The aim of this study was to determine the impact of the in situ delivery of catalase-producing lactobacilli in the DT on the development of inflammation and oxidative intestinal damage in mice. Our strategy, based on the intragastric administration of both isogenic strains of Lb. casei BL23 producing or not MnKat, allowed us to distinguish the respective effects of the strain and of the catalase activity. Both Lb. casei BL23 treatments led to anti-inflammatory effects for DSS-treated mice, suggesting that the BL23 strain itself is responsible for the protective effect. The mechanisms and bacterial components involved in these anti-inflammatory effects remain unknown.
In the tested conditions, no protection due to the catalase was observed. Two hypotheses could explain this result: i) no induction of effective oxidative stress in our moderate colitis model and/or ii) a too low catalase activity delivered in digestive tractus via Lb. casei BL23 MnKat+. No increase of intestinal oxidative stress in our model was detected using a first method (evaluation of MDA level in digestive tissues homogenates).
Using a colorimetric method, we observed for the first time a significant decrease of the total catalase activity in the intestinal lumen of all DSS-treated mice. The four-fold higher catalase activity of caecal contents compared to colonic contents is coherent with the decrease of aerobiosis from the stomach to the rectum. DSS-induced modifications of the homeostasis of the intestinal microbiota could explain this reduction of catalase activity. Dominant colon species are strict anaerobic bacteria like Bacteroides, Bifidobacterium, Clostridium coccoides and Cl. leptum whereas other groups like Enterobacteriaceae or Lactobacillus are subdominant. Among microbiota, some bacteria possess catalase activity as described for Bacteroides fragilis . A loss of diversity of Bacteroides and Enterobacteriaceae groups was observed in biopsies of patients with Crohn's disease or ulcerative colitis . In our experiments, the reason of the observed decrease is difficult to determine because the composition of the microbiota of the conventional mice used is unknown. Previous studies showed that activities of antioxidative enzymes produced by epithelial cells were modified in biopsies of patients with Crohn's disease or ulcerative colitis [12,15,16]. In this study, we showed for the first time a reduction of antioxidative defences of caecal and colonic contents during the induction of inflammation. All these data suggest a global deficiency of antioxidative enzymes during intestinal inflammation that might be involved in the mechanism of development of the disease.
The second reason of the absence of any incidence of MnKat production could be a too low catalase activity and more generally, a too low antioxidative potential. Our results showed that the administration of our MnKat-producing strain did not restore the initial catalase activity level in the digestive contents. Increasing catalase activity seems to be necessary. To optimise our antioxidative strategy, evaluation of the effects of co-administration of Lb. casei BL23 strains producing high levels of MnKat and SOD from L. lactis  as some previous studies showed the positive impact of increased SOD activity in intestinal inflammation models will be relevant [20,21,37,38].
In our test conditions, no increase of mucosal pro-inflammatory cytokines (TNF-α and IFN-γ) or decrease of IL-4 and IL-10 was observed in DSS-treated mice. The administration of both Lb. casei BL23 strains did not induce modification of mucosal cytokines. In contrast, an increase of TNF-α in DSS-induced colitis was previously observed either at the gene expression level [39,40] or at the systemic protein level [22,41]. In the DSS-colitis model, the modification of mucosal cytokine levels could occur with the apparition of chronic inflammation depending of the mice used: Melgar et al. (2005) observed an increase of IL-1β, IL-12 and IFN-γ levels between 100- and 1000-fold in C57Bl/6 mice (where the colitis progresses in chronicity) in contrast to BALB/c mice where cytokines levels did not increase . These data could explain the absence of the modulation of cytokine levels in our experiments performed with BALB/c mice. For now, the pathways responsible for inflammation development in the DSS-colitis model are not clearly understood. No modulation of mucosal cytokines by Lb. casei BL23 administration in vivo was observed. So far, such modulations were only observed in in vitro co-cultures between LAB and peripheral blood mononuclear cells [27,30]. No published data correlate cytokine modulations by LAB in in vitro and in vivo assays.
In conclusion, our results showed anti-inflammatory capacities of Lb. casei BL23 strain in a DSS-induced colitis model. This original result is in accordance with a recent one observed in TNBS-induced colitis  and previous ones obtained with other Lb. casei strains. Both Shirota and DN-114001 strains impair inflammation development in a DSS-induced [29,30] and TNBS-induced  colitis model, respectively. These two colitis models mimic Crohn's disease (for TNBS) and ulcerative colitis (for DSS) and the efficiency of these Lb. casei strains in both open interesting potential uses of them. Several studies based on host-probiotic interactions aimed to identify the still unkown mechanisms responsible for these anti-inflammatory effects: Borruel et al. first observed that Lb. casei and Lb. bulgaricus are able to interact with immune cells and to modulate pro-inflammatory cytokines using co-cultures of LAB with biopsies of patients with active Crohn's disease, but the mechanism of action is still not known . Two recent studies established a link between probiotic anti-inflammatory capacities and Toll-like receptors (TLR) pathway [27,46]. Although the involvement of TLR in the stimulation of the immune system by the bacterial components is well documented, their implication in inflammatory diseases remains confusing: they were also shown to be essential to maintain a mucosal homeostasis during intestinal inflammation, as observed with MyD88-/- knockout mice which are very sensible to the induction of DSS colitis .
The increasing availability of probiotic genomic sequences constitutes a major progress to understand molecular mechanisms involved in the anti-inflammatory capacities observed. This will be facilitated by the use in parallel of cellular and animals models and could result in a global view of probiotic-host interactions.