Influence of repeated arthrocentesis and exercise on synovial fluid concentrations of nitric oxide, prostaglandin E2 and glycosaminoglycans in healthy equine joints

REASONS FOR PERFORMING STUDY: The importance of osteoarthritis (OA) in the horse and the difficulty in its early diagnosis have led to a search for potential biomarkers of joint disease. If the levels of such markers are to be interpreted accurately, clinicians and researchers need to know whether they are influenced by environmental factors and/or interventions such as exercise and repeated arthrocentesis. OBJECTIVE: To investigate the influence of repeated arthrocentesis and exercise on nitric oxide (NO), prostaglandin E2 (PGE2) and glycosaminoglycan (GAG) concentrations in synovial fluid (SF) from normal equine joints. METHODS: SF was collected from the left metacarpophalangeal (MCP), radiocarpal and tarsocrural joints of 16 horses. Half of the horses were exercised and arthrocentesis was repeated 14, 14.5, 17 and 24 days after the start of the exercise programme, in both exercised and control horses. Nitric oxide was determined in SF from the MCP joint only and PGE2 and GAG concentrations were determined in SF from all joints. RESULTS: Repeated arthrocentesis caused an increase in NO concentration in the MCP joint on Day 145, in PGE2 concentrations in the radiocarpal and tarsocrural joints on Day 145 and the release of GAGs into SF of the MCP and radiocarpal joints on Day 17. Exercise resulted in an increase in PGE2 levels in all joints but did not influence the other parameters measured. POTENTIAL RELEVANCE: Repeated arthrocentesis is a potential confounding factor for the use of synovial NO, PGE2 and GAG concentrations as markers of joint disease. Based on this study, such a confounding effect can be avoided if one week or more separates arthrocentesis procedures. Moderate exercise causes a transient rise in PGE2 in SF.     

Collagenase-1 (MMP-1) activity in equine synovial fluid: influence of age, joint pathology, exercise and repeated arthrocentesis

REASONS FOR PERFORMING STUDY: Matrix metalloproteinases (MMPs) are considered candidate biomarkers for both physiological and pathological tissue remodelling because of their key role in articular cartilage homeostasis. As disruption of the collagenous architecture is thought to be pivotal in chronic degenerative diseases such as osteoarthritis (OA), the collagenases form an interesting subset of the MMPs. The significance of any biomarker in synovial fluid (SF) can be assessed properly only when fluctuations in patterns induced by physiological processes such as development and growth, and by external influences and interventions such as exercise and repeated arthrocentesis, are known and taken into account. OBJECTIVES: To investigate the activity of MMP-1 in equine SF at different stages of development and in joints affected by OA, and the influence of exercise and repeated arthrocentesis thereon. METHODS: MMP-1 activity was determined in SF of normal joints of fetal, juvenile and mature horses, and in SF of horses suffering from OA, using an internally quenched fluorogenic peptide substrate. MMP-1 activity was also measured in SF from horses subjected to an exercise regimen and those subjected to repeated arthrocentesis. RESULTS: An age-related decline in the SF levels of active MMP-1 was observed. MMP-1 activity was 15-fold higher in fetal than in juvenile animals, which showed significantly higher MMP-1 activity levels than mature horses. In SF of OA joints, MMP-1 activity was increased. Exercise did not affect MMP-1 activity in SF, but repeated arthrocentesis (within 60 h) increased MMP-1 activity significantly. CONCLUSIONS: The high MMP-1 activity in SF of young individuals parallels the high metabolic activity occurring during rapid growth and differentiation at early age. The elevated MMP-1 activity in SF of OA joints probably reflects pathological matrix degradation, confirming the potential of MMP-1 to serve as a biochemical marker for early joint disease. Moderate exercise is not likely to influence the outcome of MMP-1 activity measurements in equine SF, but arthrocentesis should be taken into account as a possible confounding factor. POTENTIAL RELEVANCE: Given the crucial role of the collagen matrix for tissue integrity, MMP-1 activity may be a useful tool in diagnostic, therapeutic or prognostic studies in horses suspected of OA. However, care should be taken to exclude fluctuations in MMP-1 activity induced by physiological processes such as development and growth, and by interventions such as repeated arthrocentesis.     

Influence of glucosamine on matrix metalloproteinase expression and activity in lipopolysaccharide-stimulated equine chondrocytes

OBJECTIVE: To characterize potential mechanisms of action of glucosamine inhibition of matrix metalloproteinase (MMP) expression and activity in lipopolysaccharide (LPS)-stimulated equine chondrocytes. SAMPLE POPULATION: Chondrocytes cultured from samples of metacarpophalangeal articular cartilage collected from cadaveric limbs of horses. PROCEDURE: The effect of glucosamine on MMP activity in conditioned medium from LPS-stimulated cartilage explants was determined by a colorimetric assay with azocoll substrate. Treatments consisted of negative and positive controls, glucose (50 mM), and glucosamine (50, 25, 6.25, 3, and 1.5 mM). The influence of glucosamine on MMP synthesis was determined in chondrocytes in pellet culture incubated with LPS (20 microg/mL). Concentration of MMP-13 was quantified in spent medium via ELISA; nonspecific MMP activity was determined via azocoll digestion in organomercurial-activated medium. Effects of glucosamine on MMP mRNA concentration in similarly treated chondrocytes were determined by northern blot hybridization with MMP-1, -3, and -13 probes. Statistical analyses were performed with 2-way ANOVA. RESULTS: Glucosamine had no effect on activated MMP activity but inhibited MMP protein expression, as determined by azocoll digestion (glucosamine, 3 to 50 mM) and MMP-13 ELISA (glucosamine, 1.5 to 50 mM). Resting mRNA concentrations for MMP-1, -3, and -13 mRNA were significantly lower in cultures exposed to glucosamine at concentrations of 50 and 25 mM than those of positive controls. CONCLUSIONS AND CLINICAL RELEVANCE: Glucosamine appears capable of pretranslational, and possibly also translational, regulation of MMP expression; data suggest a potential mechanism of action for chondroprotective effects of this aminomonosaccharide.     

The changes of antioxidative enzyme activities in equine erythrocytes following exercise

The change in activities of 3 major antioxidative enzymes in equine erythrocytes, superoxide dismutase (SOD), glutathione peroxidase (GSHpx), and catalase, was investigated in order to evaluate the effect of exercise. Blood samples were obtained from 11 thoroughbred horses before and immediately after vigorous exercise which induced the increase of plasma lipid peroxide (Lpx) concentration from 1.16 +/- 0.40 nmol/ml to 1.29 +/- 0.34 nmol/ml. Following the exercise, the GSHpx activity in erythrocytes was significantly reduced from 69 +/- 10 IU/gHb to 65 +/- 8 IU/gHb, whereas SOD and catalase activities were not changed. Effects of an antioxidative compound, containing selenium and vitamin E(Se-E), on the response of antioxidative enzyme activities following the exercise were examined. Seven horses were injected intramuscularly with Se-E(Se:25 mg, vitamin E:54.8 mg) and received the same vigorous exercise. After Se-E treatment, plasma Lpx levels before the exercise were decreased from 1.24 +/- 0.09 nmol/ml to 0.86 +/- 0.03 nmol/ml, however, SOD, GSHpx, and catalase activities were not varied. The Se-E treatment slightly prevented the decrease in GSHpx activity and the increase in plasma Lpx level after the exercise, having no effect on SOD and catalase activities. These results suggested that the changes of GSHpx activity in erythrocytes might reflect the protective condition against exercise-induced lipid peroxidation.     

Evaluation of the influence of prostaglandin E2 on recombinant equine interleukin-1beta-stimulated matrix metalloproteinases 1, 3, and 13 and tissue inhibitor of matrix metalloproteinase 1 expression in equine chondrocyte cultures

OBJECTIVE: To determine the effects of prostaglandin E2 (PGE2) on recombinant equine interleukin (IL)-1beta-stimulated expression of matrix metalloproteinases (MMP 1, MMP 3, MMP 13) and tissue inhibitor of matrix metalloproteinase 1 (TIMP 1) in vitro. SAMPLE POPULATION: Cultured equine chondrocytes. PROCEDURE: Stationary monolayers of first-passage chondrocytes were exposed to graduated concentrations of PGE2 with or without a subsaturating dose (50 pg/ml) of recombinant equine IL-1beta (reIL-1beta) to induce expression of MMP 1, MMP 3, MMP 13, and TIMP 1, followed by RNA isolation and northern blotting. In subsequent experiments, gene expression was similarly quantified from mRNA isolated from cultures pretreated with phenylbutazone to quench endogenous PGE2 synthesis, followed by exposure to reIL-1beta and exogenous PGE2 (5 mg/ml) with appropriate controls. RESULTS: Exogenous PGE2 (10 mg/ml) significantly reduced reIL-1beta-induced expression of MMP 1, MMP 3, MMP 13, and TIMP 1. Abrogation of cytokine induction with this dose of PGE2 was comparable to that for dexamethasone (10(-5) M) control. Similarly, pretreatment with phenylbutazone, followed by exposure to relL-1beta and PGE2 (5 mg/ml), was associated with a reduced expression of the genes of interest, an effect that was significant for MMP 1, MMP 13, and TIMP 1. CONCLUSIONS AND CLINICAL RELEVANCE: The MMP and TIMP 1 are important mediators in the pathophysiologic events in osteoarthritis. The potential for physiologically relevant regulation of expression of these genes by PGE2 is a consideration in the use of drugs that inhibit prostanoid synthesis in the treatment of equine arthropathies.     

Cloning and expression of equine insulin-like growth factor binding proteins in normal equine tendon

OBJECTIVES: To define a portion of the nucleotide sequences of each of the 6 insulin-like growth factor (IGF) binding proteins (IGFBPs) in horses and describe patterns of messenger RNA (mRNA) and protein expression for IGFBPs in normal equine tendons. ANIMALS: 7 horses. PROCEDURE: Total RNA was extracted from the tensile region of normal superficial digital flexor tendons and reverse transcribed into complimentary DNA (cDNA). The cDNA was amplified via PCR, and products representing portions of each IGFBP were cloned and sequenced. Nucleotide sequences were used to deduce the amino acid sequences, and both nucleotide and predicted amino acid sequences were compared with those published for bovine, human, mouse, and ovine IGFBPs. Gene expression was quantitated by real-time PCR assay, and protein expression was evaluated by western ligand blot (WLB). RESULTS: Clones ranged in size from 262 to 522 bp and had high degrees of sequence homology with other mammalian species. Sequence homology was highest between bovine and equine IGFBPs (86% to 95%) and amongst the IGFBP-5 sequences from the various species (92% to 95%). Message for IGFBP-2 to -6, but not IGFBP-1, was expressed in normal tendon. Protein expression for IGFBP-2, -3, and -4 was detected byWLB in normal tendon and markedly increased in damaged tendons. CONCLUSIONS AND CLINICAL RELEVANCE: Results provide basic information and tools needed for further characterization of the role of the IGF system in tendon healing and may lead to the ability to potentiate the response of healing tendon to exogenous IGF-I via concurrent manipulation of IGFBPs.     

Hypoxia and a hypoxia mimetic up-regulate matrix metalloproteinase 2 and 9 in equine laminar keratinocytes

The aim of this study was to determine if hypoxia and the hypoxia mimetic cobalt chloride regulate the activity of matrix metalloproteinase (MMP)-2 and -9 in cultures of equine hoof keratinocytes. These effects were assessed in primary cultures of laminar keratinocytes using gelatin zymography. Incubation of keratinocytes with cobalt chloride significantly increased the levels of active MMP-2 compared to untreated controls. Hypoxia significantly increased the expression of active MMP-2 and -9 in keratinocyte cultures. This up-regulation was observed after 6h and peaked at 24h. The study findings provide novel evidence of a potential link between hypoxia within the hoof and up-regulation of MMPs which may in turn result in damage to the lamellar basement membrane.     

Fever and acute phase response induced in dwarf goats by endotoxin and bovine and human recombinant tumour necrosis factor alpha

Tumour necrosis factor (TNF), a polypeptide produced by mononuclear phagocytes, has been implicated as an important mediator of inflammatory processes and of clinical manifestations in acute infectious diseases. To study further the potential role of TNF in infectious diseases, recombinant Escherichia coli (E. coli) derived human (r.HuTNF-alpha) and bovine TNF (r.BoTNF-alpha) were intravenously (i.v.) administered in dwarf goats. Rectal temperature, heart rate, rumen motility, plasma zinc and iron concentrations, and certain other blood biochemical and haematological values were studied and compared with the changes seen after E. coli endotoxin (LPS) was administered (dose: 0.1 microgram/kg i.v.). Following a single injection of 4 micrograms/kg of r.BoTNF-alpha, shivering and biphasic febrile response were observed, accompanied by tachycardia, inhibition of rumen contractions, drop in plasma zinc and iron concentrations, lymphopenia, and neutropenia followed by neutrophilia. The i.v. administration of a single injection of 4 micrograms/kg r.HuTNF-alpha induced shivering and biphasic febrile responses, accompanied by anorexia and a similar drop in plasma trace metal concentrations when compared with r.BoTNF-alpha-treated goats. The TNF-alpha-induced symptoms were essentially the same as those that occurred after LPS administration. However, the time of onset of these changes after the injection of TNF-alpha was significantly shorter than after LPS. Moreover, the r.BoTNF-alpha induced a longer lasting neutrophilic leucopenia, less neutrophilia, and a more persistent lymphopenia than after LPS injection. Neither r.BoTNF-alpha nor LPS caused severe haemo-concentration. Furthermore, no cross-tolerance between r.BoTNF-alpha and LPS could be demonstrated. We conclude that both r.BoTNF-alpha and r.HuTNF-alpha induce many of the physiologic, haematologic and metabolic changes that characterize the acute phase response to LPS. The overlapping biological activities of r.BoTNF-alpha, r.HuTNF-alpha and LPS in dwarf goats may indicate that both recombinant tumour necrosis factors have some homology with caprine TNF-alpha.     

Ketamine inhibits LPS-induced tumour necrosis factor-alpha and interleukin-6 in an equine macrophage cell line

Ketamine is widely used in equine anaesthesia. Beside its anaesthetic and analgesic properties, ketamine possesses a cytokine-modulating activity. However, to date, no data are available regarding the inhibitory effect of ketamine on the cytokine response in horses. In horses, cytokines such as tumour necrosis factor-alpha (TNF-alpha) and interleukin-6 (IL-6) play a pivotal role in the pathogenesis of equine endotoxaemia following gastrointestinal disorders. Hence, the objective of this study was to assess the influence of ketamine on LPS-induced TNF-alpha and IL-6 formation in an equine macrophage cell line (eCAS cells). The results demonstrate a cytokine-modulating activity of ketamine in an equine cell line, suggesting a beneficial role for ketamine in the treatment of equine endotoxaemia.     

Effect of gentamicin sulfate and sodium bicarbonate on the synovium of clinically normal equine antebrachiocarpal joints

The effect of gentamicin sulfate, unbuffered and buffered with sodium bicarbonate, on synovial fluid and membrane of clinically normal equine joints was evaluated. Thirty-six adult horses with clinically normal antebrachiocarpal joints were allotted to 6 treatment groups of 6 horses each. One antebrachiocarpal joint in each horse was chosen for treatment. Group-1 horses were given gentamicin (3 ml; 50 mg/ml); group-2 horses were given sodium bicarbonate (3 ml; 1 mEq/ml); group-3 horses were given gentamicin (3 ml; 50 mg/ml) and sodium bicarbonate (3 ml; 1 mEq/ml); group-4 horses were not treated; and horses of groups 5 and 6 were given polyionic physiologic solution (3 and 6 ml, respectively). Synovial fluid specimens were obtained from 5 horses of each group for cytologic analysis at postinjection hours (PIH) 0, 24, 72, and 192 and for pH determination at PIH 0, 0.25, 0.5, 1, 4, 8, 24, 72, and 192. The sixth horse of each group was euthanatized at PIH 24, and the synovial membrane of the treated and contralateral (nontreated) antebrachiocarpal joints was examined macroscopically and microscopically. After intra-articular gentamicin administration, the mean synovial fluid pH was lowest (5.98) at PIH 0.25, but by PIH 8, it was not significantly different from the control value (group-5 horses). When sodium bicarbonate was combined with gentamicin before intra-articular administration, the mean synovial fluid pH was lowest (7.07) at PIH 0.25, but by PIH 1, it was not significantly different from the control value (group-6 horses).(ABSTRACT TRUNCATED AT 250 WORDS)     

The cloning and expression of matrix metalloproteinase-2 and tissue inhibitor of matrix metalloproteinase 2 in normal canine lymph nodes and in canine lymphoma

Matrix metalloproteinase-2 (MMP-2) and its inhibitor, tissue inhibitor of matrix metalloproteinase 2 (TIMP2), are known to be important in cancer. The purposes of this study were to determine the cDNA sequence of canine MMP-2 and to investigate the expression patterns of MMP-2 and TIMP2 in normal canine lymph nodes and spontaneously arising canine lymphomas. We cloned and sequenced a PCR product containing most (1901 base pairs) of the coding sequence of canine MMP-2 that translates into a 623 amino acid protein. The cDNA and deduced amino acid sequences are highly homologous to those of other mammalian species. Canine MMP-2 and TIMP2 mRNAs were detectable in the majority of normal lymph node and lymphomatous samples evaluated. No statistical difference was identified when comparing the expression of either gene with regard to normal versus neoplastic nodes, nodal versus extranodal lymphoma, lymphoma grade, or B versus T cell immunophenotype.     

Evaluation of the inflammatory response to two intra-articular hyaluronic acid formulations in normal equine joints

Intra-articular (IA) hyaluronic acid (HA) is commonly used to treat equine arthritis. Inflammatory response or “joint flare” is a recognized potential side effect. However, the incidence and severity of inflammation following IA HA injection in horses is not well documented. This study compared the effects of two IA HA formulations of different molecular weight (MW) and a saline control on clinical signs and synovial fluid markers of inflammation in normal equine joints. Eight adult horses each had three healthy fetlock joints randomly assigned to treatment with either 1.4 mega Dalton HA, 0.8 mega Dalton HA or saline control once weekly for three weeks. Clinical evaluation and synovial fluid analysis were performed by blinded assessors. Outcomes of interest were lameness score, joint effusion score and synovial fluid white cell count and differential, total protein, viscosity and serum amyloid A. Joints injected with HA developed significant mild-to-moderate inflammatory responses often associated with lameness and joint effusion compared with saline control joints. The higher MW HA formulation elicited a significantly greater inflammatory response than the lower MW HA after the first injection. In HA injected joints, viscosity remained poor for the entire study. Both IA HA formulations in this study induced an inflammatory response in healthy equine joints. This may have implications for the use of HA in equine joints. The findings in this study are limited to the two HA formulations used. Further investigation of different HA formulations and the use of HA in normal and arthritic equine joints is warranted.     

Profiles of matrix metalloproteinase activity in equine tear fluid during corneal healing in 10 horses with ulcerative keratitis

OBJECTIVE: Levels of tear film matrix metalloproteinases (MMPs) activity are significantly elevated in horses with ulcerative keratitis and contribute to the excessive breakdown of stromal collagen. Changes in the amount of proteolytic activity in horse tear film during corneal healing and stromal remodeling have not yet been reported, but we hypothesize they should decrease. In the present study we analyzed serial tear fluid from horses with ulcerative keratitis to identify any changes in MMP activity during corneal healing and stromal remodeling. PROCEDURES: Samples of tear fluid were obtained from both eyes of 10 horses with ulcerative keratitis on the day of admission (day 1) at the hospital and then at various time points until complete healing of the cornea. Tear film MMP2 and MMP9 activity was determined by quantitative gelatin zymography. In all cases medical treatment included topical applications of equine serum, antibiotics, atropine and systemic administration of anti-inflammatory drugs. Surgical procedures were performed in several cases on day 2 in addition to the medical treatment. RESULTS: The mean total MMP activity (+/- SD) measured in relative standard units (RSU) in the tear fluid of the ulcerated eye (2.44 +/- 1.44) of the 10 horses was significantly higher than the mean in the contralateral eye (0.81 +/- 0.68) (P = 0.006), on the day of admission at the VMTH. The mean MMP activity in these ulcerated eyes significantly decreased (-82.4%) between the first day of admission and the day when the ulcer had completely healed (P = 0.0002). The activity level in the healed eye (0.43 +/- 0.17) was not significantly different to the one in the contralateral eye (0.36 +/- 0.18) on the day of complete corneal healing (P = 0.374). The level of MMP activity in the contralateral eye also decreased from 0.81 +/- 0.68-0.36 +/- 0.18 but this decrease (56%) was not significant (P = 0.069). CONCLUSIONS: Ulcerative keratitis in horses is associated with initially high levels of tear film proteolytic activity that decrease as the ulcers heal. The success of medical and surgical treatment of the corneal ulcers is reflected by the enzyme activity in tears. In horses successful treatment does lead to a rapid reduction in tear film proteolytic activity that corresponded with the improvement in the clinical signs of corneal ulceration. Measurement of MMP activity in the tear film might represent a way to monitor the progression of corneal healing in horses with ulcerative keratitis.     

The role of inflammation and matrix metalloproteinases in equine endometriosis

Equine endometriosis is a multifactorial disease considered to be a major cause of equine infertility. The purpose of this study was to evaluate the reliability of histomorphological grading for biopsy-like samples compared to entire uterine wall samples, to examine the association between the degree of endometriosis with animal age, and to investigate the role of inflammation in endometriosis and the expression of different matrix metalloproteinases in equine endometrium. Histomorphological lesions in 35 uterine samples were examined while comparing biopsy-like samples and entire-wall samples. Seventeen uterine samples were stained with antibodies against MMP-2, MMP-9, MMP-14, and TIMP-2. The morphologic evaluation results of the biopsy-like tissue and entire-wall samples were significantly correlated. Endometriosis in older mares (>12 years of age) was more severe than in young mares (2~4 years of age), confirming the positive correlation between animal age and disease severity, while inflammation was poorly related to the degree of endometriosis. MMP-2 and MMP-14 were detected in stromal cells, while MMP-9 and TIMP-2 were both found in stromal and glandular epithelial cells. There were no significant differences in MMPs expression between the two groups (young vs. old mares). Additional studies on the activity of MMPs could further define the role of these enzymes in equine endometriosis.     

Plasma leptin responses to lipopolysaccharide and tumor necrosis factor alpha in cows

Peripheral administration of bacterial lipopolysaccharide (LPS) and various inflammatory cytokines to rodents is known to raise plasma levels of leptin, a potent satiety factor secreted from adipocytes, implying a role of leptin in endotoxin-induced anorexia. We previously reported no effect of LPS on serum leptin levels in sheep, despite marked anorexia and fever. Our results suggest that leptin might not be involved in the endotoxin-induced anorexia in ruminants. To test this idea, in the present study, plasma leptin levels were measured during acute experimental endotoxemia in Holstein cows. Intravenous injection of LPS induced anorexia accompanied with increases in plasma levels of cortisol and insulin, all of which are known to stimulate leptin secretion in rodent and human, while it did not affect plasma leptin levels at all in cows. Similar results were also obtained after injection of recombinant bovine tumor necrosis factor alpha. These results indicate that plasma leptin levels in cows during acute endotoxemia are differentially regulated from those in rodents, and that leptin might not be involved in the endotoxin-induced anorexia in ruminants.     

The influence of lameness on equine stride length consistency

The aim of this study was to assess the influence of orthopaedic pain on the variation of stride length as a kinematic system-parameter in 21 horses with forelimb lameness. Data were collected while the horses were trotting on a treadmill during a minimum of 12 motion cycles, both before and after intra-articular or perineural anaesthesia. Stride length was assessed for each motion cycle, and the mean and standard deviation were calculated for each condition. Forelimb lameness was documented as percentage of asymmetry of vertical head movement. With significant decrease of forelimb lameness after regional anaesthesia, the SD of stride length increased significantly (+0.35%, P< 0.05). Our results show that in the presence of orthopaedic pain horses keep stride variability low, possibly because the lame horse employs an optimum compensatory mechanism to reduce the pain in the affected limb, and every deviation from this pattern increases pain.     

The effect of exercise and conditioning on equine red blood cell characteristics

The effect of exercise and conditioning on 2,3-diphosphoglycerate levels was studied in nine mature horses. During a 12 minute exercise bout producing heart rates of 165 bpm, 2,3-DPG was significantly increased (p<.05). In addition, exercising levels of 2,3-DPG were increased (p<.05) approximately 8% after a six-week submaximal conditioning program. These increases could not be entirely attributed to changes in erythrocyte number. Mean corpuscular volume was also increased during exercise (p<.05) but was not altered by conditioning.     

The effects of cyclo-oxygenase inhibitors on bile-injured and normal equine colon

A potential adverse effect of cyclo-oxygenase (COX) inhibitors (nonsteroidal anti-inflammatory drugs [NSAIDs]) in horses is colitis. In addition, we have previously shown an important role for COX-produced prostanoids in recovery of ischaemic-injured equine jejunum. It was hypothesised that the nonselective COX inhibitor flunixin would retard repair of bile-injured colon by preventing production of reparative prostaglandins, whereas the selective COX-2 inhibitor, etodolac would not inhibit repair as a result of continued COX-1 activity. Segments of the pelvic flexure were exposed to 1.5 mmol/l deoxycholate for 30 min, after which they were recovered for 4 h in Ussing chambers. Contrary to the proposed hypothesis, recovery of bile-injured colonic mucosa was not affected by flunixin or etodolac, despite significantly depressed prostanoid production. However, treatment of control tissue with flunixin led to increases in mucosal permeability, whereas treatment with etodolac had no significant effect. Therefore, although recovery from bile-induced colonic injury maybe independent of COX-elaborated prostanoids, treatment of control tissues with nonselective COX inhibitors may lead to marked increases in permeability. Alternatively, selective inhibition of COX-2 may reduce the incidence of adverse effects in horses requiring NSAID therapy.