Adipose tissue-derived adiponectin expression is significantly associated with increased post operative mortality in horses undergoing emergency abdominal surgery




Reasons for performing study: Adipose tissue is an important source of inflammatory cytokines (adipokines) and adiposity has been identified as having a significant effect on human morbidity and mortality. Obesity is also an emerging welfare problem in the UK horse population, but the role that it plays in secondary diseases is unclear.

Objectives: To examine the expression of inflammation-related adipokine genes in retroperitoneal adipose tissue of horses undergoing emergency abdominal surgery and to explore associations with adiposity and post operative survival.

Methods: Retroperitoneal adipose tissue samples were obtained from 76 horses undergoing emergency abdominal surgery. Real-time PCR was used to measure gene expression for leptin, adiponectin, tumour necrosis factor-alpha, macrophage chemoattractant protein-1, macrophage inhibitory factor, serum amyloid A, haptoglobin and interleukin-1. Multivariate patterns of adipokine expression were explored with principal component analysis (PCA), whilst univariable associations with post operative survival were tested in a Cox proportional hazards model.

Results: Leptin gene expression was higher in overweight and obese horses than in lean animals. Expression of mRNA encoding adiponectin mRNA in visceral adipose tissue was positively associated with increased post operative mortality (hazard ratio 1.31, 95% CI 1.05–1.65). However, PCA did not demonstrate multivariable patterns of adipokine gene expression from visceral adipose tissue associated with body mass index or with survival.

Conclusions: In horses presented with acute intestinal disease, increased adiponectin gene expression from retroperitoneal adipose tissue is associated with an increased risk of mortality. Obesity assessed by BMI had no association with increased post operative mortality in horses with primary gastrointestinal disease.

Potential relevance: Further study is warranted on the expression and effects of adipokines, particularly adiponectin, and correlation with postoperative outcome.


In addition to storage of triglycerides, white adipose tissue (WAT) is an influential endocrine organ, acting as a regulator of normal physiological processes (Trayhurn and Wood 2004). The tissue is comprised principally of adipocytes, pre-adipocytes, endothelial cells, fibroblasts, leucocytes and macrophages (Tilg and Moschen 2006). In various species, including man, dogs, cats and horses, WAT synthesises and secretes protein factors (referred to as adipokines, when produced from adipose tissue, e.g. leptin and adiponectin), some of which are involved in inflammatory pathways (Trayhurn and Wood 2004; Radin et al. 2009; German et al. 2010). Adipokines have also been associated with the development of a chronic inflammatory state in subjects which possess excessive WAT (Vachharajani and Vital 2006; Andersson et al. 2008; German et al. 2009) and, in man, obesity is associated with increased post operative morbidity and mortality (Bercault et al. 2004; Goulenok et al. 2004).

Obesity is becoming an increasing problem in horse populations around the world (Sillence et al. 2006; Thatcher et al. 2007; Salonen et al. 2009), with an estimated 45% being obese (Wyse et al. 2008). As in humans, equine obesity is associated with insulin resistance, which may predispose horses to laminitis and hyperlipaemia (Watson et al. 1992; Treiber et al. 2006).

Horses undergoing emergency intestinal surgery are commonly affected by endotoxaemia (Burrows 1981; Werners et al. 2005), and this is known to increase post operative mortality risk (Morton and Blikslager 2002; Proudman et al. 2002; Mair and Smith 2005). Sepsis is characterised by dysregulation of coagulation pathways and uncontrolled systemic inflammatory response to pathogens (Vachharajani and Vital 2006). Given that macrophages are present in WAT, it is unsurprising that the tissue has been linked to human inflammation and sepsis (Tilg and Moschen 2006). Glucose metabolism and lipolysis are also altered in both sepsis and obesity, and adipose tissue macrophage chemoattractant protein-1 (MCP-1) and interleukin-1 (IL-6) gene expression increases substantially in endotoxaemic mice (Leuwer et al. 2009). However, no equivalent information currently exists in horses with acute intestinal disease. Therefore, the aim of this study was to examine inflammatory adipokine gene expression in retroperitoneal fat from horses undergoing emergency intestinal surgery and to explore potential associations with post operative mortality.

Materials and methods

Horse population

Adult horses with a clinical diagnosis of acute intestinal disease, that presented to the Phillip Leverhulme Equine Hospital, University of Liverpool, between February 2007 and December 2008, were enrolled in the study. It was a further study requirement that each horse had undergone exploratory coeliotomy. Horses that were subjected to euthanasia or died prior to recovery from general anaesthetic were excluded, as were those with incomplete clinical records and missing data. The study was performed according to the University of Liverpool Guidelines on Animal Ethics and only used tissue that was otherwise to be discarded as clinical waste. Informed consent was obtained, in writing, from the owner of each horse in the study.

Sample and clinical data collection

Samples of retroperitoneal WAT were removed during closure of the midline coeliotomy incision, immediately frozen in liquid nitrogen and stored at -80°C until analysis. The site of collection was the ventral midline, 2–30 cm cranial to the umbilicus, a site comparable to that used in previous canine studies (Eisele et al. 2005; Ryan et al. 2010).

The age of the horse, gender and breed were recorded. Morphometric and clinical parameters collected included weight (kg), height (cm), heart rate at presentation to the hospital (recorded as beats/min), packed cell volume (PCV; measured in %), and total blood protein concentration (TP, measured in g/l). The duration of intestinal disease and the type of nonsteroidal antinflammatory drug (NSAID) administered prior to referral was also recorded. Intraoperatively, the final diagnosis and resection of intestine were noted. Post operative survival data were collected by periodic telephone questionnaires, carried out every 3 months after the horse was discharged from hospital in the first year following surgery, then every 6 months thereafter until the end of the study period (September 2009). Outcomes of interest were death of the horse and duration of post operative survival.

RNA extraction and reverse transcription polymerase chain reaction (RT-PCR)

Total RNA was extracted from the adipose tissue (50–100 mg) with TRIZOL Reagent1. The RNA concentration of each sample was determined from the absorbance at 260 nm and 0.5 mg was reverse transcribed to cDNA with anchored oligo (dT) primer using an Iscript cDNA Synthesis Kit2. One microlitre of the RT product was used as a template for PCR using PCR Master Mix2. RT and PCR were performed according to the manufacturer's protocols on PCR Express Hybaid thermal cyclers3 (Ryan et al. 2008).

Equine-specific primers were designed for leptin, adiponectin, IL-6, tumour necrosis factor-alpha (TNFα), serum amyloid A (SAA), haptoglobin, MCP-1 and macrophage inhibitory factor (MIF) using Beacon Designer 44 and synthesised commercially5. Sequence data are presented in Table 1. Equine β-actin and equine succinate dehydrogenase complex subunit A (SDHA) were validated and used as the reference genes in equine adipose tissue. Validation of equine β-actin and SDHA was performed using specific software programs BestKeeper (Pfaffl et al. 2004), GeNorm (Vandesompele et al. 2002) and NormFinder (Andersen et al. 2004).

Table 1. PCR Primer sequence data for adipokines assayed designed by Beacon Designer 44
CytokineSense sequenceAnti-sense sequence

The amplification parameters for PCR were: 94°C for 4 min for one cycle, 94°C for 20 s, then 56°C for 30 s, 72°C for 40 s for 25 cycles and 72°C for 10 min. PCR products were separated by electrophoresis on a 1% agarose gel and stained with ethidium bromide.

Real-time PCR

Quantitative changes in the level of mRNAs encoding leptin, MCP-1 and IL-6, were measured using SYBRgreen by relative quantification with the 2-ΔΔCt method (Livak and Schmittgen 2001) on a Stratagene Mx3005P cycler6. All samples were normalised to values of either equine β-actin or SDHA. Results were expressed as ‘fold changes’ in threshold cycle value (Ct) relative to controls.

Real-time PCR was performed in 96 well plates using a qPCR Core Kit5 according to the manufacturer's instructions. Primers were designed by Beacon Designer 44 and synthesised commercially5 as described above. Amplifications were performed commencing with 95°C for 10 min for one cycle, 95°C for 30 s, 60°C for 60 s for 40 cycles, then 95°C for 60 s, 55°C for 30 s, and 95°C for 30 s for one cycle. Data were collected and analysed using MxPro QPCR software6.

Data handling and statistical analysis

A database (Microsoft Access 2007)7 was constructed to record morphometric data, clinical data, RT-PCR results and survival data. In addition, body mass index (BMI) was calculated using weight (kg)/height of horse at the withers2 (cm2). Horses were categorised as lean with a BMI of <200, overweight with BMI 200–220, and obese with a BMI of >220 (Donaldson et al. 2004). Real-time PCR data were entered into the database as change in critical threshold values (ΔCT) and also transformed into a 2-ΔΔCT value to allow expression of the data as ‘fold changes’ (Livak and Schmittgen 2001). The normality of frequency distributions for age, height, weight and adipokines were assessed by visual inspection of frequency distribution histograms and by performance of the Anderson-Darling normality test.

Association in expression amongst different adipokine genes was assessed by Pearson's rank correlation coefficient. Correlation structures within the dataset were further explored with principal component analysis (PCA; SIMCA-P+)8. A PCA model was generated using unit variance scaling and without transformation of the original data, trends in adipokine gene expression were explored and class models created to investigate potential differences in correlation structure between lean and overweight/obese animals. The predictive value of adipokine gene expression was explored using an orthogonal partial least squares-discriminant analysis (OPLS-DA). This analysis provided graphical and statistical evidence of the predictive variation in levels of adipokine gene expression.

Survival analysis was performed by using survival time in a Cox proportional hazards model (S-Plus)9. Univariable associations between adipokine expression and survival time were explored; Kaplan-Meier survival curves and smoothing splines were generated to illustrate these relationships. The significance of observed associations was determined by evaluation of 95% confidence intervals (CI) and using the likelihood ratio test statistic. A critical value of 0.05 was used to infer significance for hypothesis tests.


The study population comprised 70 horses of mean ± s.d. age 10.7 ± 6 years and median age 10.5 years (range 1–28 years); median height 160 cm (range 82–178 cm) and median weight 534 kg (range 93–780 kg). Age data were normally distributed; however, height and weight data were not normally distributed, so the median values were used. Forty-two horses were male (one entire, the others castrated) and 28 were female. Twenty animals (26%) were ponies (including miniature breeds), 19 animals (25%) were Thoroughbreds or Thoroughbred crosses, 12 (15%) were Warmbloods and the remaining 19 horses represented a variety of other breeds. In the lean group of 18 horses, there were 4 horses with small intestinal lesions that had caused ischaemic intestine and 2 horses with small intestinal lesions with no ischaemic intestine. Nine horses had primary large intestinal lesions without a compromised intestinal blood supply, 2 had large colon lesions with compromised blood supply and one horse had a simple obstruction of the small colon. In the overweight group of 24 horses, 5 horses had ischaemic small intestinal lesions, 7 had simple obstructions of the small intestine, 4 had large colon lesions with compromised intestinal blood supply and 8 had large colon lesions with no compromised blood supply. In the obese group, 8 horses had ischaemic small intestinal lesions, whereas 8 horses had small intestinal lesions without compromised intestinal blood supply. Three horses had large colon lesions with compromised blood supply and 9 horses had large colon lesions without compromised intestinal blood supply (including 2 cases of colitis).

All of the adipokine genes examined (leptin, adiponectin, IL-6, TNFα, MCP-1, MIF, haptoglobin and SAA) were expressed in equine retroperitoneal adipose tissue, as demonstrated by RT-PCR (Fig 1). The signal for SAA was, however, very weak. Quantification of adipokine gene expression in the fat samples is shown in Figure 2. Leptin gene expression was significantly higher in horses with higher body mass index. There was a significant difference between the lean and overweight group and between the lean and obese group (P = 0.004). The mRNA encoding the adipokines MCP-1 and IL-6 were strongly correlated with each other (r2= 0.85), and both appeared to be more abundantly expressed in lean than in overweight and obese horses with the lowest levels being in the obese group; however, this was not significant (MCP-1, P = 0.474, IL-6, P = 0.215). There was no apparent relationship between BMI and gene expression for the remaining adipokines (adiponectin, TNFα, MIF, SAA and haptoglobin).

Figure 1.

Representative PCR gel for all eight adipokines. 100 bp ladder each side. From left to right: leptin, adiponectin, IL-6, MCP-1, haptoglobin, SAA, MIF and TNFα.

Figure 2.

Fold changes in retroperitoneal adipokine gene expression measured by RT-PCR. Data expressed as group median with 95% confidence interval. The y-axis represents expression relative to the reference group of lean horses, L = lean, OW = overweight, OB = obese (see text for definitions).

Principal component analysis revealed one strong outlier (outside the 95% CI for Hotelling's T2 range), a horse with high levels of gene expression for all adipokines measured, and this was excluded from further analysis. There was no observable clustering of data when classified by age, breed, gender or presence of ischaemic intestine, as determined at surgery, when intestine was resected. Classification of observations by BMI also demonstrated no systematic variation (Fig 3a). Separate class models were used to explore adipokine gene expression in lean horses vs. overweight and obese horses; in nonsurviving lean horses, higher adiponectin and leptin levels and lower MCP-1 and IL-6 were seen, but no such pattern was observed in either overweight or obese animals (data not shown). OPLS-DA models were generated to explore differential patterns of adipokine expression. No discrimination was evident with observations classified by BMI. Further, poor discrimination was observed when classified as survival vs. nonsurvival (Q2(cumulative) = 0.018, Fig 3b).

Figure 3.

(a) Scores plot of PCA model for all 8 cytokines measured, observations coloured by BMI classification: blue = lean, green = overweight, red = obese. (b) Scores plot for OPLS-DA model with observations classified by survival (black) vs. nonsurvival (red).

Associations between post operative survival, adipokine gene expression, and BMI were analysed in Cox proportional hazards models (Table 2). Only adiponectin gene expression, at the time of surgery, was significantly and positively associated with an increased rate of post operative mortality. Analysing adiponectin gene expression as a continuous variable in this model gave a positive hazard ratio (1.31, 95% CI 1.05–1.65) indicating an increasing risk of death with increasing levels of adipokine gene expression. Data were then categorised into 2 groups of equal size (high and low values), using the median value as the cut-point, and presented as a Kaplan-Meier plot (Fig 4a). However, there was no association between BMI and post operative survival time. High post operative mortality in horses with high adiponectin expression was particularly seen in the early post operative period (in the region of 20 days post surgery) and the continuous relationship between adiponectin mRNA expression and post operative mortality is further illustrated by the P-spline smoothing plot (Fig 4b).

Table 2. Univariable associations (Cox proportional hazards model) between adipokine expression in retroperitoneal fat and post operative survival following surgery for acute intestinal disease
AdipokineCoef.HR95% CILRTS P value
  1. HR = hazard ratio, CI = confidence interval, LRTS = likelihood ratio test statistic.

Figure 4.

a) Kaplan Meier plot illustrating the rate of death in horses categorised by adiponectin gene expression at the time of surgery. Solid line = high adiponectin, dashed line = low adiponectin. b) P-spline smoother illustrating the relationship between adiponectin gene expression and risk of post operative death. Dotted lines indicate 95% confidence intervals.


This study reports novel information about adipokine gene expression in the retroperitoneal fat of horses with acute intestinal disease. We present evidence that; 1) leptin gene expression in visceral adipose tissue is higher in horses with greater BMI, 2) IL-6 and MCP-1 gene expression decreases with increased BMI and 3) adiponectin gene expression is positively and significantly associated with post operative mortality.

Leptin was the only adipokine whose gene expression increased with increasing BMI, a finding consistent with previous studies in horses that have demonstrated increasing levels of circulating leptin associated with increasing fat mass (Buff et al. 2002; Kearns et al. 2006). Kearns et al. (2006) also reported that circulating adiponectin was inversely proportional to adiposity and, in both rodents and man, both adiponectin gene expression in adipose tissue and circulating adiponectin concentration fall with increasing adiposity (Arita et al. 1999; Więcek et al. 2007). The reasons why a similar association was not found in the current study are not known, but may relate to differences in study population (i.e. exclusive assessment of a population with disease).

A further finding of the present study was the lack of association between BMI and expression of some adipokine genes in retroperitoneal adipose tissue. Increased leptin and decreased MCP-1 and IL-6 gene expression were seen in overweight and obese horses, but, there were no clear patterns of expression for the other adipokines, including TNFα. This suggests that BMI cannot be used to predict the adipokine gene expression profile of adipose tissue in individual animals with primary gastrointestinal disease. Our search for other sources of systematic variation in adipokine gene expression revealed some differences in clustering of observations, when classified by post operative survival in lean horses, which was not observed in overweight/obese animals. This suggests that the influence of inflammatory adipokines on post operative survival may be different in horses with different levels of adiposity.

The association between adiponectin gene expression and post operative survival time was an interesting and unexpected finding, since this adipokine is frequently characterised as an anti-inflammatory factor (Więcek et al. 2007). Therefore, the anti-inflammatory function of adiponectin may be detrimental to post operative survival, especially in the early post operative period. There are similarities in some human studies; for example, asymptomatic subjects with high adiponectin levels have lower risk of artherosclerotic disease, due to the protective effect of adiponectin (Schnabel et al. 2008) and, in patients with symptomatic disease, elevated circulating adiponectin concentration is associated with an increased risk of cardiovascular events occurring (Pilz et al. 2006; Schnabel et al. 2008). Therefore, as was the case in our study, it is possible that when active disease is present, the anti-inflammatory role of adiponectin may have an adverse effect on survival. This mechanism has been termed ‘risk factor reversal’ and is due to a whole system imbalance resulting from the severity of underlying disease (Schnabel et al. 2008). We propose that this mechanism of a counter-regulatory anti-inflammatory response to inflammatory mediators, leading to higher gene expression of adiponectin, may explain the observed relationship between survival and adiponectin in horses with acute intestinal disease; however, due to very limited knowledge about the biology of adiponectin in horses, we must be careful not to draw parallels with findings from other species There may be differences in the anti-inflammatory response between horses with acute and chronic disease but this would require further study as the horses in this study were all suffering from acute abdominal diseases.

The inverse relationship between MCP-1 and IL-6 gene expression and increasing BMI values contrasts with the findings that IL-6 is elevated in obese human subjects (Miller et al. 2002). However, IL-6 is also elevated in hypertensive, slim children (Glowinska and Urban 2003) and may therefore be more closely linked to disease status than to body fat mass. Previous studies have demonstrated that increased circulating IL-6 concentration is associated with increasing age in both horses and humans (Harris et al. 1999; Adams et al. 2008). However, a similar association was not evident in our study, which may again suggest differences in methodology, study population and outcome measure (adipose tissue gene expression vs. circulating protein). Circulating MCP-1 concentration and adipose tissue gene expression are elevated in obese mice (Takahashi et al. 2003), whilst increased circulating MCP-1 concentration is associated with increasing adiposity in women with polycystic ovary syndrome (Glintborg et al. 2009) and increased cardiovascular mortality in human patients (Piemonti et al. 2009). Inflammatory cytokines are also produced by nonadipose tissue, such as monocytes and macrophages, and a complex relationship exists between expression of adipokines from adipose tissue and from these other sources. It is suggested that age related obesity has a role in dysregulation of inflammatory cytokine production from the immune system (Adams et al. 2008); however, we found no evidence for this in our study.

The current study has a number of limitations given the diverse nature of the outbred population studied and limited sampling strategy (single site of adipose tissue). Variability could feasibly have arisen from differences in age, weight, breed and type of disease; however, PCA did not reveal any systematic variation associated with these variables. Gene expression of adipokines from adipose tissue was measured; however, protein quantification was not performed. Further, errors may have been introduced with inaccuracies in determination of obesity, as assessment of equine body fat is extremely difficult and no current methods of assessing overall equine adiposity are ideal (Dugdale et al. 2010a). We used BMI as a measure of equine adiposity which, although not properly validated, correlates relatively well with body condition score (Donaldson et al. 2004). BMI calculation in the present study used weight as measured on a calibrated equine weighbridge10. Breed of horse was also recorded in the database to control for variation in conformation between breeds. BCS and ‘cresty neck’ score could also have been assessed; however, there was concern over the subjective nature of these scoring systems in this study. Deuterium oxide (D2O) dilution method is a useful measure of adiposity (Dugdale et al. 2010b) but was beyond the scope of the current study.

A further limitation of the study was the fact that adipokine gene expression was assessed in adipose tissue from a single anatomic location. Therefore, extrapolation of our findings should be made with caution, since a comprehensive picture of adipokine response to acute intestinal disease would require both measurement of circulating adipokines and gene expression in a variety of adipose tissue sites. Recently, nuchal adipose tissue, which has been shown to be a more active site for production of some inflammatory adipokines than visceral fat in horses without gastrointestinal disease (Carter et al. 2008; Burns et al. 2010).

We conclude that, in a population of horses with acute intestinal disease, increased gene expression of adiponectin from retroperitoneal adipose tissue was associated with an increased risk of mortality. Leptin gene expression in the same tissue was greater in overweight and obese horses, but multivariable patterns of adipokine expression were not identified that were significantly associated with adiposity (BMI). Further studies are recommended to determine the putative role of inflammatory adipokines in horses with acute intestinal disease.

Conflicts of interest

There are no conflicts of interest from any author.

Source of funding

The study was funded by a grant from The Horse Trust. P.T. is a member of COST BM0602. A.J.G.'s Senior Lectureship is funded by Royal Canin.


The authors wish to acknowledge the referring veterinarians for referring cases, the owners of all horses, the clinical staff at the University of Liverpool for assistance with case management and S.E. Taylor for her help with reference gene validation. The contribution made by each author is as follows: M.P. was responsible for sample collection, adipokine assays and data presentation, A.G., L.H. and P.T. supervised qPCR, C.P. was responsible for initiating the study and data analysis. All authors contributed to the design of the study and to writing the manuscript and have approved the final version.

Manufacturers' addresses

1 Invitrogen Ltd, Paisley, Renfrewshire, UK.

2 Bio-Rad, Hercules, California, USA.

3 Rhys Scientific Ltd, Chorley, Lancashire, UK.

4 Premier Biosoft International, Palo Alto, California, USA.

5 Eurogentec Ltd, Romsey, Hampshire, UK.

6 Stratagene, Agilent Technologies, Santa Clara, California, USA.

7 Microsoft UK Headquarters, Reading, Berkshire, UK.

8 Umetrics, Umea, Sweden.

9 Insightful Corp., Bagshot, Surrey, UK.

10 Horse Weigh, Llandrindod Wells, Powys, UK.