Elevated Serum Chemokine CXC Ligand 5 Levels Are Associated with ...

ORIGINAL E n d o c r i n e

ARTICLE R e s e a r c h

Elevated Serum Chemokine CXC Ligand 5 Levels Are Associated with Hypercholesterolemia But Not a Worsening of Insulin Resistance in Chinese People Zhen Yang,* Zhaoyun Zhang,* Jie Wen, Xuanchun Wang, Bin Lu, Zhihong Yang, Weiwei Zhang, Mei Wang, Xiaocheng Feng, Charlotte Ling, Songhua Wu, and Renming Hu Institute of Endocrinology and Diabetology (Zhe.Y., Z.Z., J.W., X.W., B.L., Zhi.Y., W.Z., M.W., X.F., R.H.) and Department of Endocrinology and Metabolism, Huashan Hospital, Shanghai Medical College, Fudan University, Shanghai 200040, China; Department of Clinical Sciences (C.L.), Diabetes and Endocrinology Research Unit, Clinical Research Center, Malmo¨ University Hospital, Lund University Diabetes Center, S-205 02 Malmo¨, Sweden; and Shanghai Diabetes Institute (S.W.), Shanghai No. 6 People Hospital, Shanghai Jiaotong University, Shanghai 200233, China

Objective: Recent study showed high chemokine CXC ligand 5 (CXCL5) is thought to be associated with insulin resistance in humans. However, evidence from large-scale populations about the relationship between serum CXCL5 level and metabolic phenotypes is scarce. Here we sought to evaluate serum CXCL5 distribution and its association with metabolic phenotypes among middleaged and older Chinese. Research Design and Methods: We evaluated serum CXCL5 in a cross-sectional sample of 3225 Chinese aged from 50 to 88 yr in a Shanghai downtown district by ELISA. Glucose, insulin, lipid profile, inflammatory marker, and adipokine were also measured. Results: The crude mean of serum CXCL5 concentrations were 1493.31 pg/ml for men and 2059.42 pg/ml for women (P ⬍ 0.001), respectively. After multiple adjustment, the odds ratios were substantially higher for hypercholesterolemia (odds ratio 3.26, 95% confidence interval 2.36 – 4.51) in the highest CXCL5 quartile compared with those in the lowest quartile. These associations remained significant after further adjustment for body mass index, body fat, inflammatory marker, and adipokine. However, serum resistin CXCL5 was not associated with body mass index, percent body fat, fasting glucose, insulin levels, and homeostasis model assessment index-insulin resistance (r ⫽ 0.01, 0.01, 0.01, 0.04, and 0.03, respectively; all P ⬎ 0.05). Conclusions: Elevated circulating CXCL5 concentrations were associated with higher risk of hypercholesterolemia in middle-aged and elderly Chinese independent of obesity, inflammation, adipokines, and other risk factors but not insulin resistance. (J Clin Endocrinol Metab 95: 3926 –3932, 2010)

O

besity causes insulin resistance. Over the last decade, an abundance of evidence has emerged demonstrating a close link between metabolism and immunity. It is now clear that obesity is associated with a state of chronic low-level inflammation. Chronic inflammation in fat

plays a crucial role in the development of obesity-related insulin resistance (1). Obesity is characterized by progressive macrophage infiltration in white adipose tissue (WAT) (2), which is positively correlated with insulin resistance. One mechanism through which the progressive

ISSN Print 0021-972X ISSN Online 1945-7197 Printed in U.S.A. Copyright © 2010 by The Endocrine Society doi: 10.1210/jc.2009-2194 Received October 13, 2009. Accepted May 5, 2010. First Published Online May 25, 2010 * Z.Y. and Z.Z. contributed equally to this work.

Abbreviations: ALT, Alanine aminotransferase; BMI, body mass index; CI, confidence interval; CRP, C-reactive protein; CVD, cardiovascular disease; CXCL5, CXC ligand 5; HDL, high-density lipoprotein; HOMA-IR, homeostatic model assessment of insulin resistance; IFG, impaired fasting glucose; IGT, impaired glucose tolerance; LDL, low-density lipoprotein; NGT, normal glucose tolerance; OR, odds ratio; TC, total cholesterol; WAT, white adipose tissue.

3926

jcem.endojournals.org

J Clin Endocrinol Metab, August 2010, 95(8):3926 –3932


macrophage infiltration in enlarged WAT mass is thought to influence insulin action is via secretion of adipose tissuederived factors (3, 4), such as TNF-␣, leptin, resistin, and adiponectin (5– 8). Adipose tissue-derived factors are secreted by the different cell compartments of WAT, such as adipocytes or macrophages. CXC ligand 5 (CXCL5), an adipokine with high levels of expression and secretion in the macrophage fraction of WAT belonging to the chemokine family (9), has also been suggested as a link between obesity and insulin resistance (10). Chemokines are proinflammatory cytokines that stimulate leukocyte chemoattraction and are produced in response to infectious and other inflammatory stimuli by an umber of different cell types. CXCL5 is a cytokine belonging to the family of chemokines that is mainly implicated in the chemotaxis of inflammatory cells through the generation of local concentration gradients (11). Serum CXCL5 is increased in diet-induced obese, ob/ ob, and db/db mice; the insulin sensitizer rosiglitazone has been shown to reduce CXCL5 expression of adipose tissue in these mice (12). In obese, insulin-resistant mice, the ip injection of neutralizing anti-CXCL5 antibody lowered fasting glucose (10). Furthermore CXCR2⫺/⫺ mice are protected against obesity-induced insulin resistance (10). CXCL5 protein is detectable in human serum, and its circulating levels were found dramatically increased in serum of human obese compared with lean subjects. Conversely, CXCL5 concentration is decreased in obese subjects after a weight reduction program, or in obese noninsulin-resistant, compared with insulin-resistant subjects (10). However, evidence from large-scale populations about the relationship between CXCL5 and obesity and insulin resistance is scarce. To better evaluate a possible role of CXCL5 in the development of insulin resistance, we examined the association of serum CXCL5 levels, adiposity, and insulin resistance measured by the homeostasis model assessment score in large-scale Chinese populations. Our aims were to examine the CXCL5 distribution in a cross-sectional population study of 3225 middle-aged and older Chinese subjects and to explore the association of serum CXCL5 levels with insulin sensitivity as well as other metabolic phenotypes in these subjects.

Subjects and Methods Study population In 2007 China launched a national incidence trends of metabolic syndrome study. The data presented in this article are partially based on subsamples from Shanghai in eastern China (total 2598 subjects). All studied individuals came from the Simenerlu community of the Jingan District and the Jiangninglu com-


3927

munity of the Putuo District in Shanghai, China. A multistage stratified cluster sampling method was used to select subjects from these two communities. Only one participant (ⱖ50 yr) was randomly selected from each household. In this part, subjects, all recruited subjects received the oral glucose tolerance test test except for previously diagnosed type 2 diabetes patients. Furthermore, we also chose 627 type 2 diabetes patients (ⱖ50 yr) to participate in this study from an outpatient unit. All studied individuals were of southern Han Chinese ancestry and were residing in the metropolitan area of Shanghai and aged 50 – 88 yr. A total of 3225 eligible participants (1483 men and 1742 women) were recruited (detail data show in Supplemental Table 1, published on The Endocrine Society’s Journals Online web site at http://jcem.endojournals.org). Written consent was obtained from all the participants. The study was approved by the Institutional Review Broad of Huashan Hospital.

Data collection A standardized questionnaire was used by trained physicians to collect information such as age; sex; education level (6, 7–9, or 10 yr in school); smoking (yes/no); alcohol drinking (yes/no); and self-reported diabetes, hypertension, dyslipidemia, coronary heart disease, and stroke. Physical activity level was classified as low, moderate, or high according to the International Physical Activity Questionnaire scoring protocol. According to participants’ responses to the corresponding questions, family history of diabetes or self-reported cardiovascular disease (CVD) was classified as yes or no. All subjects were assessed after overnight fasting for at least 10 h. The details of anthropometric measurements including height, weight, waist circumference, hip circumference, and blood pressure were carried by trained physicians. Body mass index (BMI) was calculated as weight in kilograms divided by the square of height in meters and categorized as normal weight (⬍24.0 kg/m2), overweight (24 –28 kg/m2), or obesity (ⱖ28.0 kg/m2), according to the criteria for Chinese individuals (12).

Biochemical measurements Peripheral venous blood samples were collected. The fasting glucose, glucose 2 h after oral glucose tolerance test, total cholesterol (TC), triglycerides, low-density lipoprotein (LDL) cholesterol and high-density lipoprotein (HDL) cholesterol were measured on an automatic analyzer (Hitachi 7080; Tokyo, Japan). Alanine aminotransferase (ALT) was measured by the UV method. Fasting insulin was determined by RIA (Linco Research, St. Charles, MO). Insulin resistance was estimated using homeostasis model assessment index-insulin resistance (HOMAIR) (13). Body fat was quantified with the TBF-410 Tanita body composition analyzer (Tanita, Tokyo). Type 2 diabetes was defined by 1998 World Health Organization criteria or previously diagnosed type 2 diabetes (14). Normal glucose tolerance (NGT), impaired fasting glucose (IFG) and impaired glucose tolerance (IGT) were defined as fasting glucose 5.6 – 6.9 mmol/liter and 7.8 –11.0 mmol, respectively (15). Of 3225 participants, 1103 had type 2 diabetes (711 previously diagnosed, 392 screen detected and treatment naive), 336 had isolated IFG (all screen detected and treatment naive), 364 had isolated IGT (all screen detected and treatment naive), 297 had IFG⫹IGT (all screen detected and treatment naive) and 1125 had NGT.

3928

Yang et al.

CXCL5, Hypercholesterolemia, and Insulin Resistance

Measurements of serum C-reactive protein (CRP), adiponectin, and CXCL5 The serum CRP, adiponectin, and CXCL5 were determined in duplicate by ELISA with Duoset kit (DY1707, DY1065, and DY254; R&D Systems, Minneapolis, MN) as recommended by the manufacturer. The ELISA system had an intraassay coefficient of variation of 3– 8% and an interassay coefficient of variation of 4 –10%, respectively.

Definition of hypercholesterolemia The hypercholesterolemia was defined based on the updated National Cholesterol Education Program Adult Treatment Panel III criteria as a blood cholesterol concentration of than equal to 6.2 mmol/liter or greater (240 mg/dl) (16).


quartiles. Analysis of covariance was used to compare CXCL5 levels between genders. Correlation coefficients between CXCL5 and metabolic features were calculated by partial correlation analysis. Multiple stepwise regression analysis was used to examine the association of serum CXCL5 and other parameters. Multivariate logistic regression models were used to estimate the odds ratios (ORs) for hypercholesterolemia. Potential confounding variables including age, gender, smoking, alcohol drinking, physical activity, educational level, self-reported CVD, family history of diabetes, CRP, adiponectin, HOMA-IR, BMI, body fat, waist circumference, and hip circumference were controlled in the regression models. When appropriate, natural log-transformed values were used for the analyses. All statistical analysis were performed with the SPSS Statistical Package (version 13.0; SPSS Inc., Chicago, IL). P ⬍ 0.05 was considered statistically significant.

Statistical analysis Normally distributed data were expressed as means ⫾ SD, whereas variables with a skewed distribution were reported as median (interquartile range) and log transformed to approximate normality before analysis. Categorical variables were represented by frequency and percentage. Analysis of covariance for continuous variables and multivariate logistic regression analysis for categorical variables were applied for the comparison according to CXCL5

TABLE 1.

Results Characteristics of participants according to CXCL5 quartiles When analyzed by quartiles of CXCL5 levels, as summarized in Tables 1 and 2, the subjects with higher CXCL5 were

Characteristics of male participants according to CXCL5 quartilesa

Characteristics Hypercholesterolemia (%) CXCL5 (pg/ml) Age (yr)b Smoking (yes) Alcohol (yes) Education (yr) 0–6 7–9 ⱖ10 Physical activity Low Moderate High Self-reported CVDc Family history of diabetesd SBP (mm Hg) DBP (mm Hg) BMI (kg/m2) Body fat (%) Waist circumference (cm) Waist to hip ratio Glucose (mmol/liter) Insulin (␮U/ml)e HOMA-IRe CRP (␮g/ml) Adiponectin (␮g/ml)e ALT (U/liter) Triglycerides (mmol/liter)e Total cholerterol (mmol/liter) LDL cholesterol (mmol/liter) HDL cholesterol (mmol/liter)

Q1 (n ⴝ 370) 1805.34 28.30 2892.72 ⫾ 885.41 62.63 ⫾ 9.14 165 (47.17) 89 (23.99)

23 (6.22) 120 (32.43) 227 (61.35)

21 (5.66) 117 (31.54) 233 (62.80)

19 (5.12) 121 (32.61) 231 (62.27)

27 (7.28) 95 (25.61) 249 (67.11)

245 (66.22) 80 (21.62) 45 (12.16) 28 (7.57) 75 (20.27) 134.31 ⫾ 18.50 82.88 ⫾ 10.01 24.86 ⫾ 3.14 24.56 ⫾ 4.99 87.01 ⫾ 9.45 0.89 ⫾ 0.09 6.80 ⫾ 2.71 8.63 (5.35–13.67) 2.34 (1.42– 4.61) 4.77 ⫾ 5.52 9.35 (6.39 –12.95) 29.17 ⫾ 19.38 1.38 (1.00 –2.09) 5.05 ⫾ 1.13 3.00 ⫾ 0.87 1.23 ⫾ 0.33

241 (64.96) 74 (19.95) 56 (15.09) 35 (9.43) 81 (21.83) 135.55 ⫾ 18.68 83.43 ⫾ 11.17 24.97 ⫾ 3.13 24.68 ⫾ 5.69 87.56 ⫾ 8.52 0.89 ⫾ 0.08 7.00 ⫾ 2.76 8.55 (5.04 –12.95) 2.27 (1.26 – 4.15) 4.86 ⫾ 6.73 7.78 (5.64 –12.27) 27.27 ⫾ 13.50 1.55 (1.05–2.24) 5.16 ⫾ 1.02 3.04 ⫾ 0.78 1.25 ⫾ 0.70

231 (62.26) 82 (22.10) 58 (15.64) 45 (12.13) 79 (21.29) 134.37 ⫾ 19.95 82.76 ⫾ 11.54 24.72 ⫾ 3.20 21.17 ⫾ 6.06 86.46 ⫾ 9.09 0.90 ⫾ 0.08 6.90 ⫾ 2.44 7.86 (4.98 –12.40) 2.19 (1.33– 4.06) 5.17 ⫾ 6.41 8.94 (5.83–12.69) 27.88 ⫾ 19.44 1.53 (1.07–2.12) 5.22 ⫾ 0.98 3.07 ⫾ 0.84 1.24 ⫾ 0.30

230 (61.99) 88 (23.72) 53 (14.29) 54 (14.76) 70 (18.87) 133.17 ⫾ 17.38 81.97 ⫾ 10.21 25.25 ⫾ 2.82 25.11 ⫾ 6.56 87.82 ⫾ 7.94 0.90 ⫾ 0.08 6.9 ⫾ 2.75 7.87 (4.95–12.77) 2.28 (1.24 – 4.53) 5.44 ⫾ 6.39 8.57 (4.90 –11.20) 28.12 ⫾ 15.39 1.64 (1.15–2.38) 5.29 ⫾ 0.99 3.09 ⫾ 0.80 1.20 ⫾ 0.27

P value ⬍0.001 ⬍0.001 0.076 0.012 0.185 0.352

0.709

0.015 0.772 0.501 0.237 0.065 0.139 0.225 ⬍0.001 0.147 0.084 0.124 ⬍0.001 0.246 0.645 ⬍0.001 ⬍0.001 ⬍0.001 0.209

Q, Quartile; SBP, systolic blood pressure; DBP, diastolic blood pressure. Data are means ⫾ SD, median (interquartile range), or number (percent); P value was calculated after adjustment for age; b not adjusted for itself; self-reported CVD including stroke and coronary heart disease; d parents or siblings had a history of diabetes; e these variables were log transformed before analysis. a c


TABLE 2.


3929

Characteristics of female participants according to CXCL5 quartilesa

Characteristics Hypercholesterolemia (%) CXCL5 (pg/ml) Age (yr)b Smoking (yes) Alcohol (yes) Education (yr) 0–6 7–9 ⱖ10 Physical activity Low Moderate High Self-reported CVDc Family history of diabetesd SBP (mm Hg) DBP (mm Hg) BMI (kg/m2) Body fat (%) Waist circumference (cm) Waist to hip ratio Glucose (mmol/liter) Insulin (␮U/ml)e HOMA-IRe CRP (␮g/ml) Adiponectin (␮g/ml)e ALT (U/liter) Triglycerides (mmol/liter)e TC (mmol/liter) LDL cholesterol (mmol/liter) HDL cholesterol (mmol/liter)

Q1 (n ⴝ 435) 2481.58 43.45 3592.93 ⫾ 1570.58 60.13 ⫾ 8.76 62 (14.25) 11 (2.53)

P value ⬍0.001 ⬍0.001 0.115 0.024 0.297 0.175

20 (4.60) 137 (31.49) 278 (63.91) 0.641

365 (83.91) 27 (6.21) 43 (9.88) 31 (7.13) 73 (16.78) 133.03 ⫾ 20.55 80.22 ⫾ 10.76 24.86 ⫾ 3.52 30.64 ⫾ 6.64 82.01 ⫾ 9.39 0.85 ⫾ 0.06 6.36 ⫾ 2.28 8.20 (5.05–13.45) 2.00 (1.17–3.44) 3.97 ⫾ 4.16 12.20 (8.87–15.79) 24.28 ⫾ 16.15 1.42 (0.98 –2.05) 5.37 ⫾ 1.04 3.01 ⫾ 0.79 1.43 ⫾ 0.38

351 (80.50) 34 (7.80) 51 (11.70) 41 (9.40) 88 (20.18) 131.63 ⫾ 19.96 80.03 ⫾ 10.09 24.55 ⫾ 3.56 31.86 ⫾ 6.12 81.91 ⫾ 9.13 0.85 ⫾ 0.06 6.91 ⫾ 2.50 7.88 (5.16 –12.08) 1.97 (1.21–3.13) 4.13 ⫾ 5.13 10.26 (7.31–14.80) 24.47 ⫾ 16.20 1.38 (0.99 –1.99) 5.46 ⫾ 0.98 3.14 ⫾ 0.78 1.42 ⫾ 0.37

347 (79.59) 36 (8.26) 53 (12.15) 48 (11.01) 86 (19.72) 128.38 ⫾ 18.69 79.10 ⫾ 10.28 24.44 ⫾ 3.45 31.25 ⫾ 5.75 81.59 ⫾ 8.65 0.85 ⫾ 0.07 6.12 ⫾ 2.16 8.03 (5.32–11.70) 2.24 (1.26 – 4.28) 4.37 ⫾ 4.40 10.62 (7.70 –13.94) 25.23 ⫾ 15.50 1.40 (0.97–2.07) 5.62 ⫾ 1.15 3.27 ⫾ 0.93 1.41 ⫾ 0.31

343 (78.85) 38 (8.74) 54 (12.41) 57 (13.10) 80 (18.39) 129.89 ⫾ 18.08 81.41 ⫾ 27.35 24.79 ⫾ 3.45 31.74 ⫾ 4.23 81.93 ⫾ 8.64 0.86 ⫾ 0.08 6.15 ⫾ 2.16 7.25 (4.59 –10.56) 1.77 (1.04 –2.78) 4.89 ⫾ 5.27 10.14 (7.42–14.24) 23.58 ⫾ 12.44 1.47 (1.04 –2.17) 5.71 ⫾ 1.04 3.34 ⫾ 0.85 1.46 ⫾ 0.58

0.028 0.567 0.085 0.141 0.213 0.777 0.085 0.007 0.152 0.375 0.092 0.017 0.229 0.455 ⬍0.001 ⬍0.001 ⬍0.001 0.513

Q, Quartile; SBP, systolic blood pressure; DBP, diastolic blood pressure. Data are means ⫾ SD, median (interquartile range), or number (percent); P value was calculated after adjustment for age; b not adjusted for itself; self-reported CVD including stroke and coronary heart disease; d parents or siblings had a history of diabetes; e these variables were log transformed before analysis. a c

more likely to be more smokers (P ⬍ 0.05), more self-reported CVD (P ⬍ 0.05), and higher waist to hip ratio (P ⬍ 0.01). With respect to metabolic parameters, the subjects in the higher CXCL5 quartiles exhibited higher levels of CRP (P ⬍ 0.05), TC (P ⬍ 0.001), and LDL cholesterol (P ⬍ 0.001). However, elevated CXCL5 levels showed no association with the levels of BMI, HOMA-IR, and adiponectin. Distribution of CXCL5 levels The crude mean (SD) of serum CXCL5 concentration was 1493.31 (1250.38) pg/ml for male and 2059.42 (1355.40) pg/ml for female (P ⬍ 0.001), respectively (Fig. 1). Association between CXCL5 and hypercholesterolemia Partial correlation analysis demonstrated the strongest correlation between CXCL5 and TC among various metabolic features (Supplemental Table 2).

Remarkably, as presented in Table 3, the ORs for hypercholesterolemia were higher with increasing CXCL5 quartiles (P ⬍ 0.001 for trend). In the highest CXCL5 quartile, the ORs were 2.86 [95% confidence interval (CI) 2.31–3.54] for hypercholesterolemia after adjusting for age, gender, lifestyle factors, educational attainment, selfreported CVD, and family history of diabetes (model 2). Interestingly, further adjustment for CRP, adiponectin, and HOMA-IR (model 3) only slightly reduced the magnitude of the ORs for hypercholesterolemia. Furthermore, the ORs for hypercholesterolemia were not substantially attenuated by additional adjustment for BMI, body fat, waist circumference, and hip circumference (model 4) (OR 2.63; 95% CI 1.79 –3.84). In addition, when analyzed separately in men and women, the ORs for hypercholesterolemia were 2.83 (95% CI 1.89 –3.85) in men and 3.17 (95% CI 2.53–3.96) in women with simple adjustment (model 1). In the multivariate adjusted analyses (model 4), the ORs were 2.35 (95% CI 1.57–3.32) and 2.84 (95% CI

3930

Yang et al.



Discussion

FIG. 1. Serum CXCL5 in male and female subjects. Data are shown as means ⫾ SE after adjustment for age. †, P ⬍ 0.001.

2.35–3.62)for men and women, respectively (detail data shown in Supplemental Table 3). There was no association of CXCL5 with HOMA-IR in the total population (P ⬎ 0.05), after controlling for age and sex (detail data shown in Supplemental Table 2). To determine whether obesity status affect the associations of CXCL5 with HOMA-IR, stratified analyses was performed. Similarly, there were no associations of CXCL5 with HOMA-IR in normal weight (⬍24.0 kg/m2), overweight (24 –28 kg/m2), or obesity (ⱖ28.0 kg/m2) group (all P ⬎ 0.05). Furthermore, to investigate whether smoking affect the association of CXCL5 with obesity status and insulin resistance, stratified analyses was also performed. We also fail to observe there are associations of CXCL5 with obesity status and insulin resistance in smoker and nonsmoker group (P ⬎ 0.05). Moreover, we observed insignificant differences in CXCL5 levels between NGT, i-IFG, i-IGT, IFG⫹IGT, and diabetes after adjustment for age and gender (P ⬎ 0.05).

TABLE 3.

In this study, we found a strong association between CXCL5 levels and the risk of hypercholesterolemia in a large-scale population study. Moreover, this association is independent of lifestyle factors; education status; family history of chronic diseases; and, remarkably, CRP, adiponectin, HOMA-IR, percent body fat, and BMI. One of the interesting findings of the present study is that the distribution of CXCL5 levels in the middle-aged and elderly population in China substantially differed by gender. The serum CXCL5 concentrations were found to be sexually dimorphic in our study population, even after adjusting for the confounders, which had previously been reported for adipokines, such as leptin and adiponectin, which might be explained by the different levels of adiposity and influences of sex hormones on adipokines (17, 18). As described previously, no difference was found in serum CXCL5 levels with respect to fat amounts or body fat percentages in the present study. Thus, it appears that differences in sex hormone status might affect CXCL5 serum levels. CXCL5 has been shown as a potent regulator of glucose up-take and insulin sensitivity by activating the Janus kinase 2/signal transducer and activator of transcription 5/ suppressor of cytokine signaling 2 pathway in muscles tissue and primary cultures of mouse embryonic fibroblasts (10). Further analysis suggests that TNF-␣, which is a well-characterized cytokine with insulin resistance (19), could trigger the expression of CXCL5 in WAT through the nuclear factor-␬B pathway (10). However, our study demonstrated that serum concentrations of CXCL5 in humans were not associated with insulin sensitivity as measured by the HOMA-IR. Moreover, although data regarding serum TNF␣ level was unavailable in our study, notably the lack of an independent association between CXCL5 and adiponectin, which is also a well-documented

Adjusted ORs and 95% CIs for hypercholesterolemia according to CXCL5 quartiles ORs (95% CI)

Hypercholesterolemia Model 1a Model 2b Model 3c Model 4d

Q1

Q2

Q3

Q4

P value for trend

1.0 1.0 1.0 1.0

1.58 (1.30 –1.93) 1.46 (0.97–2.15) 1.43 (0.96 –2.13) 1.39 (0.91–2.10)

1.94 (1.58 –2.37) 1.72 (1.17–2.54) 1.65 (1.11–2.45) 1.61 (1.03–2.35)

2.86 (2.31–3.54) 2.78 (1.90 – 4.06) 2.69 (1.83–3.96) 2.63 (1.79 –3.84)

⬍0.001 ⬍0.001 ⬍0.001 ⬍0.001

Q, Quartile. a

Model 1 adjusted for age and gender.

b

Model 2 further adjusted for alcohol drinking, smoking, education, physical activity, self-reported CVD, family history of diabetes, systolic blood pressure, and diastolic blood pressure. c

Model 3 further adjusted for CRP, adiponectin, and HOMA-IR.

d

Model 4 further adjusted for BMI, body fat, waist circumference, and hip circumference.


adipokine with potent insulin-sensitizing activity (20), suggested that CXCL5 might not be closely related to insulin sensitivity but secondary to the effects of obesity. Thus, our data indicate that in humans, unlike in rodents, there is no direct relation between circulating levels of CXCL5 and insulin action in peripheral tissues. Contrary to the previous finding that obese subjects have significant higher serum CXCL5 levels than lean subjects (10), we found no association between serum CXCL5 levels and the amount or percentage of body fat. Further investigation also found serum CXCL5 concentration no association with percentage of body fat and BMI in different obesity status. However, obese ob/ob and db/db and high-fat diet-fed mice were found to have a 2-fold higher serum CXCL5 level than lean mice, and, according to Chavey et al. (10), obese human subjects with elevated serum CXCL5 levels had unequivocally higher BMIs than lean control subjects. Moreover, we found that waist to hip ratio increased with serum CXCL5 quartiles in both male and female subjects (P ⬍ 0.01). It is possible that sc fat is more important than total body fat in determining circulating CXCL5 levels. In our study, we observed that the positive association of CXCL5 with smoking in both male and female subjects. As was best known, chronic smoking could induce lowgrade systemic inflammation. CXCL5 is an adipokine and also an inflammatory factor, and serum CXCL5 levels should be increased with the proportion of subjects with smoking. It has been demonstrated that low-grade systemic inflammation plays a key role in the process of weight gain and insulin resistance (1). Smoking may affect the association of serum CXCL5 level, obesity status, and insulin resistance. However, our further analysis indicated that the effects of serum CXCL5 levels on obesity and insulin resistance were independent of smoking. We found a significant positive association of CXCL5 levels with inflammatory factor CRP by partial correlation (P ⬍ 0.05). In addition, we found serum CRP levels increased with serum CXCL5 quartiles in both male and female subjects (P ⬍ 0.05), which indicate that CXCL5, as a innate immunity marker (20), is correlated with lowgrade systemic inflammatory response and could potentially modulate chronic inflammatory processes. Moreover, serum TC was also found to be independently associated with serum CXCL5 by multiple stepwise regression analysis (Table 4). Therefore, our data suggest that serum CXCL5 levels are more related to lipid metabolism instead of insulin sensitivity. It is well established that elevated serum CRP and TC levels were the independent risk factors for the development of CVD (21). Given that serum CXCL5 levels were significantly associated with CRP and TC levels, it is plausible to consider CXCL5


3931

TABLE 4. Multiple stepwise regression analysis showing variables independently associated with the serum level of CXCL5 Independent variables Gender CRP TC

Standardized ␤ ⫺0.341 0.190 0.172

t ⫺6.702 3.850 3.246

P value ⬍0.001 ⬍0.001 ⬍0.001

The analysis also included BMI, waist circumference, hip circumference, percentage of body fat, ALT, triglycerides, HDL cholesterol, LDL cholesterol, HOMA-IR, obesity status, and glucose tolerance status, which were all excluded in the final model.

as a promising candidate for risk assessment and a potential intervention target for CVD. Interestingly, in line with our hypothesis, we observed that incidence of self-reported CVD increased with serum CXCL5 quartiles in both male and female subjects in this study (P ⬍ 0.05). Certainly prospective studies with solid clinical end points are urgently needed to clarify whether a high CXCL5 level plays a causal role in the development of CVD. To our knowledge, this is the first study to evaluate the relationship between CXCL5 levels and hypercholesterolemia and insulin resistance in a large-scale population. Most potential confounders were carefully controlled, which limited the possibility of residual confounding effects. Furthermore, CXCL5 concentrations were measured in duplicate, and the detection was completed within 2 months to minimize seasonal influences on biomarkers. However, due to the cross-sectional nature of the present study, admittedly we could not determine whether CXCL5 plays a causal role in the pathogenesis of hypercholesterolemia. Also, it has yet to be seen whether our results in middle-aged and older Chinese subjects can be generalized to younger populations or other ethnic groups. In conclusion, we have found that elevated CXCL5 levels are strongly and independently associated with hypercholesterolemia but not insulin resistance. Although longitudinal studies are needed, our findings provide novel insights into the potential role of CXCL5 in the pathogenesis of hypercholesterolemia as well as in the prevention and management of dyslipidemia.

Acknowledgments Address all correspondence and requests for reprints to: Renming Hu, Institute of Endocrinology and Diabetology at Fudan University, Department of Endocrinology and Metabolism, Huashan Hospital affiliated with Fudan University, no. 12 Wulumuqi Middle Road, Shanghai 200040, China. E-mail: [email protected]. This work was supported by the National Natural Science Foundation of China (330670999, 30711120573, 3030770854),

3932

Yang et al.


Shanghai Science and Technology Commission (08dj1400605, 08JC1403200, 09DZ1950200), the National High Technology Research and Development Program (“863” Program) of China (2009AA022704), China Postdoctoral Science Foundation (20080440078), and Shanghai Postdoctoral Scientific Program (09R21411600). Disclosure Summary: The authors have nothing to disclose.

References 1. Wellen KE, Hotamisligil GS 2005 Inflammation, stress, and diabetes. J Clin Invest 115:1111–1119 2. Wellen KE, Hotamisligil GS 2003 Obesity-induced inflammatory changes in adipose tissue. J Clin Invest 112:1785–1788 3. Xu H, Barnes GT, Yang Q, Tan G, Yang D, Chou CJ, Sole J, Nichols A, Ross JS, Tartaglia LA, Chen H 2003 Chronic inflammation in fat plays a crucial role in the development of obesity-related insulin resistance. J Clin Invest 112:1821–1830 4. Heilbronn LK, Campbell LV 2008 Adipose tissue macrophages, low grade inflammation and insulin resistance in human obesity. Curr Pharm Des 14:1225–1230 5. Kahn BB, Flier JS 2000 Obesity and insulin resistance. J Clin Invest 106:473– 481 6. Porte D Jr., Baskin DG, Schwartz MW 2002 Leptin and insulin action in the central nervous system. Nutr Rev 60:S20 –S29 7. Steppan CM, Bailey ST, Bhat S, Brown EJ, Banerjee RR, Wright CM, Patel HR, Ahima RS, Lazar MA 2001 The hormone resistin links obesity to diabetes. Nature 409:307–312 8. Tsao TS, Lodish HF, Fruebis J 2002 ACRP30, a new hormone controlling fat and glucose metabolism. Eur J Pharmacol 440:213–221 9. Walz A, Burgener R, Car B, Baggiolini M, Kunkel SL, Strieter RM 1991 Structure and neutrophil-activating properties of a novel inflammatory peptide (ENA-78) with homology to interleukin 8. J Exp Med 174:1355–1362 10. Chavey C, Lazennec G, Lagarrigue S, Clape´ C, Iankova I, Teyssier J, Annicotte JS, Schmidt J, Mataki C, Yamamoto H, Sanches R, Guma A, Stich V, Vitkova M, Jardin-Watelet B, Renard E, Strieter R, Tuthill A, Hotamisligil GS, Vidal-Puig A, Zorzano A, Langin D, Fajas L 2009 CXC ligand 5 is an adipose-tissue derived factor that links obesity to insulin resistance. Cell Metab 9:339 –349 11. Walz A, Schmutz P, Mueller C, Schnyder-Candrian S 1997 Regu-

12.

13.

14.

15. 16.

17.

18.

19.

20.

21.


lation and function of the CXC chemokine ENA-78 in monocytes and its role in disease. J Leukoc Biol 62:604 – 611 Zhou BF 2002 Predictive values of body mass index and waist circumference for risk factors of certain related diseases in Chinese adults—study on optimal cut-off points of body mass index and waist circumference in Chinese adults. Biomed Environ Sci 15: 83–96 Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF, Turner RC 1985 Homeostasis model assessment: insulin resistance and ␤-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia 28:412– 419 Alberti KG, Zimmet PZ 1998 Definition, diagnosis and classification of diabetes mellitus and its complications, part 1: diagnosis and classification of diabetes mellitus: provisional report of a WHO consultation. Diabet Med 15:539 –553 American Diabetes Association 2005 Diagnosis and classification of diabetes mellitus. Diabetes Care 28:S37–S42 National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III) 2002 Third Report of the National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III) final report. Circulation 106: 3143–3421 Combs TP, Berg AH, Rajala MW, Klebanov S, Iyengar P, JimenezChillaron JC, Patti ME, Klein SL, Weinstein RS, Scherer PE 2003 Sexual differentiation, pregnancy, calorie restriction, and aging affect the adipocyte-specific secretory protein adiponectin. Diabetes 52:268 –276 Saad MF, Damani S, Gingerich RL, Riad-Gabriel MG, Khan A, Boyadjian R, Jinagouda SD, el-Tawil K, Rude RK, Kamdar V 1997 Sexual dimorphism in plasma leptin concentration. J Clin Endocrinol Metab 82:579 –584 Hivert MF, Sullivan LM, Fox CS, Nathan DM, D’Agostino Sr RB, Wilson PW, Meigs JB 2008 Associations of adiponectin, resistin, and tumor necrosis factor-␣ with insulin resistance. J Clin Endocrinol Metab 93:3165–3172 Gear AR, Camerini D 2003 Platelet chemokines and chemokine receptors: linking hemostasis, inflammation, and host defense. Microcirculation 10:335–350 Ridker PM, Rifai N, Rose L, Buring JE, Cook NR 2002 Comparison of C-reactive protein and low-density lipoprotein cholesterol levels in the prediction of first cardiovascular events. N Engl J Med 347: 1557–1565