The relationship between a predisposition to obesity and the development of colitis is not well understood. Our aim was to characterize the adipokine response and the extent of colitis in diet-induced obese (DIO) rats. DIO and control, diet-resistant (DR) animals were administered either saline or trinitrobenzene sulfonic acid (TNBS) to induce colitis. Macroscopic damage scores and myeloperoxidase (MPO) activity were measured to determine the extent of inflammation. Trunk blood was collected for the analysis of plasminogen activator inhibitor-1 (PAI-1) as well as leptin, ghrelin, and adiponectin. Colonic epithelial physiology was assessed using Ussing chambers. DIO rats had a modestly increased circulating PAI-1 before TNBS treatment; however, during colitis, DR animals had more than a fourfold increase in circulating PAI-1 compared with DIO rats. Circulating leptin was higher in DIO rats compared with DR animals, in the inflamed and noninflamed states. These changes in TNBS-induced adipokine profile were accompanied by decreased macroscopic tissue damage score in DIO animals compared with DR tissues. Furthermore, TNBS-treated DR animals lost significantly more weight than DIO rats during active inflammation. Colonic epithelial physiology was comparable between groups, as was MPO activity. The factors contributing to the decreased colonic damage are almost certainly multifold, driven by both genetic and environmental factors, of which adipokines are likely to play a part given the increasing body of evidence for their role in modulating intestinal inflammation.
- diet-induced obese
- body weight
- plasminogen activator inhibitor-1
inflammatory bowel disease (IBD) occurs in genetically predisposed individuals as a result of an incompletely understood gene-environment interaction, also involving the intestinal microflora (3). Similarly, obesity, or more specifically diet-induced obesity (DIO), develops as a consequence of environmental factors, diet for example, but also genetic factors because not all animals (27) or humans (7) who display a preference for high-fat diets actually develop obesity. Studies in DIO-prone rats implicate decreased central leptin sensitivity (26), hypothalamic pituitary dysfunction (32), maternal obesity (19), and postweaning exercise (24a) as among some of the factors that contribute to the DIO phenotype.
Although DIO animals display increased gene-expression profiles for several immune response genes, including those involved in antigen presentation, response to biotic stimuli, and phagocytosis (29), there is also a growing recognition that endocrine factors released from adipose tissue are involved in the pathophysiology of many chronic inflammatory conditions (17, 20, 42), including IBD (23). The accumulation of central adipose tissue and changes in the expression of the proteins synthesized and released by adipocytes (adipokines) are thought to underlie the abnormalities that characterize aspects of the metabolic syndrome including leptin resistance, chronic inflammation, and impaired fibrinolysis (17, 21). Collectively, the growing list of adipokines has redefined adipose tissue as a complex endocrine organ that regulates immunity and inflammation (17, 20, 42), in addition to energy balance (43).
As well as altering the course of colitis, adipokines can also contribute to homeostatic imbalances during IBD (11). Plasminogen activator inhibitor-1 (PAI-1) is produced by a number of tissues including adipocytes (1, 38) and is increased in obese humans (22) and rodents (39). PAI-1, which promotes a hypercoagulant state, is significantly increased in plasma taken from patients suffering from IBD (13) and may be associated with coagulation abnormalities that accompany intestinal inflammation (11).
Studies in leptin-deficient, ob/ob mice (4, 40) further support a role for adipokines as direct modulators of colonic inflammation. In these studies, it was shown that leptin infusion worsened colitis and that leptin-deficient mice developed less severe experimental colitis than wild-type controls. Leptin circulates in direct proportion to fat stores and is almost exclusively released into the circulation from adipocytes, and its receptor is expressed both in the central nervous system and periphery (16). The protective effect of the ob/ob genotype was abolished by leptin infusion, suggesting that leptin, not obesity per se, increased damage caused by experimental colitis. Other adipokines such as adiponectin may also influence the development of colitis. The role played by this adipokine during colitis in gene-deficient animals is controversial (15, 33).
A similar, somewhat ambiguous situation exists for the feeding stimulant ghrelin. Administration of this adipokine 12 h after trinitrobenzene sulfonic acid (TNBS) treatment significantly decreased TNBS-induced damage and proinflammatory cytokine production (18), indicative of an anti-inflammatory effect. However, in human colonic cell lines, ghrelin increases TNF-α-induced proinflammatory interleukin-8 production and activates NF-κB (47).
These studies demonstrate that circulating adipokines have the potential to influence the development of IBD, but to date it is unclear whether DIO in animals predisposed to this condition alters the development of intestinal inflammation. A number of factors may contribute positively or negatively to the development of inflammation, including the enteric flora, which is altered in obesity (3, 28, 45), and the extent of mesenteric fat deposition. Understanding how the mechanisms governing intestinal inflammation are altered in obesity is of importance given its increasing prevalence throughout the world (24). Therefore, the aim of the present study was to determine whether experimental colitis is altered in DIO rats.
MATERIALS AND METHODS
Animals and diet.
All methods used in this study were approved by the University of Calgary Animal Care Committee and were carried out in accordance with the guidelines of the Canadian Council on Animal Care. Animals were housed in a temperature-controlled room maintained on a normal 12-h:12-h light/dark cycle and were allowed access to food and water ad libitum.
DIO and DR animals.
Male selectively bred diet-resistant (DR) and DIO rats [strain Crl: CD(SD) DIO or DR; Charles River Laboratories, Montréal, QC, Canada] were 6 wk of age when they were exposed to a medium-high-fat diet (MHF, 31.8% fat, 16.8% protein, and 51.4% carbohydrate per Kcal; D12266B, Research Diets, New Brunswick, NJ) ad libitum for a period of 9 wk.
Plasminogen activator inhibitor knockout mice.
Four- to six-week-old wild-type male C57BL/6 mice and PAI-1−/− mice (B6.129S2-Serpine1/J) on a C57BL/6 background were obtained from Jackson Laboratories (Bar Harbor, ME) and were housed in a temperature-controlled room (22°C). Mice were maintained on a normal 12-h:12-h light/dark cycle and were allowed free access to standard lab chow and water before treatment with TNBS as outlined below.
Induction of colitis.
Randomly assigned DR and DIO rats and wild-type and PAI−/− mice were briefly anesthetized with isoflurane (induced at 4% and maintained at 2%) and subjected to administration of either TNBS [rats: 0.5 ml of 50 mg/ml in 50% (vol/vol) ethanol, mice: 0.1 ml of 40 mg/ml in 40% (vol/vol) ethanol; Caledon Laboratories, Edmonton, AB, Canada] or an equivalent volume of sterile saline (0.9%) into the lumen of the colon through a polyethylene catheter inserted rectally 7 cm (for rats) or 3 cm (for mice) proximal to the anus. Following the induction of colitis, animals were housed individually, and daily food intake and body weight were monitored. Animals were allowed to recover for a period of 7 days.
At the time of tissue collection (9:30–10:30 AM), DR and DIO rats were anesthetized with sodium pentobarbital (80–100 mg/kg ip) and decapitated, and trunk blood was collected in tubes containing EDTA. After immediate centrifugation, serum was removed and then snap frozen and stored at −80°C for subsequent assay of leptin, insulin, ghrelin, adiponectin, and PAI-1. Mice were euthanized by cervical dislocation, and the entire colon of both rats and mice was removed and assessed for macroscopic damage (31); colonic samples (rat, 82–294 mg; mice, 48–230 mg) were taken from damaged sites and snap frozen for later assessment of myeloperoxidase (MPO) activity, an enzyme found in cells of myeloid origin, and used as a marker of neutrophil infiltration as previously described (31).
Assessment of adipokines, insulin, and ghrelin.
All serum samples were sent to Linco Diagnostics (now Millipore; St. Charles, MO) for either radioimmunoassay (RIA) or LINCOplex analysis. Nonacidified samples were analyzed for total ghrelin and adiponectin by RIA. The interassay variability for the ghrelin assay was 4.1–10.0% coefficient of variation (CV), intra-assay variability 14.7–17.8% CV, and the lower limit of detection was 93 pg/ml. For the adiponectin RIA, the interassay variability was 3.7–4.4% CV, intra-assay variability 6.6–8.2% CV, and the lower limit of detection was 1 ng/ml. Serum leptin, insulin, and PAI-1 levels were measured using a rat adipokine LINCOplex assay (the characteristics of which are available via www.millipore.com).
Ussing chamber experiments.
To assess colonic epithelial physiology, rat distal colons were immediately placed in fresh Krebs solution of the following composition (in mM) 117 NaCl, 4.8 KCl, 2.5 CaCl2, 1.2 MgCl2, 25 NaHCO3, 1.2 NaH2PO4, and 11 D-glucose. Mucosal preparations (with intact submucosal plexus) were obtained by opening the colon along the mesenteric border and removing the circular and longitudinal smooth muscle layers plus myenteric plexus by blunt dissection. Preparations were subsequently mounted in Ussing chambers (exposed mucosal area of 0.6 cm2) containing 10 ml oxygenated (95% O2-5% CO2) Krebs solution that was maintained at 37°C. Tissues were voltage clamped at 0 mV using an automatic voltage clamp (EVC 4000; World Precision Instruments, Sarasota, FL), and the short-circuit current (Isc) required to maintain the 0-mV potential was monitored using DataTrax software (World Precision Instruments) and resistance calculated using Ohms law.
Tissues from the inflamed regions were fixed by overnight immersion in Zamboni's fixative, washed in PBS (3 × 10 min), cryoprotected in PBS-sucrose (20%), and embedded for cryostat sectioning in optimal cutting temperature compound. Cross sections (12–14 μm) were cut and stained with hematoxylin and eosin to reveal structural features. Scoring of sections was based on a semiquantitative scoring system taking into account the following five features: mucosal architecture, muscle thickness, presence and degree of cellular infiltration, crypt abscesses, and goblet cell mucus depletion as outlined in Table 1.
Changes in body weight and food consumption for each day were compared using a two-way ANOVA with repeated measures with phenotype and treatment as between factors (independent variables) and time from treatment (day, dependent variable) as the repeated measure. Where a significant effect of treatment was found, one-way ANOVAs with Bonferroni's test were conducted comparing the groups on each day. Macroscopic tissue damage score was analyzed by nonparametric Mann-Whitney test, and total length of damage, length of severe damage, and serum, leptin, insulin, adiponectin, ghrelin, and PAI-1 concentrations were determined using an unpaired two-tailed Student's t-test. If adipokine levels fell below the detection limit of an assay, zero was recorded for statistical analysis. In each case, data were considered statistically significant when P ≤ 0.05. Data are presented as means ± SE.
Development of DIO and DR.
After exposure to the MHF diet for a period of 9 wk, DIO rats weighed significantly more than DR animals before the induction of colitis (Fig. 1A). Although previous studies indicate increased circulating insulin and insulin resistance, in DIO-prone animals (25), fasted serum insulin levels were comparable between DR and DIO animals in our study (Fig. 1B). Consistent with previous studies (37), DIO-prone rats displayed significantly increased circulating leptin concentrations compared with DR animals (Fig. 2A). No significant differences in circulating PAI-1, ghrelin, or adiponectin (summarized in Fig. 2, B–D) were observed between noninflamed DIO and DR animals; however, DIO animals tended to have increased circulating PAI-1 (twofold), and adiponectin (1.4-fold) and decreased circulating ghrelin (1.5-fold; P = 0.06).
Serum adipokine and ghrelin measurements in TNBS-treated DIO and DR rats.
Leptin levels remained significantly increased in DIO rats after TNBS treatment (P < 0.05, Fig. 2A) relative to DR TNBS-treated rats; however, in DIO animals, leptin was reduced by ∼2.5-fold by TNBS treatment. TNBS treatment caused a greater than 20-fold increase in PAI-1 levels in DR animals and a twofold increase in DIO rats. Between DR and DIO inflamed rats, PAI-1 levels were significantly increased in DR animals (P < 0.001, Fig. 2B). Inflammation had a tendency to increase circulating total ghrelin in both DR and DIO animals; however, the relative increase in ghrelin after TNBS treatment was similar between DR and DIO animals (DR saline vs. DR TNBS, 39.2% increase in ghrelin; DIO saline vs. DIO TNBS, 41.8% increase in ghrelin). With respect to this adipokine, the difference between DR and DIO animals only reached significance in TNBS-treated DIO rats relative to DR TNBS-treated animals (Fig. 2C). Serum adiponectin, on the other hand, was significantly increased after TNBS treatment in DIO rats (Fig. 2D).
TNBS-induced damage was significantly reduced in DIO rats.
Macroscopic tissue damage scores were significantly increased in TNBS-treated DIO and DR rats compared with saline-treated controls (data not shown). Only colonic thickness contributed to the score observed in saline-treated animals (DIO, 0.4 ± 0.03 mm; DR, 0.4 ± 0.04 mm). Comparisons between TNBS-treated groups revealed a significantly lower macroscopic damage score in inflamed DIO tissues compared with DR colon (Fig. 3A, P < 0.05). Breakdown of the macroscopic damage scores indicated that the total involved length (Fig. 3B, P < 0.05) and length of severe TNBS-induced damage (Fig. 3C, P = 0.01) were significantly less in DIO compared with DR rats. MPO activity was significantly increased by TNBS treatment in both DR and DIO colon (data not shown); however, no significant difference was observed between TNBS-treated DIO and DR animals [DIO, 51.0 ± 9.6 U/g, (n = 5); DR, 62.3 ± 13.0 U/g, (n = 6)]. Microscopic tissue damage scores were also comparable between TNBS-treated DIO and DR tissues [DIO, 8.5 ± 0.8, (n = 6); DR, 10.5 ± 0.3, (n = 6); Fig. 4].
DIO rats lose significantly less weight than DR rats after TNBS treatment.
DIO rats lost significantly less weight than TNBS-treated DR animals 4 (P < 0.001) and 5 (P < 0.001) days posttreatment (Fig. 5A) and, compared with DR rats, ate significantly more food on days 2–7 after TNBS treatment (Fig. 5B, P < 0.05–0.001).
TNBS colitis in PAI-1−/− mice.
Because DR rats displayed a striking elevation in circulating PAI-1 levels postinflammation relative to DIO animals, we hypothesized that PAI-1 may play a protective role during inflammation in obese rodents by potentially limiting the extent of inflammation. In the absence of a commercially available PAI-1 antagonist, we investigated the course of TNBS colitis 7 days posttreatment in wild-type and PAI-1−/− mice. Seven days after TNBS treatment, inflammatory indices were reduced in PAI-1−/− animals (Fig. 6, A–D); however, no significant differences in any of these parameters were observed.
Colonic secretory function in saline- and TNBS-treated DIO and DR tissues.
Brun et al. (8) described alterations in the electrical resistance and permeability parameters of obese ob/ob and db/db mouse small intestine compared with their lean counterparts. Because alterations in colonic permeability and secretory function can exacerbate TNBS-induced colitis, we examined basal electrical properties in DIO and DR saline- and TNBS-treated colon.
In both DR and DIO tissues, TNBS treatment significantly decreased basal Isc (P < 0.05, Table 2); however, we did not observe any significant effect of either phenotype or treatment on the transepithelial barrier resistance, a measure of colonic permeability (Table 2).
Environmental factors, such as lifestyle and diet, have been implicated in the development of colitis (2, 5). These environmental factors very likely interact with the gut microflora in genetically predisposed individuals to shape the expression of inflammation (2). Although obesity is not frequently associated with colitis, we have demonstrated its impact on the inflammatory adipokine response during colonic inflammation.
The observation that DIO rats lost less weight and ate more food than DR rats provided the first indication that DIO rats were less susceptible to the effects of TNBS. This increased intake of a MHF diet may itself protect against the effects of TNBS colitis because oral nutrient intake (Modulen IBD; Nestle, Milan, Italy; 42 Kcal% fat), when used as a first-line treatment for Crohn's disease (CD), results in decreased endoscopic and histological damage (6). Exposure to a high-fat diet has been proposed to disrupt the balance among intraepithelial lymphocytes, increasing susceptibility to colitis in rodents relative to animals fed a normal fat content diet (30). Although we cannot rule a differential effect of diet on the inflammatory response observed in DR- and DIO-prone animals, we do not believe this was a confounding factor in our study because both groups consumed the same diet. Furthermore, DIO animals consumed more food by weight than DR animals yet displayed decreased colonic damage indicative of a protective response in DIO animals.
Changes in the expression of circulating adipokines have led to the characterization of obesity as a state of chronic inflammation (17). In our study, noninflamed DIO rats had significantly higher circulating leptin levels than age-matched DR control rats. PAI-1 also tended to be increased in DIO animals, whereas differences in circulating adiponectin and ghrelin were nonsignificant. In TNBS-treated animals, this adipokine profile was altered (Fig. 2A), and these changes were accompanied by decreased TNBS-induced damage in DIO animals without an associated change in either MPO or microscopic damage. Whether differences in MPO activity or microscopic damage occurred earlier in the inflammatory response, before euthanasia, is however unknown.
The fact that PAI-1 is significantly increased in plasma taken from patients suffering from IBD (13), is associated with coagulation abnormalities that accompany intestinal inflammation (11), and displayed the greatest changes postinflammation in our study suggests that this procoagulant factor and adipokine may influence the severity or course of colitis. As anticipated, PAI-1 tended to be higher in noninflamed DIO rats compared with DR controls. However, the opposite was observed in DR TNBS-treated animals. Deficits in the clotting system are now recognized as having a potential role in increasing the inflammatory process during IBD (11). Prothrombotic plasma PAI-1 is significantly elevated in patients with IBD (10, 13), and vascular lesions are prominent features of CD, characterized by fibrin deposits; in some cases, fibrin thrombi were found to occlude the vascular lumen of blood vessels in the ileum (14). Decreased capillary blood flow was also reported during the acute phase of TNBS colitis in rats. Because we observed an ∼26-fold increase in PAI-1 during active TNBS colitis in DR rats, it is tempting to speculate that this may lead to increased coagulation in DR inflamed tissues and potentially localized hypoxia, giving rise to the increased extent of tissue damage or necrosis.
Changes in the circulating PAI-1 profile in DR rats may be due to the decreased food intake observed in these animals. DR animals ate significantly less food over several days (days 1–7) after TNBS treatment, and food deprivation in mice, while increasing PAI-1 message in epididymal and intestinal fat, does not alter circulating PAI-1 in obese mice but does so in lean animals (34). Nonetheless, our data obtained in PAI-1−/− mice, while demonstrating a moderate protective effect for this adipokine during TNBS colitis, did not reach significance, perhaps supporting the notion that several adipokines, as observed in our models, are likely to have synergistic or antagonistic effects in terms of modulating the inflammatory response to TNBS.
Significantly elevated circulating levels of adiponectin have been identified in patients with CD, the source of which is likely to be hypertrophied mesenteric adipose tissue or the creeping fat associated with CD (36, 46). Yamamoto et al. (46) also identified an inverse relationship between adiponectin and proinflammatory IL6 production, suggesting that adiponectin may have anti-inflammatory properties. We did observe a significant increase in circulating adiponectin in DIO inflamed rats 7 days after TNBS treatment when the colon was actively inflamed; however, given the conflicting results obtained in genetically modified animals with respect to the effects of this adipokine on the severity of colitis (15, 33), it is difficult to conclude what role adiponectin plays in modulating the inflammatory response in DR or DIO animals. Seven days after TNBS treatment a significant increase in serum ghrelin in both DIO and DR rats was observed, and this was significantly greater in DR compared with DIO TNBS-treated animals. This data is consistent with decreased food intake and weight loss observed during inflammation. To date the role ghrelin plays in contributing to intestinal inflammation is not well characterized (18, 47). However, recently De Smet et al. (12) demonstrated a proinflammatory role for ghrelin during colitis in mutant mice lacking the peptide as well as in mice exogenously treated with ghrelin. Consistent with this finding, we observed a significantly decreased ghrelin response in DIO animals postcolitis that was accompanied by decreased macroscopic damage in keeping with the findings of De Smet et al.; however, the percentage change in ghrelin in DR and DIO animals after TNBS treatment was similar between the groups. Therefore, the precise role for ghrelin in contributing to decreased colonic damage in DIO animals, while plausible, is likely to be more complex in the DIO model in which several adipokines are altered in response to inflammation.
At this point, we can only speculate as to what aspect of the DIO phenotype confers protection during inflammation, but, because leptin is significantly altered in DIO animals both before and during inflammation and can alter colonic neutrophil activity during colitis through activation of the hypothalamic pituitary adrenal axis (9), as well as modulating inflammation directly though receptors on most immune cells (42), it is tempting to question the role of this adipokine in modulating the course of TNBS colitis in DIO rats. Although the proinflammatory effect of leptin (41) makes the observed decreased colonic damage in DIO animals seem counterintuitive, leptin resistance has also been observed in the immune system of DIO animals; leptin receptors on T lymphocytes taken from DIO rodents are resistant to leptin activation, suggesting that T lymphocytes develop leptin resistance during DIO (35), potentially leading to a decreased immune response to TNBS. It is therefore likely, although remains to be investigated, that the DIO rats used in our study may have developed T lymphocyte resistance to leptin and are less responsive to the proinflammatory effect induced by circulating leptin that occurs after TNBS treatment (4). Previous studies have indicated that, as early as 8 h after TNBS treatment, circulating leptin significantly increased above that of saline-treated control animals and may contribute to the initial decrease in food intake and body weight in TNBS-treated rats (4). However, by 6 days after TNBS treatment, leptin levels were comparable in saline- and TNBS-treated rats (4).
We found no preexisting defect in colonic barrier or basal secretory function that may predispose tissues to altered TNBS-induced damage. Our data in the DIO model of obesity differs from that in genetically obese, ob/ob and db/db mice, which displayed decreased epithelial resistance and increased permeability (8). This may reflect differences between obese models in that leptin deficiency or decreased signaling, in ob/ob and db/db mice, respectively, from birth may have long-term consequences on the development of tight junctions, whereas DIO rats have normal circulating levels of leptin during development and generally do not display significant increases in leptin until exposure to a MHF diet.
This is the first study to investigate the adipokine response following inflammation in a physiological model of obesity compared with that in lean animals. We have described significant changes in postinflammatory circulating adipokine profile that was accompanied by decreased colonic damage after TNBS treatment in obese animals, without an associated change in colonic physiology. The factors contributing to the decreased colonic damage are almost certainly multifold, driven by both genetic and environmental factors, of which adipokines are likely to play a part given the increasing body of evidence for their role in modulating intestinal inflammation.
This work was supported by a New Emerging Team operating grant from the Canadian Institutes of Health Research (CIHR), Institute of Nutrition, Metabolism and Diabetes (K. Sharkey; team members, Drs. W. F. Colmers, A. V. Ferguson, and D. Richard).
K. Sharkey and Q. Pittman are Alberta Heritage Foundation for Medical Research (AHFMR) Medical Scientists. K. Sharkey is the Crohn's and Colitis Foundation of Canada Chair in inflammatory bowel disease research at the University of Calgary, and Q. Pittman is a University Professor. N. Hyland was a recipient of Canadian Association of Gastroenterology/AstraZeneca/CIHR Postdoctoral Fellowship, and A. Chambers was a recipient of an AHFMR graduate studentship.
- Copyright © 2009 the American Physiological Society