There is evidence that adipocytokines play an important role in metabolism and in inflammation. Because human metabolism dramatically changes in inflammatory bowel disease (IBD) and chronic inflammation is the hallmark of the disease, we studied serum levels of leptin, adiponectin, resistin, and ghrelin in patients with ulcerative colitis (UC) and Crohn's disease (CD) in comparison with healthy controls (HC).
Leptin, adiponectin, resistin, and active ghrelin serum levels were measured in 100 IBD patients (46 UC and 54 CD) and in 60 matched HC using commercially available enzyme-linked immunosorbent assays. Leptin, adiponectin, resistin, and ghrelin levels were correlated with disease activity, type, localization, and treatment.
Mean serum leptin levels were 10.6 ± 2.0 ng/mL in UC patients, 12.5 ± 2.6 ng/mL in CD patients, and 15.0 ± 1.8 ng/mL in HC (P = .01). Mean serum adiponectin levels were 9514.8 ± 787.8 ng/mL in UC patients, 7651.1 ± 613 ng/mL in CD patients, and 7270.6 ± 559.4 ng/mL in HC (P = .05). Mean serum resistin levels were 21.2 ± 2.2 ng/mL in UC patients, 18.7 ± 1.6 ng/mL in CD patients and 11.8 ± 0.6 ng/mL in HC (P = .0002). Mean serum ghrelin levels were 48.2 ± 4.2 pg/mL in UC patients, 49.4 ± 4.6 pg/mL in CD patients and 14.8 ± 3.0 pg/mL in HC (P < .0001). Serum levels of these adipocytokines were not correlated with either C-reactive protein levels or the clinical indices of activity. No association between serum adipocytokines levels and disease localization in both UC and CD patients was found. Only serum ghrelin was significantly higher in ileal compared with colonic CD (P = .04).
Serum levels of adiponectin, resistin, and active ghrelin are increased whereas serum levels of leptin are decreased in patients with IBD. Further studies are needed to elucidate the role of adipocytokines in IBD.
Recent studies suggest that white adipose tissue (WAT), besides its ability to respond to afferent signals from traditional hormone systems and the central nervous system, also expresses and secretes factors with important functions, collectively called adipocytokines.1 There is evidence that adipocytokines are involved in inflammatory and metabolic pathways in humans. Among the adipocytokines, leptin, adiponectin, and resistin appear to play an important role.2
Anorexia, malnutrition, altered body composition, and development of mesenteric obesity (accumulation of intraabdominal WAT), are well-known features of inflammatory bowel disease (IBD), mainly of Crohn's disease (CD), indicating an important role for WAT-secreted proteins.3 Overexpression of adipocytokines such as leptin, adiponectin, and resistin in mesenteric adipose tissue of patients with CD who have been operated on has recently been reported, suggesting that mesenteric adipocytes in IBD may act as immunoregulatory cells.4,–6 Therefore, it is suggested that adipocytokines may participate in the disease pathogenesis.7
Leptin is a 16-kDa nonglycosylated protein, which belongs to the type I cytokine superfamily and is characterized by a long-chain 4-helical bundle structure.8 Leptin possesses proinflammatory as well as anti-inflammatory properties according to the experimental conditions.9 Its role in IBD has been studied, but the results are conflicting, therefore further investigation is required.10,–13
Adiponectin is an approximately 30-kDa polypeptide,14 and interestingly, the terminal structure of its globular domain bears a striking similarity to tumor necrosis factor-α (TNF-α), despite a lack of homology in primary sequence.15 Adiponectin appears to have anti-inflammatory properties because of its antagonism against TNF-α. Various experiments have demonstrated that adiponectin and TNF-α suppress each other's production and also antagonize each other's action in their target tissues.16,–19 Overexpression of adiponectin in the mesenteric adipose tissue of CD patients has been reported recently.5,6
Human resistin is a 108-amino acid peptide hormone with a molecular weight of 12.5 kDa.20 Resistin was extensively studied in patients with both rheumatoid arthritis and osteoarthritis and significant correlation with inflammation and elevated C-reactive protein (CRP) was reported.21 There is evidence that resistin is involved in inflammatory and metabolic pathways in humans and a possible role in IBD was recently postulated in certain IBD patients.6
Ghrelin is a recently discovered hormone, with a crucial role in the regulation of food intake and energy homeostasis.22 It is mainly produced at the stomach but is also expressed in WAT, albeit in trace amounts.23 Ghrelin is an endogenous ligand of the growth hormone secretagogue receptor (GHS-R) and it has been identified in T cells. Ghrelin can inhibit cytokine activation including interleukins, TNF-α, and most interestingly leptin.24 Recently, high levels of serum ghrelin were found in patients with celiac disease.25 To our knowledge, the possible role of ghrelin in IBD patients has not been examined.
The present study was designed to evaluate fasting serum levels of leptin, adiponectin, resistin, and ghrelin in IBD patients, to correlate the results with the disease activity and the clinical characteristics of the disease, and to examine the possible interaction between the estimated hormone values.
Materials and Methods
Patients
One hundred consecutive IBD patients followed up at the Department of Gastroenterology of the University Hospital of Heraklion, Crete, participated in this study (Table 1). The disease was either active or quiescent and the diagnosis was based on standard criteria.26 Disease activity in CD patients was evaluated with the CD activity index score27 and in ulcerative colitis (UC) patients with the simple clinical colitis activity index (SCCAI), as described by Walmsey et al.28 Evaluation of disease activity was done at the time of serum collection. These patients were age and sex matched with the control group, which consisted of 60 healthy controls (HC) with a mean age of 36 years (blood donors, normal hospital personnel). All patients and controls were white. Standard laboratory parameters included red and white blood cell count, hemoglobin, hematocrit, platelet count, total protein, albumin, cholesterol, triglycerides, uric acid, erythrocyte sedimentation rate, and CRP and were routinely determined in all patients. Informed consent was obtained from all of the patients and the Ethics Committee of the Medical Faculty of Crete approved the protocol of the study.
Patients with concomitant endocrine disease (mainly diabetes mellitus, hypo- or hyperthyroidism, adrenal failure), hyperlipidemia (cholesterol >230 mg/dL and triglycerides >140 mg/dL), coronary heart disease, chronic obstructive pulmonary disease, malignant hypertension, malignant obesity, and any other kind of malignancy or autoimmune disease were excluded from the study.
Laboratory Tests
Blood samples were collected in the morning after an overnight fast and after centrifugation (3500g for 15 min) 1- to 2-mL serum samples were stored at −80°C until assayed. Leptin, resistin, adiponectin, and ghrelin concentrations were measured using commercially available sandwich enzyme-linked immunosorbent assay kits (R&D Systems, Abington, UK, for leptin, adiponectin, and resistin, and Peninsula Laboratories/Bachem, Torrance, Calif, for ghrelin). Escherichia coli-expressed recombinant human leptin and resistin, NS0-expressed recombinant human adiponectin, and human octanoylated ghrelin were used as standards. The ghrelin enzyme-linked immunosorbent assay detected the active form of the hormone (Bachem H4864). A monoclonal mouse antibody specific for each hormone was employed and detected with conjugated horseradish peroxidase according to the manufacturer's instructions. The intra- and interassay coefficients of variation for leptin, adiponectin, resistin, and ghrelin were below 3.3%, 4.7%, 5.3%, and 5% and below 5.4%, 6.9%, 9.2%, and 14%, respectively (n = 20 and n = 40, respectively).
Statistical Analysis
All results are expressed as mean ± SEM. Comparisons among the 3 diagnostic groups in terms of continuous measurements were made by the Kruskal-Wallis test (nonparametric ANOVA). Post hoc multiple comparisons tests were made by Dunn's test. The same tests were used for the comparisons between the groups of disease localization. Comparisons between 2 groups (e.g., active versus nonactive disease, stenotic versus nonstenotic CD) were made by either the Student t test or the Mann-Whitney U test. The Kolmogorov-Smirnov test was used to assess the assumption that data were sampled from populations that follow Gaussian distributions. Correlations between serum adipocytokines and indices of disease activity were analyzed with the Pearson's correlation method. All analyses were 2-tailed and conducted using the computer-based statistics software program Instat version 3.0 (GraphPad Software, San Diego, Calif). A level of P < .05 was considered statistically significant.
Results
Figure 1 shows the distribution of serum leptin in patients and HC. Mean serum leptin levels were 10.6 ± 2.0 ng/mL (median: 4.7, range: 0.3-68.1) in UC patients, 12.5 ± 2.6 ng/mL (median: 5.8, range: 1.2-116.3) in CD patients, and 15.0 ± 1.8 ng/mL (median: 9.3, range: 1.4-64.3) in HC. A statistically significant difference between the mean levels of leptin of the 3 groups was found (P = .01). The multiple comparisons Dunn's test showed that UC and CD patients had similar average leptin levels (P > .05). UC patients had significantly lower leptin levels than HC (P < .05). CD patients had lower but not significantly different leptin levels compared with HC. Mean serum adiponectin levels were 9514.8 ± 787.8 ng/mL (median: 9021.7, range: 1871.2-7038) in UC patients, 7651.1 ± 613 ng/mL (median: 6582.6, range: 2371.7-26072) in CD patients, and 7270.6 ± 559.4 ng/mL (median: 6947.5, range: 1744-21320) in HC (Fig. 2). The difference between the 3 groups was on the margin of statistical significance (P = .05). Only UC patients had significantly higher adiponectin levels than HC (P = .02). Mean serum resistin levels were 21.2 ± 2.2 ng/mL (median: 16.5, range: 6.0-71.7) in UC patients, 18.7 ± 1.6 ng/mL (median: 14.6, range: 5.6-56.7) in CD patients, and 11.8 ± 0.6 ng/mL (median: 10.7, range: 5.4-23.4) in HC (Fig. 3). The difference between the 3 groups was statistically significant (P = .0002). The multiple comparisons Dunn's test showed that both UC and CD patients had significantly higher resistin levels than HC (P < .01). Mean serum ghrelin levels were 48.2 ± 4.2 pg/mL (median: 47.5, range: 0.0-108.9) in UC patients, 49.4 ± 4.6 pg/mL (median: 48.7, range: 0.0-189.1) in CD patients, and 14.8 ± 3.0 pg/mL (median: 4.6, range: 0.0-107.5) in HC (Fig. 4). The difference between the 3 groups was statistically significant (P < .0001). The multiple comparisons Dunn's test showed that both UC and CD patients had significantly higher ghrelin levels than HC (P < .001).
Distribution of leptin in HC (n = 60), UC (n = 46), and CD (n = 54) patients. Each individual patient or control is shown as a circle; bold lines are the mean values.
Distribution of adiponectin in HC (n = 60), UC (n = 46) and CD (n = 54) patients. Each individual patient or control is shown as a circle; bold lines are the mean values.
Distribution of resistin in HC (n = 60), UC (n = 46), and CD (n = 54) patients. Each individual patient or control is shown as a circle; bold lines are the mean values.
Distribution of ghrelin in HC (n = 60), UC (n = 46), and CD (n = 54) patients. Each individual patient or control is shown as a circle; bold lines are the mean values.
The mean serum level of ghrelin was significantly higher in males with IBD (55.0 pg/mL) as compared with female IBD patients (40.3 pg/mL; P = .02). There was a trend toward higher serum levels for leptin in males (13.8 ng/mL) as compared with females (9.3 ng/mL), but the difference was not statistically significant (P = .14). No other association between the examined adipocytokines and sex was found.
UC and CD patients with active disease had similar levels of the examined adipocytokines in comparison with patients with inactive disease (P > .05) The serum levels of these adipocytokines were not correlated with either CRP levels or the clinical indices of activity (simple clinical colitis activity index and CD activity index, respectively). No association between serum adipocytokines levels and disease localization in both UC and CD patients was found. Only serum ghrelin was significantly higher in ileal compared with colonic CD (P = .04). Subsequent analyses of other subgroups did not reveal any statistically significant correlation between serum levels of the examined adipocytokines and disease type (stenotic versus nonstenotic), years of diagnosis (early versus late disease), smoking habits, and the current use of medications such as 5-aminosalicylic acid, prednisone, infliximab, and azathioprine. Only patients with fistulizing CD have a trend toward lower leptin (P = .06) and higher adiponectin (P = .09) levels. Patients with a body mass index (BMI) <25 had significantly lower serum leptin levels (6.8 ng/mL) compared with patients with BMI ≥25 (18.8 ng/mL; P = .004). The other adipocytokines did not differ between these 2 groups.
Discussion
In our study, we measured the circulating levels of 4 hormones (adipocytokines) in patients with IBD. These hormones are produced by WAT either in large amounts (leptin, adiponectin, and resistin) or in trace amounts (ghrelin), and all of them are closely related to human metabolism and inflammation. Human metabolism dramatically changes in IBD, and chronic inflammation is the hallmark of the disease. To our knowledge, this is the first report examining serum levels of adiponectin, resistin, and ghrelin in IBD patients. We found significant differences in serum adipocytokines between IBD patients and HC.
Leptin levels were significantly lower in IBD patients, mainly in UC, compared with HC, suggesting that chronic intestinal inflammation may decrease circulating leptin levels. TNF-α is possibly the major cytokine involved in intestinal inflammation in IBD patients. Recently, a number of studies indicated that whereas TNF-α transiently induces acute release of intracellular pools of leptin, it decreases leptin synthesis during chronic inflammation.29,–31 Additional evidence came from the description of reduced levels of leptin in patients with long-lasting tuberculosis.32 Franchimont et al have also found an increase in leptin levels in IBD patients treated with infliximab, an anti-TNF-α agent.10 Thus, IBD patients could have lower circulating leptin levels compared with HC because of the interference by TNF-α. This conclusion contrasts with many other studies that have shown a positive relationship between leptin and inflammation.33,–36 Previous reports have also tried to correlate leptin levels with anorexia and weight loss, which are common in IBD patients. The results of these studies are conflicting.4,11,–13 We found a significant positive association of leptin with high BMI as expected, but no other significant associations of serum leptin with the disease clinical characteristics were established. Recently, inflamed colonic epithelial cells were found to express and release leptin into the intestinal lumen, and the product appears to induce epithelial wall damage and neutrophil infiltration, a characteristic histological finding in IBD.37
The role of adiponectin as an anti-inflammatory cytokine is well established. A decreased expression of adiponectin in alcohol-induced liver injury in mice16 and in patients with nonalcoholic steatohepatitis38 was reported. In both instances, elevated TNF-α was speculated to be the mediator of this reduction, via suppression of adiponectin expression in WAT and liver through a paracrine or endocrine mechanism. Administration of adiponectin suppressed TNF-α expression in liver tissue and also decreased circulating concentrations of TNF-α.16,38 Within the vascular wall, adiponectin inhibits monocyte adhesion on endothelial cells by decreasing the expression of adhesion molecules and inhibits macrophage transformation to foam cells by inhibiting expression of scavenger receptors.39 In addition, adiponectin increases nitric oxide production in endothelial cells and stimulates angiogenesis.40 The possible role of adiponectin was recently studied in IBD patients.5,6 Yamamoto et al showed that adiponectin production is enhanced in hypertrophied mesenteric adipose tissue in contact with the involved intestine of CD patients; this overexpression is higher than in UC patients and patients operated for colon cancer (control group). They speculated that the size of adipocytes is crucial (the smaller they are, the more adiponectin they produce).5 Paul et al reported similar results.6 Serum adiponectin levels in the present study are in accordance with these studies. Interestingly, the lower levels observed in CD compared with UC may be caused by the inhibition of adiponectin production by TNF-α. Alternatively, the high levels observed in UC may represent a counterregulatory mechanism to the effects of TNF-α.
The main production of resistin and ghrelin occurs in different sites, and in addition to their participation in the mechanisms of energy homeostasis, they seem to have in common a close association with inflammation, a fact that may implicate them in the pathogenesis of IBD. In our study, both have significantly higher circulating levels in IBD patients as compared with HC (P = .0002 and P < .0001, respectively). The role of these 2 hormones in inflammation has been extensively studied in recent years.
Treatment of 3T3-L1 adipocytes with TNF-α down-regulated resistin mRNA and protein levels41 or inhibited resistin gene expression by up to 70% to 90%.42 However, TNF-α as well as interleukin-6 (IL-6) significantly increased resistin mRNA expression in human peripheral blood mononuclear cells in vitro.43 Synovial resistin levels were found to be ≈10 times higher in patients with rheumatoid arthritis compared with patients with osteoarthritis.21 Bokarewa et al showed that resistin strongly up-regulated IL-6 and TNF-α, responded to TNF-α challenge, and induced arthritis when injected into healthy mouse joints.44 Another study demonstrated that human resistin acts as a proinflammatory molecule, stimulating the synthesis and secretion of TNF-α and IL-12 through the activation of NF-κB transcription factor.45 Recently, a higher expression of resistin was found in the mesenteric adipose tissue of patients operated on for CD as compared with patients being operated on for colon carcinoma.6
Ghrelin is the first natural hormone to be identified in which the hydroxyl group of one of its serine residues (Ser3) is acylated by n-octanoic acid. This acylation is essential for its biological activity, especially the binding and activation of the ghrelin receptor (GHS-R-1a). Nonacylated ghrelin, which is present in human serum in far greater quantities than acylated ghrelin, seems to be devoid of any endocrine action.46 We measured the active (n-octanoylated) form of ghrelin in our study.
Ghrelin has been shown to affect a number of different systems including growth hormone release, feeding, gastric acid secretion, gastric motility, and cell proliferation. As we mentioned above, ghrelin and its receptor have been identified in T cells. Ghrelin can inhibit cytokine activation including interleukins, TNF-α, and most interestingly leptin.24 Our results for leptin levels agree with this observation because higher ghrelin levels were correlated with lower leptin levels in IBD patients compared with HC. Patients with celiac disease have higher ghrelin levels than BMI-matched controls and ghrelin decreases on a gluten-free diet.25 A correlation between the severity of the inflammation and ghrelin levels has been suggested. In our study, no association of ghrelin with disease activity was found, but patients with ileal CD had higher ghrelin levels as compared with colonic CD. Additional studies in tissue samples are needed to further explore this finding.
Conclusions
In conclusion, we found increased serum levels of adiponectin, resistin, and active ghrelin and decreased serum levels of leptin in patients with IBD. Taken together, our results in the present study and those reported in the literature may indicate interesting interactions in IBD schematically presented in Figure 5. This observed dysregulation of protein secretion by adipose tissue may play an important role in the disease pathogenesis. Additional studies in serum and tissue are needed to elucidate the role of adipocytokines in IBD.
Possible interactions among T lymphocytes, adipocytes, and mucosal macrophages in IBD. Increased resistin production by adipocytes causes increased TNF-α and IL-6 production by macrophages. Adiponectin and ghrelin antagonize proinflammatory cytokine production by macrophages. Leptin production by adipocytes is down-regulated by both ghrelin and TNF-α.
References
Author notes
Reprints: Ioannis E. Koutroubakis, MD, Department of Gastroenterology, University Hospital Heraklion, PO Box 1352, 71110 Heraklion, Crete, Greece (e-mail: ktjohn@her.forthnet.gr)
