Ileal inducible nitric oxide synthase mRNA expression in response to stress is modified in Sprague-Dawley rats exposed to a previous intestinal inflammation

Abstract

The ability of stress to initiate or reactivate an inflammatory process seems to depend on an individual's susceptibility to stressful stimuli. The aim of this study was to establish whether previous inflammation alters the response to stress in Sprague-Dawley rats, a strain not especially susceptible to stressful stimuli. Stress exposure was performed in rats treated with indomethacin, to induce cyclic intestinal inflammation, during the inactive phase of inflammation. Both control and indomethacin-treated rats submitted to stress showed a decrease in body weight gain and blood leukocyte levels, as well as an increase in fecal pellet output. The increase in intestinal mucosal mast cell count induced by stress was similar in both groups of animals. Moreover, no differences were observed between control and indomethacin-treated rats in the degree of bacterial translocation and myeloperoxidase levels after stress exposure. Despite these similarities, differences between groups were observed in inducible nitric oxide synthase (iNOS) mRNA expression. Although ileal iNOS mRNA expression was inhibited in healthy rats submitted to stress, stress failed to modify this parameter in indomethacin-treated rats. As iNOS is another inflammatory marker, our results may allow the possibility that a previous intestinal inflammation could change the intestinal susceptibility to stress. Whether these differences in ileal iNOS expression can be indicative of a possible change in the predisposition to develop an intestinal inflammatory reaction in response to stress in Sprague-Dawley rats remains to be elucidated.

Keywords::

• Enteric bacteria
• bacterial translocation
• iNOS
• intestinal inflammation
• mast cells
• psychological stress

Introduction

The gastrointestinal tract is a well-known target for physiological changes that occur during stressful life events (Bhatia and Tandon 2005). The main effects related to stress are motility disturbances (Morrow and Garrick 1997; Taché et al. 2001; Cao et al. 2005) and epithelial barrier dysfunction (Santos et al. 2000; Soderholm and Perdue 2001).

Several animal studies have addressed assessing the role of stress as a trigger of intestinal inflammation. The results obtained clearly show that, per se, both acute (Saunders et al. 1994; Santos et al. 1999) and chronic (Santos et al. 2000) stress increase intestinal permeability, an effect that can allow the penetration of luminal bacteria into the epithelium (Kiliaan et al. 1998; Kuge et al. 2006). This bacterial translocation can cause an overstimulation of the immune system, and hence initiate the inflammatory process (Soderholm et al. 2002). However, most of these studies have been performed in rat strains highly responsive to stress that showed an increased susceptibility to develop intestinal inflammation under stress conditions (Soderholm and Perdue 2001; Mawdsley and Rampton 2005). For instance, Wistar-Kyoto rats show an enhanced hypothalamo-pituitary–adrenal (HPA) reactivity and a blunted noradrenergic stress response (Pardon et al. 2002), and these alterations can contribute to the intestinal inflammation observed in these animals after stress exposure. Recently, we have reported that stress induces mucosal mast cell hyperplasia and bacterial translocation in Sprague-Dawley rats (a strain not especially susceptible to stressful stimuli), similar to that observed in Wistar-Kyoto rats. However, unlike changes described in the Wistar-Kyoto strain, no clear signs of intestinal inflammation were observed in these rats (Jorge et al. 2010). Thus, the ability of stress to trigger an inflammatory response seems to depend upon the stress susceptibility showed by an individual at the time. Therefore, factors that can modify stress susceptibility could facilitate the development of intestinal inflammation.

It is well known that immune stimuli can activate the HPA axis, causing similar effects to those observed after stressful conditions (Shanks et al. 1998; Dunn et al. 2005). However, prolonged exposure to inflammatory cytokines can alter both basal and stress-induced HPA activity. Indeed, a HPA axis dysfunction with blunted glucocorticoid responses has been found in several inflammatory chronic diseases such as rheumatoid arthritis (Chikanza et al. 1992), systemic lupus erythematosus (Gutierrez et al. 1998), and inflammatory bowel disease (IBD) (Mawdsley and Rampton 2005). In addition, previous enteric infection may lead to the condition of the so-called post-infectious irritable bowel syndrome (PI-IBS) (Spiller and Garsed 2009). In this chronic gastrointestinal disease, as well as in IBD (where bacteria seem to play a fundamental role), stress seems to be a contributing factor modulating the time course of the illness (Monnikes et al. 2001; Bitton et al. 2003; Mardini et al. 2004).

These observations consequently prompt the question whether previous inflammation can alter the intestinal response to stress, and hence modify the predisposition to develop an inflammatory process. Thus, the main goal of this study was to elucidate whether, in the absence of a genetic predisposition for stress vulnerability, a previously inflamed intestine would show an altered response to stress that could favor the appearance of an inflammatory reaction. For this purpose, we used Sprague-Dawley rats, as a strain not especially susceptible to stress, treated with indomethacin to induce intestinal inflammation. As previously reported (Porras et al. 2004), two subcutaneous injections of 7.5 mg/kg indomethacin 48 h apart induced a characteristic cyclic oscillation of active and quiescent phases of inflammation. Active phases were characterized by intestinal hypomotility and increased paracellular and transcellular permeability associated with bacterial translocation and an increase in inflammatory markers. By contrast, intestinal hypermotility and only enhanced paracellular permeability (without alterations in the transcellular pathway) were observed during the quiescent phases, correlating with a decrease in the incidence of bacterial translocation and a return to basal levels of inflammatory markers (Porras et al. 2004, 2006). In the present study, stress was applied during the quiescent phase, when no clear signs of intestinal inflammation were present, in order to assess whether stressful stimuli were able to trigger an inflammatory response. Our hypothesis was that a previous intestinal inflammation can facilitate bacterial translocation under stress conditions and, thus, trigger an inflammatory reaction. To test this hypothesis, we compared the stress responses between a healthy intestine and a previously inflamed one by assessing the following parameters: bacterial translocation, mucosal mast cell count, inducible nitric oxide synthase (iNOS) mRNA expression and myeloperoxidase (MPO) concentration in ileum samples.

Materials and methods

Animals

Male Sprague-Dawley rats (Charles River, Lyon, France) of 8–10 weeks old and weighing 300–350 g were used. They were kept singly housed under conventional conditions in an environmentally controlled room (20–21°C, 60% humidity, 12h:12-h light–dark cycle, lights on at 07:00 h) with tap water and standard laboratory rat chow ad libitum. Rats were handled daily by the same investigator who performed the experiments for 1 week before the study, in order to minimize the stress of contact with unfamiliar humans. All experimental protocols were approved by the Ethics Committee of the Universitat Autònoma de Barcelona.

Experimental model of enteritis

Intestinal inflammation was induced by administration of two subcutaneous injections of indomethacin (7.5 mg/kg) 48 h apart, as previously described (Porras et al. 2004). As mentioned in the introduction section, this model shows a spontaneous cyclic oscillation between active and quiescent phases of inflammation, which can be identified by determining the time course of blood leukocyte (BL) levels (150 μl blood samples were taken every 2 days) throughout the experiment. In this study, BL counts reached high values during the active phase, mean ± SEM 22,954 ± 925.1 cells/mm³; the BL counts during the quiescent phase were similar to those observed in control rats (14,900 ± 462.5 cells/mm³ in indomethacin-treated rats; 12,957 ± 194.2 cells/mm³ in control rats). Figure 1 (panel A) shows this typical oscillatory pattern of BL counts in an individual rat maintained for up to 60 days after indomethacin administration. In this study, exposure to the chronic stress protocol was performed during the quiescent phase following the second peak of high BL counts (two active phases were recorded before starting stress exposure). The quiescent phase was identified when two consecutive BL counts were lower than the maximal leukocyte value reached during the second active phase (Figure 1, panel B).

Figure 1.  Representative graphs showing the cyclic oscillation of BL counts associated with active and inactive phases of intestinal inflammation induced by indomethacin administration. (A) Representative graph of BL count changes evolution in an indomethacin-treated rat and in a control rat. The cyclic oscillation of BL count was maintained for up to 60 days after drug administration. (B) Representative graph showing the timing of the start of the stress protocol. Stress exposure was performed during the inactive phase of inflammation following the second peak of high BL counts. The inactive phase was identified when two consecutive BL counts were less than the maximal BL count reached during the active phase. The gray zone of the graph corresponds to the 5 days of stress exposure using the protocol illustrated by the * stress scheme. WR, wrap restraint; WAS, water avoidance stress, each 1 h on alternate days.

Chronic stress protocol

Chronic stress was induced by submitting rats to a heterotypic chronic intermittent stress paradigm involving the physical–psychological stress of wrap restraint (WR) and the psychological stress of water avoidance stress (WAS). During the WR procedure, the limbs and the body of rats were wrapped in a cloth harness to restrict, but not to prevent, body movements. The water avoidance procedure (WAS) consisted of placing the rat on a plastic platform (7 × 7 × 10 cm) located in the middle of a plastic container (55 cm diameter) filled with warm water (25°C) to 1 cm below the platform. The chronic stress protocol was adapted from the previously described method (Million et al. 1999), and consisted of submitting rats to 1 h WR or 1 h WAS alternately for 5 consecutive days (WR–WAS–WR–WAS–WR). Control rats were kept in their home cages during the procedure. All the experiments were performed between 09:00 and 12:00 h to minimize the effect of circadian rhythm.

Experimental design

Rats were randomized into four groups:

(1)	Rats belonging to control-nonstressed group (n = 16) that received vehicle saline injections subcutaneously were kept in their own home cage during the stress procedure;
(2)	Rats belonging to control-chronic stress group (n = 16) that received saline subcutaneously were submitted to the chronic stress protocol;
(3)	Rats belonging to indomethacin-nonstressed group (n = 13) that received indomethacin subcutaneously to induce cyclical intestinal inflammation were kept in their own home cage during the stress procedure;
(4)	Rats belonging to indomethacin-chronic stress group (n = 15) that received indomethacin subcutaneously to induce intestinal inflammation were submitted to the repeated stress protocol. Exposure to the chronic stress paradigm was performed during the quiescent period following the second active phase of the BL oscillatory pattern. The stress protocol was started when two consecutive BL counts were below the maximal leukocyte value that reached during the active phase.

Figure 2 shows the study time course. Stress exposure was started in indomethacin-treated rats immediately after the quiescent phase had been identified (indomethacin-chronic stress rat). Simultaneously, the stress protocol was applied to a control rat (control-stress rat). A similar procedure was performed between indomethacin-nonstressed and control-nonstressed rats. In all cases, rats were submitted to the stress/control protocol 23 ± 6 days after indomethacin or saline administration.

Figure 2.  Overview of the experimental design used in the study. Each indomethacin-treated rat was matched to a control rat. The stress/sham sessions started when the inactive phase following the second active phase of inflammation was identified in an indomethacin-treated rat. Simultaneously, the stress/sham sessions were started in their matched control individual. After 5 days of stress/sham exposure, both indomethacin and control rats were euthanized.

Rats were euthanized 1 h after the last stress/control session by isofluorane inhalatory anesthesia and exsanguinated by heart puncture to obtain intestinal tissue samples.

Animal monitoring

Several parameters were evaluated through the entire experimental period in all groups of rats. These parameters were to assess the effectiveness of both enteritis induction by indomethacin and the chronic stress protocol and but also to evaluate the putative differences in the responses to stress between groups.

Growth

Rats were weighed daily. Body weight (BW) change was expressed as percent weight gain in relation to the first day of the stress protocol.

Food intake

Food consumption was recorded daily through the stress protocol period and results were expressed as mean grams per day.

Defecation

Defecation, as an indirect index of colonic propulsive activity (Maillot et al. 2000) was determined at each session during the stress protocol. Pellets expelled during 1 h stress/sham session were counted and expressed as the number of fecal pellets per 1 h.

Blood leukocytes

BL levels (BL) were monitored every 2 days through the study. Blood samples were taken by the tail-nick procedure. This procedure consisted of gently wrapping the rats with a cloth, making a 2 mm incision at the end of the tail veins and then massaging the tail while collecting 150 μl of blood into EDTA capillary tubes (Starsted, Granollers, Spain). Once the stress/control protocol was started, blood samples were obtained 30 min after the stress/control session. BL were counted using a Neubauer chamber and expressed as cells/mm³. BL change was expressed as percent (%) of BL count increase compared with the first day of the stress protocol.

Bacterial translocation

Bacterial translocation from the lumen of the intestine was determined by detection of viable enteric bacteria in mesenteric lymph nodes (MLN). MLN from the ileocecal region of each rat were removed aseptically immediately after euthanasia. Specimens were frozen in liquid nitrogen, powdered in a mortar, and weighed. Samples were homogenized in 10 parts of milk (Difco, Detroit, MI, USA) and incubated under aerobic conditions onto blood agar and McConkey agar for 48 h at 37°C. After incubation, colonies were identified by studying their morphological and biochemical properties. The incidence of bacterial translocation was expressed as the number of positive cultures in the total number of samples in each group.

Rat mast cell protease II immunochemistry and mucosal mast cell count

Distal ileum samples were fixed in 10% buffered formalin and embedded in paraffin wax. Transverse sections were cut (3 μm) and immunodetection of rat mast cell protease II (RMCP II) was carried out using a monoclonal antibody (1:500; Moredun Animal Health, Edinburgh, UK). Detection was performed with avidin/peroxidase (Vectastain ABC kit; Vector Laboratories, Burlingame, CA, USA). Sections were counterstained with hematoxylin and counted at 400 × magnification. Positively stained mast cells were counted in three to five sections per rat. Seven to ten well-oriented villus–crypt units (VCU) were examined per section. Analysis of all morphological data was performed blinded to prevent observer bias. Estimation of mast cell numbers was expressed as cells per VCU.

RT-PCR studies

Distal ileum samples of each rat were taken immediately after euthanasia. Tissue segments were frozen in liquid nitrogen and stored at − 80°C until use.

RNA extraction

Isolation of total RNA was performed using TRI Reagent (Ambion, Madison, WI, USA) according to the manufacturer's instructions and treated with DNA-free (Ambion) for 30 min at 37°C to remove any genomic DNA contamination.

Rt-pcr

First-strand cDNA was synthesized from 5 μg total RNA in a reaction mixture of 50 μl containing 0.5 μg of oligo18 (dT) primer (Ambion), 2 mM dNTP (Ecogen, Barcelona, Spain), and 10 units Moloney murine leukemia virus (Ambion). The resultant cDNA was amplified in a total volume of 50 μl with 1 unit of taqDNA (Ecogen), 1 mM dNTP mixture, and 1.2 μM primers (Proligo, Sigma Aldrich, St Louis, MO, USA). The sequences of sense and antisense primers for rat iNOS and glyceraldehyde 3-phosphate dehydrogenase (GAPDH; a housekeeping gene used as an internal control) are listed in Table I. Thermal cycling conditions were as follows: denatu-ration for 4 min at 95°C, then 35 cycles PCR with denaturation at 95°C for 1 min, annealing at 62°C (iNOS) or 50°C (GAPDH) for 1 min, extension at 72°C for 1 min, and a final extension at 72°C for 5 min. Amplified products were electrophoresed on 1.5% agarose gel in TAE buffer, stained with ethidium bromide, photographed under ultraviolet light, and quantified using image-analyzing software (PC-BAS 2.0). For semiquantification, the ratio of the optical den-sity of each PCR product and GAPDH was determined.

Table I.  Primer sequences for RT-PCR.

Gene	Primer sequence	Amplicon length (bp)
iNOS
Sense	5′-ACAACAGGAACCTACCAGCTCA-3′	651
Antisense	5′-GATGTTGTAGCGCTGTGTGTCA-3′
GAPDH
Sense	5′-CCGCCCCTTCCGCTGATGCC-3′	140
Antisense	5′-ATGAGCCCTTCCACGATGCC-3′

GAPDH, glyceraldehyde 3-phosphate dehydrogenase and iNOS, inducible nitric oxide synthase.

Tissue myeloperoxidase determination

Proteins were extracted from the ileal tissue of each rat using a lysis buffer containing protease inhibitors (Minicomplete tablet, Roche Diagnostics, Mannheim, Germany). MPO concentration was determined using a specific enzyme-linked immunosorbent assay (HyCult Biotechnology, Uden, The Netherlands), with a minimal detectable concentration of 1 ng/ml.

Statistical analyzis

Data are expressed as means ± SEM. Statistical analysis was performed using Minitab 15 Statistical Software (Coventry, UK). Differences between groups were compared using a two-way ANOVA (factors: inflammation and stress) and Tukey post hoc analysis. Results for bacterial translocation were analyzed using Chi-square test. In all cases, results were considered to be statistically significant when p < 0.05.

Results

Effects of stress on animal monitoring parameters

As seen in Figure 3 showing the percent of BW gain during the first 10 days after indomethacin administration, rats with induced intestinal inflammation showed a transient loss of BW gain, whereas a linear increase in this parameter was observed in the control group. This loss of BW gain induced by indomethacin together with the higher BL counts recorded in these rats (19,264 ± 791 and 20,975 ± 729.2 cells/mm³ 2 and 4 days after indomethacin administration [all indomethacin-treated rats n = 28] vs. 12, 957 ± 194.2 cells/mm³ in control rats [all control rats n = 32]) are consistent with the development of intestinal inflammation. Exposure to stress during the quiescent phase of inflammation induced a significant decrease in BW gain, which was similar to that observed in stressed control rats (F_1,55 = 13.47, p < 0.001) (Figure 4). This stress-induced effect on BW gain rate was not related to differences in food intake either in healthy (24.71 g/day ± 0.51 [n = 16] vs. 26.07 ± 0.56 in control-nonstressed group [n = 16]) or in indomethacin-treated rats (25.09 g/day ± 0.58 [n = 15] vs. 26.84 ± 0.6 in inflamed-nonstressed group [n = 13]).

Figure 3.  Effects of indomethacin on percent of BW gain relative to the first day of drug administration. Data are mean ± SEM, n = 13–16 rats/group. A decrease in BW gain was observed after indomethacin administration (p < 0.001, ANOVA), whereas control rats showed a linear increase in this parameter.

Figure 4.  Effects of stress on BW gain in control and indomethacin-treated rats. Data are mean ± SEM, n = 13–16 rats/group. (A) Time course of the percent weight gain relative to the first day of the stress/sham protocol in control and indomethacin-treated rats. (B) Percent of BW gain in control and indomethacin groups 5 days after stress exposure. Stress significantly decreased BW gain (two-way ANOVA, p = 0.0005). *p < 0.05 vs. control-nonstressed group; +p < 0.05 vs. indomethacin-nonstressed group.

Regarding the defecation frequency, the inflammatory process was associated with an increase in pellet output when compared to the healthy conditions (F_1,55 = 8.81, p = 0.004). Moreover, stress exposure also caused a significant increase in defecation frequency, independent of whether the intestine had been previously inflamed or not (F_1,55 = 118.48, p < 0.0001) (Figure 5).

Figure 5.  Effect of stress exposure on fecal pellet output in control and indomethacin-treated rats. Date are mean ± SEM, n = 13–16 rats/group. Indomethacin-treated rats showed an increase in defecation frequency when compared with the healthy rats (two-way ANOVA, p = 0.004). Stress significantly increase pellet output when compared to the unstressed rats (two-way ANOVA p < 0.0001). ***p < 0.001 vs. control-nonstressed group: +++p < 0.001 vs. indomethacin nonstressed rats.

In accordance with previously reported data (Porras et al. 2004), nonstressed indomethacin-treated rats showed the oscillatory BL count pattern characteristic of this experimental model, associated with the active and quiescent phases of chronic intestinal inflammation. Chronic stress was applied during the remission phase of intestinal inflammation, when BL counts (15,435 ± 502.1 cells/mm³, n = 28) were similar to those observed in control rats (12,727 ± 410 cells/mm³, n = 32). Exposure to stress resulted in a significant decrease in BL count both in inflamed (n = 15) and control groups (n = 16) when compared to their respective counterparts (inflamed-nonstressed group n = 13; control-nonstressed group n = 16) (F_1,55 = 78.39, p < 0.001). This decrease was not related to the inflammatory state in indomethacin-treated rats (F_1,55 = 2.60, p = 0.1125). These results are shown in Figure 6.

Figure 6.  Effect of stress on BL counts in both control and indomethacin-treated rats. (A) Representative graph showing the time course of BL counts in an individual rat of each group. In each graph, the time that stress/sham exposure was initiated is indicated (rectangle in unstressed rats; oval in stressed rats). (B) Change in BL count 5 days after stress exposure expressed as percent of BL count recorded on the first day of the stress/sham protocol. Data are mean ± SEM, n = 13–16 animals/group. Stress significantly decreased BL levels (two-way ANOVA, p < 0.0001). This decrease was not related to the inflammatory state in indomethacin-treated rats (two-way ANOVA, p = 0.1125). ***p < 0.001 vs. control-nonstressed group; +++p < 0.001 vs. indomethacin-nonstressed group.

Effects of stress on bacterial translocation

Bacterial translocation is defined as the migration of bacteria from the intestinal lumen to MLN or other extra-intestinal sites (Berg 1995).

As expected, no viable enteric bacteria were found in MLN of nonstressed control rats (n = 14). By contrast, unstressed rats with intestinal inflammation (n = 13), even though they were in the quiescent phase of the inflammatory process, showed bacterial translocation.

As previously reported (Jorge et al. 2010), a significant bacterial translocation was found in control rats submitted to stress (n = 16). By contrast, exposure to stress did not significantly increase the total bacterial translocation in the inflamed group (n = 15). However, it is important to note the remarkable translocation of Enterobacteriaceae, which was present in 53% of MLN of indomethacin-treated rats exposed to chronic stress. This result was statistically significant when compared with the control group. These results are summarized in Table II.

Table II.  Number of animals with bacterial translocation and incidence for a specific micro-organism.

	Enterobacteriaceae	Enterococcus/Streptococcus spp.	Lactobacillus spp.	Bacterial translocation
Control-nonstressed	0/14	0/14	0/14	0/14
Control-stressed	3/16	3/16	3/16	8/16*
Indomethacin-nonstressed	3/13	3/13	1/13	6/13*
Indomethacin-stressed	8/15*	5/15^†	0/15	11/15*

Data are expressed as number of positive cultures from MLN for the total animals in each group. *,^†, p < 0.05 and 0.01 respectively, compared with control-nonstressed group (Chi-square test).

Effects of stress on mucosal mast cells

The immunochemical detection of RMCP II, a chymase predominantly expressed by rat intestinal mucosal mast cells, is commonly used to determine mucosal mast cell count.

Rats with small intestinal inflammatory disease selected during the remission phase showed a mild, but statistically significant, increase in mucosal mast cell count (F_1,50 5.16, p = 0.0274). Stress exposure for 5 consecutive days increased this parameter in control (n = 12) as well as in inflamed rats (n = 13) when compared with their respective counterparts (control-nonstressed group n = 14; inflamed-nonstressed group n = 15) (F_1,50 = 46.01, p < 0.001). The magnitude of the stress-induced increment in mast cell count was similar in both groups (Figure 7).

Figure 7.  Effect of stress on the number of mucosal mast cells per VCU. (A–D): Microphotographs show RMCPII immunopositive cells (mucosal mast cells: dark staining) in the intestinal ileum mucosa of (A) control-nonstressed, (B) control-stressed, (C) indomethacin-nonstressed and (D) indomethacin-stressed animals. (E) Bar diagram representing the number of mucosal mast cells per VCU in ileum mucosa of each group. Three to five sections were counted per rat, 7–10 well-oriented VCUs were examined per section. Data are mean ± SEM, n = 13–16 rats/group. Indomethacin-treated rats showed an increase in mast cell count per VCU compared to the control group (two-way ANOVA, p = 0.0274). Stress also increased mast cell count per VCU both in control and indomethacin-treated rats (two-way ANOVA, p < 0.001). ***p < 0.001 vs. control-nonstressed group; +++p < 0.001 vs. indomethacin-nonstressed group.

Effects of stress on ileal iNOS mRNA expression

As seen in Figure 8, healthy rats exposed to chronic stress (n = 16) showed a significant decrease in iNOS mRNA expression, when compared to the nonstressed rats (n = 15) (Student t-test, p < 0.05). This decrease in ileal iNOS mRNA expression induced by stress was not observed in rats with intestinal inflammation (inflamed-nonstressed group n = 13; inflamed-stressed group n = 13).

Figure 8.  Effect on iNOS mRNA expression in ileum of both control and indomethacin-treated rats. (A) Representative photograph of agarose gel showing RT-PCR products for inducible isoform of iNOS mRNA in ileum. C − , negative PCR control; CTRL (ns: control-nonstressed rat; s: control-stressed rat), and INDO (ns: indomethacin-nonstressed rat; s: indomethacin-stressed rat). (B) Bar diagram showing semiquantitative analysis by RT-PCR of iNOS mRNA expression. GAPDH, glyceraldehyde 3-phosphate dehydrogenase. Data are mean ± SEM, n = 13–16 animals/group. Stress induced a decrease in iNOS mRNA expression in control rats exposed to stressful stimuli. *p < 0.05 vs. control-nonstressed group, Student t-test.

Effects of stress on tissue myeloperoxidase

MPO, an enzyme found in granulated cells, has been widely used as a reliable index of inflammatory activity (Smith and Castro 1978). As previously reported (Porras et al. 2006), MPO levels in unstressed indomethacin-treated rats (1.16 × 10^{− 3}% of total protein, n = 13) were similar to those observed in unstressed control group (0.97 × 10^{− 3}% of total protein n = 13), confirming that they were in the quiescent phase of the inflammatory process. Exposure to stress did not modify MPO levels either in control (0.92 × 10^{− 3}% of total protein, n = 13) or in inflamed rats (1.35 × 10^{− 3}% of total protein, n = 12) (F_1,47 = 0.11, p = 0.7450).

Discussion

The rationale of this experimental approach was to establish whether previous inflammation alters the response to stress in Sprague-Dawley rats, a strain not especially susceptible to stressful stimuli, and to assess whether these putative changes can trigger an inflammatory reaction.

Besides the role of stress as a trigger of intestinal inflammation, several studies have been performed to evaluate the involvement of stress in the exacerbation of the inflammatory process. It has been described that stress increases the severity of subsequent colitis induced by dextran sulfate sodium (DSS) as well as by 2,4,6-trinitrobenzene sulfonic acid (TNBS) (Israeli et al. 2008; Reber et al. 2008). Moreover, exposure to stress during the acute phase of inflammation also enhances the inflammatory response (Gué et al. 1997; Million et al. 1999). However, few studies have assessed whether stress can induce or reactivate an inflammatory response in an intestine that has apparently recovered from a previous inflammation. To address this question, the experimental model of indomethacin-induced small intestinal inflammation was used. It has been postulated that depletion of endogenous prostaglandins (Fang 1997), increased activity of iNOS (Konaka et al. 1999b), and vascular dysfunction (Piepoli et al. 2005) may be involved in the inflammation induced by indomethacin. Moreover, the initial inflammatory reaction has been associated with bacterial overgrowth and wall invasion by luminal bacteria (Yamada et al. 1993; Porras et al. 2004). In a previous study, we demonstrated that two injections of 7.5 mg/kg indomethacin 48 h apart induce an oscillatory pattern with active and inactive phases of inflammation that can be observed for up 60 days after drug administration (Porras et al. 2004). Although no clear signs of intestinal inflammation were present during the inactive phases in this model, enhanced paracellular permeability and low-grade bacterial translocation were still present (Porras et al. 2006). The recurrent appearance of active phases in this model must be attributed to these remaining alterations rather than to a direct effect of indomethacin, which was administrated several days before. Thus, this experimental model can represent a useful tool to study the pathophysiological mechanisms of gastrointestinal illnesses in which the equilibrium between microbial flora and host response seems to be deranged, as in IBD and PI-IBS. In both cases, persistent enteric infection associated with increased gut permeability have been proposed as factors contributing to the development of the disease (Spiller and Garsed 2009; Kalischuk and Buret 2010).

As previously reported, chronic stress induced a decrease in BW gain (Velin et al. 2004; Yang et al. 2006), increased colonic motor activity (Monnikes et al. 2001), and resulted in a decrease in BL counts in control rats (Dhabhar and Mcewen 1997). Similar effects were observed in indomethacin-treated rats exposed to stress during the quiescent phase of inflammation, indicating that the effectiveness of the stress protocol used was similar in both experimental groups. Notably, as described in IBS and IBD patients during the remission phase of inflammation (Simren et al. 2002; Ansari et al. 2008), unstressed indomethacin-treated rats showed an increased defecation frequency when compared to the unstressed control group, indicating that the motility disturbances induced by the inflammatory process were still present.

In keeping with similar results (Velin et al. 2004; Zareie et al. 2006), stress induced bacterial translocation in control rats. Although the results obtained in inflamed rats exposed to stress did not reach statistical significance when compared to their unstressed counterparts (probably reflecting the low n in the Chi-square test), some observations can be made. Whereas bacterial translocation was observed in 46% of unstressed indomethacin-treated rats, the incidence after stress exposure was around 73% of inflamed rats. Recently, we have reported that the changes induced by stress in the small intestinal motor activity can facilitate bacterial translocation in control Sprague-Dawley rats, even in the absence of clear signs of villus epithelial permeability dysfunction (Jorge et al. 2010). Thus, the effects of stress exposure can be added to the motor and intestinal permeability disturbances caused by the inflammatory process, hence increasing the passage of bacteria to the MLN. Another observation can be made in relation to the specific bacterial species found in MLN. It has been reported that host-mediated inflammatory responses induced by an infecting agent and/or a chemical trigger can disturb the intestinal microbiota balance and allow an overgrowth of Enterobacteriaceae (Lupp et al. 2007). In agreement, an increase in luminal enterobacteria occurred during the active phases of inflammation in our experimental model (Porras et al. 2004). This can be related to our findings showing that enterobacteria were isolated from MLN in 53% of stressed indomethacin-treated rats, whereas the incidence in the stressed control group was 19%. More studies are needed to elucidate whether these observations represent a difference in the stress response between a previously inflamed intestine and a healthy one.

It is well known that mast cells are involved in stress-induced intestinal barrier function disturbances (Soderholm et al. 2002; Yang et al. 2006) that can allow bacterial translocation. Although exposure to stress of control SD rats increased mast cell counts, in a previous study, we demonstrated that this increase was not associated with villus epithelial barrier dysfunction (Jorge et al. 2010). However, other studies indicate that mast cells are also involved in the stress-induced barrier disruption of follicle-associated epithelium (Keita et al. 2010), providing a possible explanation for the bacterial translocation seen in these rats in our study. Also, in our study the bacterial translocation observed in unstressed inflamed rats was accompanied by an increase in ileal mast cells. This rise in mast cell number might be involved in the paracellular permeability increase observed during the quiescent phase of inflammation in this experimental model as reported previously (Porras et al. 2006). Exposure to stress further increased mast cell number in indomethacin-treated rats, an event that could have led to a worsening of the permeability disturbances and hence to have facilitated the passage of bacteria to the MLN.

The differences observed in iNOS mRNA expression after stress exposure support the hypothesis that a previous inflammation alters the intestinal response to stressful stimuli. Under nonstress conditions, control and indomethacin-treated rats showed similar levels of iNOS mRNA expression. However, the decrease in iNOS mRNA expression induced by stress in control rats was not observed in indomethacin-treated rats. It has been reported that bacterial translocation is the first step required for activation of various factors such as neutrophils and iNOS (Konaka et al. 1999a), and that prolonged exposure to high NO levels can cause a breakdown in the gut barrier function (Nadler et al. 1999). However, glucocorticoids exert an inhibitory action on iNOS expression (Korhonen et al. 2002; Hamalainen et al. 2008). It can be hypothesized that the iNOS mRNA decrease observed in control rats exposed to stress could represent a defensive mechanism triggered by the release of corticosteroids in order to avoid NO overproduction and limit the degree of bacterial translocation by maintaining the intestinal barrier integrity. In keeping with the findings that prolonged exposure to inflammatory cytokines can alter the activity of the HPA axis (Shanks et al. 1998), the failure of stress to decrease iNOS expression in indomethacin-treated rats could be due to a change in stress susceptibility resulting from a previous intestinal inflammatory process. Another possibility could be related to the ability of enterobacteria to directly induce iNOS expression (Kolios et al. 2004). In this way, the iNOS expression induced by the persistent bacterial translocation that occurs in this experimental model could counterbalance the inhibitory effects exerted by glucocorticoids, explaining why iNOS mRNA levels remained unchanged in stressed indomethacin-treated rats.

As we have previously reported, stress in control animals is not enough to trigger intestinal inflammation (Jorge et al. 2010). Exposure to stress in indomethacin-treated rats did not modify MPO levels. However, taking into account that iNOS is another inflammatory marker and that stress in indomethacin-treated rats failed to decrease iNOS mRNA expression as it did in control animals, we cannot conclude that stressful stimuli were able to trigger an inflammatory reaction. It is well known that the effects of stress depend on the quality and duration of the stressor. The stress protocol used in the present study can be considered as a mildly stressful when compared with other chronic stress protocols such as chronic subordinate colony housing (Reber et al. 2008) or 10 days of WAS stress exposure (Soderholm et al. 2002). In a previous study performed in a murine model of DSS-induced colitis, a trend towards a reactivation of the inflammatory process was observed after 7 days of WAS exposure, which was enhanced by the administration of a sub-threshold dose of DSS (Melgar et al. 2008). Collins et al. (1996) found some changes associated with inflammation after stress exposure in TNBS-induced colitis in rats, although they failed to observe clinically significant colitis. Thus, differences between studies can be related to both the stress protocol and the experimental model used to induce inflammation.

In summary, our results show differences in ileal iNOS mRNA expression after exposure to stress between a healthy intestine and a previously inflamed one. Although a decrease in this parameter was observed in control animals submitted to stress, stressful stimuli did not change ileal iNOS expression in indomethacin-treated rats. Although no clear signs of inflammatory response were observed in these animals, the failure of stress to inhibit ileal iNOS expression opens the possibility that a previous intestinal inflammation alters the “intestinal susceptibility” to trigger an inflammatory reaction in Sprague-Dawley rats after stress exposure. More studies are needed to assess to what extent these differences in ileal iNOS mRNA expression are indicative of a change in the predisposition to develop an intestinal inflammatory reaction in response to stress.

Acknowledgments

The authors are grateful to A. C. Hudson for the editorial revision of the manuscript, L. Abarca, for the help for the microbiological studies and A. Acosta for the care of the animals. This work was supported by FPU grant AP-2004-5574 from Ministerio de Ciencia e Innovación (financial support to E. Jorge), Grant BFU2007-62794 from Dirección General de Investigación, Ministerio de Ciencia e Innovación. CIBEREHD is funded by the Instituto de Salud Carlos III.

Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.