Prostaglandin E₂ and reactive oxygen metabolite damage in the cecum in a pony model of acute colitis

Abstract

The objective of this project was to determine early tissue biochemical events associated with increased colonic secretion during the acute stage of castor-oil-induced colitis by measuring cecal mucosal and submucosal malondialdehyde (MDA) and prostaglandin E₂ (PGE₂), levels in ponies. Intestinal tissue (inflamed or healthy) samples were obtained from 4 age- and sex-matched Shetland ponies. Biochemical methods were used to determine MDA and PGE₂ levels in intestinal tissue samples from inflamed and healthy equine intestine. Inflamed tissue MDA and PGE₂ levels increased with time after castor oil challenge and correlated with granulocyte infiltration, as determined by myeloperoxidase levels in a companion study. Elevated intestinal tissue MDA levels suggest that lipid peroxidation could be attributed to reactive oxygen metabolites (ROM) released from stimulated, recruited, and resident granulocytes. Tissue levels of MDA and PGE₂ suggest a role for granulocyte-derived mediators of intestinal inflammation in the massive secretory response in cases of acute equine colitis. Tissue MDA and PGE₂ levels may be useful laboratory tools to quantify and characterize intestinal secretory inflammatory responses in acute inflammatory conditions in the equine colon.

Introduction

Inflammatory cells, particularly phagocytic granulocytes, play an important role in mucosal pathophysiology in cases of colitis (1,2,3,4). Large numbers of these cells are observed on histopathologic examination of tissues from human and animal cases of colitis (5,6,7,8,9). Products of cell activation, in particular reactive oxygen metabolites (ROM), stimulate direct and indirect secretory responses in intestinal cells and tissues (10,11,12,13,14). Therefore, these products may amplify the inflammatory signal or have effects on other target cells in intestine, such as enterocytes and smooth muscle cells.

Previous studies have demonstrated that in ponies, moderate to severe histopathologic changes and clinical signs begin by approximately 24 h following nasogastric intubation and castor oil administration (6,15). Determination of myeloperoxidase (MPO) activity and malondialdehyde (MDA) and prostaglandin E₂ (PGE₂) levels have been used extensively in other species to indicate or suggest phagocyte involvement in inflamed tissue secretory responses (phagocyte presence, ROM release, and eicosanoid liberation, respectively) (10,11,13,14). Our objective was to determine early tissue biochemical events associated with increased colonic secretion during the acute stage of colitis in the pony model. Myeloperoxidase activity was measured as a quantitative index of tissue granulocyte (predominantly eosinophils) infiltration, (16,17) and was reported in an earlier companion paper (18). Malondialdehyde, a product arising from lipid peroxidation of tissues, was measured to determine the relative degree of oxidant-mediated tissue damage (19,20). Prostaglandin E₂ levels were analyzed for a purported role in mediating colonic hypersecretion directly or in response to reactive oxygen metabolites (13,21,22,23,24). Therefore, changes in PGE₂ and MDA levels during the first 24 h following chemical induction of typhlitis were examined.

Studies on natural and induced forms of acute typhlocolitis suggest that complex interactions exist among mesenchymal and immune cells, enteric neurons, and epithelial cells. The goal of this study was to determine the relationship between tissue granulocyte infiltration, prostaglandin production, and lipid peroxidative damage during the early stage of castor oil-induced acute equine colitis.

Materials and methods

Horses

This project was approved by the North Carolina State University Animal Care and Use Committee. Tissues analyzed were the same as those used in a parallel experiment and results have been reported previously (granulocyte infiltration based on a histopathology grading system and myeloperoxidase levels) (18). Four ponies, with average body weight of approximately 200 kg were studied. All ponies were housed in individual, 4-m × 4-m, shavings-bedded box stalls, were fed timothy hay and mixed concentrated pelleted grain ration, and had ad libitum access to a salt/mineral block and water. Ambient temperature was maintained at 22 ± 3°C, and the light:dark interval was 12 h. Ponies were dewormed by use of an ivermectin product 3 wk prior to the experiment. All ponies were seronegative by agar gel immunodiffusion assay for equine infectious anemia virus, and were vaccinated at least 2 mo prior to the beginning of the study against the equine forms of encephalitis, influenza, tetanus, rhinopneumonitis, and strangles.

Model

A silastic cannula (2 cm internal diameter) was surgically placed at the midbody of the cecum in each of the 4 ponies via a right flank laparotomy performed under general anesthesia. A minimum of 3 wk was allowed for acclimation before experiments began. Physical examination was performed at 6-hour intervals. Each pony received a single dose of castor oil at 2.5 mL/kg to induce localized typhlitis. Castor oil was administered directly into the cecum through the cecal cannula. Feed and hay were withheld 12 h prior to castor oil administration, after which feed was withheld for an additional 24 h (to enable biopsy specimen collection). Castor oil remained in the cecum for 30 min and was removed by siphoning, using approximately 2 L of 0.9% saline solution for rinsing. Cecal mucosa was diffusely reddened following the 30-minute exposure period to castor oil, resulting in diffuse localized typhlitis. Cecal peristalsis results in constant mixing of cecal contents, resulting in a homogenous mixture with the exogenously administered castor oil. The cecal mucosa was thus diffusely exposed to the inflammation-inciting chemical. Biopsy samples were obtained using a 1-m-long flexible fiberoptic endoscope and a 3-mm biopsy punch. Adjacent samples were taken within 10 to 15 mm of each other. Ponies did not have adverse systemic effects, except for mild transient fever and mild diarrhea, which lasted approximately 12 h.

Tissue sampling

Mucosal biopsy specimens were obtained via the cannulation site, using a 1-m-long flexible fiberoptic endoscope and a 3-mm biopsy punch. Specimens were taken at 0 (control), 2, 4, 6, 8, 12, 16, and 24 h after castor oil administration. Thirteen specimens were obtained from closely adjacent sites at each time, yielding a single specimen for histologic assessment (reported previously) (18) and providing specimens for biochemical assays, which were placed in aluminum foil, flash frozen in liquid nitrogen, and stored at −70°C. Specimens were stored no longer than 2 wk. Three frozen specimens from each respective time were pooled to provide a larger sample size. Pooled specimens were then analyzed for myeloperoxidase (18) (reported previously), malondialdehyde, and PGE₂ levels.

Prostaglandin assay

Methods for tissue extraction and purification have been described in detail (25). Briefly, tissues were thawed and homogenized in 80% cold ethanol. Sample extracts were loaded on to separation columns (Waters C-18 Sep Pak columns; Waters-Millipore, Milford, Massachusetts, USA) and the eluant was used for PGE₂ radioimmunoassay (Advanced Magnetics Inc., Cambridge MA).

Lipid peroxidation

Malondialdehyde (MDA) synthesis was used as a marker for lipid peroxidation of tissues. Samples were homogenized in 1.15% KCl and mixed with sodium dodecyl sulfate (SDS), thiobarbituric acid (TBA), and potassium phosphate buffer then heated for 60 min in a 95°C water bath. Water and n-butanol pyridine were added, the mixture was centrifuged, and the organic layer (containing TBA- reactive MDA) was read spectrophotometrically at a wavelength of 532 nm.

Statistics

Differences in PGE₂ levels, MDA formation, and catalase activity between control (time 0) and inflamed tissue (2, 4, 6, 8, 12, 16, 24 h) were expressed as the mean ± SEM. Statistical significance was calculated using Student's t-test for paired or unpaired data where appropriate. The level of statistical significance was taken as P 0.05.

Results

Prostaglandin E₂ content

Prostaglandin E₂ content increased in pooled cecal mucosal biopsy samples at 2 h after cecal lumenal administration of castor oil. The values were significantly greater from 8 through 24 h after castor oil treatment (Figure 1).

Figure 1. Cecal mucosal biopsy PGE₂ levels over 24 h (n = 4). Significant increases were found at 8 h through 24 h. PGE₂ was measured in nmol/mg protein, and results were expressed as % of control value. ^a P 0.05 for 0 vs 8 h and 16 h; ^b P 0.01 for 0 vs 12 h and 24 h.

Malondialdehyde formation

Malondialdehyde formation was significantly less at 2 h post intraluminal cecal castor oil administration compared to time 0, then subsequently showed a dramatic increase at 6, 12, and 16 h after castor oil treatment (Figure 2). Malondialdehyde levels were 6 times greater at 24 h than at time 0, though not showing a statistically significant difference.

Figure 2. Cecal mucosal biopsy MDA formation over 24 h. Significant decrease occurred at 2 h after castor oil treatment (n = 3). MDA was measured in nmol/mg protein, and results were expressed as % of control value. ^a P 0.05 for 0 vs 2 h, 6 h, and 12 h; ^b P 0.01 for 0 vs 16 h.

Myeloperoxidase levels

Myeloperoxidase from these tissues (reported previously) were low in control cecal samples (0.049 ± 0.012 U/mg of protein). After castor oil administration, MPO activity gradually increased until 12 h (0.138 ± 0.04 U/mg of protein; Figure 3). There was a positive correlation between cecal MPO activity and time (r = 0.88) until the 16-hour sample collection period when values decreased for the final 24-hour sample (not shown).

Figure 3. Cecal mucosal biopsy MPO activity over 24 h. Significant decrease occurred at 6 h and 16 h after castor oil treatment (n = 3). MPO was measured in U/mg protein, and results were expressed as % of control value. ^a P 0.05 for 0 vs 6 h and 16 h.

Figure 4 depicts the correlation between all tissue biochemical parameters over the 24-hour sampling period on the same graph. The first data point is equal to 100% at time 0 (for all biochemical values) and all measurements were subsequently compared to the 100% value. The marked increase in PG activities at 2 h and 4 h were not statistically greater than at time 0 due to a large standard error among ponies, suggesting extreme variability between ponies in prostaglandin response following castor oil-induced epithelial barrier damage. The data points that show significance correspond to tissue MPO activity at 6 and 16 h; MDA formation at 6, 12, and 16 h; and PGE₂ levels at 8, 12, 16, and 24 h after castor oil administration.

Figure 4. Comparison between all biochemical assays expressed as % of control value over 24-hour biopsy collection period. ^a Statistically significant differences.

Discussion

Initial studies of equine colitis revealed that MPO activity is elevated in large intestinal tissue in ponies with naturally occurring and castor oil-induced acute colitis compared with normal control animals (18). This period of increased intestinal MPO activity corresponded to the presentation of adverse clinical signs, including hemoconcentration, profuse watery diarrhea, and other signs compatible with intestinal endotoxemia. This is consistent with the clinical pattern reported by Roberts (15) and coincides with epithelial cell injury and granulocyte infiltration in the pony (6). Studies have shown that castor oil, a C-18 hydroxy fatty-acid, causes active anion secretion with subsequent intraluminal fluid and electrolyte sequestration (26,27). The specific mechanism of castor oil-induced diarrhea is unclear although several mediators have been shown to be involved, including E-type prostaglandins (28), platelet-activating factor (29), histamine and serotonin (28), and nitric oxide (30).

Malondialdehyde was assayed as an index of lipid peroxidation, reflecting free radical reaction in the intestinal mucosa and submucosa. Granulocyte activation with the release of peroxidases (predominantly eosinophil peroxidase) is the most likely explanation linking ROM presence to lipid peroxidation. Increased intestinal biopsy MDA formation observed from 6 through 16 h corresponded to increased biopsy MPO levels from 6 to 16 h, suggesting that lipid peroxidation could be attributed to ROM released from stimulated, recruited, and resident granulocytes.

As expected, PGE₂ levels were elevated from 8 to 24 h in biopsy samples, supporting Racusen and Binder's demonstration of cyclic AMP (cAMP) involvement in the intestinal secretory response following ricinoleic acid treatment (31). The actual mechanism of PGE₂ involvement remains unclear and further study is necessary.

The findings from this study and the companion paper (18) provide evidence that ROM and PGE₂ may play an important role in the secretory response during the early stages of acute equine colitis in this model and that eosinophils (18) may be the major ROM- contributing granulocyte. Tissue inflammatory cells may produce ROMs, which may contribute directly to increased enterocyte chloride secretion or may contribute to amplification of prostaglandin-mediated secretion by enteric neurons or cAMP enterocyte activation, the latter 2 via mesenchymal cell stimulation and subsequent prostaglandin production (14).

The significance of tissue eosinophils as reflected by levels of MPO activity in the companion study remains unclear. Although MPO levels were increased, along with increased MDA formation and PGE₂ levels, the mere presence of eosinophils does not explain their role in this response, as was shown in the previously reported study (18). However, the consistent increase in MDA formation with corresponding histopathologic score (18) suggests an important role for ROM in the pathophysiology of acute colitis. Further studies are necessary to elucidate the pathways involved and other key regulatory intermediates proximal and distal to the production of ROM for their specific role in colitis-associated intestinal secretion.

Control samples taken for a 24-hour duration from fasted ponies given a saline control may have been useful and simple to obtain, but were not done. The time 0 sample from each pony was thought to be a better control in this study for comparison purposes since ponies could serve as their own controls and since fasted ponies do not appear to have a cecal or colonic mucosal inflammatory response during or following a 36-hour fast.