EXPERIMENTAL COLITIS

Male or female mice weighing 20–30 g were used in all experiments. Colitis was induced by intrarectal administration of 0.1 ml trinitrobenzene sulphonic acid (TNBS; 60 mg/ml in 30% ethanol), through a trochar needle approximately 3–4 cm proximal to the anus. This model has been described previously in detail.16 ,17Healthy (no treatment), wild type or iNOS deficient mice were also studied as controls. Animals were killed one, three, seven, or 14 days after the induction of colitis.

ANALYSIS OF iNOS mRNA EXPRESSION BY RT-PCR

Colon tissue samples from mice one, three, seven, and 14 days post-TNBS were rinsed in saline, weighed, and placed in guanidinium isothiocyanate to extract total RNA.18 The final RNA concentrations were determined by absorbance using a GeneQuant spectrophotometer (Pharmacia, Piscataway, New Jersey, USA). The reverse transcriptase (RT) and polymerase chain reaction (PCR) steps were performed as described by Wong et al.19 Briefly, complementary DNA (cDNA) was generated using a RT reaction by incubating 2 μg of total RNA, 1× PCR buffer (10 mM Tris-HCl, pH 9.0, 50 mM KCl, 1.5 mM MgCl), 1 nM each of deoxynucleotide triphosphates (dATP, dGTP, dCTP, and dTTP), 30 units of placental ribonuclease inhibitor (RNA guard, Pharmacia, Piscataway, New Jersey, USA), 200 units of Superscript reverse transcriptase (GIBCO-BRL, Burlington, Ontario, Canada), and 2000 pmol of random hexamer oligodeoxynucleotides (Pharmacia, Piscataway, New Jersey, USA). Reaction mixtures were preincubated for 10 minutes at 21°C prior to DNA synthesis. The RT reactions were carried out for 50 minutes at 42°C and were heated to 95°C for five minutes to terminate the reaction. Reactions were performed in an Amplitron I Thermal Cycler (Barnstead/Thermolyne, Dubuque, Iowa, USA).

Multiple PCR steps were performed using the primer dropping method as described by Wong et al 19 in 50 μl reaction volumes containing 2 μl of RT reaction product as template DNA, 1× PCR buffer, 80 μM of each dinucleotide, and 20 pmol of each specific 5′ and 3′ starter primers. The following primer sequences (5′–3′) were used for iNOS: the sense TCACTGGGACAGCACAGAAT and antisense TGTGTCTGCAGATGTGCTGA with a final PCR product size of 499 base pairs (bp). Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) cDNA was coamplified as an internal control using the following primer sequences (5′–3′): sense CGGAGTCAACGGATTTGGTCGTAT and antisense AGCCTTCTCCATGGTGGTGAAGAC14 with a final PCR product size of 302 bp. Two units of Taq DNA polymerase (Pharmacia) were added to each tube during the first denaturation step and equal aliquots of secondary primer sets (20 pmol GAPDH) were added to the appropriate cycle number by the dropping method. Each PCR cycle consisted of a heat denaturation step at 94°C for one minute, a primer annealing step at 55°C for 30 seconds, and a polymerisation step at 72°C for one minute in an Amplitron I Thermal Cycler (Barnstead/Thermolyne). PCR cycle numbers were chosen to ensure that the amplification of PCR products was in the exponential range: 26 cycles for iNOS and 22 for GAPDH. Aliquots of PCR products (approximately 10 μl) equalised to given equivalent signals from the GAPDH mRNA were electrophoresed through 2% agarose gels (Ultrapure, Pharmacia) containing 0.5 μg/ml ethidium bromide. Gels were visualised under ultraviolet light and photographed with Kodak film. The optical density of each band for iNOS and GAPDH was quantified using NIH image 1.6 software and the results expressed as a ratio of iNOS to GAPDH.

ASSESSMENT OF SEVERITY OF COLITIS

Mice were killed by cervical dislocation, the colon excised, and the severity of colonic damage assessed using criteria previously established for TNBS induced colitis (table 1).20 This scoring system includes features of clinical colitis, macroscopically visible damage, the presence and severity of adhesions, the presence of diarrhoea (defined as loose, watery stool), presence of stricture formation, and maximal bowel wall thickness in millimetres. After gross macroscopic scoring, samples of colonic tissue were fixed in neutral buffered formalin and processed for subsequent histological examination. In addition samples of colon were taken for estimation of myeloperoxidase activity and immunohistochemical studies as described below.

HISTOLOGICAL SCORING

After overnight fixation in formalin, tissues were dehydrated (graded alcohols) and cleared (xylene) before being embedded in paraffin wax. Sections of tissue were cut and stained with haematoxylin and eosin and the sections scored blindly. Histological scoring was based on a semiquantitative scoring system where features were graded as follows: extent of destruction of normal mucosal architecture (0, normal; 1, 2, and 3, mild, moderate, and extensive damage respectively), presence and degree of cellular infiltration (0, normal; 1, 2, and 3, mild, moderate, and transmural infiltration respectively), extent of muscle thickening (0, normal; 1, 2, and 3, mild, moderate, and extensive thickening respectively), presence or absence of crypt abscesses (0, absent; 1, present) and the presence or absence of goblet cell depletion (0, absent; 1, present). The scores for each feature were summed with a maximum score possible of 11.

DETERMINATION OF TISSUE MYELOPEROXIDASE ACTIVITY

Samples of distal colon were weighed, frozen on dry ice, and processed for determination of myeloperoxidase (MPO) activity. Myeloperoxidase is an enzyme found in cells of myeloid origin, and has been used extensively as a biochemical marker of granulocyte (mainly neutrophil) infiltration into gastrointestinal tissues.20 ,21 The samples were stored at −20°C for no more than one week before the MPO assay was performed. MPO activity was determined using an assay described previously,21 but with the volumes of each reagent modified for use in 96 well microtitre plates. The rate of change in absorbance at 450 nm over a 90 second period was determined using a kinetic microplate reader (Molecular Devices, Canada). One unit of MPO activity was defined as that degrading one micromole of hydrogen peroxide per minute at 25°C. Values are expressed as units of MPO activity per milligram of tissue sampled (U/mg tissue).

COLONIC EPITHELIAL PERMEABILITY

To provide further a quantitative measure of epithelial dysfunction, in separate experiments, colonic epithelial permeability was determined in inflamed wild type or iNOS deficient mice at 24 hours, 72 hours, and seven days after TNBS treatment. Healthy mice from both groups were studied as controls. Mice were anaesthetised by intraperitoneal injection with a mixture of ketamine hydrochloride 200 mg/kg (Rogar/STB Inc., Montreal, Quebec, Canada) and xylazine 10 mg/kg (MTC Pharmaceuticals, Cambridge, Ontario, Canada). The left jugular vein was cannulated to administer anaesthetic and a laparotomy was performed. Both renal pedicles were ligated. The entire large intestine was isolated by fitting an inflow tube at the ascending colon-caecal junction and an outflow tube from the anus. The colonic loop was perfused with warm Tyrodes solution.

Chromium-51 labelled EDTA was injected intravenously to obtain counts per minute of at least 25 000/ml (100–150 μCi/kg).⁵¹Cr-EDTA is a 300 molecular weight marker which is restricted by the mucosal barrier but not the vasculature.22 Therefore, any movement of⁵¹Cr-EDTA is a direct reflection of changes across the mucosal not vascular barrier. ⁵¹Cr-EDTA activity in plasma and 2 ml aliquots of perfusate was measured in an autogamma counter (WIZARD 3", model 1480; Wallac Oy, Turku, Finland). At the end of the experiment the colonic loop was removed, rinsed and weighed. The plasma to lumen clearance of ⁵¹Cr-EDTA was calculated as follows: where clearance of ⁵¹Cr-EDTA is given in ml/min/100 g, cpm_p is counts/min/ml perfusate, pr is the perfusion rate, cpm_pl is counts/min/ml plasma, and wt is weight of the intestinal segment in g. The animals were allowed a 30 minute control period to equilibrate after ⁵¹Cr-EDTA administration. Samples of perfusate were collected every 10 minutes for one hour and the average ⁵¹Cr-EDTA clearance was obtained.

IMMUNOHISTOCHEMISTRY

Colonic samples of tissue seven days post-TNBS were used for immunohistochemistry; they were fixed overnight in Zamboni's fixative (pH 7.4) at 4°C, rinsed (3 × 10 minutes) in phosphate buffered saline (PBS), transferred to PBS containing 20% sucrose (pH 7.4), and stored at 4°C overnight. They were embedded in OCT embedding medium (Sakura Finetek Inc., Torrance, California, USA), cryostat sectioned at 12 μm, thaw mounted onto poly-d-lysine coated slides, and dried.

To assess the expression of iNOS protein in colonic tissue, frozen sections were rehydrated in PBS containing 0.1% Triton X-100 (PBS-T), and incubated with 2% normal goat serum in PBS for 30 minutes at room temperature, to block non-specific binding, before being placed in primary antibody. Sections were then incubated for 48 hours at 4°C with a rabbit polyclonal antibody raised against iNOS (1/500; N32030, Transduction Laboratories, Lexington, Kentucky). The specificity of iNOS antibody was verified by substitution of the same concentration of normal rabbit immunoglobulin G (IgG) as was used for the anti-iNOS IgG.

To assess the possible cellular source of iNOS, sections were incubated with primary antibody for iNOS (as above) in combination with a monoclonal antibody raised in rat against mouse neutrophils (1/500; clone 7/4, AMU0021, Camarillo, California) or a monoclonal antibody raised in rat which recognises tissue macrophages (1/50 clone M1/70; MCA 74, Serotec Ltd, Kidlington, Oxford, UK, which recognises Mac-1 differentiation antigen “CD11b”). The monoclonal rat antimouse CD11b antibody primarily recognises macrophages.23 The monoclonal rat antimouse neutrophil (polymorphic) antibody labels specifically all neutrophils and does not stain inflammatory macrophages, lymphocytes, eosinophils, mast cells, and mature erythroid cells in haemopoietic or lymphoid tissues.24 All antibodies were diluted in PBS-T containing 0.1% bovine serum albumin. Immunohistochemical controls routinely performed involved incubation with blocking solution and diluent in place of the primary antibody.

Sections were rinsed (3 × 10 minutes) with PBS-T and incubated for one hour at room temperature with sheep antirabbit IgG conjugated to CY3 (1/100) alone or combined with donkey antirat IgG conjugated to fluorescein isothiocyanate (FITC, 1/50). After a final wash (3 × 10 minutes) with PBS-T, sections were mounted in bicarbonate buffered glycerol (pH 8.6) and examined using a Zeiss Axioplan fluorescence microscope. Sections were photographed using Kodak TMax film.

STATISTICAL ANALYSIS

Data are expressed as mean (SEM). Groups of data were compared using the non-parametric Mann-Whitney U test or Kruskal-Wallis one way analysis of variance (ANOVA) followed by a Dunns multiple comparison test. Probabilities of 5% or less were considered statistically significant.

iNOS mRNA EXPRESSION

Levels of iNOS mRNA were measured by RT-PCR in wild type non-inflamed and inflamed mice at 24 and 72 hours, and seven and 14 days after TNBS treatment. Non-inflamed and inflamed iNOS deficient mice were also studied at seven days. Figure 1A shows a representative band for each time point in wild type mice. Interestingly, some message for iNOS was observed in untreated wild type mice. The RT-PCR product bands were quantified using non-linear densitometry, and the ratio of the iNOS to internal standard (GAPDH) was determined. The mean optical density ratio of wild type mice that did not receive TNBS (n=4) was taken as 1 and each band expressed as a fold increase of this mean. Figure 1B shows the mean fold increase for four mice. At 24 hours after TNBS administration no significant increase in iNOS message was observed relative to control. However, by 72 hours post-TNBS administration there was a fivefold increase in the mean iNOS mRNA levels in wild type mice. This level was maintained at seven days (p<0.05), highlighting the dramatic increase in iNOS at this time point. Messenger RNA levels had returned to control values by 14 days. No detectable levels of RT-PCR products were observed in TNBS treated (seven days) and untreated iNOS deficient mice (fig1A).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

(A) A representative 2% agarose gel of RT-PCR products. Base pair markers (M) denoting DNA size are shown in the far left lane. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH; internal marker) and inducible nitric oxide synthase (iNOS) RT-PCR products are shown for non-inflamed (without trinitrobenzene sulphonic acid (−TNBS)) or inflamed (+TNBS) wild type (iNOS^+/+) and iNOS deficient (iNOS^−/−) tissue. (B) Mean data (n=4) after semiquantitative analysis of the iNOS band intensity expressed as a ratio of the GAPDH band for the various groups. The intensity of iNOS band in non-inflamed mice was equated to one and the other bands expressed as fold increase over this level. *Significantly greater than wild type non-inflamed group.

ACUTE COLITIS: FIRST 72 HOURS

Figure 2 illustrates macroscopic damage scores and MPO activity in wild type and iNOS deficient mice, 24 and 72 hours after TNBS administration. A significant increase in macroscopic damage score and granulocyte infiltration was observed at both time points in wild type mice. In iNOS deficient animals, macroscopic damage score (fig 2A) was significantly increased compared with their wild type counterparts at 24 hours, although granulocyte infiltration (fig 2B) was not significantly different. These data correlate well with data previously reported by our laboratory in an acute intestinal inflammatory response (acetic acid induced colitis in mice).25 At this time point colonic damage has been shown to be independent of granulocyte infiltration in acetic acid induced26 and TNBS induced colitis.27 Substantial granulocyte infiltration was confirmed histologically with both wild type and iNOS deficient mice. Significant destruction of mucosal architecture (including epithelial cell layer), goblet cell depletion, and mucosal and submucosal oedema were also noted. Histological scoring of this tissue in wild type and iNOS deficient groups yielded mean scores of 7.5 (0.2) and 8 (0) respectively (see table2).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

(A) Macroscopic damage score, (B) myeloperoxidase (MPO) activity, and (C) blood to lumen leakage of ⁵¹Cr-EDTA leakage as an indication of mucosal permeability in colonic tissue. n⩾6 in all groups. *Significant increase from healthy mice; δ significant decrease from 24 hour time point; τ significant increase from wild type mice. iNOS, inducible nitric oxide synthase.

Macroscopic damage remains significantly elevated in iNOS deficient mice 72 hours after the induction of colitis compared with the wild type inflamed group. At this time, granulocyte infiltration remained elevated in iNOS deficient mice; however, in wild type mice there was a significant drop in the level of neutrophil infiltration into colonic tissue (although this was not significantly different from iNOS deficient mice at this time point). Transmural cellular infiltration was observed histologically with crypt abscess formation and mild to moderate loss of mucosal architecture. Haemorrhage was also noted in the mucosa in areas of frank ulceration. Mean histological scores from tissue sections from wild type and iNOS deficient mice were 4.8 (1.0) and 5.1 (0.7) respectively (table 2). These histological scores show that sections of tissue from both wild type and iNOS deficient have a similar degree of inflammation. It is extremely important to note that histological score does not reflect the extent of damage along the length of colon which was much larger as assessed macroscopically (fig2A). For example, the mean length of tissue involved in iNOS deficient mice was 2–2.5 cm whereas only 0.5 cm was observed in wild type tissue.

Next, we used the technique of blood to lumen clearance of⁵¹Cr-EDTA to measure mucosal permeability. A very large elevation in ⁵¹Cr-EDTA leakage was observed in both wild type and iNOS deficient mice 24 hours post-TNBS administration (fig2C). This level of mucosal dysfunction was maintained at 72 hours. No significant differences in mucosal permeability were observed in non-inflamed or inflamed wild type and iNOS deficient mice at either time point. This is not surprising as frank haemorrhage suggests no restriction to blood components across the mucosa at these time points.

CHRONIC COLITIS

TNBS induced inflammation develops into a chronic phase by seven days onwards in rats20 ,28 and mice.16 ,17 In this model mast cell hyperplasia, macrophages, and lymphocytes were noted up to two weeks post-TNBS.16 In this study we observed a 5.5-fold increase in macrophage numbers over non-inflamed control tissue seven days post-TNBS administration (see figs 4B and6B). Figure 3 illustrates macroscopic damage scores (fig 3A) and myeloperoxidase activity (fig 3B) in wild type and iNOS deficient mice seven days after TNBS treatment. In wild type mice a significant, 12-fold increase in macroscopic damage score was observed over that in healthy controls. This damage score represents the presence of at least two or more areas of ulceration, diarrhoea, and occasional adhesions and was accompanied by a significant elevation in granulocyte infiltration (fig 3B).

• Download figure
• Open in new tab
• Download powerpoint

Figure 4

Fluorescence micrographs of inducible nitric oxide (iNOS) immunoreactivity (A and C) in healthy wild type animals. The same sections were double labelled to examine the cellular source of iNOS (B and D). Little or no iNOS was present in normal animals in macrophages (B). However, occasional small clusters of iNOS immunoreactive neutrophils (D) were observed in otherwise healthy tissue. Scale bar: 100 μm.

• Download figure
• Open in new tab
• Download powerpoint

Figure 6

Fluorescence micrographs of inducible nitric oxide (iNOS) immunoreactivity (A, C, and E) in wild type animals seven days after treatment with trinitrobenzene sulphonic acid. The same sections were double labelled to examine the cellular source of iNOS (B, D, and F). Though there was an overlapping distribution, there were relatively few macrophages that expressed iNOS immunoreactivity (B). In contrast, iNOS immunoreactivity was extensively colocalised in neutrophils (D and F) in inflamed tissues. Scale bar: 100 μm (A–D), 50 μm (E and F).

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

(A) Macroscopic damage score, (B) myeloperoxidase activity (MPO), and (C) blood to lumen leakage of ⁵¹Cr-EDTA leakage as an indication of mucosal permeability in colonic tissue from healthy or inflamed wild type or inducible nitric oxide synthase (iNOS) deficient mice. n⩾6 in all groups. *Significant increase from healthy mice. There were no significant differences between iNOS deficient and wild type animals after seven days of inflammation.

Histologically, a significant amount of damage was observed in both iNOS deficient and wild type groups (table 2). Tissue sections displayed varying degrees of damage ranging from complete destruction of the mucosal surface with muscle thickening, transmural cellular infiltration, and the presence of granulomas, to areas of distinct ulceration, with cellular infiltration and goblet cell depletion limited to the surrounding tissue. The extent of damage as measured by macroscopic damage scores was similar in both groups. No significant differences were observed in damage (macroscopic or microscopic) or granulocyte infiltration at this time point in iNOS deficient mice compared with inflamed wild type mice despite high levels of iNOS mRNA. As no frank haemorrhage was noted macroscopically some barrier function was regained. Indeed, the mean value for ⁵¹Cr-EDTA leakage into the lumen of the bowel was 65% lower (p<0.05) at seven days (fig3C) after TNBS treatment than at 72 hours (fig 2C). Nevertheless, a sevenfold increase in mucosal permeability in inflamed wild type mice was noted compared with healthy controls (fig 3C). This highly sensitive quantitative assay of mucosal permeability revealed no difference between iNOS deficient and wild type mice; a significant, and equally large, increase in mucosal permeability was observed in iNOS deficient mice.

Therefore seven days post-TNBS, iNOS deficient mice developed a similar level of colonic damage and dysfunction (as measured by macroscopic damage, microscopic damage, cellular infiltration, and epithelial permeability) despite the inability to produce NO from iNOS.

TNBS induced inflammation was also studied in wild type and iNOS deficient mice 14 days after TNBS treatment. As previously indicated there was no increase in message for iNOS at this time point post-TNBS. Only very mild ulceration, no stricture formation or diarrhoea (3.2 (1.3)) and little neutrophil infiltration (7.6 (4.4) U/mg tissue) was observed at 14 days in wild type mice. No significant differences were observed in these parameters between wild type and iNOS deficient mice at this time point (damage score 1.7 (1.7), p=0.28; MPO activity, 2.2 (0.9) U/mg tissue, p=0.25), suggesting that the early enhanced injury in iNOS deficient animals, did not impact on the development of TNBS induced inflammation between seven and 14 days.

IMMUNOHISTOCHEMISTRY

Figure 4A and C illustrates inducible NOS immunoreactivity in healthy, wild type mouse colon. It is clear that in healthy (macroscopic damage scores, MPO activity, and histology were all normal) mouse colon there were one or two very small areas of iNOS protein expression. Weakly immunoreactive epithelial cells and occasional patches of immunoreactive cells in the mucosal lamina propria were consistently observed in normal tissue. These data correlate well with a small amount of message for iNOS in healthy wild type mice. Double labelling for tissue macrophages (fig 4B) or neutrophils (fig 4D) revealed that the few iNOS positive cells in the lamina propria were neutrophils. Resident populations of macrophages in the healthy colon did not express iNOS immunoreactivity under normal conditions. The small amount of iNOS in the colon is not surprising as the colon is always exposed to an abundance of bacteria.

Figure 5A and B illustrates iNOS immunoreactivity in normal (fig 5A) and inflamed (fig 5B) colon seven days after TNBS treatment. These data illustrate an impressive increase in iNOS immunoreactivity in mouse colon seven days after TNBS treatment. Moreover, iNOS deficient mice provided a useful control to support the specificity of the staining in the wild type animals. There was no iNOS immunoreactivity in iNOS deficient mice under basal or inflamed conditions (fig 5C and D respectively). All other controls were negative, also supporting the specificity of the antibodies used in this study. The increase in iNOS immunoreactivity in the wild type animal correlates well with the increased levels of iNOS mRNA which were measured using RT-PCR (fig1).

• Download figure
• Open in new tab
• Download powerpoint

Figure 5

Fluorescence micrographs of inducible nitric oxide (iNOS) immunoreactivity in representative sections from a healthy wild type mouse (A), healthy iNOS deficient mouse (C), inflamed wild type mouse (B), and inflamed iNOS deficient mouse (D). There was a substantial increase in iNOS immunoreactivity in inflamed mouse colon seven days after treatment with trinitrobenzene sulphonic acid in wild type mice (B). No iNOS immunoreactivity was observed in the iNOS deficient mice (C and D). Scale bar: 100 μm.

The cellular source of iNOS was studied by double labelling immunohistochemistry. Figure 6 illustrates the colocalisation of iNOS immunoreactivity with inflammatory cell infiltrate in wild type animals seven days after TNBS treatment. Though there was considerable overlap in the distribution of iNOS (fig 6A) and macrophages (fig 6B), macrophages were not the major source of iNOS immunoreactivity in inflamed tissues. In fact, neutrophils which were the most abundant infiltrating cells (fig 6D) were consistently found to express iNOS immunoreactivity (fig 6C). At higher magnification iNOS immunoreactivity (fig 6E) can clearly be seen within neutrophil labelled cell bodies (fig 6F).