Nonsteroidal anti-inflammatory drugs can adversely affect the colon. In people, one third of life-threatening gastrointestinal complications from NSAIDs involve the distal portion of the gastrointestinal tract.1 A wide spectrum of adverse effects have been reported in people, including isolated colonic ulcers, diffuse colonic ulcers with or without bleeding and perforation, diffuse colitis,2-4 and reactivated inflammatory bowel disease.2-5 Adverse effects of NSAIDs on the distal portion of the gastrointestinal tract are believed to result from direct local damaging effects after oral administration; a recurrent local effect resulting from enterohepatic recirculation of drugs; and systemic effects, such as inhibition of prostaglandin synthesis, after absorption.1,6-8
Carprofen is an NSAID approved for use in dogs at a dosage of 2.2 mg/kg twice daily. Ulcers of the stomach or duodenum with or without bleeding and perforation are a common potential adverse effect of NSAIDs in dogs.9 Bleeding in the stomach with emesis and melena are potential adverse gastrointestinal effects described by the distributor of carprofen.10 In contrast to humans, there are few reports of complications from NSAIDs specifically involving the colon in dogs. The distributor of carprofen reported for a safety study10 that there was 1 dog with redness of the colonic mucosa after it had been given 6.6 mg of carprofen/kg twice daily for 6 weeks.
It is not known whether NSAIDs adversely affect the colon of dogs. The objective of the study reported here was to determine whether carprofen compromises the integrity and barrier function of the colonic mucosa of dogs. Our intent was to measure electrical conductance and permeability to mannitol of the colonic mucosa of dogs during incubation with carprofen in Ussing chambers. Histologic appearance of the colonic mucosa after exposure to carprofen was evaluated. We hypothesized that carprofen would increase electrical conductance and permeability to mannitol and cause deleterious effects evident in the histologic appearance of the colonic mucosa of dogs.
Materials and Methods
Sample population—Thirteen mature mixed-breed dogs were the source for tissue sections. The dogs appeared healthy on the basis of physical examinations performed immediately prior to anesthesia and euthanasia. The dogs were anesthetized and euthanized for reasons unrelated to the study reported here. General anesthesia was induced with 5% thiopental (5 mL/kg, IV, to effect) and maintained with 1% isoflurane in oxygen via an endotracheal tube. Dogs were euthanized by administration of an overdose of sodium pentothal. The dogs were used in accordance with the Louisiana State University Institutional Animal Care and Use Committee policy.
Preparation of tissue sections—The entire colon was harvested from each dog exactly at the time of euthanasia. The harvested colon was incised along the mesenteric side, placed in ice-cold Krebs-Ringer bicarbonate buffer solution, and transported to our laboratory. Sections of colonic mucosa were harvested from 6 dogs for the control sections and 7 dogs for the treatment sections. Time from harvest to mounting in Ussing chambers was < 30 minutes.
Colons were placed in stripping pans filled with 400 mL of ice-cold oxygenated (95% oxygen and 5% carbon dioxide) Krebs-Ringer bicarbonate buffer solution. Each colon was divided into 3 regions. The transverse colon extended from the cecocolic junction to the middle colic vein. The second and third regions comprised the descending colon from the middle colic vein to the pelvic inlet, which was divided equally into the proximal and distal portions of the descending colon, respectively. The colonic mucosa was separated from the seromuscular layer by use of blunt and sharp dissection. Three sections of mucosa were obtained from each region of the colon. An additional section of mucosa was obtained from a randomly chosen region of the colon and placed in neutral-buffered 10% formalin for subsequent examination for histologic evidence of preexisting colonic disease; this section of mucosa was presumed to represent the entire colon.
Ussing chamber—All sections from each dog were processed at the same time. Each section of mucosa was randomly assigned to 1 of 9 Ussing chamber units (aperture of 3.14 cm2). The mucosa was clamped flat between the 2 halves of the acrylic chamber.
Each hemichamber was filled with 15 mL of Krebs-Ringer bicarbonate buffer solution (pH, 7.4). The Krebs-Ringer bicarbonate buffer solution was continuously oxygenated (95% oxygen and 5% carbon dioxide) and circulated in water-jacketed reservoirs. Temperature of the solution was maintained at 37°C. Carprofen was added to the bathing solution of treated sections at a concentration of 400 μg/mL 30 minutes after mounting. For control sections, an equivalent amount (120 μL) of Krebs-Ringer bicarbonate buffer solution was added to the bathing solution. The pH of the Krebs-Ringer bicarbonate buffer solution with carprofen was maintained at 7.4.
Electrical measurements—Transepithelial potential difference was measured by use of agar bridges connected to silver–silver chloride voltage electrodes. When the transepithelial potential difference was between −1.0 and 1.0 mV, tissues were current clamped at 100 MA for 5 seconds and the transepithelial potential difference was recorded.
Transepithelial potential difference was short-circuited through the voltage electrodes by use of a voltage clamp that corrected for fluid resistance. Short-circuit current was measured by use of a separate pair of agar bridges connected to silver–silver chloride electrodes (current electrodes).
Transepithelial potential difference and short-circuit current were recorded at 15-minute intervals for 240 minutes. Electrical conductance was calculated at each time point by use of Ohm's law such that electrical conductance is equal to short-circuit current divided by transepithelial potential difference in accordance with the following equation:
Mannitol—Fifteen minutes after tissue sections were mounted, 3H-mannitol (10 MCi/mL) was added to the mucosal bathing solution. The same concentration of nonradiolabelled mannitol was added to the serosal solution. A sample (0.1 mL) was collected from the mucosal solution 30 minutes after addition of 3H-mannitol. Samples (0.5 mL) were collected from the serosal solution 60, 120, 180, and 240 minutes after addition of 3 H-mannitol. Samples were assessed for B emission, and mucosal-to-serosal flux of mannitol was calculated for each of three 1-hour periods (60 to 120 minutes, 120 to 180 minutes, and 180 to 240 minutes).
Histologic examination—At the end of the experiments, sections of mucosa were removed from the Ussing chambers and placed in neutral-buffered 10% formalin for subsequent histologic examination. Fixed sections of mucosa were trimmed, embedded in paraffin, and sectioned at a thickness of 5 μm. Tissue sections were mounted on slides and stained with H&E.
Tissue sections were evaluated by a single investigator by use of light microscopy. The investigator was not aware of the source or treatment for each tissue section.
Inflammation, edema, sloughing of cells from surface epithelium, erosions, sloughing of epithelial cells within mucosal glands, and any additional findings were recorded. Inflammation was categorized on the basis of the percentage of the surface area of lamina propria infiltrated by inflammatory cells. Normal background inflammation was defined as < 20% of the surface area of the lamina propria infiltrated by inflammatory cells. Mild inflammation was defined as 20% to < 40% of the surface area infiltrated by inflammatory cells, moderate inflammation was defined as 40% to ≤ 60% of the surface area infiltrated by inflammatory cells, and severe inflammation was defined as > 60% of the surface area of the lamina propria infiltrated by inflammatory cells. Edema was categorized as not evident when mucosal glands were adjacent to each other (ie, not separated from each other by clear fluid), mild when mucosal glands were < 50 μm apart, moderate when mucosal glands were 50 to 150 μm apart, and severe when mucosal glands were > 150 μm apart. Sloughing of cells from surface epithelium was defined as detachment of surface epithelial cells without discontinuity of surface epithelium. Sloughing of cells from surface epithelium was categorized as minimal when < 10% of the surface area was affected, mild when 10% to < 20% of the surface area was affected, moderate when 20% to ≤ 50% of the surface area was affected, and severe when > 50% of the surface area was affected. Erosions were defined as discontinuity of surface epithelium. Erosions were categorized as not evident when < 10% of the epithelial surface was involved or evident when ≥ 10% of the surface epithelium was involved. Sloughing of cells within mucosal glands was categorized as not evident when no glands had detached epithelial cells in their lumen, mild when < 5% of the mucosal glands had detached cells, moderate when 5% to ≤ 10% of the mucosal glands had detached cells, and severe when > 10% of the mucosal glands had detached cells in their lumen.
Statistical analysis—Statistical analysis was conducted for data on electrical conductance, mannitol flux, and histologic examination. Commercially available softwarea was used for the analysis. Results were considered significant at values of P ≤ 0.05, unless stated otherwise.
ELECTRICAL CONDUCTANCE
Data from 0 to 15 minutes (equilibration period) were not used for analysis. Electrical conductance from 30 to 240 minutes was graphed against time for each Ussing chamber, and the AUC was calculated by use of the trapezoid method.11 Sections from the same region of the colon within a dog were considered replicates. The AUC was the response variable used for the statistical analysis. Mean ± SEM AUC (ie, electrical conductance × time) for each region (transverse colon, proximal portion of the descending colon, and distal portion of the descending colon) was calculated. Data for electrical conductance × time were normally distributed, as verified by failure to reject the null hypothesis of normality at P ≤ 0.05 (Shapiro-Wilk statistic). Data from control sections were examined for a fixed effect of region by use of a mixed-effect general linear model that included the random variance of dog among regions. When there was a significant (P ≤ 0.05) interaction, ad hoc comparisons were made with the Scheffe adjustment to maintain A = 0.05. On the basis of results for the control sections, data for treated sections were compared with data for control sections to detect a fixed effect of region and treatment by use of a mixedeffect linear model that included the random variance of dog nested within treatments and dog among regions. When there was a significant interaction, ad hoc comparisons were made by use of a Scheffe adjustment to maintain A = 0.05.
MANNITOL
Mucosal-to-serosal flux of mannitol was calculated for three 1-hour periods (60 to 120 minutes, 120 to 180 minutes, and 180 to 240 minutes). Sections from the same region of the colon within a dog were considered replicates. Mucosal-to-serosal flux of mannitol was the response variable used for the statistical analysis. Mean ± SEM mucosal-to-serosal flux of mannitol for each period and for each region of the colon (transverse colon, proximal portion of the descending colon, and distal portion of the descending colon) was calculated. Data were normally distributed, as verified by failure to reject the null hypothesis of normality (Shapiro-Wilk statistic). Data for control sections were initially examined for a fixed effect of region and period by use of a mixed-linear model that included the random variance of dog among regions and periods. When there was a significant interaction, ad hoc comparisons were made with the Scheffe adjustment to maintain A = 0.05. On the basis of results for control sections, data for treated sections were compared with data for control sections to detect a fixed effect of region, treatment, and period by use of a mixed-effect linear model that included the random variance of dog nested within treatment and dog among regions and periods. When there was a significant interaction, ad hoc comparisons were made by use of a Scheffe adjustment to maintain A = 0.05.
HISTOLOGIC EXAMINATION
Frequency distribution of histologic categories for control and treated sections were compared by use of a χ2 analysis or Fisher exact test. When the frequency was 0 in a category for control and treated sections, the category was deleted. When there were 3 or fewer categories, a Fisher exact test was performed. Where there were 4 categories, a χ2 analysis was performed, with a 0.5 correction used when a single cell had a frequency of 0.
Results
Electrical conductance—Typically, electrical conductance remained relatively stable over time for all control sections, with a slight but gradual increase toward the end of the experiments for the transverse colon (Figure 1). The mean ± SEM electrical conductance × time for control sections of the transverse colon, proximal portion of the descending colon, and distal portion of the descending colon was 3,115 ± 304.8 (mS/cm2) × min, 2,367 ± 147.0 (mS/cm2) × min, and 2,449 ± 156.7 (mS/cm2) × min, respectively. Mean electrical conductance × time for control sections of the transverse colon was significantly higher than for the proximal and distal portions of the descending colon. However, mean electrical conductance × time for the proximal and distal portions of the descending colon did not differ significantly.
Mean electrical conductance over time for control (A) and carprofen-treated (400 μg/mL; B) sections of mucosa obtained from the transverse colon (squares), proximal portion of the descending colon (circles), and distal portion of the descending colon (inverted triangles) of dogs.
Citation: American Journal of Veterinary Research 69, 2; 10.2460/ajvr.69.2.174
Electrical conductance typically had a gradual then more pronounced increase over time for all treated sections, with little difference among regions (Figure 1). Mean ± SEM electrical conductance × time for treated sections of the transverse colon, proximal portion of the descending colon, and distal portion of the descending colon was 3,902 ± 232.5 (mS/cm2) × min, 3,829 ± 309.6 (mS/cm2) × min, and 3,975 ± 294.3 (mS/cm2) × min, respectively. Mean electrical conductance × time for treated sections of the transverse colon, proximal portion of the descending colon, and distal portion of the descending colon did not differ significantly. However, the mean electrical conductance × time for treated sections was significantly higher than the mean for control sections for all regions of the colon.
Mannitol flux—Mean ± SEM mucosal-to-serosal flux of mannitol for control sections of the transverse colon, proximal portion of the descending colon, and distal portion of the descending colon at 60 to 120 minutes was 0.18 ± 0.016 (μmol/cm2) × h, 0.15 ± 0.018 (μmol/cm2) × h, and 0.15 ± 0.016 (μmol/cm2) × h, respectively; 0.21 ± 0.014 (μmol/cm2) × h, 0.19 ± 0.029 (μmol/cm2) × h, and 0.18 ± 0.018 (μmol/cm2) × h, respectively, at 120 to 180 minutes; and 0.23 ± 0.013 (μmol/cm2) × h, 0.20 ± 0.020 (μmol/cm2) × h, and 0.19 ± 0.018 (μmol/cm2) × h, respectively, at 180 to 240 minutes. Mean mucosal-to-serosal flux of mannitol for control sections of the transverse colon at 180 to 240 minutes was significantly higher than the flux at 60 to 120 minutes. Mean mucosal-to-serosal flux of mannitol for control sections of the proximal and distal portions of the descending colon did not differ significantly among the periods. Mean mucosal-to-serosal flux of mannitol for control sections from the transverse colon, proximal portion of the descending colon, and distal portion of the descending colon did not differ significantly at any time period.
Mean ± SEM mucosal-to-serosal flux of mannitol for treated sections of the transverse colon, proximal portion of the descending colon, and distal portion of the descending colon at 60 to 120 minutes was 0.10 ± 0.016 (μmol/cm2) × h, 0.10 ± 0.011 (μmol/cm2) × h, and 0.09 ± 0.013 (μmol/cm2) × h, respectively; 0.23 ± 0.037 (μmol/cm2) × h, 0.21 ± 0.027 (μmol/cm2) × h, and 0.22 ± 0.037 (μmol/cm2) × h, respectively, at 120 to 180 minutes; and 0.43 ± 0.061 (μmol/cm2) × h, 0.38 ± 0.039 (μmol/cm2) × h, and 0.36 ± 0.056 (μmol/cm2) × h, respectively, at 180 to 240 minutes. Mean mucosal-to-serosal flux of mannitol for treated sections of the transverse colon, proximal portion of the descending colon, and distal oprtion of the descending colon at 180 to 240 minutes was significantly higher than the values at 120 to 180 minutes, which were significantly higher than the values at 60 to 120 minutes. Mean mucosal-to-serosal flux of mannitol for treated sections from the transverse colon, proximal portion of the descending colon, and distal portion of the descending colon did not differ significantly among periods. Mean mucosal-to-serosal flux of mannitol for treated sections of the transverse and distal portion of the descending colon was significantly lower than values for corresponding control sections at 60 to 120 minutes; flux for treated sections of the proximal portion of the descending colon did not differ significantly from values for the control sections at 60 to 120 minutes. Mean mucosal-to-serosal flux of mannitol for treated sections of the transverse colon, proximal portion of the descending colon, and distal portion of the descending colon did not differ significantly from values for corresponding control sections at 120 to 180 minutes. However, mean values for mucosal-to-serosal flux of mannitol for treated sections of the transverse colon and proximal and distal portions of the descending colon were all significantly higher than values for corresponding control sections at 180 to 240 minutes.
Histologic examination—Six control sections of colonic mucosa (1 from each dog) were examined. Normal background inflammation (lymphocytes and plasma cells) was detected in all sections, but edema was not evident in any of the sections. Sloughing of cells from the surface epithelium was minimal in 5 sections and mild in 1 section. Erosions were not evident in 5 sections and involved < 10% of the surface epithelium in 1 section. There was no sloughing of cells within mucosal glands in any sections.
Seven sections of treated colonic mucosa (1 from each dog) were examined. Normal background inflammation (lymphocytes and plasma cells) was detected in 6 sections, and a small focus of neutrophilic inflammation was detected in 1 section. Edema was not evident in any sections. Sloughing of cells from the surface epithelium was minimal in 4 sections, mild in 1 section, and moderate in 2 sections. Erosions were not evident in 4 sections, involved < 10% of the surface epithelium in 1 section, and involved ≥ 10% of the surface epithelium in 1 section. There was no sloughing of cells within the mucosal glands in any sections. The distribution of histologic categories did not differ significantly between treated and control sections.
Fifty-four control sections of colonic mucosa (9 from each dog) were examined after being bathed in Ussing chambers. Normal background inflammation (lymphocytes and plasma cells) was detected in all sections. Edema was not evident in 9 of 54 (16.7%) sections but was rated as mild in 25 (46.3%) sections, moderate in 14 (25.9%) sections, and severe in 6 (11.1%) sections. Sloughing of cells from the surface epithelium was minimal in 7 of 54 (13.0%) sections, mild in 23 (42.6%) sections, moderate in 16 (29.6%) sections, and severe in 8 (14.8%) sections (Figure 2). Erosions were not evident in 34 of 54 (63.0%) sections but involved < 10% of the surface epithelium in 19 (35.2%) sections and involved ≥ 10% surface epithelium in 1 (1.8%) section. There was no sloughing of cells within mucosal glands in 25 of 54 (46.3%) sections; sloughing of cells within mucosal glands was mild in 20 (37.0%) sections, moderate in 6 (11.1%) sections, and severe in 3 (5.6%) sections.
Photomicrographs of a control (A) and carprofentreated (B) section of colonic mucosa at the end of the study. In the control section (panel A), there is mild sloughing of cells from the surface epithelium (arrowheads) but no sloughing of epithelial cells within the mucosal glands or erosions of the surface epithelium. In the carprofen-treated section (panel B), there is mild edema, severe sloughing of cells from the surface epithelium (arrowheads), and erosions of the surface epithelium (arrow). H&E stain; bars = 200 μm.
Citation: American Journal of Veterinary Research 69, 2; 10.2460/ajvr.69.2.174
Sixty-three sections of treated colonic mucosa (9 from each dog) were examined after being bathed in Ussing chambers. Normal background inflammation (lymphocytes and plasma cells) was detected in 62 of 63 (98.4%) sections, and mild inflammation was detected in 1 (1.6%) section. The distribution of inflammation categories did not differ significantly between control and treated sections. Edema was not evident in 16 of 63 (25.4%) sections but was rated mild in 37 (58.7%) and moderate in 10 (15.9%) sections (Figure 2). The distribution of edema categories differed significantly (P < 0.001) between control and treated sections; treated sections typically had moderate or severe edema, compared with control sections, which had mostly mild or moderate edema. Sloughing of cells from the surface epithelium was mild in 5 of 63 (7.9%) sections, moderate in 17 (27.0%) sections, and severe in 41 (65.1%) sections. The distribution of categories for sloughing of surface epithelium differed significantly (P < 0.001) between control and treated sections; treated sections typically had moderate or severe sloughing, compared with control sections, which had mostly minimal sloughing. Erosions were not evident in 1 of 63 (1.6%) sections but involved < 10% of the surface epithelium in 27 (42.8%) sections and ≥ 10% of the surface epithelium in 35 (55.6%) sections. The distribution of erosion categories differed significantly (P < 0.001) between control and treated sections; treated sections typically had erosions involving ≥ 10% of the surface epithelium, compared with control sections, which mostly had no evidence of erosion or had erosion that involved < 10% of the surface epithelium. There was no sloughing of cells within mucosal glands in 39 of 63 (61.9%) sections. Sloughing of cells within mucosal glands was mild in 20 (31.7%) sections, moderate in 3 (4.8%) sections, and severe in 1 (1.6%) section. The distribution of categories for sloughing of cells in the mucosal glands did not differ between treated and control sections.
Discussion
Carprofen increased electrical conductance and permeability to mannitol and caused sloughing of cells and erosions of canine colonic mucosa in vitro. Analysis of these findings suggests compromise of the integrity and loss of barrier function of the colonic mucosa.
Sloughing of cells from the surface epithelium and erosions likely contributed to the increase in electrical conductance and flux of mannitol by allowing indiscriminate passage of ions and molecules of mannitol through areas devoid of epithelial cells. In addition, increased electrical conductance may reflect changes in transcellular and paracellular transport of ions across the epithelium of the colonic mucosa or it may reflect compromise of the functional integrity of intercellular tight junctions.12,13 According to Ohm's law, electrical conductance of the conductor (eg, colonic mucosa) increases proportionally with short-circuit current flowing through the conductor, assuming the transepithelial potential difference across the conductor is unchanged.14 Transport of chloride ions, sodium ions, and potassium ions account for most of the in vitro short-circuit current across the epithelium of the colonic mucosa. Carprofen could have increased electrical conductance by modifying transport of 1 or more of these ions across the epithelium of the colonic mucosa.
It has been clearly determined that prostaglandins can alter the transport of ions across the epithelium of the intestinal mucosa, thereby affecting electrical conductance. Prostaglandins are important secretagogues in the colon of mammals and promote secretion of chloride ions into the lumen. This increases short-circuit current and electrical conductance. Thus, prostaglandins increase electrical conductance across the epithelium of the colonic mucosa by promoting secretion of chloride ions. In the distal portion of the colon of guinea pigs, exogenous prostaglandin E2 increases secretion of chloride ions, resulting in increased electrical conductance.14 In monolayers of human colonic cells, prostaglandin E1 potentiates carbachol-induced secretion of chloride ions and increases electrical conductance.15 On the basis of this information, the effect of carprofen on electrical conductance is unlikely to be attributable to inhibition of prostaglandins through their effect on ion transport because this should result in a decrease in electrical conductance. Whether the effect of carprofen on electrical conductance is mediated by inhibition of prostaglandins through mechanisms other than their effect on ion transport is unknown and was not addressed in the study reported here.
Carprofen could have increased electrical conductance of the colonic mucosa and increased the flux of mannitol by compromising functional integrity of intercellular tight junctions. Mannitol is passively transported along a concentration gradient by paracellular and transcellular pathways in the colonic mucosa of dogs.16 Because there is no active transport of mannitol in the colonic mucosa of dogs, increased permeability to mannitol is likely a result of compromised functional integrity of intercellular tight junctions of the colonic mucosa. Oral administration of indomethacin, an NSAID, uncouples oxidative phosphorylation in the mitochondria of epithelial cells of the intestinal mucosa of rats.7 This causes a decrease in production of ATP by mitochondria, which in turn causes release of calcium from cytoplasmic storage vesicles. The release of calcium adversely affects the functional integrity of intercellular tight junctions.7 Oral administration of indomethacin was associated with increased permeability to radiolabelled EDTA (51Cr-EDTA, a molecular marker of paracellular permeability) in the intestinal mucosa of rats.7 These findings are consistent with another study17 in which NSAIDs increased permeability of human colon to 51Cr-EDTA. A similar increase in permeability was also detected with dinitrophenol, an agent that uncouples oxidative phosphorylation in the mitochondria of epithelial cells of the intestinal mucosa without affecting the activity of cyclooxygenase enzymes. Uncoupling with the release of calcium was also reported in a study8 in which an increase in the intracellular concentration of calcium ions in a monolayer of human colonic cells compromised functional integrity of intercellular tight junctions. A simultaneous increase in electrical conductance and permeability to mannitol was also detected. Thus, carprofen could have uncoupled oxidative phosphorylation in the mitochondria of epithelial cells of the colonic mucosa of dogs, thereby compromising functional integrity of intercellular tight junctions and consequently increasing electrical conductance and permeability to mannitol.
Carprofen-treated sections had significantly more sloughing of cells from the surface epithelium and more erosions. These histologic findings are consistent with compromise of the integrity and loss of barrier function of the treated sections and may explain, in part, the aforementioned changes in electrical conductance and permeability to mannitol. This severe cellular disruption may represent an extension of the cellular damage and demise caused by NSAID-induced uncoupling of phosphorylation and depletion of ATP. Investigators examined the relationship between permeability to mannitol and histologic changes in the intestinal mucosa of rats during experiments conducted by use of Ussing chambers and detected a gradual increase in the permeability of the duodenal and jejunal mucosa to mannitol during the experiment (180 minutes).18 This gradual increase in permeability to mannitol was attributed to sloughing of cells from the surface epithelium of the intestinal mucosa in the face of an extensive repair process. Repair of the mucosal epithelium was indicated by a reduction in villi index and reduced nucleoapical distance, which denotes a reduction in the number of cells in the epithelium with the remaining cells flattening out to compensate for cell loss.8 In the study reported here, sloughing of cells from the surface epithelium is likely to have contributed to the gradual increase in permeability to mannitol of the carprofen-treated sections during the first 180 minutes, which then became more severe for the period from 180 to 240 minutes as sloughing likely worsened and erosions developed. Electrical conductance would be affected because indiscriminate passage of ions through areas devoid of epithelial cells (erosions) would preclude establishment of a concentration gradient necessary for normal transport of ions across the epithelium.
Electrical conductance was higher in the untreated sections of the transverse colon, compared with electrical conductance for the proximal and distal portions of the descending colon. The effect of carprofen overwhelmed any regional differences. Regional differences in electrical conductance of the colonic mucosa have been reported in humans,19 mice,20 and rabbits.21 In rabbits, electrical conductance is higher in the proximal portion of the colon, compared with electrical conductance for the distal portion of the colon,21 which is consistent with our findings for dogs. Regional differences in electrical conductance in rabbits are attributable to a Na+–Cl− cotransport pathway that exists in the proximal portion of the colon but not in the distal portion of the colon.21 Electrical conductance was lower in the ascending colon of humans, compared with electrical conductance in the transverse and descending colon.19 Comparing the findings for the study reported here with those of human studies is complicated by differences in anatomic divisions of the canine and human colon.22,23 Regional differences in electrical conductance in people have been attributed to an aboral gradient of increasing transport of sodium ions.19 Regional differences in electrical conductance of the colonic mucosa of dogs may reflect differences in transport of ions across the epithelium. Hypothetically, greater secretion of chloride ions in the mucosa of the transverse colon, compared with secretion in the descending colon, could explain regional differences. A depletion of ATP inhibits secretion of chloride ions to a greater extent in the distal portion of the colon of rats than in the proximal portion of the colon.24 Similarly for our study, a gradual depletion of ATP in the Ussing chamber environment would reduce secretion of chloride ions in the mucosa of the descending colon, which would cause a lower electrical conductance, compared with the electrical resistance in the transverse colon. Alternately, greater transport of sodium ions in the mucosa of the descending colon, compared with transport in the transverse colon, could explain regional differences in electrical conductance that were detected in our study. Systems for the transport of ions across the epithelium of the colonic mucosa have not been extensively studied in dogs, and additional studies are needed to investigate these hypotheses.
Regional differences in electrical conductance of the canine colonic mucosa could reflect regional variations in intercellular tight junctions. The seal formed by intercellular tight junctions is relative, and despite their name, tight junctions are the most permeable element of the epithelium of the mucosa. The permeability of intercellular tight junctions to molecules varies.25,26 In mice, variations in the expression of the molecular constituents of tight junctions result in differences in paracellular transport of ions across the epithelium of the mucosa in various regions of the intestinal tract.27,28
Permeability to mannitol gradually increased during the three 1-hour flux periods in the control sections of mucosa of the transverse colon but not in the proximal or distal portions of the descending colon. This result, together with regional differences in electrical conductance, suggested that the colonic mucosa of the transverse colon of dogs is distinct from the mucosa of the descending colon, at least in an in vitro environment. Preservation of integrity and barrier function of tissues in a Ussing chamber are always a concern. The in vitro environment of Ussing chambers will eventually fail to sustain tissues. The integrity and barrier function of the colonic and ileal mucosa of rats are better preserved in Ussing chambers, compared with preservation of the duodenal and jejunal mucosa.18 A greater increase in permeability to mannitol over time, as well as histologic evidence of mucosal damage in the duodenum and jejunum compared with the colon and ileum, have been reported.18 Similarly, regional differences for the colonic mucosa of dogs may exist, and the mucosa of the transverse colon may be more vulnerable than other regions to an in vitro environment of a Ussing chamber.
Histologic examination of all sections of colonic mucosa after being bathed in a Ussing chamber was performed to verify viability and preservation of structural integrity in the control sections and to compare the histologic findings from the carprofen-treated sections. Sloughing of cells from the surface epithelium, sloughing of epithelial cells within the mucosal glands, and edema in the colonic mucosa of rats after 60 minutes of immersion in a Ussing chamber were detected in another study.18 These changes gradually worsened throughout the duration of the experiment (180 minutes). In that study,18 mild histologic changes were not associated with increased permeability, whereas severe histologic changes were associated with increased permeability. Mild sloughing of cells from the surface epithelium and sloughing of epithelial cells within the mucosal glands may represent loss of epithelial cells within the limits of repair of the epithelium and therefore does not indicate complete loss of tissue integrity or barrier function. Edema does not result in loss of continuity or function of the epithelium of the mucosa and was not interpreted as loss of tissue integrity or barrier function in the aforementioned study.18 In the study reported here, the epithelium was intact (without erosions) in most (63%) control sections, and edema was also not evident or mild in 63% of the control sections. These mild histologic changes, along with the relatively stable conductance and permeability throughout the experiment, suggested that integrity and barrier function were adequately preserved for the control sections.18 Thus, the severe damage to the epithelium of the carprofentreated sections (sloughing of cells from the surface epithelium and erosions) can be attributed to the effect of carprofen.
One section of mucosa placed in formalin immediately after harvesting from the colon of each dog was examined for histologic evidence of preexisting colonic disease. Although it is presumed that the findings in this section represented the entire colon, we cannot determine whether any changes evident in 1 section represented a focal change limited to that section or were representative of diffuse disease. Isolated sections had some mild changes, but there was no significant difference in the distribution of the frequency of these findings between control and carprofen-treated sections. Thus, for the constraints of the experimental design, we believe that preexisting disease was not a factor in the results.
The concentration of carprofen in the solution was 400 mg/mL. Because, to our knowledge, the effect of carprofen on the colonic mucosa of dogs has not been investigated, we began our examination with a high concentration of carprofen to maximize the confidence of obtaining a significant result. Although a plasma concentration cannot be directly extrapolated to a solution concentration in a tissue bath, we used the information currently available on plasma concentrations after administration of carprofen in dogs as a guideline to choose a solution concentration for the study reported here. The maximum plasma concentration after an orally administered dose of 25 mg of carprofen in Beagles is 18.7 μg/mL.29 This orally administered dose corresponds to approximately 2.2 mg/kg, half the total recommended daily dose of carprofen,30 assuming the dogs were of expected body weight and size for the breed.31 The plasma concentration of carprofen is directly proportional to the dose.29 Whether plasma concentrations similar to 400 μg of carprofen/mL in a Ussing chamber can be achieved in vivo and whether the dose-plasma concentration relation remains linear at such high plasma concentrations is not known. If so, 400 mg/mL represents a concentration of carprofen approximately 10 times that achieved with a full single daily dose in vivo, which is not outside the realm of a clinical overdose. Luminal, intracellular, and interstitial concentrations sof carprofen in the colon of dogs would have provided more representative information from which to extrapolate the solution concentration of carprofen, but this information was not available. The minimum concentration of carprofen resulting in damage to the colonic mucosa of dogs has not been determined. Whether the effect of a specific plasma concentration of carprofen on the colonic mucosa of dogs in vivo is the same as the effect of the corresponding concentration in vitro is unknown. Whether repeated dosing with carprofen has a cumulative damaging effect on the colonic mucosa of dogs in vivo is also unknown.
Carprofen increased electrical conductance and permeability to mannitol and caused sloughing of cells and erosions of the canine colonic mucosa in vitro. Together, these findings suggested compromise of integrity and loss of barrier function of the colonic mucosa.
The mechanisms by which carprofen at a concentration of 400 μg/mL compromised the integrity and barrier function of the colonic mucosa of dogs in vitro are unknown. Studies involving measurement of concentrations of prostanoids and inhibitors of oxidative phosphorylation and transport of radiolabelled ions would help clarify these mechanisms. The effects of carprofen on the colonic mucosa of live dogs have not been investigated. On the basis of our findings, additional investigation is warranted.
The mucosa of the transverse colon of dogs appears to be distinct from the mucosa of the descending colon, at least in an in vitro environment. Regional differences in electrical conductance and permeability to mannitol between the transverse and descending colon may reflect differences in transport of ions across the epithelium of the colonic mucosa or variations in intercellular tight junctions. Regional differences in electrical conductance and permeability to mannitol should be taken into account when designing studies of the colonic mucosa of dogs.
ABBREVIATIONS
| NSAID | Nonsteroidal anti-inflammatory drug |
| AUC | Area under the curve |
| mS | Millisiemen |
SAS, version 9.1, SAS Institute Inc, Cary, NC.
References
- 1.↑
Lanas A, Panes J, Pique JM. Clinical implications of COX-1 and/or COX-2 inhibition for the distal gastrointestinal tract. Curr Pharm Des 2003;9:2253–2266.
- 2.↑
Kurahara K, Matsumoto T, Iida M, et al. Clinical and endoscopic features of nonsteroidal anti-inflammatory drug-induced colonic ulcerations. Am J Gastroenterol 2001;96:473–480.
- 3.
Hall RI, Petty AH, Cobden I, et al. Enteritis and colitis associated with mefenamic acid. Br Med J (Clin Res Ed) 1983;287:1182.
- 4.
Pierrugues R, Fontes J. Hemorrhagic angiodysplasia of the right colon after ingestion of non-steroidal anti-inflammatory agents. Gastroenterol Clin Biol 1994;18:1041–1042.
- 5.
Evans JM, McMahon AD, Murray FE, et al. Non-steroidal anti-inflammatory drugs are associated with emergency admission to hospital for colitis due to inflammatory bowel disease. Gut 1997;40:619–622.
- 6.
Somasundaram S, Rafi S, Hayllar J, et al. Mitochondrial damage: a possible mechanism of the “topical” phase of NSAID induced injury to the rat intestine. Gut 1997;41:344–353.
- 7.↑
Somasundaram S, Sigthorsson G, Simpson RJ, et al. Uncoupling of intestinal mitochondrial oxidative phosphorylation and inhibition of cyclooxygenase are required for the development of NSAID-enteropathy in the rat. Aliment Pharmacol Ther 2000;14:639–650.
- 8.↑
Tai YH, Flick J, Levine SA, et al. Regulation of tight junction resistance in T84 monolayers by elevation in intracellular Ca2+: a protein kinase C effect. J Membr Biol 1996;149:71–79.
- 9.↑
Curry SL, Cogar SM, Cook JL. Nonsteroidal antiinflammatory drugs: a review. J Am Anim Hosp Assoc 2005;41:298–309.
- 11.↑
Stewart J. Approximate integration. In: Stewart J, ed. Calculus early transcendentals. 3rd ed. Pacific Grove, Calif: Brooks/Cole Publishing Co, 1995;457–465.
- 12.
Nedergaard S, Larsen EH, Ussing HH. Sodium recirculation and isotonic transport in toad small intestine. J Membr Biol 1999;168:241–251.
- 13.
Mlodzik-Danielewicz N, Tyrakowski T. Effects of amiloride and bumetanide on hyperpolarization after movement across the distal colon epithelium. Pharmacol Rep 2005;57:489–497.
- 14.↑
Rechkemmer G, Frizzell RA, Halm DR. Active potassium transport across guinea-pig distal colon: action of secretagogues. J Physiol 1996;493.2:485–502.
- 15.↑
Dharmsathaphorn K, Pandol SJ. Mechanism of chloride secretion induced by carbachol in a colonic epithelial cell line. J Clin Invest 1986;77:348–354.
- 16.↑
Nejdfors P, Wang Q, Ekelund M, et al. Increased colonic permeability in patients with ulcerative colitis: an in vitro study. Scand J Gastroenterol 1998;33:749–753.
- 17.↑
Jenkins AP, Trew DR, Crump BJ, et al. Do non-steroidal anti-inflammatory drugs increase colonic permeability? Gut 1991;32:66–69.
- 18.↑
Polentarutti BI, Peterson AL, Sjoberg AK, et al. Evaluation of viability of excised rat intestinal segments in the Ussing chamber: investigation of morphology, electrical parameters, and permeability characteristics. Pharm Res 1999;16:446–454.
- 19.↑
Sellin JH, DeSoignie R. Ion transport in human colon in vitro. Gastroenterology 1987;93:441–448.
- 20.↑
Charney AN, Egnor RW, Alexander-Chacko JT, et al. Effect of E. coli heat-stable enterotoxin on colonic transport in guanylyl cyclase C receptor-deficient mice. Am J Physiol 2001;280:G216–G221.
- 21.↑
Sellin J, DeSoignie R. Rabbit proximal colon: a distinct transport epithelium. Am J Physiol 1984;246:G603–G610.
- 22.
Strombeck DR. Small and large intestine: normal structure and function. In: Guilford WGC, Center SA, Strombeck DR, et al, eds. Strombeck's small animal gastroenterology. 3rd ed. Philadelphia: WB Saunders Co, 1996;318–350.
- 23.
Evans HE. The large intestine. In: Evans HE, ed. Miller's anatomy of the dog. 2nd ed. Philadelphia: WB Saunders Co, 1993;444–450.
- 24.↑
Mayol JM, Alarma-Estrany P, Adame-Navarrete Y, et al. Effects of luminal ATPase inhibitors on electrogenic ion transport in rat distal colon. J Surg Res 2005;129:85–89.
- 25.
Alberts B, Bray D, Lewis J, et al. Junctions cellulaires, adherence cellulaire et matrice extracellulaire. In: Alberts B, Bray D, Lewis J, et al, eds. Biologie moléculaire de la cellule. 3rd ed. Paris: Flammarion, 1994;950–962.
- 26.
Amasheh S, Schmidt T, Mahn M, et al. Contribution of claudin-5 to barrier properties in tight junctions of epithelial cells. Cell Tissue Res 2005;321:89–96.
- 27.
Muresan Z, Paul DL, Goodenough DA. Occludin 1β, a variant of the tight junction protein occludin. Mol Biol Cell 2000;11:627–634.
- 28.
Fujita H, Chiba H, Yokozaki H, et al. Differential expression and subcellular localization of claudin-7, -8, -12, -13, and -15 along the mouse intestine. J Histochem Cytochem 2006;54:933–944.
- 29.↑
Clark TP, Chieffo C, Huhn JC, et al. The steady-state pharmaco-kinetics and bioequivalence of carprofen administered orally and subcutaneously in dogs. J Vet Pharmacol Ther 2003;26:187–192.
- 30.↑
Fox SM, Johnston SA. Use of carprofen for the treatment of pain and inflammation in dogs. J Am Vet Med Assoc 1997;210:1493–1498.
- 31.↑
Deavers S, Huggins RA, Smith EL. Absolute and relative organ weights of the growing beagle. Growth 1972;36:195–208.
