Improvement of an experimental colitis in rats by lactic acid bacteria producing superoxide dismutase

ORIGINAL ARTICLE

Improvement of an Experimental Colitis in Rats by Lactic Acid Bacteria Producing Superoxide Dismutase Wei Han, MD,* Annick Mercenier, PhD,0z Afifa Ait-Belgnaoui, PhD,* Sonia Pavan, PhD,0 Florence Lamine, PhD,* Iris I. van Swam, BS,1 Michiel Kleerebezem, PhD,1 Christel Salvador-Cartier, PhD,* Michael Hisbergues, PhD,0 Lionel Bueno, PhD,* Vassilia Theodorou, PhD,* and Jean Fioramonti, PhD*

Key Words: colitis, superoxide dismutase, lactic acid bacteria

Abstract: The use of superoxide dismutases (SODs) in inflammatory diseases is hampered by their short circulatory half-life. To determine whether a bacterial supply of SOD into the colon might improve an experimental colitis, the effects of oral treatment with live recombinant lactic acid bacteria producing different amounts of SOD and those of colonic infusion of SOD were compared. Wistar rats were fitted with a catheter in the proximal colon through which TNBS was administered to induce colitis. Animals received a continuous intracolonic infusion of bovine SOD (40 U per rat per day) for 4 days after TNBS or were treated orally with live recombinant Lactococcus lactis or Lactobacillus plantarum strains (109 colony-forming units (CFU)/d), producing or not producing SOD, for 4 days before and after TNBS. SOD activity of bacterial extracts was 0, 26, 74, and 624 units/109 CFU for L. plantarum, L. lactis, L. lactis SOD+, and L. plantarum SOD+, respectively. Four days after TNBS, macroscopic and microscopic damage, myeloperoxidase (MPO) activity, and nitrotyrosine immunostaining were evaluated. TNBS induced macroscopic and microscopic damages, an increase in MPO activity, and intense immunostaining for nitrotyrosine. Macroscopic damage and MPO activity were reduced by bovine SOD. These parameters and microscopic damages also were reduced by L. lactis, L. lactis SOD+, and L. plantarum SOD+, but not by L. plantarum. Nitrotyrosine immunostaining was attenuated after treatment with the 4 bacterial strains. Although not all of the antiinflammatory effects could be attributed directly to SOD, our results suggest that SOD-producing lactic acid bacteria open a novel approach in inflammatory bowel disease treatment.

Received for publication October 25, 2005; accepted June 25, 2006. From the *Neurogastroenterology and Nutrition Unit, Institut National de la Recherche Agronomique, Toulouse, France; 0Bacteriology of Ecosystems, Institut Pasteur de Lille, Lille, France; 1Wageningen Centre for Food Sciences, NIZO Food Research, Ede, the Netherlands; and zDepartment of Nutrition and Health, Nestle´ Research Centre, Lausanne, Switzerland. Supported in part by the QLK3-CT2000-0146 project LABDEL granted by the EU. Dr Han was supported by a fellowship from INRA. Reprints: Jean Fioramonti, PhD, Neurogastroenterology and Nutrition Unit, INRA, 180 Chemin de Tournefeuille, BP3, F-31931 Toulouse Cedex 9, France (e-mail: [email protected]) Copyright * 2006 by Lippincott Williams & Wilkins

1044

(Inflamm Bowel Dis 2006;12:1044–1052)

T

he pathogenesis and molecular mechanisms of inflammatory bowel disease (IBD), including ulcerative colitis and Crohn`s disease, are not yet elucidated and are considered to be multifactorial.1Y3 A growing body of data indicates that excessive amounts of reactive oxygen metabolites (ROMs) such as superoxide (O2.j), nitric oxide (NO .), and hydroxyl radicals (OH.) have a role in mediating intestinal damage in IBD.4,5 By causing damage to lipids, DNA, and proteins, ROMs are known to disrupt epithelial cell integrity, with subsequent loss of fluids and electrolytes.6 They also are involved in various intracellular signaling pathways.7 They contribute to the perpetuation of the inflammatory process by activating redox-sensitive transcription factors that regulate the expression of a variety of inflammatory cytokines, adhesion molecules, and enzymes.8 Therefore, the reduction or elimination of adverse oxidant effects corresponds to a promising therapy for IBD. Superoxide dismutases (SODs), converting superoxide anion to hydrogen peroxide (H2O2),9 are considered a primary defense against oxidative stress. However, SOD-based therapy has not been used yet in IBD treatment.10 Therapeutic uses of SOD are hampered by its short circulatory half-life of only 5 to 10 minutes,11 and efforts are being made to find suitable delivery vehicles for these enzymes.12 Accumulating evidence suggests that probiotics, mainly lactic acid bacteria (LAB), represent promising candidates for the prevention and control of IBD.13 Data from animal models of colitis have indicated that specific LAB strains may prevent and treat intestinal inflammation.14 Lactococcus lactis is an extensively studied LAB used widely in food fermentation, especially in dairy products. It is a nonpathogenic Grampositive bacterium that does not persist in the gastrointestinal tract of mammals.15 Lactobacillus plantarum, another wellstudied LAB species, is found in a variety of fermented foods Inflamm Bowel Dis

& Volume 12, Number 11, November 2006

Copyr ight © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

Inflamm Bowel Dis

& Volume 12, Number 11, November 2006

of plant origin and is common in the human gastrointestinal tract. It differs from L. lactis by its higher and longer survival capacity in the murine and human intestine.16,17 A novel strategy for IBD treatment consists of engineering these bacteria for local delivery of therapeutic compounds. For example, it has been demonstrated that intragastric administration of a recombinant L. lactis strain secreting murine interleukin (IL)-10 prevents or treats inflammation in mouse colitis models.18 More recently, murine trefoil factors also were successfully delivered using a similar strategy.19 The aim of this study was to determine whether a bacterial supply of SOD into the colon might improve an experimental colitis. To this end, we compared the effects of intragastric treatment with L. lactis and L. plantarum strains engineered to produce SOD on various parameters of a colonic inflammation induced by 2,4,6-trinitrobenzenesulphonic acid (TNBS) in the rat. We also analyzed the effects of a continuous infusion of bovine SOD in the colonic lumen.

MATERIALS AND METHODS Bacterial Preparations Two LAB strains were used as Bwild-type^ reference strain. L. lactis NZ9800 is a derivative of the dairy starter MG1363,20 and L. plantarum NCIMB8826 Int-1 is a derivative of the human isolate NCIMB8826.21 Both strains carry a chromosomal copy of the nisRK regulatory genes. L. lactis NZ9800 and L. plantarum NCIMB8826 Int-1 strains were transformed with the pNZ804sodA plasmid. This plasmid contains the L. lactis MG1363 sodA gene encoding an MnSOD, cloned under the control of the nisA promoter, and confers chloramphenicol resistance (5 to 10 mg/mL). Electroporation of L. lactis and L. plantarum was performed according to previously described protocols.22,23 L. lactis was grown at 30-C in M17 medium (Fisher Scientific Labosi, Elancourt, France) supplemented with 0.5% glucose (GM17); L. plantarum was grown at 37-C in MRS medium (Fisher Scientific Labosi). L. lactis and L. plantarum strains harboring pNZ8048sodA were grown on medium supplemented with 10 mg/mL chloramphenicol. Nisin was used to induce SOD production, as previously described for L. lactis20 and L. plantarum.21 Briefly, strains were subcultured and induced with 5 (L. lactis) or 20 (L. plantarum) ng/mL nisin. After further incubation for 3 or 5 hours, respectively, the cultures were harvested by centrifugation at 3000g for 10 minutes and washed twice with phosphate-buffered saline. Strains were resuspended at 109 colony-forming units (CFU)/mL in 0.2 M NaHCO3 buffer containing 2% glucose. Bacterial suspensions were prepared daily, and 1 mL per dose was administered intragastrically to the rats. For measurement of SOD activity in bacteria, 10-mL cultures induced with nisin were harvested, washed with Tris 10 mM, pH 8, and mechanically broken (French-Press,

SOD-Producing Lactic Acid Bacteria and Colonic Inflammation

Bioritech, Chamarande, France). The supernatant containing the soluble cytoplasmic proteins was recovered after centrifugation for 10 minutes at 3000g at 4-C. The SOD activity was determined as previously described24 and expressed as SOD units/109 CFU or SOD units/mg protein. One SOD unit corresponds to the activity of 100 ng bovine CuZnSOD in 1 mL. The SOD activities in bacterial extracts were equal to 0, 26, 74, and 624 units/109 CFU or 0, 54, 272, and 602 units/mg soluble cytoplasmic proteins for L. plantarum, L. lactis, L. lactis SOD+, and L. plantarum SOD+, respectively. SOD activity also was measured in colonic content harvested in Tris 10 mM, pH8, homogenized, sonicated, and centrifuged for 10 minutes at 10,000g at 4-C. SOD activity was determined in the supernatant as described for bacteria and expressed in SOD units/mL colonic content.

Animal Preparation and TNBS Colitis Male Wistar rats (weight, 200 to 250 g) were obtained from Harlan (Gannat, France). They had free access to tap water and standard laboratory food (U.A.R., Epinay/Orge, France). They were housed individually in polypropylene cages in a room with controlled temperature (21 T 1-C) and light-dark cycle (12Y12 hours). The Local Animal Care and Use Committee approved all experimental protocols described in this study. Rats were anesthetized by intraperitoneal injection of acepromazine (Calmivet, Vetoquinol, Lure, France) and ketamine (Imalgene 1000, Rhoˆne-Me´rieux, Lyon, France) at doses of 0.6 and 120 mg/kg, respectively. A polyethylene catheter (outer diameter, 0.7 mm; inner diameter, 0.3 mm; length, 90 cm) was implanted in the proximal colon 2 cm from the cecocolonic junction, attached to the abdominal muscle wall, and exteriorized at the back of the neck. Five days after surgery, colitis was induced by administration of TNBS (80 mg/kg in 0.2 mL of 50% ethanol) through the intracolonic catheter. Controls received 0.2 mL saline.

Experimental Designs In a first series of experiments, 4 groups of 8 rats were used. They received saline (groups 1 and 2) or TNBS (groups 3 and 4) through the catheter inserted into the proximal colon. Four hours after saline or TNBS administration, rats were continuously (24 h/24 h) infused intracolonically for 4 days with saline (0.25 mL/h; groups 1 and 3) or bovine SOD (40 U per rat per day; 0.25 mL/h; groups 2 and 4). After treatment, animals were killed, and samples of proximal colon were collected to assess inflammation. In a second set of experiments, 6 groups of rats were used. Group 1 received saline intracolonically, and the other groups received TNBS. All animals were fed daily with a gastric feeding tube with 1 mL of 0.2 M NaHCO3 buffer containing 2% glucose and 109 CFU bacteria per rat (groups 3 through 6) or with 1 mL buffer with glucose (groups 1 and 2). These

* 2006 Lippincott Williams & Wilkins

Copyr ight © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

1045

Inflamm Bowel Dis

Han et al

treatments lasted 8 days, starting 4 days before and continuing for 4 days after TNBS or saline administration. Groups 3 and 5 received the reference L. lactis and L. plantarum strains, respectively. Groups 4 and 6 received the L. lactis SOD+ and L. plantarum SOD+ strains, respectively. Animals were killed at the end of treatments, and samples of proximal colon were collected to assess inflammation. In a last series of experiments, 5 groups of 6 rats were used. A control group received 1 mL buffer with glucose daily for 1 week. The other groups received L. lactis, L. lactis SOD+, L. plantarum, or L. plantarum SOD+ at 109 CFU/d. Rats were killed at the end of treatments, and colonic contents harvested for determination of SOD activity.

& Volume 12, Number 11, November 2006

diluted in potassium phosphate buffer containing 0.167 mg O-dianisidine dihydrochloride/mL and 0.0005% (vol/vol) H2O2. MPO from human neutrophils (Sigma, Saint Quentin Fallavier, France; 0.1 U/100 mL) was used as a standard. Changes in absorbance at 450 nm were recorded with a spectrophotometer (mc2UV, Safras, Monaco) every 10 seconds over 2 minutes. One unit of MPO activity was defined as the quantity of MPO degrading 1 mmol H2O2Iminj1ImLj1 at 25-C. Protein concentrations (g/mL) were determined using a modified method of Lowry (Detergent Compatible Assay, BioRad, Ivry/Seine, France), and MPO activity was expressed as MPO units/g protein.

Immunohistochemistry for Nitrotyrosine Body Weight The animals were weighed on the day of intracolonic administration of TNBS or saline and on the day they were killed. The body weight variation was expressed as the percentage of change between the 2 days.

Macroscopic and Microscopic Damage Immediately after death, the colon was removed and rinsed with saline. Intestinal damage was scored according to a modified Wallace et al25 scale, which takes into account the severity and extent of macroscopic lesions, the presence and degree of adhesion, and the presence of liquid feces in the distal colon. Scores were blindly evaluated by 2 independent experimenters. To assess microscopic damage, a segment of proximal colon was resected and fixed in neutral phosphate-buffered 4% formalin. Transverse slices (5 mm thick) were stained with hematoxylin and eosin and were examined by light microscopy in a blinded manner. The histological grading scale described by Fabia et al26 was used to evaluate the severity of inflammation.

Myeloperoxidase Activity Assay The activity of the enzyme myeloperoxidase (MPO), a marker of polymorphonuclear neutrophil primary granules, was determined in proximal colon tissue according to a previously described technique.27 Immediately after death, a colonic segment (1 cm long) was taken at 3 cm from the cecocolonic junction. It was suspended in potassium phosphate buffer (KH2PO4 44 mM, K2HPO4 6 mM, pH 6.0), homogenized on ice with a Polytron (PCU-2, Kinematica GmbH, Lucerne, Switzerland), and submitted to 3 cycles of freezing and thawing. Homogenates were then centrifuged at 9000g for 15 minutes at 4-C. The pellets were resuspended in hexadecyl trimethylammonium bromide buffer [0.5% (wt/vol) in potassium phosphate buffer] to release MPO from polymorphonuclear neutrophil primary granules. These suspensions were sonicated (Bu¨chi, Flawil, Switzerland) on ice and centrifuged at 9000g for 15 minutes at 4-C. Supernatant fractions were

1046

Tyrosine nitration, an index of the nitrosylation of proteins by peroxynitrite and/or oxygen-derived free radicals, was determined by immunohistochemistry. Colonic tissue samples were fixed for 12 hours in 4% formalin, dehydrated through graded ethanol, and embedded in paraffin. Sections (5 mm) were rehydrated and submerged in antigen retrieval solution (citrate buffer, 10 mM, pH 6, 95-C, 3 minutes). After inhibition of endogenous peroxidases with 0.6% H2O2 in 100% methanol for 30 minutes and incubation in blocking solution (phosphate-buffered saline containing 1% bovine serum albumin and 2% normal serum), sections were incubated with rabbit nitrotyrosine antibody (5 mg/mL, 1.5 hours at room temperature) and then with a goat anti-rabbit IgG immune serum coupled to dextran polymer carrying peroxidase (30 minutes at room temperature). Antigen-antibody complexes were revealed using 3-3 diaminobenzidine. Hemalum was used as a counterstain. Negative controls were sections treated similarly but not incubated with primary antibody. To quantify the degree of nitrotyrosine staining, a grading system was used28 in which 0 indicates no staining and 1 through 3 indicate increasing degrees of staining. These degrees were estimated for the epithelium, lamina propria, and submucosa. The final score corresponded to the mean of the degrees in the 3 layers. In each experimental group, 5 sections were

TABLE 1. SOD Activity in Colonic Content after a 1-week Treatment with the Different Bacteria (109 CFU/d) Treatment Control L. lactis L. lactis SOD+ L. plantarum L. plantarum SOD+

SOD Activity, units/mL 329.5 399.4 450.2 323.0 492.9

(14.6) (29.9) (14.1)* (13.0) (10.7)*

Values are means (SEM). *P G 0.05 versus Control.

* 2006 Lippincott Williams & Wilkins

Copyr ight © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

Inflamm Bowel Dis

& Volume 12, Number 11, November 2006

SOD-Producing Lactic Acid Bacteria and Colonic Inflammation

Fallavier, France). Rabbit nitrotyrosine antibody, goat antirabbit IgG, diaminobenzidine kit, and hemalum were purchased from Upstate (Lake Placid, NY), Dako Cytomation (Glostrup, Denmark), ICN Pharmaceutical (Costa Mesa, Calif), and Labo Moderne (Paris, France), respectively.

Statistics Data are expressed as mean T SEM. Comparisons between different treatments were performed with analysis of variance, followed by Tukey’s multiple-comparison test for unpaired data. A value of P G 0.05 was considered statistically significant.

RESULTS One-week treatment with L. lactis SOD+ or L. plantarum + SOD at a dose of 109 CFU/d significantly increased (P G 0.05) SOD activity in colonic content compared with controls. A tendency to an increase, without reaching statistical significance (P = 0.06), was observed after treatment with L. lactis strain not overproducing SOD (Table 1).

FIGURE 1. Effect of a continuous colonic infusion of bovine SOD (40 U per rat per day) for 4 days after TNBS on (A) body weight gain, (B) colonic macroscopic damage, and (C) MPO activity. Rats received an intracolonic administration of either saline or TNBS and were infused intracolonically with either saline (saline, saline; TNBS, saline) or SOD (saline, SOD; TNBS, SOD) *P G 0.05 versus saline, saline; 0P G 0.05 versus TNBS, saline.

evaluated by 2 independent observers blinded to the experimental protocol.

Chemicals Bovine SOD; 2,4,6-trinitrobenzene sulfonic acid; nisin; chloramphenicol; KH2PO4; K2HPO4; hexadecyl trimethyl ammonium bromide; hydrogen peroxide; and O-dianisidine hydrochloride were purchased from Sigma (Saint Quentin

FIGURE 2. Absence of effect of a continuous colonic infusion of bovine SOD (40 U per rat per day) for 4 days after TNBS on (A) colonic histological damage score and (B) nitrotyrosine immunostaining score. Rats received an intracolonic administration of either saline or TNBS and were intracolonically infused with either saline (saline, saline; TNBS, saline) or SOD (saline, SOD; TNBS, SOD) *P G 0.05 versus saline, saline.

* 2006 Lippincott Williams & Wilkins

Copyr ight © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

1047

Inflamm Bowel Dis

Han et al

Effects of Colonic Infusion of Bovine SOD Four days after TNBS administration, a body weight loss of 1.1 T 2.1% was observed in the treated group, whereas control rats gained 11.1 T 0.7% in weight. Infusion of bovine SOD for 4 days after TNBS administration reversed the weight loss into a gain of 6.8 T 0.8% (Fig. 1A). The colonic macroscopic damages induced by TNBS consisted of mucosal necrotic zones over 1 to 5 cm, a dark red to brown color, and apparent hyperemia and ulcers. The macroscopic damage score estimated 4 days after TNBS administration was significantly reduced (P G 0.05) in SODinfused rats (Fig. 1B). Consistent with this observation, the MPO activity rose to 3000 U/g protein after TNBS administration, whereas it was significantly (P G 0.05) reduced by half after SOD infusion (Fig. 1C).

& Volume 12, Number 11, November 2006

Microscopically, TNBS administration resulted in submucosal and mucosal infiltration with numerous inflammatory cells, extensive ulceration, and dilated blood vessels, corresponding to a damage score of 6.3 T 1.9. Despite a positive trend (score, 3.7 T 0.9), SOD treatment did not reduce significantly this score (P = 0.07) (Fig. 2A). TNBS also induced an intense immunostaining for nitrotyrosine in epithelial cells, lamina propria, and submucosa. Again, SOD infusion led to a reduced immunostaining score (1.4 T 0.3 versus 2.1 T 0.4) (Fig. 2B), but the difference was not statistically significant (P = 0.07).

Effects of Treatment with LAB In the second series of experiments, weight loss, mucosal macroscopic and microscopic lesions, and the

FIGURE 3. Effect of intragastric treatment with LAB (109 CFU per rat per day) for 4 days before and after TNBS on (A) body weight gain, (B) colonic macroscopic damage, and (C) MPO activity. Rats received intracolonic administration of either saline or TNBS and were treated with vehicle of bacteria (saline, vehicle; TNBS, vehicle) or L. lactis (lact.), L. lactis SOD+ (lact. SOD+), L. plantarum (plant.), or L. plantarum SOD+ (plant. SOD+). *P G 0.05 versus saline, vehicle; 0P G 0.05 versus TNBS, vehicle.

1048

* 2006 Lippincott Williams & Wilkins

Copyr ight © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

Inflamm Bowel Dis

& Volume 12, Number 11, November 2006

SOD-Producing Lactic Acid Bacteria and Colonic Inflammation

FIGURE 4. Effect of intragastric treatment with LAB (109 CFU per rat per day) for 4 days before and after TNBS on (A) colonic histological damage score and (B) nitrotyrosine immunostaining score. Rats received intracolonic administration of either saline or TNBS and were treated with vehicle of bacteria (saline, vehicle; TNBS, vehicle) or L. lactis (lact.), L. lactis SOD+ (lact. SOD+), L. plantarum (plant.), or L. plantarum SOD+ (plant. SOD+). *P G 0.05 versus saline, vehicle; 0P G 0.05 versus TNBS, vehicle.

FIGURE 5. Representative hematoxylin-and-eosinYstained sections from proximal colon (magnification 200). A, Normal colon from a control rat. B, TNBS colitis with thickening of the colonic wall, large edema (asterisk), and immune cell infiltration. C, D, Sections of TNBS-treated rats after bovine SOD infusion (C) or L. lactis SOD+ treatment (D). Note the reduction in mucosa thickness and the absence of edema. * 2006 Lippincott Williams & Wilkins

Copyr ight © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

1049

Inflamm Bowel Dis

Han et al

increase in MPO activity in TNBS-infused rats were more pronounced than in the first series. Treatment with 109 CFU/d L. lactis, L. lactis SOD+, and L. plantarum SOD+ for 4 days before and after TNBS administration significantly ameliorated these 4 parameters (Figs. 3 and 4). In contrast, the L. plantarum strain exerted no significant beneficial effect (P > 0.05), even though a positive trend was observed. Notably, the 3 SOD-producing bacteria were equally efficient in reducing the colitis. For example, the weight loss observed 4 days after TNBS treatment was equal to 5.8 T 0.6%, whereas it was limited to 1.1 T 1.2%, 1.0 T 0.9%, and 1.8 T 0.5% after treatment with L. lactis, L. lactis SOD+, and L. plantarum SOD+, respectively (Fig. 3A). Macroscopically, the mucosa remained reddish and edematous after treatment with the 3 beneficial bacteria, but the number of ulcerations was reduced and the macroscopic damage score was significantly lower (P G 0.05) than in vehicle-treated rats (Fig. 3B). The lowest MPO activity was observed after treatment with L. lactis SOD+ (906 T 359 units/g protein), but there was no significant difference (P > 0.05) in MPO activities after treatment with L. lactis or L. plantarum SOD+ (1865 T 563 and 1780 T 901 units/g protein, respectively) (Fig. 3C). Similarly, the lowest microscopic damage score was observed after L. lactis SOD+ treatment, which again did not differ significantly from values obtained with the L. lactis or L. plantarum SOD+ strains (Figs. 4A and 5). In contrast to the parameters described above, the increase in nitrotyrosine immunostaining score induced by TNBS was strongly reduced (P G 0.05) after treatment with each of the 4 bacterial strains (Fig. 4B). Notably, it was completely abolished after treatment with L. lactis SOD+ and L. plantarum SOD+, whereas it was only attenuated after treatment with the L. lactis strain not overproducing SOD and L. plantarum strain.

DISCUSSION Accumulating evidence, including the observed increase in lipid peroxides in rectal biopsy specimens from patients with ulcerative colitis,29,30 suggests that ROM and nitrogen metabolites play a key role in the pathogenesis of IBD.4,5 SODs are key antioxidant enzymes in the cellular protection against ROM, and complex alterations of SOD activity have been described in IBD. In Crohn`s disease as in ulcerative colitis, colonic levels of Mn-SOD have been found to be increased, whereas those of CuZn-SOD decreased,31 but a proportion of Mn-SOD was present in an enzymatically inactive form.31 These changes in SOD activity in IBD cannot be dissociated from alterations of other enzymes such as MPO or catalase.32 Globally, the colonic mucosa of IBD patients is characterized by an imbalanced and inefficient endogenous antioxidative response.

1050

& Volume 12, Number 11, November 2006

Experiments conducted in rodents point to beneficial effects of SOD treatment against an experimental colitis. SOD or SOD derivatives, administered subcutaneously or intraperitoneally, have been found to reduce the severity of colonic inflammation induced by TNBS or acetic acid.33,34 However, the doses of SOD used in these studies were extremely high compared with the dose of bovine SOD infused into the colon in our study. Transgenic mice overproducing CuZnSOD have been shown to develop less severe lesions and a decreased MPO activity after induction of a mild, but not a severe, colitis by dextran sodium sulfate35 compared with wild-type mice. On the contrary, in another study, the same transgenic mice were reported to develop more severe colitis in response to dextran sodium sulfate than their wild-type littermates.36 In humans, encouraging results were obtained several years ago with SOD treatment in Crohn`s disease, but these data remained preliminary and have not been confirmed.37,38 Nevertheless, some mechanisms involved in the anti-inflammatory action of SOD have been elucidated. Neutrophils are capable of producing large amounts of superoxide anion,39 and SOD contributes to the regulation of neutrophil infiltration by reducing the expression of intercellular adhesion molecule-1 and P-selectin in endothelial and epithelial cells.40,41 Moreover, SOD has been found to inhibit the production of the proinflammatory cytokines tumor necrosis factor-a and IL-1b in TNBS-induced colitis.41 We did not observe any relationship between the amount of SOD produced by the bacteria and the antiinflammatory effect. The reason may be that SOD transforms superoxide anion into hydrogen peroxide, which may contribute to further free radical reaction. To expect any doseeffect relationship, it also would be necessary to add catalase to remove hydrogen peroxide. The effects of different LAB, including those used in this study, have been studied previously. Although we found an antiinflammatory effect of L. lactis in our rat TNBS model, Steidler et al18 found no effect on dextran sodium sulfateY induced inflammation in mice or spontaneous colitis developed in IL-10Ydeficient mice. The difference between these observations is likely to be attributed to the use of different animal models and experimental settings. The absence of effect of the L. plantarum NCIMB8826 strain confirms the results of a previous study using the rat TNBS model and a similar dose of another L. plantarum strain (L. plantarum 299V).42 On the contrary, a protective effect of L. plantarum 299V has been described in the spontaneous colitis developed by IL10Ydeficient mice,43 highlighting the importance of the colitis model used to study the effects of a treatment. Because of the very short half-life of SOD, several chemical strategies have been proposed to improve its bioavailability.12 In this work, we pursued an approach similar to that of Steidler et al,18 who engineered an LAB to deliver an anti-inflammatory cytokine or trefoil factors19 in * 2006 Lippincott Williams & Wilkins

Copyr ight © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

Inflamm Bowel Dis

& Volume 12, Number 11, November 2006

the gastrointestinal tract. Our results indicate that oral administration of LAB expressing SOD improves a TNBS experimental colitis in rats. Among the 4 strains studied, the 3 able to produce SOD spontaneously (L. lactis) or after engineering (L. lactis SOD+ and L. plantarum SOD+) were found to be protective. Of note, the only bacterial strain that does not synthesize SOD (i.e., L. plantarum) was ineffective in reducing most of colonic alterations. Intracolonic infusion of bovine SOD also reduced the inflammation. However, differences between the effects of this treatment and the oral administration of the 3 beneficial strains were observed. For example, bovine SOD infusion did not ameliorate significantly histological and nitrotyrosine scores, whereas the 3 Bactive^ strains did. Several factors could explain these differences. The treatment with bacteria was preventive and curative, whereas bovine SOD was used only in a curative way because of the technical difficulty of maintaining a continuous intracolonic infusion for several days. The dose of infused bovine SOD is hardly comparable to that delivered by the bacteria because different half-life times can be expected between free bovine SOD and bacterially protected SOD. In addition, it is difficult to correlate the SOD bacterial activities determined in optimal in vitro conditions with the amount of enzymatically active SOD effectively released by bacteria in the colonic lumen. Moreover, the bacterial and bovine SODs belong to different enzyme classes, requiring different cofactors. Our data indicate that SOD activity was increased in the colonic content after a 1-week treatment with L. lactis SOD+ or L. plantarum SOD+, whereas only a tendency to an increase, which did not reach statistical difference, was observed after treatment with L. lactis not overproducing SOD. However, we found no significant increase in SOD activity of the colonic content in the absence of sonication (data not shown), suggesting that SOD is not secreted by the bacteria but becomes available in the colonic content after lysis of the bacteria. Such a hypothesis is reinforced by data obtained with LAB engineered to express tetanus toxin. In L. plantarum or L. lactis engineered to express tetanus toxin intracellularly, mutants more susceptible to lysis were found to be more immunogenic.44 In another study, it has been shown that immunization was greater with L. plantarum engineered to produce the toxin in the cytoplasm than with L. plantarum engineered to produce lower amounts of toxin secreted in the culture medium or anchored to the cell wall.45 We cannot conclude that all observed anti-inflammatory effects are mediated through SOD directly. Indeed, it might be speculated that SOD overexpression may improve the protection of the bacteria against ROM and consequently promote its growth. For example, it has been shown in mice46 that treatment with LAB increases the number of IgM-positive cells in the lamina propria of the small intestine, which may help protect the mucosa. Peroxynitrite-mediated cell injury,

SOD-Producing Lactic Acid Bacteria and Colonic Inflammation

detected by immunostaining of nitrotyrosine, was found to be increased in colonic epithelium in IBD patients47 and in rodent TNBS-induced colitis.28 In our study, nitrotyrosine immunostaining did not correlate with other parameters of inflammation. It was not significantly affected by infusion of bovine SOD, which, however, reduced macroscopic damage and MPO activity. In contrast, it was significantly reduced after treatment with L. plantarum, which did not ameliorate the other inflammatory parameters. This bacterial species does not produce SOD but contains micromolar concentrations of manganese in a dialyzable form that is able to scavenge superoxide anion as effectively as micromolar levels of SOD present in other bacteria.48 A recent article described the heterologous expression of a Streptococcus thermophilus sodA in 4 Lactobacillus strains with upper levels of SOD activity reaching 61 U/mg protein. The authors studied the impact of increased SOD levels by analyzing the in vitro ability of the recombinant strains to resist oxidative stress.49 The 4 recombinant strains were indeed better protected against an H2O2challenge. In our work, by overexpressing the L. lactis sodA gene, SOD activities up to 272 and 602 U/mg protein in L. lactis and L. plantarum were obtained, respectively. To the best of our knowledge, we have demonstrated for the first time the protective effect of recombinant strains producing SOD on intestinal inflammation in a rat colitis model. Our results support a beneficial effect of SOD and validate the approach taken. In the scope of IBD treatment, bacterial delivery of SOD offers clear-cut advantages over the use of free SOD, if only in terms of administration protocols. Future developments may include the engineering of naturally anti-inflammatory probiotic strains to produce enzymes involved in oxidative stress defense mechanisms. REFERENCES 1. Scholmerich J. New developments in aetiological mechanisms of inflammatory bowel disease. Eur J Gastroenterol Hepatol. 2003;15: 585Y586. 2. Hibi T. Molecular biological studies of the pathogenesis in inflammatory bowel disease. Intern Med. 2003;42:285Y287. 3. O’Neil D, Steidler L. Cytokines, chemokines and growth factors in the pathogenesis and treatment of inflammatory bowel disease. Adv Exp Med Biol. 2003;520:252Y285. 4. Keshavarzian A, Banan A, Farhadi A, et al. Increases in free radicals and cytoskeletal protein oxidation and nitration in the colon of patients with inflammatory bowel disease. Gut. 2003;52:720Y728. 5. Kruidenier L, Kuiper I, Lamers CB, et al. Intestinal oxidative damage in inflammatory bowel disease: semi-quantification, localization, and association with mucosal antioxidants. J Pathol. 2003;201:28Y36. 6. Grisham MB, Gaginella TS, von Ritter C, et al. Effects of neutrophilderived oxidants on intestinal permeability, electrolyte transport, and epithelial cell viability. Inflammation. 1990;14:531Y542. 7. Nose K. Role of reactive oxygen species in the regulation of physiological functions. Biol Pharm Bull. 2000;23:897Y903. 8. Jourd’heuil D, Morise Z, Conner EM, et al. Oxidants, transcription factors, and intestinal inflammation. J Clin Gastroenterol. 1997;25:S61YS72. 9. Zelko IN, Mariani TJ, Folz RJ. Superoxide dismutase multigene family: a comparison of the CuZn-SOD (SOD1), Mn-SOD (SOD2), and

* 2006 Lippincott Williams & Wilkins

Copyr ight © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

1051

Inflamm Bowel Dis

Han et al

10. 11. 12. 13. 14. 15. 16. 17.

18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30.

EC-SOD (SOD3) gene structures, evolution, and expression. Free Radic Biol Med. 2002;33:337Y349. Salvemini D, Riley DP, Cuzzocrea S. SOD mimetics are coming of age. Nat Rev Drug Discov. 2002;1:367Y374. Turrens JF, Crapo J, Freeman BA. Protection against oxygen toxicity by intravenous injection of liposome-entrapped catalase and superoxide dismutase. J Clin Invest. 1984;73:87Y95. Nakaoka R, Tabata Y, Yamaoka T, et al. Prolongation of the serum halflife period of superoxide dismutase by poly(ethylene glycol) modification. J Control Release. 1997;46:253Y261. Shanahan F. Probiotics in inflammatory bowel disease: therapeutic rationale and role. Adv Drug Deliv Rev. 2004;56:809Y818. Mack DR, Lebel S. Role of probiotics in the modulation of intestinal infections and inflammation. Curr Opin Gastroenterol. 2004;20:22Y26. Kimoto H, Nomura M, Kobayashi M, et al. Survival of lactococci during passage through mouse digestive tract. Can J Microbiol. 2003;49: 707Y711. Pavan S, Desreumaux P, Mercenier A. Use of mouse models to evaluate the persistence, safety, and immune modulation capacities of lactic acid bacteria. Clin Diagn Lab Immunol. 2003;10:696Y701. Vesa T, Pochart P, Marteau P. Pharmacokinetics of Lactobacillus plantarum NCIMB 8826, Lactobacillus fermentum KLD, and Lactococcus lactis MG 1363 in the human gastrointestinal tract. Aliment Pharmacol Ther. 2000;14:823Y826. Steidler L, Hans W, Schotte L, et al. Treatment of murine colitis by Lactococcus lactis secreting interleukin-10. Science. 2000;289:1352Y1355. Vandenbroucke K, Hans W, Van Huysse J, et al. Active delivery of trefoil factors by genetically modified Lactococcus lactis prevents and heals acute colitis in mice. Gastroenterology. 2004;127:502Y513. de Ruyter PG, Kuipers OP, de Vos WM. Controlled gene expression systems for Lactococcus lactis with the food-grade inducer nisin. Appl Environ Microbiol. 1996;62:3662Y3667. Pavan S, Hols P, Delcour J, et al. Adaptation of the nisin-controlled expression system in Lactobacillus plantarum: a tool to study in vivo biological effects. Appl Environ Microbiol. 2000;66:4427Y4432. Josson K, Scheirlinck T, Michiels F, et al. Characterization of a grampositive broad-host-range plasmid isolated from Lactobacillus hilgardii. Plasmid. 1989;21:9Y20. Wells JM, Wilson PW, Le Page RW. Improved cloning vectors and transformation procedure for Lactococcus lactis. J Appl Bacteriol. 1993; 74:629Y636. Marklund S, Marklund G. Involvement of the superoxide anion radical in the autoxidation of pyrogallol and a convenient assay for superoxide dismutase. Eur J Biochem. 1974;47:469Y474. Wallace JL, Keenan CM, Gale D, et al. Exacerbation of experimental colitis by nonsteroidal anti-inflammatory drugs is not related to elevated leukotriene B4 synthesis. Gastroenterology. 1992;102:18Y27. Fabia R, Ar`Rajab A, Johansson ML, et al. The effect of exogenous administration of Lactobacillus reuteri R2LC and oat fiber on acetic acid-induced colitis in the rat. Scand J Gastroenterol. 1993;28:155Y162. Bradley PP, Priebat DA, Christensen RD, et al. Measurement of cutaneous inflammation: estimation of neutrophil content with an enzyme marker. J Invest Dermatol. 1982;78:206Y209. Zingarelli B, Szabo C, Salzman A. Reduced oxidative and nitrosative damage in murine experimental colitis in the absence of inducible nitric oxide synthase. Gut. 1999;45:199Y209. Simmonds NJ, Allen RE, Stevens TR, et al. Chemiluminescence assay of mucosal reactive oxygen metabolites in inflammatory bowel disease. Gastroenterology. 1992;103:186Y196. Bhaskar L, Ramakrishna BS, Balasubramanian KA. Colonic mucosal antioxidant enzymes and lipid peroxide levels in normal subjects and

1052

31. 32. 33.

34. 35. 36.

37. 38. 39. 40. 41. 42. 43. 44. 45.

46. 47. 48. 49.

& Volume 12, Number 11, November 2006

patients with ulcerative colitis. J Gastroenterol Hepatol. 1995;10: 140Y143. Kruidenier L, Kuiper I, van Duijn W, et al. Differential mucosal expression of three superoxide dismutase isoforms in inflammatory bowel disease. J Pathol. 2003;201:7Y16. Kruidenier L, Kuiper I, Van Duijn W, et al. Imbalanced secondary mucosal antioxidant response in inflammatory bowel disease. J Pathol. 2003;201:17Y27. Segui J, Gironella M, Sans M, et al. Superoxide dismutase ameliorates TNBS-induced colitis by reducing oxidative stress, adhesion molecule expression, and leukocyte recruitment into the inflamed intestine. J Leukoc Biol. 2004;76:537Y544. Keshavarzian A, Morgan G, Sedghi S, et al. Role of reactive oxygen metabolites in experimental colitis. Gut. 1990;31:786Y790. Kruidenier L, van Meeteren ME, Kuiper I, et al. Attenuated mild colonic inflammation and improved survival from severe DSS-colitis of transgenic Cu/Zn-SOD mice. Free Radic Biol Med. 2003;34:753Y765. Krieglstein CF, Cerwinka WH, Laroux FS, et al. Regulation of murine intestinal inflammation by reactive metabolites of oxygen and nitrogen: divergent roles of superoxide and nitric oxide. J Exp Med. 2001;194: 1207Y1218. Niwa Y, Somiya K, Michelson AM, et al. Effect of liposomal-encapsulated superoxide dismutase on active oxygen-related human disorders: a preliminary study. Free Radic Res Commun. 1985;1:137Y153. Emerit J, Pelletier S, Tosoni-Verlignue D, et al. Phase II trial of copper zinc superoxide dismutase (CuZnSOD) in treatment of Crohn`s disease. Free Radic Biol Med. 1989;7:145Y149. Weiss RH, Fretland DJ, Baron DA, et al. Manganese-based superoxide dismutase mimetics inhibit neutrophil infiltration in vivo. J Biol Chem. 1996;271:26149Y26156. Yang H, Shi M, Richardson A, et al. Attenuation of leukocyteendothelium interaction by antioxidant enzymes. Free Radic Biol Med. 2003;35:266Y276. Cuzzocrea S, Mazzon E, Dugo L, et al. Protective effects of M40403, a superoxide dismutase mimetic, in a rodent model of colitis. Eur J Pharmacol. 2001;432:79Y89. Kennedy RJ, Hoper M, Deodhar K, et al. Probiotic therapy fails to improve gut permeability in a hapten model of colitis. Scand J Gastroenterol. 2000;35:1266Y1271. Schultz M, Veltkamp C, Dieleman LA, et al. Lactobacillus plantarum 299V in the treatment and prevention of spontaneous colitis in interleukin-10-deficient mice. Inflamm Bowel Dis. 2002;8:71Y80. Grangette C, Muller-Alouf H, Hols P, et al. Enhanced mucosal delivery of antigen with cell wall mutants of lactic acid bacteria. Infect Immun. 2004;72:2731Y2737. Reveneau N, Geoffroy MC, Locht C, et al. Comparison of the immune responses induced by local immunizations with recombinant Lactobacillus plantarum producing tetanus toxin fragment C in different cellular locations. Vaccine. 2002;20:1769Y1777. Perdigon G, Vintini E, Alvarez S, et al. Study of the possible mechanisms involved in the mucosal immune system activation by lactic acid bacteria. J Dairy Sci. 1999;82:1108Y1114. Singer II, Kawka DW, Scott S, et al. Expression of inducible nitric oxide synthase and nitrotyrosine in colonic epithelium in inflammatory bowel disease. Gastroenterology. 1996;111:871Y885. Archibald FS, Fridovich I. Manganese, superoxide dismutase, and oxygen tolerance in some lactic acid bacteria. J Bacteriol. 1981;146:928Y936. Bruno-Barcena JM, Andrus JM, Libby SL, et al. Expression of a heterologous manganese superoxide dismutase gene in intestinal lactobacilli provides protection against hydrogen peroxide toxicity. Appl Environ Microbiol. 2004;70:4702Y4710.

* 2006 Lippincott Williams & Wilkins

Copyr ight © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.