Mucosal injury and inflammatory cells in response to brief ischaemia and reperfusion in the equine large colon



    Corresponding author
    1. Transplant Center, Department of Surgery, College of Medicine, Shands at the University of Florida, Gainesville, USA.
    Search for more papers by this author
  • A. J. MORTON,

    1. Island Whirl Equine Colic Research Laboratory, Department of Large Animal Clinical Sciences; Department of Infectious Diseases and Pathology, College of Veterinary Medicine
    Search for more papers by this author
  • A. S. GRAHAM,

    1. Island Whirl Equine Colic Research Laboratory, Department of Large Animal Clinical Sciences; Department of Infectious Diseases and Pathology, College of Veterinary Medicine
    Search for more papers by this author

    1. Gastroenterology, Hepatology and Nutrition Faculty, Department of Medicine, College of Medicine
    Search for more papers by this author
  • J. R. ABBOTT,

    1. Island Whirl Equine Colic Research Laboratory, Department of Large Animal Clinical Sciences; Department of Infectious Diseases and Pathology, College of Veterinary Medicine
    Search for more papers by this author
  • M. M. R. POLYAK,

    1. Transplant Center, Department of Surgery, College of Medicine, Shands at the University of Florida, Gainesville, USA.
    Search for more papers by this author

    1. Island Whirl Equine Colic Research Laboratory, Department of Large Animal Clinical Sciences; Department of Infectious Diseases and Pathology, College of Veterinary Medicine
    Search for more papers by this author



Reason for performing study: Intestinal ischaemia and reperfusion (I/R) can activate inflammatory cells in the equine colon, although effects on different types of inflammatory cells have received little attention.

Objectives: To assess early mucosal injury, the reaction of mucosal neutrophils, eosinophils, mast cells and macrophages, and cyclooxygenase (COX)-1 and -2 expression in response to I/R in the equine large colon.

Methods: Large colon ischaemia was induced for 1 h (1hI) followed by 4 h of reperfusion in 6 horses, and mucosal biopsies were sampled before and after ischaemia, and after 1, 2 and 4 h of reperfusion. Semithin sections (500 nm) of epon-embedded biopsies were stained with toluidine blue for histomorphometric evaluation. The number and distribution of mucosal macrophages (CD163), neutrophils (calprotectin), eosinophils (LUNA) and mast cells (toluidine blue) were determined, and mucosal COX-1 and -2 expression was identified.

Results: Ischaemia caused epithelial cell and nuclear swelling (mean ± s.e. nuclear width; control: 2.7 ± 0.2 µm vs. 1hI: 4.2 ± 0.2 µm; P<0.01), subepithelial oedema (control: 0.2 ± 0.1 µm vs. 1hI: 3.2 ± 0.2 µm; P<0.01) and increased epithelial apoptosis (control: 14.3 ± 4.1 apoptotic cells/mm mucosa vs. 1hI: 60.4 ± 14.0 apoptotic cells/mm mucosa; P<0.01). COX-2 expression (P<0.01) was evident after ischaemia. Reperfusion caused paracellular fluid accumulation (control: 0.9 ± 0.1 µm vs. 1hI: 0.6 ± 0.6 µm vs. 1hI + 4hR: 1.6 ± 0.2 µm; P<0.05). Epithelial repair started at 1 h of reperfusion (P<0.001), followed by migration of neutrophils into the mucosa after 2 h (control: 72.3 ± 18.4 cells/mm2 mucosa vs. 1hI + 2hR: 1149.9 ± 220.6 cells/mm2 mucosa; P<0.01). Mucosal eosinophils, mast cells and macrophages did not increase in numbers but were activated.

Conclusions: Epithelial injury and COX-2 expression caused by short-term hypoxia were followed by intense inflammation associated with epithelial repair during reperfusion.

Potential relevance: Equine colonic mucosa subjected to a brief period of ischaemia can repair during reperfusion, despite increased mucosal inflammation.


Ischaemia causes mucosal damage in horses with large colon volvulus, resulting in translocation of bacteria and toxins, endotoxaemic shock and possible death (Snyder et al. 1989). It also sets conditions for generation of reactive oxygen species, damage to the vasculature and activation of granulocytes after reoxygenation of the tissue (Rowe and White 2002). Moore et al. (1994a,b) found that 3 h of reperfusion exacerbated colonic mucosal injury and increased neutrophil influx in a low-flow ischaemia model in horses. However, reperfusion did not affect production of prostaglandins and cytokines in equine colon (Moore et al. 1995a) and response to antioxidants was not typical of reperfusion injury (Moore et al. 1995b). Also, prostaglandins, produced by cyclooxygenase (COX) during reperfusion, could contribute to tissue damage or assist repair (Crofford 2001; Wallace and Devchand 2005; Little et al. 2007) and early evidence of repair by restitution was evident in Moore's reperfusion model (Moore et al. 1994a). The low-flow ischaemia study by Moore et al. (1994b) in equine colon focused on the role of neutrophils as the major inflammatory cell, although eosinophils and other cells seemed to be involved to an undetermined extent (Moore et al. 1994b). Therefore, the reaction of inflammatory cells and the expression of COX during colonic ischaemia and reperfusion (I/R) warrant further study.

Cells of the innate immune system are the main initiators of acute inflammatory reactions. At the beginning of injury, resident macrophages recognise damage-associated signals, invading bacteria and toxins. Alerted by these signals, macrophages attract large numbers of neutrophils to the site of injury to assist in recognising, ingesting and destroying the invading agents (Smith et al. 2005). Neutrophil infiltration is a crucial step in the I/R cascade, and much of the tissue injury that occurs upon reperfusion is thought to result from neutrophilic radicals and proteolytic enzymes (Gayle et al. 2000). However, neutrophils also play a key role in controlling the infection, sterilising the wound and generating pro-reparative signals (Serhan and Savill 2005; Nathan 2006). Additionally, resident eosinophils and mast cells are potent immunomodulatory cells that are activated by similar signals (Rothenberg et al. 2001; Galli et al. 2008). Once activated, mast cells release toxic metabolites, initiate inflammation and recruitother immune cells to the site of injury (Marshall 2004). Eosinophils are frequently found in association with activated mast cells and are thought to manipulate the inflammatory response triggered by mast cell degranulation (Munitz and Levi-Schaffer 2004).

Although mucosal neutrophils, eosinophils, mast cells and macrophages can interact and contribute to mucosal inflammation and injury (Santos et al. 2001; Chen et al. 2004; Furuta et al. 2005), the responses of these cells in colonic I/R in the horse are unknown. The purpose of the present study was to determine the number and tissue distribution of neutrophils, eosinophils, mast cells and macrophages, and the expression of COX-1 and -2 in response to I/R in the equine colon. The hypothesis was that colonic I/R can induce an intense inflammatory response that causes further damage to the mucosa.

Materials and methods


Six horses used in this study were of mixed breeds with a mean age of 16 years and a mean bodyweight of 548 kg. They were donated for research purposes and they were free of gastrointestinal diseases. The following study was performed with approval and under guidelines of the Institutional Animal Care and Use Committee of the University of Florida. Horses were fed grass hay (2% of their bwt/day), and water was provided ad libitum. Horses were adapted to their diet and environment for at least one week before the study.

Experimental procedures

A 14 gauge, 13.3 cm Teflon catheter was inserted into the left jugular vein for administration of anaesthetic drugs and isotonic fluids. Horses were placed under general anaesthesia according to the following protocol: xylazine (1.0 mg/kg bwt, i.v.) to provide sedation and then general anaesthesia was induced with diazepam (0.1 mg/kg bwt, i.v.) to effect followed by ketamine (2.2 mg/kg bwt, i.v.) as a bolus injection. General anaesthesia was maintained with isoflurane (1–3%) in 100% oxygen. Horses were mechanically ventilated at 6 breaths/min. Isotonic polyionic fluids were infused i.v. continuously at 2.5–5 ml/kg bwt/h. Mean arterial blood pressure was monitored through a 20 gauge, 5.1 cm Teflon catheter in the facial artery and was maintained at or above 60 mmHg. Other monitoring tools used during anaesthesia included electrocardiography, blood gas analysis, capnography and pulse oximetry.

Horses were positioned in dorsal recumbency and prepared for an aseptic ventral midline celiotomy. The large colon was exteriorised and placed on a plastic drape on the ventral abdomen. To induce ischaemia, a 40 cm segment of colon at the pelvic flexure was subjected to transmural compression by intestinal clamps at each end of the selected segment, and combined venous and arterial occlusion was achieved with umbilical tape ligatures. After induction of ischaemia, the colon, colonic vasculature and associated mesentery were surgically divided at the pelvic flexure so that 2 segments of colon comparable in size (dorsal and ventral) did not communicate. The colon was then replaced in the abdomen and the abdominal incision closed temporarily with towel clamps. After 1 h of ischaemia, the colon was re-exteriorised and one of the 2 ischaemic segments was resected for histological evaluations and in vitro experiments (Graham et al. 2011). At the same time, the clamps and ligatures were removed from the remaining segment of colon and it was replaced in the abdomen to allow resumption of blood flow for 4 h of reperfusion under general anaesthesia. Small mucosal biopsies (1–2 cm2) were taken before (control), after ischaemia (1hI), and after 1 (1hR), 2 (2hR) and 4 h of reperfusion (4hR). After the reperfused tissues were sampled, the horses were humanely subjected to euthanasia with an overdose of sodium pentobarbital (88 mg/kg bwt, i.v.) while under anaesthesia. The same investigator performed all surgeries and sampling (D.E.F.).

Sample preparation

Samples were fixed in formalin, subsequently embedded in paraffin, and cut into 5 µm sections. After deparaffinising and rehydration, sections were stained with LUNA (Luna 1963) to assess eosinophils within the colonic mucosa. Paraffin-embedded sections were also processed for immunohistochemistry.

Additionally, four 2 mm punch mucosal biopsies of control tissues and tissues after 1hI, 1hR, 2hR and 4hR were processed for toluidine blue staining (histomorphometry, mast cells). Briefly, biopsies were fixed in glutaraldehyde and paraformaldehyde at 4°C overnight, washed in cacodylate buffer and placed in osmium tetroxide for 1 h at room temperature. After washing, samples were dehydrated in graded concentrations of acetone and embedded in epon. Tissues were then cut into semithin sections (500 nm) with an ultramicrotome equipped with a glass knife. Sections were mounted on glass slides and stained with toluidine blue.


The neutrophil-marker calprotectin was detected using 1:100 monoclonal mouse anti-human macrophage antibody MAC3871 according to a previously published protocol (Grosche et al. 2008). Although calprotectin is also expressed in activated macrophages, the correlation between the numbers of calprotectin-positive cells and neutrophils is strong in I/R models in the horse (Grosche et al. 2008). For detection of macrophages, the macrophage surface protein CD163 was stained with 10 µg/ml of a monoclonal mouse anti-human macrophage surface antibody (clone AM-3 K)2 and a commercially available ABC detection kit with DAB as chromogen (Faleiros et al. 2010). Cytoplasmic COX-1 and -2 were stained with polyclonal goat anti-human antibodies3 based on a previously described protocol (Morton et al. 2009).

For antigen retrieval, all tissues underwent heat pretreatment using a pressure cooker and retrieval buffer with a pH of 6.0. After cooling and washing with Dulbecco's PBS, sections were processed according to manufacturer's instructions5. When the desired stain intensity has developed, all tissues were counterstained with Mayer's haematoxylin in a routine manner before processing for mounting. Instead of primary antibodies, Dulbecco's PBS only was added to experimental samples as negative controls. Positive stained cells appeared brown.


For evaluation of the images obtained by light microscopy, Image-Pro Express 5.04 was used. Routine histomorphometric measures (epithelial height, epithelial width, percentage of denuded epithelium) were determined in toluidine blue-stained sections of 3 randomly defined mucosal segments from each tissue with a length of 217.5 µm (equal to the length of one image using the 40× objective) according to a previously described protocol (Rötting et al. 2003). Additional histomorphometric measurements included the height and width of 15 epithelial cell nuclei, the width of the paracellular space between 15 epithelial cells and the dimension of the subepithelial space on 15 locations. Swollen/necrotic epithelial cells, apoptotic cells characterised by nuclear chromatin margination and condensation, and presence of apoptotic bodies were manually counted within the epithelium of each segment and their number was calculated/mm mucosal length. Additionally, the length of the epithelium that displayed features of ongoing regeneration and repair (covering of denuded or injured mucosal areas by detached epithelial cells, membrane extensions of intact neighbouring cells or flattened epithelial cells) was expressed as percentage of repaired epithelium.

Neutrophils, eosinophils, mast cells and macrophages

The mean numbers of calprotectin-positive neutrophils/mm2 cross-sectional venule area, mucosal area and mucosal zones of equal size (M1–M5) were counted and calculated (Grosche et al. 2008). In LUNA-stained tissues, the numbers of eosinophils/mm2 mucosal area and mucosal zones of equal size (M1–M5) were counted as described above for calprotectin-positive neutrophils.

Mast cells were characterised by their red-purple-stained granules after staining with toluidine blue. Three randomly defined segments of colonic mucosa with a length of 217.5 µm (equal to the length of one image using the 40× objective) and the full height of the mucosa were photographed. The numbers of mast cells within the mucosa were counted manually, and the mean numbers of mast cells/mm2 mucosa, and /mm2 upper and lower half of the mucosal lamina propria were determined. The same procedure was used to determine the number and distribution of CD163-positive macrophages.

COX-1 and -2 expression

For evaluation of mucosal COX expression, the epithelium, upper and lower lamina propria, and crypts were scored from 0–3 (Morton et al. 2009). Grade 0 was assigned when stained cells were absent or single stained cells were observed after careful inspection. Grades 1, 2 and 3 were assigned if accumulation of stained cells wassubjectively assessed as mild, moderate and marked, respectively. Histomorphometric examinations, quantification of eosinophils, mast cells, and calprotectin- and CD163-positive cells within tissues were performed blindly by one investigator (A.G.).

Statistical analysis

Data were expressed as means ± s.e. Values of P<0.05 were considered as significant. Kruskal-Wallis test was used to compare nonparametric data during different I/R time periods and with controls. Whenever a significant P value for ischaemia and reperfusion was identified, Mann-Whitney U test was used for pair-wise comparison.



Ischaemia for 1 h resulted in minor but significant histomorphometric changes of the colonic epithelium (Fig 1b) compared to controls (Fig 1a). Ischaemic injury was characterised by cellular oedema (increased epithelial cell and nuclear width), subepithelial fluid accumulation and chromatin condensation and margination, a characteristic of apoptosis (Figs 2a–c). Single epithelial cells became necrotic, although their number was not significantly increased during reperfusion (Table 1, Figs 2a–d). Epithelial injury was not exacerbated by reperfusion (Figs 1c, d). Dominant changes during reperfusion were enlargement of paracellular spaces, formation of large subepithelial clefts that were covered with shortened epithelial cells or membrane extensions between neighbouring cells (Figs 3b and 4a,b). Although the cell structure of epithelial cells did not appear completely normal after 4hR, small epithelial defects were re-epithelialised and large subepithelial clefts were covered by a continuous layer of shortened epithelial cells with apical membrane connections that appeared tightly adhered to each other (Figs 3a–c, 4b–d). This was first evident at 1hR (Table 1). During reperfusion, apoptotic bodies became evident (Figs 3a–d, 4a–d), although it was not possible to distiguish between apoptotic epithelial cells and intraepithelial immune cells (lymphocytes, neutrophils, eosinophils).

Figure 1.

Colonic mucosal tissues during I/R (toluidine blue; x400). a) Control; b) 1 h of ischaemia; c) 1 h of ischaemia followed by 2 h of reperfusion; d) 1 h of ischaemia followed by 4 h of reperfusion. Ischaemia was characterised by epithelial cell swelling, subepithelial vacuolisation and fluid accumulation, detachment from the basement membrane, small epithelial defects and single cell necrosis (b). No further exacerbation of epithelial damage was present after reperfusion (c, d). Neutrophils, lymphocytes and eosinophils infiltrated the damaged epithelium and moved into the intestinal lumen (c). Epithelial defects were covered by a continuous layer of shortened epithelial cells that appeared tightly adhered to each other (d).

Figure 2.

Characteristic cellular features of the colonic epithelium and subepithelium after 1 h of ischaemia (toluidine blue; x1000, oil immersion). One hour of ischaemia resulted in epithelial swelling, subepithelial vacuolisation and fluid accumulation, detachment from the basement membrane, single cell necrosis (inline image) and small epithelial defects (a–d). Epithelial cell nuclei were rounded and characterised by chromatin margination (; a–c). Some apoptotic bodies were present in the epithelium and subepithelium (−; a–c). Subepithelial mast cells appeared degranulated characterised by decrease of cytoplasmic granules which were located near the cellular plasma membrane (−; a–c). Resident macrophages displayed phagocytic activity characterised by large cytoplasmic vacuoles containing cell debris and apoptotic bodies (−▸; a–d). −apoptotic body;chromatin margination;mast cell;− degranulated mast cells;−▸phagocytic active macrophage;inline imagenecrotic cell; E eosinophil; L lymphocyte.

Table 1. Histomorphometric measurements during I/R; mean ± s.e.; Mann-Whitney-U test; P<0.05 (different letters represent significant differences between conditions)
 Control1hI1hI + 1hR1hI + 2hR1hI + 4hRKruskal-Wallis
Epithelial height (µm)27.9 ± 1.625.6 ± 2.023.3 ± 1.824.7 ± 1.423.9 ± 1.40.363
Epithelial width (µm)3.8 ± 0.1a5.5 ± 0.3b5.1 ± 0.6abc4.5 ± 0.2c4.6 ± 0.3c0.007
Nuclear height (µm)6.5 ± 0.46.8 ± 0.36.5 ± 0.36.8 ± 0.26.2 ± 0.20.615
Nuclear width (µm)2.7 ± 0.2a4.2 ± 0.2b3.4 ± 0.3abc3.4 ± 0.2c3.3 ± 0.2c0.003
Intercellular space (µm)0.9 ± 0.1ab0.6 ± 0.2ac1.6 ± 0.6ab1.5 ± 0.2b1.6 ± 0.2b0.027
Subepithelial edema (µm)0.2 ± 0.1a3.2 ± 0.2b2.2 ± 0.8bc1.2 ± 0.4c1.1 ± 0.5c0.002
Apoptotic cells/mm mucosa14.3 ± 4.1a60.4 ± 14.0b80.1 ± 11.4b78.4 ± 23.8b83.7 ± 23.7b0.010
Necrotic cells/mm mucosa2.6 ± 0.834.3 ± 7.330.2 ± 13.912.8 ± 5.720.2 ± 11.00.066
Denuded epithelium (%)0 ± 02.3 ± 1.42.2 ± 2.20 ± 00.7 ± 0.50.172
Repaired epithelium (%)0 ± 0a0 ± 8a29.3 ± 8.5b10.1 ± 3.6ac19.9 ± 7.1bc0.001
Figure 3.

Characteristic cellular features of the colonic epithelium and subepithelium after 1 h of ischaemia and 2 h of reperfusion (toluidine blue; x1000, oil immersion). After 2 h of reperfusion, epithelial cell swelling declined and epithelial cells appeared shorter resulting in the enlargement of paracellular spaces and subepithelial cleft formation (b–d). Neutrophils, eosinophils and lymphocytes infiltrated subepithelial clefts and moved into the intestinal lumen (a–d). Small epithelial defects were sealed by adjacent neighbouring cells ({; a), or they were covered by a continuous layer of shortened detached epithelial cells ({; c) or membrane extensions between neighbouring cells (; b). Chromatin margination was only seen in single epithelial cells (; b), and more apoptotic bodies were present in the epithelium (−; a–d). The number and distribution of subepithelial mast cell granules appeared normal (▸; a–d), but subepithelial macrophages were phagocytic active (−▸; d). −apoptotic body;chromatin margination;mast cell;−▸phagocytic active macrophage;inline imagenecrotic cell;coverage of subepithelial clefts and increased paracellular spaces by membrane extensions from neighbouring cells; { apical closure of epithelial defects by detached epithelial cells or intact neighbouring cells; E eosinophil; L lymphocyte; N neutrophil.

Figure 4.

Characteristic cellular features of the colonic epithelium and subepithelium after 1 h of ischaemia and 4 h of reperfusion (toluidine blue; x1000, oil immersion). After 4 h of reperfusion, detached epithelial cells remained adhered to each other at the luminal surface by membrane extensions preserving coverage of large paracellular spaces and subepithelial clefts (; a, b). Apoptotic epithelial cells characterised by formation of apoptotic bodies (−▹; c), and single necrotic cells appeared to be replaced by migration of neighbouring cells after extrusion into the intestinal lumen (inline image; c), and further covering of epithelial defects by re-epithelialisation proceeded ({; c, d). More neutrophils, eosinophils and lymphocytes infiltrated the injured epithelium (a–c). Cell debris and apoptotic bodies were phagocytosed by subepithelial macrophages (−▸; a, b, d). −apoptotic body;chromatin margination;mast cell;−▸phagocytic active macrophage;inline imagenecrotic cell;coverage of subepithelial clefts and increased paracellular spaces by membrane extensions from neighbouring cells; { apical closure of epithelial defects by detached epithelial cells and intact neighbouring cells; E eosinophil; L lymphocyte; N neutrophil.

Neutrophils, eosinophils, mast cells, macrophages and COX-1/-2 expression

After ischaemia, calprotectin-positive neutrophils accumulated in submucosal venules with further progession after 1hR (Table 2). Neutrophils migrated into the lamina propria and moved towards the epithelium at 1hR, and they moved into the intestinal lumen at 2hR (Table 2, Fig 1c). Mucosal mast cells and macrophages appeared activated during I/R, but their numbers remained unchanged (Table 2, Figs 2–5). With light microscopic examination of toluidine blue-stained semithin sections, subepithelial mast cells were evident and they contained a decreased number of cytoplasmic granules that were located in close proximity to the cellular membrane after ischaemia (Figs 2a–c) compared to reperfused tissues (Figs 3a–d, 4a–c). Such changes could be interpreted as degranulation and release of proinflammatory molecules into the interstitium. In addition, subepithelial macrophages displayed increased phagocytic activity during I/R, characterised by large phagocytic vacuoles in the cytoplasm containing debris and apoptotic cells (Figs 2b–d, 3d, 4a, c, d). Significantly more COX-2 was expressed by epithelial cells, and by lamina propria immune cells, especially lymphocytes, eosinophils and neutrophils, after ischaemia and after 1hR compared with controls, 2hR and 4hR (Table 3; Fig 6).

Table 2. Calprotectin-positive neutrophils, and eosinophils per mm2 mucosal and submucosal venule (neutrophils only) area, and within mucosal zones M1-M5 during I/R; mean ± s.e.; Mann-Whitney U test; P<0.05 (different letters represent significant differences between conditions)
 Control1hI1hI + 1hR1hI + 2hR1hI + 4hRKruskal-Wallis
Calprotectin-pos. cells/mm2 mucosa72.3 ± 18.4a215.0 ± 73.4ab484.1 ± 177.6bc1149.7 ± 220.6c857.0 ± 179.4c0.000
Calprotectin-pos. cells/mm2 M112.8 ± 3.7a58.7 ± 19.6b134.8 ± 57.3bc359.8 ± 83.9c196.5 ± 47.0c0.000
Calprotectin-pos. cells/mm2 M230.3 ± 8.8a88.4 ± 39.39ab179.1 ± 85.4bc303.0 ± 49.0c211.9 ± 58.6c0.002
Calprotectin-pos. cells/mm2 M322.1 ± 7.8a45.5 ± 16.5ab94.4 ± 27.6bc220.8 ± 48.8d144.4 ± 27.3cd0.001
Calprotectin-pos. cells/mm2 M44.6 ± 1.4a16.9 ± 7.1ab51.5 ± 19.2bc143.6 ± 41.8c72.9 ± 11.6c0.000
Calprotectin-pos. cells/mm2 M52.6 ± 1.7a5.4 ± 2.0ab24.4 ± 8.7b122.7 ± 26.5c231.3 ± 78.3c0.000
Calprotectin-pos. cells/mm2 venule469.7 ± 193.2a1209.9 ± 377.6b4068.0 ± 644.0c3861.0 ± 779.4cd1757.8 ± 384.0bd0.001
Eosinophils/mm2 mucosa709.9 ± 132.0530.5 ± 111.9652.5 ± 166.2684.8 ± 145.6581.2 ± 119.60.731
Eosinophils/mm2 M1355.1 ± 66.5279.1 ± 63.9284.5 ± 68.2269.2 ± 71.2295.3 ± 69.60.880
Eosinophils/mm2 M2227.1 ± 39.2169.2 ± 35.3223.7 ± 55.2244.4 ± 71.6202.0 ± 42.40.737
Eosinophils/mm2 M3110.6 ± 28.865.8 ± 20.6100.4 ± 39.0109.1 ± 38.668.2 ± 17.80.759
Eosinophils/mm2 M415.8 ± 4.814.3 ± 5.732.6 ± 15.436.5 ± 10.59.0 ± 2.40.255
Eosinophils/mm2 M51.3 ± 0.5a2.2 ± 1.1ab11.2 ± 4.5c25.6 ± 10.3c6.6 ± 2.3bc0.011
Figure 5.

Number of macrophages (top left) and mast cells (bottom left)/mm2 lower lamina propria (LLP) and upper lamina propria (ULP) during I/R; mean; Kruskal-Wallis; not significant. Top right: CD163-positive subepithelial macrophages (large brown cells; 10 µg/ml mouse anti-human macrophage surface antibody; x400); Bottom right: subepithelial mast cells (cells with numerous red-purple granules; toluidine blue; x400).

Table 3. Scores for COX-positive cells in the mucosa, epithelium, upper lamina propria (ULP), lower lamina propria (LLP), and crypts during I/R; mean ± s.e.; Mann-Whitney U test; P<0.05 (different letters represent significant differences between conditions)
 Control1hI1hI + 1hR1hI + 2hR1hI + 4hRKruskal-Wallis
COX-1 mucosa (0–12)3.8 ± 0.44.9 ± 0.65.4 ± 0.74.2 ± 0.73.5 ± 0.60.259
COX-2 mucosa (0–12)5.4 ± 0.6a8.2 ± 0.6b7.1 ± 1.1ab4.7 ± 0.4a5.6 ± 0.5a0.018
COX-1 epithelium (0–3)0.3 ± 0.11.0 ± 0.21.0 ± 0.20.7 ± 0.30.7 ± 0.20.127
COX-1 ULP (0–3)1.1 ± 0.11.3 ± 0.21.5 ± 0.21.3 ± 0.11.1 ± 0.20.485
COX-1 LLP (0–3)1.6 ± 0.21.8 ± 0.21.8 ± 0.21.3 ± 0.21.2 ± 0.10.067
COX-1 crypt (0–3)0.8 ± 0.20.8 ± 0.21.1 ± 0.20.8 ± 0.20.6 ± 0.20.565
COX-2 epithelium (0–3)0.8 ± 0.2a2.0 ± 0.1b1.7 ± 0.3b0.8 ± 0.2a1.5 ± 0.2b0.002
COX-2 ULP (0–3)1.6 ± 0.22.0 ± 0.11.9 ± 0.31.4 ± 0.21.6 ± 0.10.169
COX-2 LLP (0–3)1.7 ± 0.2a2.2 ± 0.2ab1.8 ± 0.2ac1.2 ± 0.1c1.3 ± 0.2ac0.014
COX-2 crypt (0–3)1.4 ± 0.22.0 ± 0.31.7 ± 0.41.3 ± 1.11.2 ± 0.10.307
Figure 6.

COX-expression in the colonic epithelium and upper lamina propria during I/R (positive cells appear brown); (a–c) COX-1 (polyclonal goat anti-human antibody, x400); (d–f) COX-2 (polyclonal goat anti-human antibody, ×400); (a, d) control; (b, e) 1 h of ischaemia; (c, f) 1 h of ischaemia and 4 h of reperfusion. COX-1 expression was mainly evident in lamina propria cells in control tissues, and did not change during I/R. COX-2 expression increased during ischaemia, especially in the epithelium and lamina propria, but declined after reperfusion.


Although ischaemia caused mild mucosal injury in the present study, the injury to mucosal barrier function was sufficient to decrease transepithelial resistance (TER) and increase transmucosal mannitol flux in the same colonic tissues (Graham et al. 2011). Ischaemia increased apoptosis and COX-2 expression, a possible response of the colonic epithelium to prevent further damage and initiate early recovery after reoxygenation (Savill et al. 2002; Karrasch et al. 2006). In the present study, reperfusion did not exacerbate epithelial damage, consistent with the findings of Moore et al. (1994a), using a different I/R model and time schedule. However, differences between ischaemia and reperfusion were seen in other histomorphometric criteria not evaluated in the present study, such as the depth of mucosal loss and mucosal cellular debris index (Moore et al. 1994a). Although morphology of the epithelium did not return completely to normal at the end of reperfusion in our study, 4hR was sufficient to initiate epithelial repair. In this rapid repair process, defects created by loss of individual cells were covered by apical membrane extensions from intact neighbouring cells (Figs 4a,b). This morphological expression of repair could explain recovery of TER from 69.6 Ω·cm2 after ischaemia to 127.9 Ω·cm2 after 4hR in the same mucosal tissues mounted in Ussing chambers (Graham et al. 2011).

Studies have identified apoptosis as a major cause of cell death after intestinal I/R in equine small and large intestine (Rowe et al. 2003; Grosche et al. 2010). However, it is still unclear whether apoptosis contributes to epithelial injury or resolves inflammation and hastens epithelial repair (Ramachandran et al. 2000). In the present study, semithin sections embedded in epon and stained with toluidine blue clearly demonstrated nuclear chromatin condensation and margination, characteristics of apoptotic cells that we previously identified with the TUNEL method after I/R in equine colon (Grosche et al. 2010). Chromatin condensation and margination, and formation of apoptotic bodies, were increased by ischaemia and persisted during reperfusion in the present study. Apoptosis could be a mechanism that protects the tissue from harmful exposure to inflammatory and immunogenic cell contents during reperfusion (Maderna and Godson 2003).

Prostaglandins, synthesised by COX enzymes, play a key role in regulating inflammatory reactions, and COX-2 is known to be induced rapidly in sites of inflammation in the colon of horses (Matyjaszek et al. 2009; Morton et al. 2009). Although COX enzymes regulate the production of potent proinflammatory prostaglandins (Crofford 2001), evidence is growing that COX-2 expression may contribute to resolution of gastrointestinal inflammation and might be crucial in regulating mucosal healing (Blikslager et al. 1999; Wallace and Devchand 2005). COX-2 expression in the epithelium and by lamina propria immune cells was significantly upregulated after ischaemia and after 1hR in the present study, a time point where epithelial repair began. Shifflett et al. (2004) found enhanced recovery of barrier function in porcine ischaemia-induced ileal mucosa mediated by upregulation of COX-2 and activation of neutrophils. Thus, expression of COX-2 could be involved in controlling mucosal damage and recovery of epithelial function after I/R.

Studies have shown that neutrophils are critical elements in the cascade of intestinal I/R injury and barrier dysfunction (Gayle et al. 2002; Blikslager et al. 2007), and neutrophils are involved in the inflammatory response after colonic I/R in horses (Grosche et al. 2008). Blikslager et al. (1997) found massive infiltration of neutrophils during initial stages of epithelial repair in porcine intestinal ischaemia. They hypothesised that mucosal injury is more likely to be triggered by physical damage to the repairing epithelium by migrating neutrophils at this stage. In the present study, the peak neutrophil infiltration was seen after 2hR, when epithelial repair had started. We cannot rule out possible short-term damage to the epithelium by migrating neutrophils in the present study, although mucosal TER recovered fully after 4hR (Graham et al. 2011), despite the intense influx of neutrophils. Thus, the overall effect of neutrophils after I/R could be beneficial for tissue repair after colonic I/R if the severity of ischaemic damage is mild (Serhan and Savill 2005; Nathan 2006).

Although the number of mucosal eosinophils, mast cells and macrophages remained unchanged in the present study, their role as potential effector immune cells during intestinal I/R has been demonstrated in many studies (Kanwar and Kubes 1994; Boros et al. 1999; Chen et al. 2004; Furuta et al. 2005). Intestinal mast cells are thought to contribute to mucosal permeability alterations during reperfusion in canine small intestine, but might play only a minor role in I/R-induced structural changes (Szabo et al. 1997). In contrast, Boros et al. (1999) found that mast cell degranulation can initiate tissue injury after I/R and release of histamine contributes largely to the severity of mucosal damage. Eosinophilic granulocytes can be seen in close proximity to mucosal mast cells, and degranulation of mast cells is thought to be triggered by eosinophilic toxic proteins (Piliponsky et al. 1999). Although mast cell granule release was observed during I/R in the present study, a general effect of eosinophils and mast cells on epithelial injury and barrier dysfunction after colonic I/R could not be identified. In addition, resident macrophages are well-established effector cells with pro- and anti-inflammatory activities that could also contribute to the inflammatory response and, therefore, affect mucosal injury. Activated macrophages inhibit formation of enterocyte gap junctions in vitro (Anand et al. 2008), but influence early mucosal damage during intestinal I/R in rats by expression of myeloperoxidase, Egr-1 gene and proinflammatory cytokines before neutrophil infiltration occurs (Chen et al. 2004). Macrophages are also crucial for recognition and clearance of necrotic debris and apoptotic neutrophils, an essential step in resolving inflammation (Serhan and Savill 2005). Histological evaluation of the tissues in the present study demonstrated an increased phagocytic activity of resident subepithelial macrophages, suggesting a possible role during colonic I/R in horses.

In conclusion, mild ischaemic injury in the equine colonic mucosa was accompanied by increased apoptosis and epithelial COX-2 expression, which could facilitate early epithelial repair after reoxygenation. Epithelial repair characterised by regeneration and re-epithelialisation was associated with influx of neutrophils to the site of injury during intestinal I/R. Additionally, resident mast cells and macrophages may become activated in response to I/R, but their exact role in colonic I/R in horses requires further study. Despite the intense inflammation observed during reperfusion after 1hI in the present study, the equine colonic mucosa did not incur additional injury characteristic of I/R, but actually recovered according to morphologic and functional measures of repair.

Authors' declarations of interest

No conflicts of interest have been declared.

Source of funding

The study was funded in part by the American College of Veterinary Surgeons and the DeedieWrigley-Hancock Fellowship.

Manufacturers' addresses

1 Serotec, Raleigh, NC, USA.

2 TransGenic Inc., Kumamoto, Japan.

3 Santa Cruz Biotech., Santa Cruz, CA, USA.

4 Media Cybernetics, Bethesda, MD, USA.

5 Vector Laboratories, Burlingame, CA, USA.