced colon tumorigenesis. This is the first report showing that prodrugs of GOFA/β-CD and AUR/β-CD exert cancer chemopreventive ability in colitis-related colon carcinogenesis.

In our study, several proinflammatory cytokines were expressed in the colonic tumors and the inflammatory mononuclear cells infiltrated the tumors both internally and peripherally. As the expression of these cytokines may be involved in tumor growth,52–54 we evaluated the effects of dietary GOFA/β-CD and AUR/β-CD on their expression in adenocarcinomas developed in Groups 1–5. The treatment with GOFA/β-CD and AUR/β-CD significantly lowered colonic inflammation induced by DSS. Chronic inflammation is involved in oncogenesis in certain tissues, including the large bowel. Therefore, the suppression of chronic inflammation through the modulation of expression of several proinflammatory gene products that mediate several events of carcinogenesis may result in cancer chemoprevention.55 The modulation of inflammation and expression of cyclooxygenase (COX)-2 and inducible nitric oxide synthase (iNOS) in the colon results in the suppression of colitis-related colon carcinogenesis of mice.56 Several molecular targets for the suppression of inflammation-associated carcinogenesis were proposed.36 In addition to the highly expressed levels of COX-2 and iNOS of colonic adenocarcinomas in our study (data not shown), the proinflammatory cytokines, such as NF-κB, Tnf-α, Stat3, Nrf2, IL-6 and IL-1β, were strongly expressed in adenocarcinomas that developed in the colon of the mice that received AOM and DSS. Moreover, dietary feeding with GOFA/β-CD and AUR/β-CD suppressed their expressions. The Nrf2-deficient mice are susceptible to DSS-induced colitis.57 IL-6 and IL-1β are involved in the development of IBD and IBD-related colon cancer.58, 59 These proinflammatory cytokines are thus molecular targets for the chemoprevention of inflammation-related carcinogenesis.31, 37, 55, 60, 61 They are candidate biomarkers of colon tumorigenesis,62, 63 because the expression of NF-κB, Tnf-α and IL-1β is involved in colonic tumorigenesis by affecting proliferation and apoptosis.64–67 The activation of NF-κB, a transcription factor that is activated by several cytokines released during inflammation and is responsible for many of their proinflammatory effects, was shown to promote the growth of the colon tumors in experimental models.31, 55, 60, 68 Because of the strong link of NF-κB to different stress signals, including cigarette smoke, NF-κB has been called a “smoke-sensor” of the body.69 In this context, the findings that a tobacco-specific carcinogen enhances AOM/DSS-induced colon carcinogenesis70 are of interest. In addition, Stat3 expression is an important factor in colon carcinogenesis, tumor invasion71 and survival/proliferation of the colonic preneoplastic cells.72 In addition, the anti-inflammatory potential of melatonin through the suppression of the expression of NF-κB and chemokines (IL-8 and monocyte chemoattractant protein) in a rat colitis model73, 74 is of interest, and it is important to further investigate the cancer chemopreventive ability of this bioactive substance, as was done in our study.

In our study, the treatment with both compounds in the diet significantly lowered colonic inflammation induced by DSS. As chronic inflammation involves tumorigenesis and accelerate carcinogenic steps, the suppression of chronic inflammation through the modulation of the expression of several proinflammatory gene products that mediate a critical role in several events of carcinogenesis may result in the inhibition of cancer development, and it may also serve as cancer chemoprevention.55 AUR and GOFA possess anti-inflammatory activities.20, 75 In addition, we previously reported on the cancer chemopreventive ability of AUR22, 76 and a prodrug, GOFA (called GAP in the study23) of the secondary metabolite of ferulic acid in colitis-associated colon carcinogenesis.23, 76 Several molecular targets for the suppression of inflammation-associated carcinogenesis were proposed.36 Our recent study demonstrated that the modulation of inflammation and the expression of COX-2, iNOS and other proteins in the colon contribute to the suppression of colitis-related colon carcinogenesis.10, 56

CDs (cyclic oligomers of glucose) that have the properties of forming inclusion complexes with lipophilic drugs have been widely used in therapy to improve water solubility and bioavailability of drugs. Target tissue bioavailability is an important determinant of these efficacies of chemopreventive agents.77 In our study, we selected β-CD, which is soluble in water and organic solvents. When compared the chemopreventive efficacy of the inclusion complexes of GOFA and AUR with β-CD in our study to that of previous studies,22, 23 GOFA with β-CD was superior to GOFA23 and AUR with β-CD was less effective in comparison to AUR.22 This may be related to the differences between the thermal stability of GOFA/β-CD and AUR/β-CD. A thermogram of the AUR/β-CD complex showed that this coumarin derivative melting endotherm had a substantial reduction in peak area, thus implying that the molecular arrangement of AUR in the solid complex was different from the pure crystal compound.25, 27, 45, 46 Also, the different effects of these compounds with or without b-CD in the activity of matrix metalloproteinases of inflamed colon.78

In conclusion, the novel prodrugs of GOFA/β-CD and AUR/β-CD are effective in inhibiting colon cancer development in a two-stage colitis-related mouse colon carcinogenesis through modulation of inflammation, proliferation and the expression of several proinflammatory cytokines (NF-κB, Tnf-α, Stat3, Nrf2, IL-6 and IL-1β) in the inflamed colon of the mice that received AOM and DSS. Our findings therefore support the development of novel site-specifically delivered prodrugs for colon cancer prevention in the inflamed colon.

Acknowledgements

The authors thank Ms. Veronica Jimenez for the structural characterization of GOFA/β-CD and AUR/β-CD. This work was supported in part by a Grant-in-Aid (Nos. 18592076 to T.T., 17015016 to T.T. and 18880030 to Y.Y.) for Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology of Japan and a grant (H2007-12 to T.T. and S2006-9 to Y.Y.) for the Project Research from the High-Technology Center of Kanazawa Medical University.

Ancillary