Materials and Methods

Horses—After informed consent was obtained from each horse owner, a blood sample was collected from a jugular vein of each of 50 horses (≥ 1 year old) that were admitted to the University of Georgia Large Animal Teaching Hospital for treatment of acute GIT disease. During that same period, a blood sample was collected from each of 10 university-owned healthy adult horses (4 to 12 years old) that lacked clinical signs of any substantial inflammatory process, had no history of GIT disease for ≥ 6 months, and had not received any medication for > 30 days. These 10 healthy horses were housed indoors during the day and on pasture during the night. The study was approved by the Animal Care and Use Committee and the College of Veterinary Medicine Clinical Research Committee at the University of Georgia.

For horses with GIT disease, information regarding clinical history, clinical findings recorded during admission to the hospital, treatment, outcome, and necropsy findings (when applicable) were recorded from medical records. This information was used to categorize horses on the basis of GIT disease type (ie, colic of undetermined etiology, GIT inflammation or strangulation, or nonstrangulating GIT obstruction) and disease outcome (ie, discharge from the hospital, death or euthanasia because of disease severity, or euthanasia because of financial constraints). The criteria for categorizing horses as having colic of undetermined etiology were the absence of clinical findings (eg, profuse diarrhea, mass in the large colon detected during per rectal palpation, or high cell count in a peritoneal fluid per sample) indicative of GIT inflammation or strangulation or nonstrangulating obstruction and rapid recovery after treatment with fluids, drugs, or both. The criteria for categorizing horses as euthanized because of financial constraints were the presence of clinical findings (ie, peritoneal fluid contaminated with plant particles and bacteria or a segment of nonviable intestine observed during surgery or necropsy) indicative of a terminal disease process or poor prognosis and notation in the records that euthanasia was performed because of financial constraints. All other horses that were euthanized were categorized as euthanized because of disease severity.

Blood sample collection—The timing of blood sample collections from the 10 healthy horses was distributed temporally such that a sample from 1 healthy horse was obtained after each fifth clinical case was admitted. The skin over the needle insertion site was aseptically prepared by scrubbing the skin with chlorhexidine and isopropyl alcohol. Sixty milliliters of blood was obtained from a jugular vein of each horse via venipuncture with a sterile 18-gauge hypodermic needle and a sterile syringe containing 1.5 mL of 0.1M EDTA.

Leukocyte isolation and RNA extraction—Within 30 minutes after blood sample collection, the syringe containing the sample was maintained vertically on its plunger for 20 minutes to allow for erythrocyte sedimentation. Leukocyte-rich plasma was expressed into a sterile tube, diluted with an equal volume of Dulbecco PBS solution, and centrifuged at 20°C for 10 minutes at 800 × g. The supernatant was discarded, and erythrocytes were removed by hypotonic lysis. The remaining leukocytes were washed with PBS solution and concentrated via centrifugation at 20°C for 5 minutes at 800 × g. The supernatant was discarded, leukocytes were suspended in 20 mL of PBS solution, and a 50–ML aliquot of this cell suspension was transferred to a vial containing 450 ML of a 0.04% solution of trypan blue dye^a for viability assessment on the basis of dye exclusion by the leukocytes. Viable (ie, nonstained) leukocytes were quantified microscopically by use of a hemacytometer (the mean number of cells in 4 large squares of the hemacytometer was used). On the basis of the cell count, leukocytes were suspended in RPMI 1640 medium (without phenol red)^b containing 1% equine serum,^c 2mM l-glutamine,^d and gentamicin sulfate^e (50 μg/mL) to achieve a final concentration of 2 × 10⁷ viable cells/ mL. One-milliliter aliquots of this leukocyte suspension were transferred to sterile microcentrifuge tubes. Tubes were immediately centrifuged in a microcentrifuge for 1 minute at 14,000 × g. The supernatant was discarded, and the microcentrifuge tubes containing the pelleted cells were stored at −80°C until RNA extraction. Total RNA was extracted from thawed leukocyte samples by use of a commercial kit^f in accordance with the manufacturer's protocol with the addition of a deoxyribonuclease I^g digestion step to remove residual genomic DNA. Assessment of RNA concentration and quality was performed by spectrophotometry.^h

Primer design and real-time qRT-PCR assay conditions—Expression of 14 genes associated with systemic inflammatory responses was targeted for quantification as previously described.³¹ These 14 genes encoded an adenosine A_2A receptor, cyclooxygenase-2, interleukin-1B, interleukin-1 receptor antagonist, interleukin-6, interleukin-8, interleukin-10, leukocyte-selectin molecule, monocyte chemotactic protein-3, matrix metalloproteinase-9, platelet-selectin molecule, mitochondrial superoxide dismutase, Toll-like receptor 4, and tumor necrosis factor-A. Gene sequences for these targets were identified and downloaded from the GenBank nucleotide sequence database. Optimal primers were designed by use of a commercial software systemⁱ and synthesized (Appendix). Full validation of the SYBR green real-time qRT-PCR assay for 12 of the 14 genes was previously performed³¹ with RNA isolated from LPS-stimulated equine leukocytes.

Complementary DNA strands were synthesized from extracted RNA samples by use of a commercial kit^j and thermal cycler.^k The real-time qRT-PCR assay was performed by use of SYBR green^l nucleic acid stain in a sequence detection system.^m Sample denaturation was conducted at 95°C for 15 seconds followed by a combined annealing and extension phase at 60°C for 60 seconds. A commercial 18S ribosomal RNA kitⁿ was used as an endogenous control template. The results of the real-time qRT-PCR assay for each sample were provided as follows:

View Expanded

where ΔC_T is the difference of the C_T of the gene of interest (C_Tgene) and the C_T of the endogenous control template (C_T18S). The qRT-PCR assay was performed 3 times for each cDNA sample aliquot.

Collection of climatological data—Daily weather records (ie, humidity and minimum and maximum temperatures) for the study period (July to September of 2007) were retrieved from the University of Georgia Climatology Research Laboratory. Daily air pollen count from the nearest counting station (Atlanta) was obtained from a database.^o

Statistical analysis—Results for 10 clinical variables (ie, Hct, plasma albumin and fibrinogen concentrations, serum creatinine and glucose concentrations, total leukocyte and neutrophil counts, heart rate, rectal temperature, and duration of signs of colic) were selected for statistical analysis. These variables were selected because each variable was recorded for all horses and because studies^{1–4,13–15,32} have shown that these variables are associated with severity of disease, prognosis, or both in horses with colic. The ΔC_T values and results for the 10 clinical variables in healthy horses and horses with GIT diseases were categorized by the type of GIT disease and disease outcome. Horses with GIT disease that were euthanized because the owner declined surgery were excluded from comparisons of the different disease outcomes. Comparisons were made by use of a 1-way ANOVA; to limit the overall type I error rate to 5%, post hoc comparisons were made by use of the Bonferroni procedure. Normality of the distribution of results was evaluated graphically through the creation of histograms. Fatality proportions of different groups were compared by use of the Fisher exact test. All testing assumed a 2-sided alternative hypothesis. A value of P < 0.05 was considered significant for all analyses. Analyses were conducted by use of a commercially available statistical software program.^p

Results

The duration of the study was 66 days. During this time, environmental temperature ranged from 17.5° to 40.5°C (mean, 27.6°C), relative humidity ranged from 22% to 100% (mean, 68.3%), and pollen count ranged from 3 to 16 U/m³ (mean, 7.9 U/m³). Among the 50 horses with GIT disease, there were 28 males and 22 females. Age range was 1 to 36 years (median, 10 years). Breeds of the affected horses were Quarter Horse (n = 15), Thoroughbred (9), Arabian (4), pony (4), warmblood (4), Paint Horse (3), Hanoverian (2), Passo Fino (2), Tennessee Walking Horse (2), American Saddle (1), Andalusian (1), Belgian (1), Gypsy Vanner (1), and Haflinger (1). Among the 10 healthy horses, there were 5 males and 5 females. Age range was 6 to 12 years (median, 9 years). Breeds of the healthy horses were Quarter Horse (n = 7 horses), Paint Horse (2), and mixed-breed horse (1).

Clinical signs of the GIT diseases were first observed 4 to 72 hours (median, 8 hours) before admission to the hospital. All horses with GIT disease had been treated at least 1 time with analgesic drugs (eg, flunixin meglumine, xylazine hydrochloride, or butorphanol tartrate) or other medications (eg, laxatives and fluids) before blood sample collection. Among the 50 horses with GIT disease, 8 horses had colic of undetermined etiology. Among the remaining 42 horses with GIT disease, 11 had GIT inflammation, 8 had GIT strangulation, and 23 had nonstrangulating GIT obstruction (Table 1). Among the 19 horses with GIT inflammation or strangulation, the primary lesion was located in the small intestine (n = 8 horses), large intestine (8), both the small and large intestines (2), or stomach (1). In 2 horses with GIT inflammation (colitis [n = 1 horse] or gastric impaction and rupture [1]), there was diffuse peritonitis in addition to the primary cause of GIT disease. Among the 23 horses with nonstrangulating GIT obstruction, the small intestine and large intestine were affected in 6 and 17, respectively. Among the 50 horses with GIT disease, 22 underwent a celiotomy procedure, 23 received medical treatment only, and 5 were euthanized because the owners declined surgery. Among the 45 horses receiving treatment as recommended by the attending clinician or clinicians, 38 were discharged from the hospital and 7 died or were euthanized. Necropsy was performed in 3 of the 5 horses that were euthanized because surgery was declined by the owners and in those 7 horses that died or were euthanized because of disease severity. Mortality rate was not significantly (P = 0.68) greater in horses with GIT inflammation or strangulation (4/17 [23.5%] horses), compared with the mortality rate in horses with nonstrangulating GIT obstruction (3/20 [15.0%] horses).

Among the 10 variables included in the statistical analysis, plasma albumin concentration was significantly decreased and serum glucose concentration was significantly increased in horses with GIT inflammation or strangulation, compared with plasma albumin and serum glucose concentrations in horses with colic of undetermined etiology. The Hct, serum creatinine concentration, heart rate, and rectal temperature were significantly increased in horses that died or were euthanized because of the severity of GIT disease, compared with values in horses that were discharged from the hospital. Furthermore, plasma albumin concentration was significantly (P = 0.02) decreased and duration of the clinical signs of colic was significantly (P = 0.04) longer in those horses that died or were euthanized, compared with results for horses discharged from the hospital (Table 2).

Mean ± SEM ΔC_T results from real-time qRT-PCR assay analysis for expression of genes by GIT disease type and disease outcome were summarized (Table 3). No significant differences were detected between healthy horses and horses with GIT disease (regardless of type or disease outcome) for the expression of genes encoding an A_2A receptor, cyclooxygenase-2, interleukin-1B, interleukin-1 receptor antagonist, interleukin-6, interleukin-10, or monocyte chemotactic protein-3. No significant differences were detected between healthy horses and any of the disease outcome groups for the expression of genes encoding an A_2A receptor, cyclooxygenase-2, interleukin-1B, interleukin-6, interleukin-10, or monocyte chemotactic protein-3.

Gene expressions were not significantly different in horses with colic of undetermined etiology for any of the genes examined, compared with gene expressions in healthy horses. Gene expressions of interleukin-8, matrix metalloproteinase-9, and Toll-like receptor 4 were significantly increased and expression of tumor necrosis factor-A was significantly decreased in horses with GIT inflammation or strangulation and horses with nonstrangulating GIT obstruction, compared with expression of these genes in healthy horses. Gene expressions of leukocyte- and platelet-selectin molecules and mitochondrial superoxide dismutase were significantly increased in horses with nonstrangulating obstruction, compared with expression of these genes in healthy horses. When compared among the categories of GIT disease type, no significant difference was detected among the mean ΔC_T results for all genes assayed.

Gene expression of interleukin-8 and matrix metalloproteinase-9 was significantly increased and expression of tumor necrosis factor-A was significantly decreased for dead horses or horses that were discharged from the hospital, compared with the expression of these genes in healthy horses. Gene expression of mitochondrial superoxide dismutase, Toll-like receptor 4, and leukocyte- and platelet-selectin molecules was also significantly increased in horses discharged from the hospital, compared with expression of these genes in healthy horses. The only difference observed among the disease outcomes was for the expression of the interleukin-1 receptor antagonist, which was significantly increased in horses discharged from the hospital, compared with the expression of this gene in dead horses (ie, those that died or were euthanized) and horses discharged from the hospital.

Discussion

In the group of horses that died or were euthanized because of the severity of GIT disease, Hct, serum creatinine concentration, heart and respiratory rates, and duration of colic before admission were significantly greater and plasma albumin concentration was significantly less than the findings in affected horses that survived until discharge from the hospital, as hypothesized. These changes have been reported³² to be positively associated with disease severity and fatality rate in horses with GIT disease. The lack of detectable differences with respect to mortality rate and values of clinical variables compared between horses with GIT inflammation or strangulation and those with nonstrangulating GIT obstruction suggests that the proportion of severely ill horses in each of these groups was not different. On the basis of the criteria used in this study to categorize horses by GIT disease type, it is reasonable to expect that horses with colic of undetermined etiology had less severe disease than did the horses in the other GIT disease groups. Thus, it was not surprising that gene expression in horses with colic of undetermined etiology was not different from that in healthy horses. Conversely, horses with GIT inflammation or strangulation and horses with nonstrangulating GIT obstruction had changes in the expression of several genes associated with a systemic inflammatory response. These changes in horses with nonstrangulating GIT obstruction contrast with the results of a study²⁰ in which circulating neutrophils from horses with this type of GIT disease were not activated. However, the findings of the present study were consistent with the results of another study³³ in which experimentally induced intraluminal colonic obstruction caused damage to the GIT mucosa and induced inflammatory lesions in remote organs. Furthermore, evidence of the activation of circulating leukocytes in horses with naturally occurring nonstrangulating obstruction has been reported.¹⁸

Because of the paucity of information regarding systemic inflammatory responses in horses with naturally occurring disease, interpretation of the different patterns of leukocyte gene expression observed in the present study was compared with results obtained in studies of humans and rodents. Interleukin-8 is a proinflammatory cytokine, which is primarily produced by macrophages and has a critical role in recruitment and activation of neutrophils.³⁴ An increase in interleukin-8 expression has been detected^35,36 in human patients with sepsis, and increased serum concentrations of interleukin-8 are associated with an increased mortality rate. Furthermore, an increase in the expression of the interleukin-8 gene in umbilical cord blood is a reliable indicator of sepsis in human neonates.³⁷ An increase in the expression of the interleukin-8 gene has also been detected³⁰ in sick equine neonates; however, no difference in interleukin-8 expression was detected between foals with sepsis and those with nonseptic conditions. In the study reported here, expression of the interleukin-8 gene was increased in all groups of horses with GIT disease, except in horses with colic of undetermined etiology. No difference in the expression of the interleukin-8 gene could be detected between horses that were discharged from the hospital and those that died or were euthanized.

Matrix metalloproteinase-9 is an enzyme expressed by neutrophils, monocytes, and other cell types.³⁸ This enzyme is important for extracellular matrix remodeling and degradation. Additionally, matrix metalloproteinase-9 contributes to vascular permeability and leukocyte diapedesis.^38–40 Plasma concentrations of matrix metalloproteinase-9 are increased in rats and humans with experimentally induced endotoxemia^41,42 and in humans with sepsis, compared with concentrations in unaffected individuals.⁴⁰ In a study⁴³ involving horses with colic, plasma matrix metalloproteinase-9 concentrations assayed immediately before celiotomy were not different from those in healthy horses. However, more studies are needed before conclusions can be made because only 5 horses with colic were included in that study.⁴³ Conversely, increases in plasma concentrations of matrix metalloproteinase-9 in horses with experimentally induced laminitis caused by carbohydrate overload have been reported.⁴⁴ In agreement with these observations, expression of mRNA for matrix metalloproteinase-9 was increased in horses with GIT disease in the present study, except in horses with colic of undetermined etiology. In agreement with the observations of the present study, plasma concentrations of matrix metalloproteinase-9 are not associated with disease outcome in humans with sepsis.⁴⁰

Toll-like receptor 4 is a cell surface receptor expressed by monocytes and macrophages. This receptor functions as a signal transducing molecule of the LPS-receptor complex, which has a role in detection of and response to gram-negative bacterial infection.⁴⁵ In humans with sepsis as well as in mice with experimentally induced peritonitis, expression of the Toll-like receptor 4 gene in monocytes is increased.^46,47 Findings of a recent study⁴⁸ in mice further illustrate the importance of Toll-like receptor 4 in the inflammatory response triggered by LPS because both the response to experimentally induced sepsis caused by a gram-negative bacterial infection and mortality rate were reduced in wild-type mice treated with anti–Toll-like receptor 4 antibodies and in Toll-like receptor 4 knockout mice. Similar to the situation in humans, expression of Toll-like receptor 4 is increased in foals with sepsis relative to expression observed in healthy foals; however, no difference could be detected when gene expressions in surviving and non-surviving foals were compared.²⁹ In the study reported here, the increases in expression of the Toll-like receptor 4 gene in horses with GIT inflammation or strangulation and horses with nonstrangulating GIT obstruction were in agreement with the findings of studies in foals²⁹ and humans.^46,47

Selectin molecules are expressed by endothelial cells (ie, endothelial- and platelet-selectin molecules), platelets (ie, platelet-selectin molecule), and leukocytes (ie, leukocyte-selectin molecule).^49–52 Other investigators have claimed that the platelet-selectin molecule is also expressed by murine peritoneal macrophages⁵³ and macrophages in mice with injury to the carotid artery.⁵⁴ The function of selectin molecules is to mediate cell adhesion, which is a crucial step for both thrombosis and diapedesis.^49,50,52 In humans, administration of LPS leads to platelet-monocyte aggregation,⁵⁵ which is evidence of increased gene expression of the platelet-selectin molecule.^50,56 In human patients with inflammation and sepsis, an increase in the expression of the leukocyte-selectin molecule gene in circulating leukocytes has also been detected.^57,58 Therefore, in the study reported here, it was not surprising to find an increase in the expression of both platelet- and leukocyte-selectin molecule genes in horses with GIT obstruction. Because expression of the platelet-selectin molecule gene has not been detected in circulating leukocytes,⁴⁹ expression in the samples collected from healthy horses and horses with GIT disease in the present study suggests that the leukocyte-rich samples contained platelets.

Superoxide dismutase is a ubiquitously distributed enzyme that reduces superoxide to hydrogen peroxide and protects against damage caused by free radical ions (eg, superoxide anions). As observed in other mammals,⁵⁹ horses have at least 3 isoforms of superoxide dismutase (ie, mitochondrial superoxide dismutase, cytosol superoxide dismutase, and extracellular surface superoxide dismutase).^60,61 As indicated by death shortly after birth in mice that are deficient in mitochondrial superoxide dismutase (but not the other 2 isoforms), that isoform is likely the most important.⁶² Mitochondrial superoxide dismutase gene expression increases in human mononuclear cells in response to proinflammatory stimuli (eg, oxidative stress, LPS, and interleukin-1).^63,64 Although similar studies have not been conducted with equine mononuclear cells to our knowledge, increased expression of mitochondrial superoxide dismutase in laminar tissues has been identified⁶⁵ in horses with experimentally induced acute laminitis. Therefore, it is not surprising that an increase in the expression of the mitochondrial superoxide dismutase gene was observed in horses with nonstrangulating GIT obstruction and horses that were discharged from the hospital in the study reported here. The small number of horses with GIT inflammation or strangulation and horses that died or were euthanized because of disease severity may have prevented detection of any difference in gene expression, compared with expression in healthy horses.

Tumor necrosis factor-A is a proinflammatory cytokine expressed primarily by monocytes and T cells.⁶⁶ Although increases in tumor necrosis factor-A concentration in blood^5,67 and peritoneal fluid⁵ have been detected in horses with GIT disease, expression of the tumor necrosis factor-A gene was reduced in leukocytes of horses with GIT disease in the present study. However, this finding in the present study is in agreement with the reduced expression of the tumor necrosis factor-A gene in the blood of sick foals with septic or nonseptic disease, compared with expression in a group of healthy foals.³⁰ Also in agreement with the findings of the present study was the fact that the expression of the tumor necrosis factor-A gene did not differ between surviving and nonsurviving foals.³⁰ The discrepancies in these results may be explained (in part) by the fact that the tumor necrosis factor-A gene is expressed early during disease progression and has a short half-life of elimination.⁶⁸

The interleukin-1 receptor antagonist is a protein that has an affinity for the interleukin-1 receptor and functions as a competitive antagonist of interleukin-1 at this receptor site. Therefore, the interleukin-1 receptor antagonist has anti-inflammatory properties and functions to modulate the inflammatory response.⁶⁹ Contrary to results observed in humans⁷⁰ with sepsis and in laboratory animals⁷¹ with experimentally induced sepsis, the expression of the interleukin-1 receptor antagonist gene in horses with GIT disease that were discharged from the hospital was increased in the present study, compared with expression in healthy horses. Furthermore, a similar increase in the expression of the interleukin-1 receptor antagonist gene was not detected in affected horses that died. In other animals, an increase in the expression of the interleukin-1 receptor antagonist gene is associated with greater morbidity and increased mortality rates.^70,71 To evaluate these differences, it is important to consider the function of the interleukin-1 receptor antagonist and the etiopathogenesis of the systemic inflammatory response syndrome among animals and specific type of experimentally induced disease. In a study⁷² that involved the use of 2 genetically modified strains of mice (1 strain that does not express and 1 strain that overexpresses the interleukin-1 receptor antagonist gene), susceptibility to experimental endotoxemia and infection with Listeria monocytogenes was affected by the expression of the interleukin-1 receptor antagonist gene. Mice that did not express the interleukin-1 receptor antagonist gene were more resistant to bacterial infection but were more susceptible to the effects induced by endotoxemia. The converse was evident in the mice that overexpressed the interleukin-1 receptor antagonist gene.⁷² Analysis of these findings suggests that the anti-inflammatory effect of the expression of the interleukin-1 receptor antagonist gene may be beneficial for limiting the response induced by endotoxemia but may compromise the response by the host's immune system against infection. Therefore, it can be hypothesized that an increased expression of the interleukin-1 receptor antagonist gene may be beneficial in adult horses with severe GIT disease because these horses often develop endotoxemia without bacteremia.^73,74 In contrast, increased expression of the interleukin-1 receptor antagonist gene may be detrimental in humans^75–77 and foals⁷⁸ with sepsis who commonly develop bacteremia. Further evidence of the detrimental effect of the expression of the interleukin-1 receptor antagonist gene on the response to bacterial infection is that treatment with the interleukin-1 receptor antagonist resulted in an increased mortality rate in mice with polymicrobial-induced sepsis caused by cecal ligation and puncture.⁷⁹

It is important to acknowledge the main limitations inherent to the study reported here. Samples were collected from a small number of horses when considering the diverse group of GIT disease types included in the present study. Furthermore, the time that had elapsed between the observation of the first signs of GIT disease and blood sample collection could not be standardized and was quite variable. Owners and referring veterinarians made the decision of when these horses would be admitted to the referral hospital, and the duration of the trip to the hospital was variable. Additionally, it was impossible to standardize the treatments administered by the referring veterinarians or attending clinicians before and after sample collection. Depending on the medical condition of each horse and the treatment administered by the referring veterinarian, treatment before admission ranged from 1 dose of an analgesic drug to a more extensive treatment regimen, which included several doses of analgesic drugs, fluids, and laxatives. It was not possible to standardize treatments administered after the horses were admitted to the referral hospital because several attending clinicians were responsible for the treatment of horses included in the study reported here. Thus, it is possible that individual preferences for the treatments administered by the attending clinicians could have affected the results and outcome. Finally, gene expression was assessed only once, which only provided insight to the status of the inflammatory response occurring in horses with GIT disease during admission to the referral hospital for treatment. In future studies, investigation of larger sample populations of horses with each type of GIT disease and evaluation of the expression of proinflammatory and anti-inflammatory genes in circulating leukocytes at multiple time points after admission may enable better characterization of inflammatory responses associated with disease type and refinement of prognosis assessment.