The COX-2 inhibitor nimesulide suppresses superoxide and 8-hydroxy-deoxyguanosine formation, and stimulates apoptosis in mucosa during early colonic inflammation in rats


As we have shown previously [Tardieu,D., Jaeg,J.P., Cadet,J., Embvani,E., Corpet,D.E. and Petit,C. (1998) Cancer Lett, 134, 1–5], a 48-h treatment of 6% dextran sodium sulphate (DSS) in drinking water led to a reproducible 2-fold increase of the mutagenic oxidative lesion 8-oxo-7,8-dihydro-2′-deoxyguanosine (8-oxodGuo) in colonic mucosa DNA of rats in vivo. The aim of this study was to test the effect of nimesulide, a preferential COX-2 inhibitor, on the DSS-induced 8-oxodGuo increase. We show that nimesulide when administered orally, simultaneously with DSS at 5 mg/kg/day, not only totally prevents 8-oxodGuo formation but also suppresses the 5-fold increase of superoxide induced by DSS in the colonic mucosa. This was measured by in vivo formazan blue precipitation (P < 0.01 in the Wilcoxon test). Moreover, nimesulide enhances apoptosis by ~30% as compared with the already high level induced by DSS treatment (P < 0.01). It is suggested that the significant increase in mutagenic oxidative DNA damage, produced by mild acute colonic inflammation, could be important in the initiation of colon cancer in both animals and man. These effects may explain at least partly the well-documented protective action towards colon cancer by preferential COX-2 inhibitors, either xenobiotics such as nimesulide or natural nutrients.



There is evidence that oxidative stress may be involved in the initiation and promotion of colon carcinogenesis (13), as well as in cancers from other organs. Among the factors involved in oxidative events leading to DNA damage in colonocytes, inflammation may be important, which is attested by at least three facts. (i) Chronic colitis leads to an increased frequency of colon cancers in animals and humans (4,5). (ii) Non-steroidal anti-inflammatory drugs (NSAID) protect man against colon cancer by ~2-fold and COX-2 selective reduces by 50% the cancer incidence in rats (68). (iii) As we have shown previously, even a mild acute inflammation can double the amount of 8-oxo-7,8-dihydro-2′-deoxyguanosine (8-oxodGuo) in the DNA of rat colonic mucosa within 48 h (9). In addition, the role of oxidative DNA damage has been recently emphasized as a major mutagenic event in eukaryotic cell (10).

Nimesulide, a NSAID that is both a preferential COX-2 inhibitor (1113) and at higher concentrations, an antioxidant (14), is able to suppress the promotion of colonic aberrant crypt foci (15), intestinal polyps (16) or colonic tumours (17) in rats. Hence, we wished to know whether nimesulide could prevent early oxidative DNA damage and superoxide formation in rat colonic mucosa in vivo under inflammatory conditions. We show here that nimesulide completely suppresses the increase of 8-oxodGuo, a mutagenic oxidative DNA damage observed during the early dextran sodium sulphate (DSS)-induced inflammation. This may both explain the protective role of NSAID towards colon cancer and suggest a role of mild acute inflammation in colonic genotoxicity.

Materials and methods

Animal treatments

Sixty-six female, 4-week-old, Fisher 344 rats (Iffa credo, Lyon, France), were acclimatized for 12 days to the animal colony, fed ad libitum a low fat rodent chow (UAR, Villemoisson, France) and randomized to three groups of 22 rats (six animals in one cage). A control group was given drinking water ad libitum. Two experimental groups received 6% DSS (ICN Biomedicals, Aurora OH) in distilled drinking water for 2 days (`DSS-treated group'). One of these groups received 5 mg/kg per os nimesulide (Novartis Santé Animale, Rueil-Malmaison, France) in drinking water 1 day before and during the DSS treatment. Animals were killed 2 days after the start of DSS administration by cervical dislocation.

DNA extraction and 8-oxodGuo assay

The rat colons (10 animals in each group) were dissected between ileocaecal junction and the proximal rectum, placed on a non-absorbent surface, and flushed with cold NaCl 0.9% (4°C). The colons were opened longitudinally and the mucosa removed by scratching the distal surface with a lancet blade (5 cm at the rectal end). The samples were rinsed with NaCl 0.9% and centrifuged at 3000 g for 10 min. The pellets were immediately frozen individually at –80°C for the 8-oxodGuo assay, which was performed within 2 days.

The Quiagen kit genomic DNA (Quiagen, Courtaboeuf, France) was used to extract DNA at ambient temperature from cellular pellets. After homogenization with a potter, the samples were treated with proteinase K (1 mg/ml) at 50°C for 2 h in buffer G2 (800 mM guanidine–HCl, 30 mM EDTA, 30 mM Tris–HCl, 5% Tween-20, 0.5% Triton X-100, pH 8.0). Lysates (10 ml) were then applied to the equilibrated (750 mM NaCl, 50 mM MOPS, 0.15% Triton X-100, pH 7) Quiagen genomic tips. After washing tips with QC buffer (1.0 M NaCl, 50 mM MOPS, 15% ethanol, pH 7.0), the purified genomic DNA was eluted with QF buffer (1.25 M NaCl, 50 mM Tris–HCl, 15% ethanol, pH 8.5). The DNA was precipitated by 2-propanol, washed with 70% ethanol, air dried and diluted in 100 μl sterile water before analysis.

DNA in aqueous solution was hydrolysed to nucleotides using 10 U nuclease P1 (Sigma-Aldrich, St Louis, USA) for 2 h at 37°C in acetate buffer (pH 5.3) according to Floyd (18). Dephosphorylation was achieved using 2 U alkaline phosphatase (Sigma) in Tris–HCl buffer (pH 8.0) for 1 h at 37°C. Proteins were precipitated with chloroform and the aqueous layer was analysed by high pressure liquid chromatography (HPLC) with electrochemical (EC) detection.

The HPLC system consisted of a pump (M 2200, Bischoff, Leonberg, Germany); an autosampler 738 (ICS, Toulouse, France) equipped with a 12.5×4.6 mm Spherisorb ODS 2 column (Bischoff); a detector focus (Spectra-Physics, Les Ulis, France) set at 280 nm. And a EC detector Coulochem II (ESA Chelmsford, MA), set at +200 and +400 mV for electrodes 1 and 2 respectively, monitored unmodified and modified nucleosides. An isocratic eluent, 50 mM phosphate buffer (pH 5.5) containing 10% methanol, was used at a flow rate of 1.0 ml/min. Oxidative DNA damage was quantified with the software PIC 3 (ICS) using calibration curves realized with authentic standards (Sigma-Aldrich).

Measurement of myeloperoxydase activity

Myeloperoxydase (MPO), a polymorphonuclear neutrophil marker, was assessed in colon samples (six in each group) according to a method modified from Bradley et al. (19).

Colons were homogenized on ice in 2 ml 50 mM potassium phosphate buffer (pH 6) for 30 s using an ultraturrax. Three cycles of freeze–thaw in liquid nitrogen were realized. Suspensions were then centrifuged at 10 000 g for 15 min at 4°C. Pellets were resuspended with 500 μl hexadecyl trimethyl ammonium bromide buffer (0.5% HTAB, w/v in phosphate buffer, pH 6) in order to release MPO from the polymorphonuclear neutrophil granules. Samples were sonicated for 30 s at 4°C and centrifuged at 10 000 g for 15 min at 4°C. An aliquot of 50 μl supernatant from each sample was used to determine protein concentration by the Bradford method (Biorad, Ivry sur Seine, France) and the samples were assayed for MPO activity by visible spectrophotometry.

Samples were diluted in potassium phosphate buffer (pH 6.0) containing 0.167 mg/ml of O-dianisidine dihydrochloride and 0.0005% of hydrogen peroxide (Sigma-Aldrich). Absorbency at 450 nm was recorded during 2 min with an Uvikon 860 spectrophotometer (Kontron instruments, St Quentin en Yvelines, France). MPO from human neutrophils was used as a reference and MPO activity was expressed as MPO U/mg protein.

In situ detection of superoxide

Detection of superoxide was performed using the method described by Hagen et al. (20) to assay superoxide in liver using a nitro blue tetrazolium (NBT) perfusion. In the presence of superoxide, NBT forms insoluble blue formazan precipitates within the tissues. The method was adapted as follows. The animals (six in each group) were anaesthetized with chloroform, a midline laparotomy was realized and a solution of 8 ml 1% NBT in NaCl 0.9% at 37°C containing 0.2 U/ml heparin was slowly (within 2 min) perfused in the dorsal aorta. This was followed by a perfusion of 8 ml NaCl 0.9% at 37°C. The animals were killed immediately, the colons were dissected between ileocaecal junction and the proximal rectum and placed on a non-absorbent surface. They were then flushed with cold NaCl 0.9% (4°C) and fixed in PBS–formalin. Paraffin-embedded slices were stained by haematoxylin–eosin for histological examination. The counting of the precipitates was performed in situ by light microscopy under a magnification of ×400, using a Malassez grid to estimate the surface of mucosa. A total of ~270 mm2 colonic mucosa was counted for each rat and the results were standardized as number of cells containing formazan precipitates/100 mm2 of sliced mucosa.

In situ detection of apoptosis

To determine the relative number of apoptotic nuclei, we used the TdT-Frag EL DNA fragmentation detection kit according to the manufacturer's protocol (Oncogene research products, Cambridge, USA). The animals (six in each group) were different from those used for superoxide detection in order to avoid possible interaction with NBT.

Longitudinal sections of paraffin-embedded colons were treated with 20 μg/ml proteinase K in 10 mM Tris–HCl (pH 8) for 20 min at room temperature and rinsed with Tris-buffered saline (1× TBS, 20 mM Tris, pH 7.6, 140 mM NaCl).

Endogenous peroxidases were inactivated by immersion in 3% H2O2 for 5 min, followed by a TBS wash. Samples were then labelled by incubating with TdT labelling reaction mixture for 1.5 h at 37°C. The detection reaction was started by an incubation of 10 min with a blocking buffer, followed by another incubation with a diaminobenzidine–H2O2–urea solution for 10 min. The specimens were finally counter-stained with a methyl green solution.

For each section all crypt units were examined and the results were expressed as the number of apoptotic nucleus/100 mm2 of mucosa. Positive and negative controls were performed, and the morphology of the apoptotic nuclei was checked using histological analysis. The counting was performed using the same method as described above for formazan precipitates.


Data were first analysed using one-way analysis of variance, then the Student's t-test (8-oxodGuo measurement) or Wilcoxon test (others) were used to compare experimental and control group means.


Effects of DSS inflammation and nimesulide treatment on 8-oxodGuo content in colonic DNA

As we have described in a previous, independent experiment (9), 8-oxodGuo content in colonic enterocyte DNA was increased 50% in DSS-treated rats as compared with the control. Nimesulide treatment reset the 8-oxodGuo content to control levels (Figure 1), indicating a complete protective effect against oxidative DNA damage: although DSS-induced diarrhoea was not alleviated (not shown).

MPO activity

MPO activity reflects the migration of polymorphonuclear cells leading to the chronic phase of inflammation (19). In our conditions, the MPO measurement did not show any significant difference between the three groups (0.47, 0.47 and 0.48 U MPO/mg protein for control, DSS and DSS + nimesulide, respectively). This suggests that, due to the short treatment duration, no, or very little, migration of polynuclear cells occurred at the site of inflammation. In addition, the amount of white cells in the colonic mucosa of DSS-treated rats was not significantly increased as appreciated by histology. The degree of inflammation that we obtained in our conditions can thus be described as acute since histopathological signs of inflammation were present (not shown).

Oxidative stress in colonic mucosa

NBT specifically forms insoluble formazan blue precipitates in the presence of superoxide (20). In the colon of rats, NBT perfusion led to formazan precipitates, strictly localized inside the goblet cells of the colonic mucosa (Figure 2). While those precipitates were rarely observed in control animals, their number was increased 5-fold in DSS-treated rats. Nimesulide treatment reduced the number of precipitates to 1.5 times the background, which was not significantly different from the control value in the Wilcoxon test (Table I).


Since COX-2 inhibitors have been found to modulate apoptosis (21,22), we wished to know what were their effects in colonic mucosa under early acute inflammatory conditions.

Apoptosis was rarely observed in the colonic mucosa of control animals. The number of apoptotic cells dramatically increased in DSS-treated rats, and was further significantly increased by another 30% in nimesulide-treated animals (Table I).


In this study we wished to know whether DSS-induced oxidative DNA damage in colonocytes in vivo was prevented by the COX-2 inhibitor nimesulide. Since we found the answer to be clearly positive, we eventually assayed superoxide formation and showed that it was induced by DSS and was returned to the control level by the NSAID. Finally, we tested apoptosis in colonic mucosa and found that it was further increased by nimesulide even in inflammatory conditions.

DSS is an extensively studied model for both acute and chronic colon inflammations, depending on the dose and treatment length (23). It has long been shown to induce colon cancer in rats (24). In our conditions, DSS produced a moderate inflammation, along with mild diarrhoea within 48 h. Early acute inflammatory conditions did not lead to significant migration of polynuclear cells, nor MPO concentration increase in the colon. However, the colonic mucosa of DSS-treated rats was significantly inflamed, and histological examination showed erosions in the mucosa, mild vessel dilatation and mild oedema (not shown). Furthermore, it has recently been shown that COX-2 was induced in the colon under inflammatory condition (25).

The mild acute inflammation induced by DSS led to a doubling of 8-oxodGuo concentration in colonocytes DNA, as we have shown both previously (9) and in this work. An identical 2-fold increase was documented by other workers in human gastric mucosa infected by Helicobacter pylori (26) or in lungs of cigarette smokers (27), two cancer-prone situations in which a role for oxidative DNA damage has been suggested. In the gastritis induced by Helicobacter, COX-2 is induced also (28).

It has been demonstrated that 8-oxodGuo is a highly mutagenic lesion leading to transversions GC-TA (10). In the yeast Saccharomyces, a deletion of the OGG1 gene coding for the glycosylase responsible for the repair of 8-oxodGuo leads to a strong increase of transversions, even in normal culture conditions (29). Hence, 8-oxodGuo is virtually able to play a role in mutagenic events linked to both cancer initiation and promotion.

The DSS-induced 8-oxodGuo formation in colonic mucosa was accompanied by oxidative stress as measured by superoxide-precipitated formazan blue, that was about five times higher in DSS-treated rats and returned nearby the basal level through NSAID treatment. We observed the same protection with another NSAID, olsalazine (unpublished data). It is thus likely that early oxidative DNA damage is somehow linked with superoxide formation in the colonic mucosa independently from polynuclear migration that would occur later in the chronic state.

Finally, apoptosis was dramatically enhanced by DSS treatment, and nimesulide led to a further enhancement that was similarly observed with olsalazine (not shown). Apoptosis enhancement may have a beneficial consequence for eliminating heavily damaged mutation-prone cells in which DNA repair runs the risk of being suppressed. Although an increase in apoptosis linked to COX-2 inhibition has already been documented in colonic cells it was in non-inflammatory conditions (21,30).

Taken together, our data suggest that early acute inflammation, independent from the accumulation of polynuclear cells, is able to induce significant oxidative stress in colonic mucosa leading to oxidative DNA damage. A NSAID such as nimesulide, a preferential COX-2 inhibitor, alleviates the oxidative events and stimulates apoptosis.

Although it is not clear at this time whether the protection by nimesulide is primarily due to the inhibition of COX-2 or some associated mechanism(s), such as malondialdehyde formation (22), the overall results could explain at least partly the protection by nimesulide-like molecules against colon carcinogenesis.

Furthermore, our present data support the fact that mild inflammation may be an under-estimated risk factor for colon carcinogenesis. The protective role played by natural antioxidants in diets, such as green tea (3133), flavonoids (3) or that of natural COX-2 inhibitors such as curcumin (34), berberine (35) or resveratrol (36), may possibly be understood in this respect. It is also noteworthy that n-3 polyunsaturated fatty acids, which are likely cyclooxygenase inhibitors (3739), both protect against aberrant crypt foci promotion and induce apoptosis in rats (40).


The authors wish to thank Dr I.Raymond for examining the morphology of apoptotic nuclei. This work was supported by the Direction Générale de l'Enseignement et de la Recherche du Ministère de l'Agriculture, France.