Use of quantitative real- time PCR to determine the local inflammatory response in the intestinal mucosa and muscularis of horses undergoing small intestinal resection

Background: Studies in rodents and humans have demonstrated that intestinal manipulation or surgical trauma initiates an inflammatory response in the intestine which results in leucocyte recruitment to the muscularis externa causing smooth muscle dysfunction. Objectives: To examine the intestinal inflammatory response in horses undergoing colic surgery by measuring relative differential gene expression in intestinal tissues harvested from surgical colic cases and control horses. Study design: Prospective case- control study. Methods: Mucosa and muscularis externa were harvested from of horses undergoing abdominal surgery that necessitated intestinal resection. By measuring differential gene expression of a set of target genes ( IDO1 , IL6 , IL1 β , TNF , CCL2 , NOS2 and PTGS2 ), we demonstrated the presence of intestinal inflammation in horses undergoing colic surgery. This gene list selection was based on previous reports of their upregulation in the muscularis externa of murine POI models 14 and in humans undergoing laparotomy. 7 In rodent intestinal manipulation- induced POI models, activation of MM results in cytokine and chemokine release, followed by infiltration of leucocytes into the muscularis externa . These infiltrating leucocytes inhibit smooth muscle function via the secretion of leucocytic products, such as nitric oxide (NO) and prostaglandins. This macrophage activation, which occurs within hours of surgery, has also been demonstrated in human patients. 7 As well as playing a key role in the initiation of inflammation, macrophages also have a protective function and are pivotal to the tissue repair process. At the level of the intestine, this protective function was recently demonstrated in a rodent model of infection- induced inflammation, in which muscularis externa macrophages protected the enteric neurons against post-infection neuronal death. IL1 CCL2 TNF upregulated in both the mucosa and muscularis externa in colic cases, whereas PTGS2 was upregulated in the mucosa only. These findings demonstrate the presence of active inflammation within intestinal tissue visually assessed as devitalised, a finding consistent with data derived from the muscularis externa of humans undergoing laparotomy. In addition, we also identified an inflammatory cytokine and mediator response within the mucosa, a finding which contrasts with the results derived from a study of people undergoing laparotomy which upregulation IL6 , PTGS2 and


| INTRODUC TI ON
Studies on the margins of surgically resected equine intestine from horses undergoing colic surgery identified a generalised stress response in the smooth muscle and myenteric plexus, characterised by an increase in apoptotic smooth muscle cells and apoptotic neurons and glial cells. 1 Ischaemia and manipulation of the equine jejunum also induces a post-operative neutrophilic infiltration in all tissue layers of the jejunum and eosinophilic infiltration of the jejunal submucosa. 2,3 In an equine model where the jejunum was distended and then decompressed, there was also an increase in the number of neutrophils in the mucosa and the serosa. 4 This inflammatory infiltrate is likely a downstream consequence of resident intestinal inflammatory cell activation in the muscularis externa that has been described in rodents and humans. [5][6][7] Intestinal manipulation triggers the activation of muscularis externa macrophages (MM) by pathogen-associated molecular patterns such as lipopolysaccharide. 8 This results in cytokine and chemokine release followed by the infiltration of leucocytes, predominantly neutrophils, monocytes and mast cells to the muscularis externa. These infiltrating leucocytes secrete products, such as nitric oxide and prostaglandins which can impair smooth muscle function. 9,10 Genes associated with inflammation of the muscularis externa induced by manipulation or surgery in rodent models and human studies include the pro-inflammatory cytokines interleukin (IL) 6, IL1β, tumour necrosis factor (TNF), nitric oxide synthase 2 (NOS2), the chemokine C-C motif chemokine ligand 2 (CCL2) and the immunomodulatory enzyme mediator prostaglandin-endoperoxide synthase 2 (PTGS2). 7,11 This inflammatory response within the muscularis externa is strongly implicated as a significant causative factor in the pathogenesis of post-operative ileus (POI). 5 Despite the high prevalence of POI in horses following colic surgery, there are no equine studies to date evaluating the inflammatory response genes known to be associated with the development of POI in rodent and human studies.
We hypothesised that an increase in inflammatory gene expression of IL6, IL1β, CCL2, TNF, PTGS2 and indoleamine 2,3-dioxygenase (IDO1) will be present in the resected margins of horses undergoing small intestinal resection. Furthermore, we hypothesised that inflammatory gene expression will be greater in horses that subsequently develop post-operative reflux (POR) (a characteristic feature of POI) following abdominal surgery, compared with those that do not.

| Study design
This was a prospective case-control two-centre study between October 2014 and February 2017. Control horses were presented for elective euthanasia for a variety of reasons unrelated to diseases and disorders of the intestinal tract; these included poor temperament, chronic orthopaedic conditions, recurrent laminitis and suspected dental disease. Additionally, financial constraints of the owners contributed towards the decision for elective euthanasia in some of these cases. Control horses were excluded if any gross gastrointestinal lesions were identified at post-mortem examination. Horses were euthanased with secobarbital sodium 400 mg/ mL and cinchocaine hydrochloride 25 mg/mL (Somulose; Dechra Veterinary Products) and samples were obtained approximately 30 minutes following euthanasia. The colic case group comprised horses over 1 year of age undergoing exploratory celiotomy for colic unresponsive to medical management that required resection of a segment of the small intestine. Case inclusion was also dependent on the feasibility of appropriate sample collection and processing by the surgeon on duty. One experienced surgeon from each centre performed the exploratory celiotomies, intestinal resections and sample collection. All colic cases received the same pre-and postoperative care (Data S1).
Signalment was recorded in both control horses and colic cases.
For each colic case, site of lesion and resection, duration of colic prior to surgery, presence of pre-operative reflux, resection length, shortterm survival (defined as surviving to hospital discharge) and the presence of POR were recorded. Pre-operative reflux was defined as the presence of >2 L of reflux obtained on passing a nasogastric tube prior to surgery and POR was defined as the presence of >2 L of reflux on more than 2 intubations in the post-operative period. Both owner consent and ethical approval were obtained for the collection of tissues; ethical approval was provided by the Royal (Dick) School of Veterinary Studies Veterinary Ethical Review Committee.

| Sample collection
Control samples were obtained from an antimesenteric location in the region of the mid jejunum. For the colic cases, the resected intestine was deemed by the surgeon to have healthy grossly viable resection margins. Samples were obtained from both proximal and distal resection margins of either jejunum or ileum. All proximal resection margins were obtained from the jejunum and distal resection margins were either obtained from the jejunum (n = 9) or ileum (n = 3).

| RNA extraction and cDNA synthesis
Tissue sections were thawed, then homogenised in the presence of β-mercaptoethanol and transferred to a gDNA Eliminator column (Qiagen) before extraction of total RNA using an RNeasy Plus Mini kit (Qiagen). RNA was quantified and purity was assessed using a Nanodrop spectrophotometer (ThermoFisher Scientific) and integrity was measured using the TapeStation System (Agilent Technologies). An RNA integrity number (RIN e ) greater than 7 was considered sufficient for downstream analysis 12 and cDNA was synthesised from 1 µg of total RNA using the SuperScript III First-Strand synthesis system (Invitrogen).
It was not possible to extract RNA of adequate quality from all the colic cases and control samples. A total of 9 muscularis externa samples (5 proximal and 4 distal) and 16 mucosal samples (8 proximal and 8 distal) from colic cases and 4 muscularis externa and 6 mucosal samples from control horses were used for RT qPCR analysis. For colic cases, samples from the proximal and distal resection margins were analysed. After preliminary analysis comparing relative gene expression between proximal and distal resection margins with control tissues, the proximal mucosa (n = 8) and distal mucosa (n = 8) samples were combined as were the proximal and distal muscularis externa sections ( Figure 1). Proximal and distal margins obtained from the same case were not used. Where a proximal and distal sample existed for a colic case, the proximal section was selected as all proximal resection margins were from the jejunum and therefore better matched the control samples (Table S1).
For RT-qPCR, cDNA was amplified with Power SYBR Green PCR Master Mix using the 7500 fast Real Time PCR system (ThermoFisher Scientific). Primer efficiency was validated with a standard curve of five serial dilution points (three for NOS2) with efficiency ranging between 92.68% and 100.74%. Amplification reactions were performed in triplicate and the melting curve analysed to check the annealing temperature and to ensure that there was no primer dimer formation. Efficiency and reactions were analysed using 7500 Software v2.3 (ThermoFisher Scientific). Details of primer sequences and design are included in Table S2.
The expression of target gene mRNA relative to the housekeeping gene, glyceraldehyde 3-phosphate dehydrogenase (GAPDH), was calculated for each sample using the 2^ (−delta threshold cycle (CT)) (2^− ΔΔCT ) method. 13 Relative gene expression was assessed in the mucosa and muscularis externa of samples from colic cases and control horses.
F I G U R E 1 Flow chart describing process of sample selection following quality control and preliminary analysis. *For each case, either a proximal or distal sample was selected

| Data analysis
Where appropriate, data are reported as median and interquartile range. All statistical analyses were performed using GraphPad Prism

(GraphPad Software; GSL Biotech LLC). A non-parametric
Mann-Whitney U test was applied to determine significant differences in relative gene expression between proximal and distal margins of the mucosa and muscularis externa; the mucosa samples obtained from the colic cases and the control horses; the muscularis externa samples obtained from the colic cases and the control horses; horses that did and did not survive to discharge; and horses with and without pre-and post-operative reflux. Linear regression analysis was used to analyse the relationship of age, duration of colic and resection length with relative gene expression. Normality was measured using D'Agostino's K 2 test. Significance was assumed at P < .05.

| RE SULTS
The colic cases comprised 12 horses (median age 19.5 years, range:

| Relative expression of target genes in the mucosa and muscularis externa of colic cases and controls
No significant differences in relative gene expression of target genes were observed between proximal and distal resection margins for both the mucosa and muscularis externa prior to the samples being combined ( Figures S1 and S2). NOS2 was removed from the set of target genes as no mRNA was detectable in any of the mucosa and muscularis externa samples.
Relative expression of the target genes was evaluated in the mucosa and muscularis externa of colic cases and compared with the mucosa and muscularis externa of control horses ( Figure 2

| Analysis of relative expression in colic cases with relation to age, duration of colic, resection length, short-term survival and the presence of pre-and post-operative reflux
There were no significant associations in either the mucosa or muscularis externa between inflammatory gene relative expression and the following factors: presence of pre-operative reflux; length of resected intestine; survival to discharge and duration of colic.
In the mucosa, relative expression of all genes, except for TNF, reduced with age. There was a significant decrease in the relative expression of PTGS2 and CCL2 with increasing age (P < .05) ( Figure S3). In the muscularis externa, all genes, with the exception of IDO1, showed a trend of increased relative expression with increasing age; however, this apparent association was not statistically significant ( Figure S4). Mucosal samples from colic cases that developed POR demonstrated a significantly greater relative expression of TNF when compared with mucosal samples from colic cases that did not develop POR (Figure 4). There was no statistically significant difference in relative gene expression in the muscularis externa of colic cases that developed POR when compared with colic cases that did not develop POR ( Figure 5); however, those that developed POR showed a greater median relative expression of IL1β, IL6, PTGS2, TNF and CCL2.

| D ISCUSS I ON
The principal aim of this study was to examine the intestinal inflammatory response in both the mucosa and muscularis externa F I G U R E 2 Relative gene expression of IL1B, IL6, PTGS2, TNF, CCL2 and IDO1 in the mucosa and muscularis externa of horses undergoing colic surgery. Scatter plots showing relative mRNA expression of target genes relative to Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) in the distal mucosa (n = 12) and combined proximal and distal muscularis externa (n = 9) samples compared with controls (mucosa n = 6, muscularis externa n = 4). Significance of relative gene expression levels between control and surgical samples using a Mann-Whitney U test. ns = not significant, *P < .05, **P < .01, ***P < .001, ****P < .0001   lipopolysaccharide, and did not upregulate any genes involved in arginine metabolism. 22 These data, and the data from this study suggest that NO production from macrophages does not contribute to the pathophysiology of equine POI.
The cytokines IL1β and TNFα reduce intestinal motility via a reduction in smooth muscle function, either by direct action on the smooth muscle cell or an indirect suppression of neurotransmitter release. 23,24 IL6 also suppresses motility, although the exact F I G U R E 5 Relative gene expression of IL1B, IL6, PTGS2, TNF, CCL2 and IDO1 in the muscularis externa of horses with and without postoperative reflux. Scatter plots showing mRNA expression of target genes relative to Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) in muscularis externa of colic cases (n = 7) and horses that developed post-operative reflux (n = 3). Significance of relative gene expression levels in horses with pre-operative reflux was assessed using a Mann-Whitney U Test  Several factors such as age and the presence of pre-operative nasogastric reflux have been associated with an increased incidence of POI in horses. 32 In rodent models, a greater inflammatory response is associated with more "severe" POI. 33 Analysis was performed to determine if any association with pre-, intraoperative and post-operative factors was present. In the intestinal mucosa, there was a trend towards a reduction in the relative expression of all target genes, with the exception of TNF with age. The reduction in the relative expression of PTGS2 and CCL2 was statistically significant.
In contrast, the relative expression of all target genes in the muscularis externa increased with age although 6 out of the 12 colic cases were diagnosed with pedunculated lipomas, the incidence of which is significantly greater in older horses. 34 In contrast, duration of colic was associated with a non-significant decrease in relative gene expression across all genes in both the mucosa and muscularis externa. The outlying data points were predominantly from four cases.
There was no evidence of any differences in disease process or signalment to account for these phenomena. The study of larger group sizes may have revealed specific factors associated with a greater inflammatory response, thus accounting for the response observed in the outlying individuals. It is also possible that host genetics may play a role in the severity of inflammation and, like that seen in humans, some individuals may be genetically predisposed of developing a more severe inflammatory response. 35 The main limitations of the current study were the small group sizes. The inclusion of larger group sizes would be required to definitively explore the validity of certain association trends identified by the current study. Based on our preliminary data, power calculations revealed that a minimum group size of 25-60 horses (depending on gene of interest) would be required to detect a significant correlation between relative gene expression and horses that do develop POR.
Although the use of combined proximal and distal resected margin samples could also be considered as a limitation, preliminary analysis failed to reveal any statistically significant differences in relative gene expression between these different locations. In the colic case group, there was also variation in the length of resected intestine, the underlying disease process, the disease duration prior to sample collection and the age of the horses. Although these sources of variation will inevitably impact upon the interpretation of the results, they are difficult to control within the constraints of the field study nature of the study design. Lastly, some of the distal resection margins were obtained from the ileum (3 out of 12 cases), which may have different resident inflammatory cell numbers compared with the control jejunal samples. While we have previously shown a difference between distal jejunum and ileum with respect to macrophage numbers in the submucosa, this difference was not evident in other layers including the muscularis externa. 36 That said, quantification of other resident immune cells in the equine intestine has not been performed and remains a significant knowledge gap. A further limitation of the study relates to the inherently restricted inferences which can be derived solely from consideration of gene expression data; this may not fully mirror the protein transcription profile at the tissue level. Expansion of this work to include proteomic profiling and pathway analyses would be warranted to more fully elucidate the factors involved in the initiation and propagation of the inflammatory response.
While correlation of relative gene expression with several perioperative factors (e.g. development of POR) would have added further clinical relevance to the study, this was considered an adjunct to the primary objective of the study; namely, to extend findings from rodent and human studies to samples obtained from horses undergoing abdominal surgery. In this respect, our data confirmed the presence of an inflammatory response, characterised by differential inflammatory gene expression, in the intestine of horses undergoing colic surgery. As such, these findings justify continued research in this area and provide a methodological platform for such. The inclusion of a larger cohort of horses would greatly facilitate any future efforts to further assess whether associations exist between the magnitude of the inflammatory response and certain clinically relevant factors such as age, resection length, duration of colic, short-and long-term survival and the development of POR. Furthermore, further work to determine the roles of different cell types such as neutrophils and macrophages in intestinal inflammation is also warranted.

| CON CLUS IONS
These data demonstrate an inflammatory response within the intestine of horses undergoing colic surgery during which a small intestinal resection was performed. The upregulated genes encode proteins with the capability of inhibiting smooth muscle contractility and disrupting normal intestinal motility, ultimately resulting in functional POI. Additionally, the absence of NOS2 upregulation highlights inter-species variation in this inflammatory pathway, thus emphasising the importance of considering the target species when developing potential therapeutic targets. This study provides a foundation for future work to improve our understanding of the inflammatory response in horses undergoing colic surgery.

ACK N OWLED G EM ENTS
We thank Craig Pennycook and Chandra Logie at R(D)SVS Pathology for their assistance with tissue collection and to colleagues at the R(D)SVS Equine Hospital and Bell Equine Veterinary Clinic for their help in obtaining samples and Darren Shaw for his input on statistical methods. Additionally, we thank the owners of all the horses who gave permission for the collection of samples.

CO N FLI C T O F I NTE R E S T S
No competing interests have been declared.

I N FO R M E D CO N S E NT
Owners gave consent for their animals' inclusion in the study.

DATA ACCE SS I B I LIT Y S TATE M E NT
The data that support the findings of this study are available from the corresponding author upon reasonable request.