Equine Intestinal Mucosal Pathobiology

Abstract

The equine intestinal mucosa is intimately involved in maintaining homeostasis both on a systemic level by controlling extracellular fluid movement and at the local level to maintain barrier function. Horses are particularly susceptible to the clinical syndrome of colic, with the most severe cases involving strangulating obstruction that induces ischemia. Because of the mucosal vascular architecture, the mucosal epithelium is particularly susceptible to ischemic injury. The potential for reperfusion injury has been investigated and found to play a minimal role. However, inflammation does affect mucosal repair. Mechanisms of repair, including villus contraction, epithelial restitution, and tight junction closure, are critical to reforming the mucosal barrier. Nonsteroidal anti-inflammatory drugs have an impact on this repair, particularly at the level of the tight junctions. Completion of mucosal regeneration requires proliferation, which is now being actively studied in equine enteroids. All of these aspects of equine mucosal pathobiology are reviewed in depth.

PHYSIOLOGICALLY CRITICAL FEATURES OF THE EQUINE DIGESTIVE TRACT

Horses use the digestive strategy of hindgut fermentation, during which ingested forage is rapidly passed from the stomach along the small intestine (in excess of 20 m in the average 500-kg adult horse, with a transit time of approximately 120–180 min) and into the cecum and large colon (1). This may be less efficient than the digestive strategy of foregut fermentation, noted in ruminants, where forage that has been partially digested in the forestomachs, including the rumen, can then enter the small intestine for efficient absorption of nutrients (2). However, selection pressure for speed in horses to run from predators (top speed of approximately 70 km/h) likely resulted in development of a hindgut, where a large gut chamber would not impede running. The equine hindgut is voluminous and folded to accommodate its length within the abdomen into a double-horseshoe configuration, with a ventral colon (the lower horseshoe) primarily functioning as the fermentation chamber and the dorsal colon (the upper horseshoe) taking on a predominant role in water absorption (Figure 1) (3). The ventral and dorsal colons of the horse are sometimes referred to as the ascending colon because of terminology used for the human colon, but anatomically this is problematic because the equine proximal colon both ascends and descends (4). The pelvic flexure, which connects the ventral and dorsal colons, includes a hairpin turn from the larger-diameter ventral colon to the smaller-diameter dorsal colon. Suboptimal management of captive horses, such as feeding of poor-quality forage, may lead to ingesta obstructions (impactions), which are most commonly found at the pelvic flexure (5). Impactions of indigestible luminal contents, most notably sand, tend to obstruct at a more distal flexure present between the right dorsal colon and the transverse colon (6). This is also the site where enteroliths, typically composed of struvite surrounding a small central mineral nidus (7), may develop and obstruct the flow of ingesta.

Figure 1.

Enterosystemic cycle of fluid in a 100-kg pony. Proximal secretions approximate the animal’s entire extracellular fluid volume during a 24-h period. In animals fed twice daily (meal fed), periods of net secretion alternate with net absorption (L) in the large colon, as shown by the sets of opposite pointing arrows. Adapted from Reference 1.

The rapid passage of digesta along the small intestinal tract requires relatively large volumes of fluid, which are primarily secreted by the salivary glands, pancreas, and small intestinal mucosa (1). The relative fluid volumes entering the equine digestive tract have been studied (Figure 1) and show that horses secrete approximately one extracellular volume of fluid into the digestive tract on a daily basis (1, 3). Clarke and colleagues (3, 8–11) have studied in depth the control of NaCl absorption and associated water movement across the equine colon at both the local and systemic levels. The endocrine-gut axis is critical to maintaining homeostasis considering the massive shift of fluid into and back out of the colon that occurs during the feeding of a meal, and this is accomplished by activating the renin-angiotensin-aldosterone system interacting with select colonic gut transporters (8, 9, 11). This system is maximally activated during feeding of meals (Figure 1), as might occur during intensive management of horses (rather than multiple meals or continuous grazing) (3). In response to a single meal comprising the full ration for a horse in the form of hay-grain pellets, plasma renin activity was increased by 0.5 h, followed by elevations in plasma aldosterone by 3 h. The renin release was attributed to transient hypovolemia, as suggested by increases in plasma protein and packed cell volume (10). This systemic hypovolemia is a consequence of fluid moving into the digestive tract. These investigators went on to show that an increase in aldosterone in the horse resulted in a doubling of Na⁺ absorption in the proximal colon (ventral and dorsal colons) and a tripling of Na⁺ absorption in the distal colon as a mechanism to retrieve the large volumes of fluid that are secreted into the digestive tract (11). From a clinical perspective, meal feeding places horses at risk of conditions such as colonic distension, which can lead to colonic displacement or volvulus. The latter is the most fatal form of colic, requiring rapid surgical intervention to increase the likelihood of survival (12).

THE EQUINE MICROBIOME

Considering the fact that horses are herbivores, and have an extensively developed hindgut for the purposes of fiber digestion, there has been recent interest in the microbiome and how it might be changed under differing management circumstances. For example, mares are at risk of colic in the peripartum period, possibly related to factors such as changes in feeding of lactating mares (13). Numerous differences that preceded colic have been described in the fecal microbiota of postpartum mares. In particular, one study found associations between changes in Firmicutes and Proteobacteria, most notably Lachnospiraceae and Ruminococcaceae, with the development of colic, and the authors suggested that select changes in the fecal microbiota could possibly be used to predict higher risk of colic (14). Although studies on equine microbiota have often focused on fecal microbiota, they are not reliably indicative of colonic microbiota (15). Nonetheless, another study evaluating the effect of concentrate diet on colonic microbiota collected at the time of slaughter has shown some similar results to the aforementioned fecal microbiota study. In particular, a feeding change to dietary concentrate resulted in a progressive and significant increase in Lachnospiraceae, as had been seen in the fecal microbiota of horses fed a concentrate diet (16). Another study showed that horses fed concentrate diet versus a grass-only diet had an increase in lactic acid-producing microbiota (the Bacillus-Lactobacillus-Streptococcus group) in their colonic contents (obtained following euthanasia), resulting in higher concentrations of lactic acid. Interestingly, the shift to lactic acid-producing microbiota was also associated with horses that developed colonic distension or impaction (16). Despite changes that appear to be selective enough to be used to detect horses at risk of colic, a great deal of variability in individual horse microbiota even when managed under similar circumstances has been shown (15), suggesting caution should be used when interpreting microbiome-related risk factors in individual or small groups of horses.

THE EQUINE STEM CELL NICHE

The interaction between intestinal mucosa and microbiota has become an area of great interest in human health-related research (17), but intense study of the makeup of the mucosa has been undertaken as a precursor to fully understanding host interactions with luminal contents. The luminal surface of the mucosa is composed of columnar epithelial cells that have the conflicting functions of water and nutrient transport, while simultaneously guarding against invasion of noxious luminal bacteria and bacterial toxins (18). Down the entire length of the small and large intestine, the epithelia are arranged into invaginations called crypts of Lieberkühn. The small intestine is further composed of fingerlike projections, called villi, which extend into the intestinal lumen, thereby increasing the absorptive surface area. This anatomic arrangement is referred to as the crypt-villus axis. At the base of each crypt are undifferentiated stem cells interspersed between Paneth cells. Immediately adjacent to these cells are partially differentiated progenitor cells, and collectively, this region of crypts is termed the stem cell niche (19).

To maintain a continuous epithelial barrier in the face of constant threat from luminal contents, the cells remain in a dynamic and rapid state of cellular turnover that continues throughout life. This requires a tightly regulated balance of cellular loss and cellular renewal and results in the creation of a largely new intestinal lining approximately every five to seven days. This capacity for self-renewal is attributed to the adult or somatic stem cells that reside within the intestinal mucosal lining deep within the base of each crypt. Intestinal epithelial stem cells were first described in the scientific literature by transmission electron microscopy in 1974 (20). Currently two populations of stem cells are thought to exist: fast-cycling crypt base columnar stem cells and slower-cycling, reserve quiescent stem cells (21, 22). Despite evidence that distinguishes between these two cell types, cells expressing biomarkers attributed to either fast- or slow-cycling stem cells clearly have the capacity to differentiate into all four postmitotic, mature epithelial cell types (22). These cell lineages include absorptive enterocytes, goblet cells, enteroendocrine cells, and Paneth cells. The latter do not ascend the crypt villus axis but remain parked near the crypt base and are thought to have a critical role in supporting crypt-based columnar cells (CBC, fast-cycling stem cells) via release of factors involved in CBC homeostasis, such as Wnt (23). In addition, Paneth cells release antibacterial peptides called defensins (24). In a study focused on identification of epithelial cell lineages in horses, all intestinal epithelial cell types were identified by means of immunostaining and morphological characterization with transmission electron microscopy (19).

Intestinal epithelial stem cells are distinct from mesenchymal stem cells commonly used in equine orthopedic research and therapy owing to a restricted capacity to differentiate. Although they are undifferentiated and self-renew (a prerequisite of all stem cells), they do not give rise to multiple tissue types, as is the case for mesenchymal stem cells, which may turn into bone, cartilage, connective tissue, and fat cells. Instead, intestinal epithelial stem cells differentiate only into those of intestinal epithelial lineage. Many recent and exciting advances in the field of intestinal stem cell biology have enabled the detailed study of the stem cell niche (Figure 2) as a potential source of novel therapeutic targets to enhance intestinal mucosal regeneration (25, 26). Stem cells are a renewable source of mature epithelial cells in the gut, and understanding their function is critical to developing clinical applications that may improve outcome in cases of severe mucosal injury. In addition, recent work on equine colonic tissues has focused on the level of proliferative cells that may be required to regenerate tissue following large colon volvulus (27). As-yet-unpublished data show that reduced numbers of stem cells expressing phosphohistone 3, a marker of the G2-M phase of the cell cycle (28), predict case fatality when counted in biopsies from horses with naturally occurring large colon volvulus.

Figure 2.

Appearance of the stem cell niche in equine small intestine and colon. Cells undergoing proliferation can be identified using several markers, including proliferating cell nuclear antigen (PCNA) and phosphohistone 3 (PH3). The latter marks cells in the G2-M phase of the cell cycle, an important marker of colonic viability in the horse. Note pink staining in each image, indicative of co-localization of blue staining of nuclei with bisbenzamide and the respective stem cell marker. Red and yellow staining within the lamina propria is attributable to autofluorescence. Adapted with permission from a figure originally published in Reference 19.

COLIC AND ISCHEMIC INJURY IN HORSES

According to a US national survey conducted in 2001, approximately 4% of horses have an episode of colic annually, with a case fatality rate of 11% (29). Strangulating obstruction, a disease process in which simultaneous occlusion of the intestinal lumen and vasculature results in ischemic mucosal injury, is arguably the principal cause of colic-associated deaths in horses. For example, a study conducted in the United Kingdom in 1992 showed that 14 of 200 (7%) horses evaluated for colic over 2 years in a general equine practice died, and 12 of those 14 deaths were associated with intestinal strangulation (30). A more recent set of studies by the same authors has confirmed that horses with strangulating lesions had consistently higher mortality rates, a finding that was reflected by the fact that resection length and length of surgery had considerable prognostic value. Furthermore, a pattern of mortality was identified in which a high fatality rate was noted within the first few days after surgery, with continued mortality at lower rates after discharge from the hospital (31, 32). Horses with intestinal diseases other than strangulating obstruction may also suffer from ischemic disease. For example, horses suffering from simple obstruction ultimately succumb to ischemic necrosis if not attended to medically and sometimes surgically as increasing intraluminal pressure progressively occludes the circulation within the intestinal wall (33, 34). In addition, horses may suffer from nonstrangulating infarction, which is most commonly associated with parasite migration (35). However, this disease process has become relatively rare with the advent of highly effective broad-spectrum anthelmintics.

The degree of mucosal injury that is noted in horses with ischemic disease of the intestine is highly variable, depending on the type and duration of intestinal obstruction (36, 37). Although most equine research to date has focused on either low-flow ischemia, during which intestinal blood flow is uniformly reduced to a static level (typically 20% of baseline flow) (38–40), or complete ischemia, in which the intestinal vasculature is completely occluded (41, 42), most natural cases of strangulating obstruction fit into neither of these categories. Instead, horses with intestinal strangulating obstruction are most likely to have a disparity in blood flow in which the venous circulation is occluded before the intestinal arterial blood supply because of the differences in vessel wall thickness and compliance. This results in a hemorrhagic lesion in which the arteries continue to supply blood to the tissues for a variable length of time before they ultimately collapse as a result of increases in tissue interstitial pressure (43). This explains the appearance of tissues commonly encountered during strangulating obstruction, including significant thickening of the tissue and marked discoloration, as a result of tissue filling with deoxygenated blood. Less commonly, strangulation may be so rapid in onset and so tightly twisted that the veins and arteries are occluded simultaneously. This type of lesion, known as ischemic strangulating obstruction, may be encountered in lesions like large colon volvulus, particularly those of rapid onset exceeding 360° of rotation (43). This lesion is pale, with little tissue edema, because of the complete lack of blood flow. The appearance of these distinct lesions can be deceiving when it comes to judging the degree of ischemic mucosal injury at surgery. The level of mucosal injury is a critical clinical judgement, because it is highly likely that the integrity of the mucosa largely dictates whether or not a horse can survive with a particular lesion (44). During the early stages of hemorrhagic strangulating obstruction, the degree of mucosal injury is reduced as compared with ischemic strangulating obstruction lesions despite the marked abnormal coloration of these lesions, most likely because of continued provision of oxygenated blood from an arterial blood supply that is initially patent during the early phases of obstruction. This has been modeled in equine intestine by occluding the veins (45), and in porcine intestine by 360° twisting of the jejunum and partial ligation of the twisted loop with a Penrose drain to create a segmental volvulus (46). In the case of porcine volvulus, the level of injury was time-dependently less severe than complete occlusion of the bowel and vasculature, based on histological findings. Conversely, ischemic strangulating lesions do not appear severe; they may readily become pink during reperfusion but may nonetheless have severe mucosal injury because of an absolute cessation of arterial blood flow during the ischemic phase.

The ultimate diagnostic test to ascertain the degree of injury is mucosal histology, which can be accomplished during surgery with frozen sections (47) but is not possible in most hospitals because of the emergency nature of colic cases. In the absence of such a definitive test, caution is suggested at surgery in suspected cases of ischemic strangulating obstruction in favor of resecting intestine when there is concern that the mucosa may be severely compromised. Conversely, in horses with hemorrhagic strangulating obstruction in which resection is not possible, such as obstructions involving the terminal ileum, surgeons should realize that the ischemic injury may appear more severe than it actually is and that many of these horses can be saved with the intestine left in situ. This has been highlighted recently in clinical studies, where horses with strangulated small intestine that might have typically been resected have recovered well without resection (48). The judgment of whether to resect intestine in cases of hemorrhagic strangulating obstruction has recently been based on a grading system, which reduces subjectivity (48). In addition, in a study of biopsies from horses with large colon volvulus, histological analyses showed that the most predictive tissue index linked to case fatality was hemorrhage score (Figure 3), with a hemorrhage score of ≥3 placing horses in excess of eightfold at risk of nonsurvival (44). This may become clinically useful in the future in a methodology to quantify or semiquantify the level of hemorrhage within the tissue at surgery.

Figure 3.

Hemorrhage score within the lamina propria is assigned a score from 0 to 4: 0, no hemorrhage; 1, few individual red blood cells (RBCs) within the lamina propria; 2, increased number of individual RBCs; 3, appearance of clumps of RBCs; and 4, no clear demarcation between RBCs, obscuring the lamina propria. Adapted from a figure originally published in Reference 44.

MECHANISMS OF ISCHEMIC MUCOSAL INJURY

A reduction in blood flow during ischemia results in cellular necrosis as tissues are starved of oxygen (49). Mechanisms of intestinal mucosal injury are more complex than this, however. In particular, the design of the mucosal vascular supply in the small intestine, which is species specific, may predispose distinct regions to greater degrees of injury during ischemia (50). The surface area of the small intestinal mucosa is amplified by fingerlike projections called villi; in the horse, these villi have a central arteriole that arborizes near the tip of the villus, draining into peripheral venules (51). This vascular architecture sets up a countercurrent exchange mechanism in which oxygen diffuses from the arteriole into the tissue and adjacent peripheral venules, resulting in a relatively hypoxic villus tip. When arterial blood flow is reduced, for example, during the early phases of strangulating obstruction, the villus tip becomes progressively more hypoxic, resulting in epithelial injury (50). Therefore, ischemic mucosal injury results in time-dependent sloughing of epithelium from the tip of the villus toward the crypts (52). Although there are no villi in the large intestine, surface epithelium is lost initially during ischemia, followed by crypt epithelial injury as ischemia progresses (42, 53). A great deal of recent work has extensively characterized experimental ischemic strangulating obstruction in the colon (53–55), which has revealed evidence of epithelial swelling and detachment from the basement membrane during the early stages of ischemia (within 1 h). Epithelial cells also had ultrastructural evidence of organelle swelling, and some were clearly undergoing either apoptosis or autophagy. Interestingly, there was also extensive evidence of dilatation of the tight junctions (53), the most critical component of the epithelial barrier, prior to complete epithelial sloughing (56).

REPERFUSION INJURY

The concept of reperfusion injury centers on experimental findings in which mucosal injury is worsened by reperfusing an ischemic lesion. This concept seems counterintuitive given the fact that ischemic mucosal injury results largely from a lack of oxygenated blood supply. Nevertheless, reperfused lesions rapidly become infiltrated with neutrophils, and neutrophils resident within the mucosa become activated in rodent models of ischemia/reperfusion injury, which injure tissues as a result of release of reactive oxygen metabolites and proteases (57). The reason neutrophils seem to infiltrate or become activated in postischemic mucosa is the swift generation of chemoattractants in response to oxidants released by tissue enzymes, most notably, xanthine oxidase (50). This enzyme is typically present in its dehydrogenase form but becomes converted to xanthine oxidase during ischemia by tissue proteases (58). As the tissue is reperfused, xanthine oxidase begins to metabolize hypoxanthine that has accumulated as a result of end-stage metabolism of adenosine triphosphate, using oxygen as an electron acceptor (50). This results in superoxide formation. Although superoxide is not in itself particularly damaging because it is not highly lipid soluble, it triggers the formation of chemoattractants, such as leukotriene B₄ and platelet-activating factor (59). There are also other possible sources of oxidant-producing enzymes in intestinal tissues that may initiate injury. For instance, in the large colon of the horse, although there are low concentrations of xanthine dehydrogenase/xanthine oxidase, other oxidant-generating enzymes, such as aldehyde oxidase, may be present (57). Oxidant-induced release of chemoattractants results in rapid aggregation of neutrophils within the postischemic lesion. Neutrophils are the final common mediator of intestinal mucosal reperfusion injury, which is why inhibition of neutrophil infiltration with antibodies directed toward neutrophil adhesion molecules markedly attenuates reperfusion injury in rodent models (50).

Horses differ from rodents in several important ways. Firstly, levels of xanthine dehydrogenase/xanthine oxidase in the small intestine are approximately 10-fold lower than those of rodents, and there are also progressive reductions in mucosal xanthine oxidase levels from the duodenum to the ileum (58, 60). In addition, horses have fewer mucosal neutrophils than rodents (Table 1) (60). This is important because resident neutrophils that are already present in the mucosal lamina propria are responsible for most of the reperfusion injury noted in rodents (61). Finally, most of the literature on reperfusion injury uses a low-flow model, rather than a strangulating obstruction model. This appears to maximally prime tissues for reperfusion injury by triggering conversion of xanthine dehydrogenase to xanthine oxidase, with minimal damage to the tissues. However, most injury in strangulating obstruction occurs during the ischemic phase (62).

Table 1.

Comparative levels of oxidant enzymes in intestinal tissues^a

Species	Intestinal segment	Total XO/XDH (mU/g tissue)	Myeloperoxidase (U/g tissue)
Cat (adult)	Jejunum	80	12
	Ileum	-	-
Rat (adult)	Jejunum	405–523	1.9
	Ileum	150	-
Pig (6–8 weeks)	Jejunum	3.4 (0)b	-
	Ileum	0.4 (0.9)b	2.2
Horse (adult)	Jejunum	100–131 (60)b	0.02
	Ileum	30–48 (0)b	0.1
Human (adult)	Jejunum	29–56 (0)b	-

^aLevels of xanthine oxidase/xanthine dehydrogenase (XO/XDH), the enzyme system thought to be responsible for triggering reperfusion injury, and myeloperoxidase as a surrogate marker for neutrophils, the immune cells responsible for amplifying reperfusion injury.

^bNeonatal levels.

It is important to note that there are probably other mechanisms of reperfusion injury aside from injury induced by reactive oxygen metabolites and neutrophils. For example, one study in horses showed marked mucosal metabolic abnormalities induced by 3 h of ischemia, including mitochondrial dysfunction, which persisted during reperfusion because of the inability of tissues to regain normal function (49). However, in another equine study, ultrastructural evidence of ischemic epithelial injury in the colon was rapidly followed by evidence of epithelial repair during reperfusion, although the period of ischemia was brief (1 h) (53). Another population of inflammatory cells has received extensive attention in horses: eosinophils. Under normal circumstances, eosinophils are predominantly located beneath the muscularis mucosa (63). However, during ischemia/reperfusion, eosinophils have been shown to migrate from the intestinal lumen through the mucosa to the luminal surface, though this does not happen after ischemia alone in experimental studies (64). Eosinophils have also been shown to accumulate in the mucosa in horses with naturally occurring strangulating obstruction (64). In experimental trials in horses using a large colon model of colonic strangulating obstruction, eosinophil production of nitrotyrosine, which contributes to reactive oxygen metabolite-induced injury, was associated with colon epithelial apoptosis during the recovery period. This suggests that eosinophils also contribute to disruption of repairing intestinal epithelium (65).

MECHANISMS OF MUCOSAL REPAIR

Acute Mechanisms of Repair

In the small intestine, the acute phase of repair involves marked contraction of the villi, which can dramatically reduce the surface area of the denuded segment of the intestinal mucosa (Figure 4). This villus contraction is regulated by enteric nerves, because experimental studies in guinea pigs have shown the ability of the nonspecific neurotoxin tetrodotoxin to abolish contraction (66). Furthermore, there seems to be a role for prostanoids in villus contraction (67), although it is uncertain if prostaglandin (PG)-stimulated contraction results in more rapid mucosal repair (68). This may result from the fact that there are two components to villus contraction: contraction of contractile elements adjacent to the central arteriole and contraction of myofibroblasts adjacent to the epithelial basement membrane. The latter is apparently not influenced by PGs (67) and is intimately involved in restitution (66). Although all of these findings are in species other than the horse, equine studies have shown reproducible evidence of villus contraction that is initiated during ischemia, and which continues during early mucosal repair as blood flow is restored (69–71). However, the relative contribution of villus contraction to repair in ischemic-injured small intestine is unknown.

Figure 4.

Diagram of the reparative events that occur following ischemic injury in horses. (a) Initially, the villi in the small intestine contract (arrows), reducing the surface area of the mucosal wound. (b) At the same time, healthy cells adjacent to the wound flatten and crawl across intact basement membrane using a process called epithelial restitution. (c) Finally, interepithelial junctional structures (tight junctions and adherens junctions) are reformed to seal the intestinal barrier. Adapted from artwork originally provided by North Carolina State University.

Contraction of the villi is accompanied by concurrent epithelial restitution, during which intact epithelial cells shouldering the mucosal wound progressively flatten and crawl across denuded basement membrane that appears intact following ischemic injury. This was first described in the equine literature during studies on ponies (72) and involves amoeba-like movement of epithelia (73). Restitution is initiated by integrins expressed on the cell membrane of migrating epithelia that interact with critical elements within the basement membrane, including collagen, hyaluronic acid, and proteoglycans (74, 75). In experimental in vitro cell studies, epithelial migration was hastened by providing the appropriate types of collagen (75) and by adding select proteoglycans and hyaluronic acid (76). The latter treatments (proteoglycans and hyaluronic acid) have been used extensively in equine orthopedic regenerative medicine but have not been tested in equine intestinal mucosal repair. Restitution is also strongly stimulated by specific growth factors, which are peptides released locally by mucosal elements that interact with specific receptors expressed on epithelium. In particular, transforming growth factor β (TGFβ), TGFα (which is not related to TGFβ), insulin-like growth factor, and fibroblast growth factors have been found to speed epithelial migration in vitro and in vivo (77). Furthermore, a group of peptides called trefoil factors, because of their cloverleaf-like structure, strongly stimulate epithelial migration (78). Provision of these factors has hastened the repair process in cell and murine studies but has not been tried on equine injured mucosal tissues. Other factors that facilitate epithelial migration are a group of compounds called polyamines. These have been found to be critical to the process of epithelial restitution, although their mechanism of action is unclear. One possibility is that the long positively charged nature of these compounds might facilitate repair of the basement membrane (79). Components of feed may also stimulate mucosal repair. For example, in one study in which a lesion similar to that noted with ischemic injury was induced by addition of bile salts, restitution was hastened by addition of the amino acid arginine (67). However, these factors have not been extensively investigated in horses or equine tissues.

Although restitution was once thought to be the defining event in mucosal repair, studies have shown that the mucosa remains leaky to clinically relevant molecules such as lipopolysaccharide until the tight junctions have been repaired and closed. This process was initially described in pigs (80) and has been shown to involve rapid reinsertion of tight junction proteins, including occludin and sealing claudins at the region of the apical lateral membranes of adjacent epithelial cells (81). The primary mediator of tight junction repair is PGE₂, which became of particular interest to equine researchers because prostanoids are typically inhibited in horses with intestinal injury with the use of the nonselective cyclooxygenase inhibitor and visceral analgesic flunixin meglumine (82, 83). The target of PGE₂ is a relatively understudied chloride channel called ClC-2, which unlike other chloride channels is expressed in villus epithelium and within the tight junctions. Its mechanism of action is thought to involve enhanced trafficking of tight junction proteins to the apical lateral membrane, although the precise mechanism of action remains under study (84).

Role of Acute Inflammation in Mucosal Injury and Repair

Evaluation of mucosal repair following ischemic injury in pigs revealed a different time course of neutrophil infiltration as compared with murine and feline studies. In particular, ischemia in rodents and cats showed peak infiltration of neutrophils within 1 h of reperfusion injury (61, 85–87), whereas in porcine studies, peak neutrophil infiltration occurred 6 h after ischemic injury (88). The importance of this finding was that rather than exacerbating injury, this delayed neutrophil infiltration may hamper restitution, particularly tight junction repair, as neutrophils coursed through the paracellular spaces of restituting epithelium. Of translational significance, inhibiting neutrophil adhesion or treating with superoxide dismutase could restore optimal repair (88). Based on these findings, an equine model of jejunal strangulating obstruction was developed in horses in which discrete segments of the jejunum were subjected to 2 h of complete segmental jejunal ischemia by cross-clamping the bowel and clamping the associated vasculature, followed by removal of the clamps, recovery of the horse from general anesthesia, and euthanasia for collection of tissues after 18 h (89). Evaluation of intestinal tissues revealed all of the elements of acute mucosal repair, including villus contraction and restitution, as well as evidence of inflammation (90). As has been previously stated, neutrophil infiltration during mucosal repair is different from reperfusion injury shown in rodents, during which neutrophils maximally infiltrate within 1–3 h of reperfusion, causing further mucosal injury prior to reparative events (60). The equine model has proven invaluable in researching equine intestinal repair, including its use by other investigators to study colonic ischemic injury and mucosal recovery (83).

Impact of Clinical Medications on Equine Acute Mucosal Repair

The role of PGs in reassembly of interepithelial tight junctions shown in basic science work performed principally in porcine ischemic-injured tissues (80) suggested that administration of nonsteroidal anti-inflammatory drugs (NSAIDs), which is commonplace in horses with the clinical syndrome of colic, may be problematic (91). Ex vivo trials in horses showed that the nonselective COX inhibitor flunixin meglumine did indeed retard recovery of barrier function (92–94). Subsequently, in vivo trials using the 2-h ischemia/18-h recovery model in horses showed that a clinical dosage regime of flunixin meglumine [1.1 mg/kg, intravenously (IV), every 12 h (q12h)] consistently reduced barrier function recovery (70, 71, 89). This was not associated with measurable histological differences in mucosal repair, including villus height and percentage of mucosal epithelialization. Instead, flunixin meglumine’s effect appeared to be paracellular in nature (i.e., associated with the integrity of tight junctions) based on reduced recovery of transepithelial electrical resistance and increased permeability to the paracellular probe mannitol (89). To heighten the potential clinical relevance of these findings, mucosal-to-serosal-labeled Escherichia coli lipopolysaccharide fluxes were performed, which confirmed that flunixin meglumine also resulted in a permeability defect after the 18-h recovery period (70, 71). Interestingly, flunixin meglumine also caused a significant increase in mucosal inflammation, measured by the numbers of counted neutrophils within the lamina propria beneath repairing epithelium (90). The cause of this inflammation was presumed to be the increased leakiness of the recovering mucosa. In contrast, use of NSAIDs with selectivity for COX-2 allowed for repair of the mucosa, either ex vivo (deracoxib, robenacoxib) (92, 94) or in vivo (firocoxib) (70). In addition, meloxicam, which has an approximately threefold selectivity for COX-2, also allowed full mucosal recovery. To confirm COX selectivity in vivo, equine trials comparing either meloxicam or firocoxib with flunixin meglumine showed that the latter inhibited both TXB₂ (from COX-1) and PGE₂ (from COX-2) in whole blood taken during the trial, whereas meloxicam and firocoxib did not inhibit TXB₂. In other words, the COX-2 selective medications were COX-1 sparing. This property is presumed to be the reason that ischemic-injured equine mucosa repairs unabated in the presence of medications that principally inhibit COX-2, with the presence of reparative prostanoids elaborated by mucosally expressed COX-1 (95). In a recent as-yet-unpublished randomized clinical trial, horses with small intestinal strangulating obstruction were randomly assigned to treatment with either flunixin meglumine (1.1 mg/kg, IV, q12h) or the COX-2 inhibitor firocoxib (2.7-mg/kg loading dose, 0.9 mg/kg, IV, q24h). Horses receiving flunixin meglumine were at greater risk of having excessive levels of the sepsis biomarker soluble CD14 (96), suggesting measurable differences in recovery of barrier function. How this will translate to clinical outcome has not yet been determined.

The mucosal inflammation that had been increased in the presence of flunixin meglumine was significantly diminished in the presence of COX-2-selective NSAIDs (70, 71). In addition, trials evaluating the effects of systemically administered lidocaine in horses with the 2-h ischemia/18-h recovery (1.3-mg/kg loading dose, 0.05 mg/kg/min controlled rate infusion) revealed that this medication prevented mucosal inflammation in horses concurrently treated with flunixin meglumine (69). This is of particular clinical relevance because horses are frequently treated with both of these medications following surgery for strangulating small intestinal obstruction, with lidocaine being used to reduce the prevalence and duration of postoperative ileus. However, the mechanism for the action of lidocaine, in terms of its effects on mucosal inflammation, remains elusive. In vitro trials assessing the effect of lidocaine on equine neutrophils showed no effect of lidocaine on either adhesion or migration (97). One possibility is that lidocaine alters the function of intestinal epithelial electrolyte transporters to reduce epithelial injury and potentially enhance repair (98). However, this will take additional work to fully explore.

Subacute Mucosal Repair and Epithelial Proliferation

Once tissues have fully restituted, epithelium that has been sloughed during the initial injury must be replaced to restore normal mucosal architecture. During this subacute period of repair, cellular proliferation, differentiation, and the onset of inflammation predominate (99). Epithelial proliferation in the small and large intestine is derived from stem cells within the crypt compartment, with escalation of newly formed cells along the crypt-villus axis. In homeostatic conditions, a newly formed cell may take anywhere from 3 to 7 days to reach the tip of the villus or the surface epithelial compartment in the colon, during which time the epithelial cells progressively mature and attain specific functions related to surface epithelium. For example, epithelial secretion takes place largely in crypts, whereas mature epithelium at the surface has a more dominant absorptive role. Such absorption includes electrolytes and nutrients, including glucose and amino acids (in the small intestine) and short-chain fatty acids (in the colon). After mucosal injury, the crypt compartment is greatly enlarged, with dramatic increases in crypt proliferation. The lag phase before an increased proliferative rate is approximately 12 to 18 h, so that by 48 to 72 h, the crypt may be many times the size of the crypt immediately following injury (99). This process is likely driven by elaboration of increased concentrations of local growth factors, although the signals involved in these events continue to be an area of intensive study. After amplification of the crypt, the villus gradually reforms, so that the surface area of the villus has been restored by 4 to 6 days after a severe mucosal ischemic event (77). The process of reformation of normal mucosal structure may be more rapid in the colon, where there are no villi. Factors known to stimulate mucosal epithelial proliferation include peptide growth factors, such as TGFα, polyamines, and gut-specific nutrients like glutamine in the small intestine and butyrate in the colon (56). Thus, as with epithelial restitution, early refeeding likely hastens restoration of normal mucosal structure and function.

EQUINE STEM CELL CULTURE AND ORGANOIDS

Recent advances in the field of stem cell biology have enabled the growth of intestinal stem cells in culture from murine, porcine, human, and most recently horse intestinal epithelium (100–103). The process of stem cell culture begins with the isolation of intestinal crypts. In short, crypts are dissociated and isolated from whole tissue through a process of shaking the tissue immersed initially in ethylenediaminetetraacetic acid. The tissue is then repeatedly transferred, with continued shaking, into multiple solutions of ice-cold phosphate-buffered saline until crypt/villi units separate from the underlying basement membrane. This process minimizes the amount of debris and maximizes the number of crypts present in the final solution. The remnant intestinal tissue is removed and the solution is filtered through a 100-micron sterile cell strainer to remove any remaining villi. The final solution is then centrifuged, and the pelleted crypts are resuspended directly into Matrigel supplemented with multiple growth factors.

The intestinal stem cells in these cultures develop into three-dimensional structures that consist of all of the components of native intestine. These include outwardly budding crypts, all mature epithelial cell types, and a pseudo-lumen where dead cells are extruded (Figure 5). They are therefore referred to as mini-guts but may also be called enteroids or organoids. These culture systems provide an advanced tool to study intestinal epithelial dynamics during homeostasis and disease. To date, they have been used to model tissue morphogenesis, stem cell lineage selection, tissue plasticity, cellular signaling, and both bacterial and viral infection (104). Furthermore, these mini-guts have been shown to adhere and form crypt structures within denuded mucosa when infused into the intestinal lumen of a mouse with colitis (105). Despite these significant advances in the field, intestinal stem cells have yet to be used clinically. Furthermore, studies that evaluate the impact of injury and the regenerative response of stem cells are few.

Figure 5.

Equine enteroid culture. (a) Time course crypt culture 0–120 h postplating. (b) Whole mount imaged enteroid with fluorescently labeled enteroendocrine cells (red), proliferative cells (green), and general nuclear stain (blue). Outwardly budding crypts are visible peripherally with formation of a pseudolumen centrally. Three-dimensionality creates multiple visual planes with cells visible both in and out of focus. (c) Formalin-fixed, paraffin-embedded, sectioned, and hematoxylin- and eosin-stained 7-day cultured enteroid.

At this time, work in the field of equine intestinal stem cells and their culture remains limited (106). In horses, a treatment that hastens mucosal repair may help to prevent the common sequela of severe intestinal injury, including sepsis, bacteremia, laminitis, ileus, and diarrhea. Together with existing treatments, novel therapies derived from intestinal stem cell research and the use of the stem cell culture system may help shorten recovery times and facilitate enhanced repair following severe mucosal damage.