DOI: 10.1111/eci.12247

ORIGINAL ARTICLE Interleukin 10 and clustering of metabolic syndrome components in pediatrics Jung-Su Chang*, Chyi-Huey Bai†, Zu-Chieh Huang*, Eddy Owaga*, Kuo-Ching Chao‡,§, Chun-Chao Chang‡,§ and Hung-Yi Chiou¶ *School of Nutrition and Health Sciences, College of Public Health and Nutrition, Taipei Medical University, Taipei, Taiwan, † Department of Public Health, College of Medicine, Taipei Medical University, Taipei, Taiwan, ‡Division of Gastroenterology and Hepatology, Department of Internal Medicine, Taipei Medical University Hospital, Taipei, Taiwan, §Department of Internal Medicine, School of Medicine, College of Medicine, Taipei Medical University, Taipei, Taiwan, ¶School of Public Health, Taipei Medical University, Taipei, Taiwan

ABSTRACT Background Interleukin 10 (IL-10) has multifaceted anti-inflammatory properties that are known to regulate insulin sensitivity and atherosclerotic development. However, studies in children are limited and have yielded conflicting results. The aim of this study was to evaluate whether changes in this circulating anti-inflammatory cytokine is a marker for metabolic syndrome. Materials and methods This cross-sectional study involved children and young adolescents from eight elementary schools and two junior high schools located in Taipei and New Taipei City. A total of 553 children ages 8, 11 and 13 years old were included in the analysis. Parameters for obesity, anti- and pro-inflammatory cytokines, and metabolic risk profiles were evaluated. Results Overweight/obese children had lower serum IL-10 concentrations compared with normal weight children in the same age group (all P < 0
001). IL-10 quartiles were negatively associated with body mass index (BMI) and percentage (%) body fat (all P < 0
05). Multivariate regression analysis showed significant inverse relationship between IL-10 concentrations and % body fat (b = 0
009, P < 0
0001), and total cholesterol (b = 0
726, P = 0
003), and a small positive correlation between IL-10 and systolic blood pressure (b = 0
980, P = 0
027). In normal weight children, IL-10 concentrations were independently associated with fasting plasma insulin (b = 0
2912, P = 0
001) and waist circumference (b = 0
0069, P = 0
022). By contrast, % body fat (b = 0
016, P = 0
0009) was independently associated with IL-10 concentrations in overweight and obese children. Association between IL-10 and fasting plasma insulin concentrations was weaker in overweight/obese children compared with normal weight (b = 0
283, P = 0
011 vs. b = 0
2912, P = 0
001). Conclusion Our data indicate that changes in circulating IL-10 concentrations are marker of metabolic risk in children. Keywords Fasting plasma insulin, interleukin 10, metabolic syndrome, obese children, percentage body fat, total cholesterol. Eur J Clin Invest 2014; 44 (4): 384–394

Introduction Interleukin 10 (IL-10) is a pleiotropic cytokine with important immunosuppressive effects [1]. IL-10 is produced by many cells including T cells, nature killer cells and antigen presenting cells (APCs). When insults occur, the predominant function of IL-10 is to prevent tissue damage by limiting effective immune response [2]. Broadly speaking, IL-10 functions at two distinct stages: immune priming and host immune response. In the early phase of immune response, IL-10 secreting

384

dendritic cells (DC) polarizes na€ıve T cells to become regulatory T cells. By acting on APCs such as DC or macrophage, IL-10 inhibits the development of T helper 1 or T helper 2 cells [1]. In the initial phase of chronic inflammation, IL-10 concentrations may be elevated, which represents an attempt to suppress inflammatory response. However, IL-10 concentrations may decrease in the late stage of chronic inflammation, which suggests an attempt to control inflammation but it fails. Animal model showed the absence of IL-10 is associated with uncontrolled immune response and chronic gut inflammation

ª 2014 Stichting European Society for Clinical Investigation Journal Foundation

INTERLEUKIN 10 IN OBESE CHILDREN

suggesting IL-10 is required to maintain the immune homoeostasis [3]. There is evidence linking IL-10 to metabolic syndrome (MetS). Studies in adults showed circulating IL-10 might exert some beneficial metabolic effects [4–7]. However, the role of IL10 in childhood obesity remains poorly characterized [8–11]. Whether IL-10 concentrations are altered in overweight and obese children remains controversial [8,9,11]. Two individual reports showed elevated IL-10 concentrations in overweight/ obese children and young adolescents [9,11]. By contrast, Gozal et al. [8] reported that obese children with obstructive sleep apnoea had low plasma IL-10 concentrations. Thus, the clinical significance of IL-10 in childhood obesity remains to be explored. Obesity is characterized by a low-grade inflammation. Infiltrating macrophages and immune cells that reside in adipose tissue are the key regulators of the chronic inflammation. Therefore, the development of inflammation in childhood obesity involves a panel of inflammatory mediators, including adipokines, chemokines and endocrine peptide. Currently, biomarkers of childhood obesity include adiponectin, highmobility group protein B1 (HMGB1) and pro-inflammatory cytokines (e.g. TNF-a, IL1-b, IL-6). Elevated HMGB1 level, a potent pro-inflammatory inducer, has recently been reported as a biomarker of MetS in obese children [12]. Adiponectin concentrations are decreased in obese children and have been identified as early marker of atherosclerosis in obese children [13]. IL-10 is the only known cytokine that are able to downregulate pro-inflammatory immune response. Adiponectin can induce IL-10 synthesis [14]; hence, it has been suggested that the protective effect of adiponectin is mediated through, at least in part, IL-10 [15]. Currently, the relationship between IL-10 and HMGB1 is not understood. The purpose of this study was to assess the role of IL-10 in obese children and to investigate its diagnostic profile in identifying metabolic risk profiles among 553 school children in Taiwan.

Subjects and methods Study design and participants This cross-sectional study involved children and young adolescents from eight elementary schools and two junior high schools located in Taipei and New Taipei City. From September 2009 to November 2011, 161 children (ages 7
95  0
59 years, 81 boys and 80 girls), 160 children (ages 10
88  0
54 years, 87 boys and 73 girls) and 340 young adolescents (ages 13
33  1
10 years, 182 boys and 158 girls) were enrolled in the study. Inclusion criteria were as follows: (i) children who were enrolled in the selected schools and were in grade 1 and grade 4 (elementary school) or grade 7 (junior high school); (ii) adhered to the requirement not to drink or eat after midnight or exercise 24 h prior to data collection; and (iii) free of medical conditions.

Exclusion criteria were as follows: (i) individuals with missing data for clinical biochemistry and anthropometry; (ii) individuals with serum ferritin concentrations > 500 ng/mL; (iii) the presence of acute or chronic inflammation; (iv) use of medications; and (v) secondary obesity syndrome. A total of 544 children (291 boys and 253 girls) were included in the analysis. Informed parental written consent was obtained prior to enrolment into the study. The study was approved by the Research Ethics Committee of Taipei Medical University (201204011). The design and presentation of this study conform to STROBE and EQUATOR guidelines [16].

Data collection Data were collected from the subjects by the same medical staff from the Taipei Medical University Hospital using the similar methods and tools. Children were advised not to drink or eat after midnight or exercise 24 hours prior to data collection. Measurements of body weight and height, waist circumference and blood pressure were conducted as described elsewhere [17]. Waist circumference measurements were taken at the midpoint between the lower edge of the rib cage and the top of the iliac crest [18]. Percentage of body fat was estimated with bioelectrical impedance (Omron Body Fat Analyzer HBF-306).

Blood biochemistry examination Fasting plasma samples were collected by disposable vacuum blood collection tube. All blood samples were separated into red blood cells and plasma and stored at 80 °C until analysis. Low-density lipoprotein cholesterol (LDL), high-density lipoprotein cholesterol (HDL), total cholesterol and triglyceride (TG) were determined by autoanalyser (Hitachi 737, EquipNet, Inc, Canton, MA, USA). Fasting plasma glucose (FPG) concentration was detected using a glucose oxidase method (YSI 203 glucose analyser, YSI Incorporated Life Sciences, Yellow Springs, OH, USA). Homoeostasis model assessment-estimated insulin resistance (HOMA-IR) was used as an index of insulin resistance (IR) in this study [19]. HOMA-IR was calculated using fasting blood glucose (mmol/L) 9 insulin (uIU/mL)/ 22
5. Cytokines concentrations (IL1-b, TNF-a, IL-10, IFN-c) were determined by Enzyme-Linked Immunosorbent Assay kit (Procarta Cytokine Assay Kit; Affymetrix, Inc., Santa Clara, CA, USA) according to the manufacturer’s instructions. Nitric oxide (NO) concentration in the serum was determined by Griess reagent system.

Definition of metabolic syndrome (MetS) and its individual components Age–sex-specific cut-off point for body mass index (BMI) was used to define overweight and obesity in children according to guidelines of the Department of Health in Taiwan [11,20]. BMI was calculated as mass (Kg)/[height(m)]2. Children with

European Journal of Clinical Investigation Vol 44

385

J.-S. CHANG ET AL.

www.ejci-online.com

BMI > 85th percentile of age–sex-specific value were grouped as overweight while those with BMI > 95th were classified as obese. Although there is no consensus definition for MetS in paediatrics, it is well accepted that the risk factors associated with MetS are similar between adults and children. To estimate the metabolic risk in Taiwanese children, children with values of individual components of MetS exceeding the 90th percentile for age and gender were defined as abnormal. HDL cholesterol concentration below the 10th percentile for age and gender in the fasting levels of lipids was considered abnormal. Table 1 shows the age- and gender-specific cut-off values for MetS. Individuals with the presence of ≥ 3 criteria listed in Table 1 were classified as MetS [21]. FPG ≥ 126 mg/dL was diagnosed as type 2 diabetes mellitus (DM), and FPG ≥ 110 mg/dL was defined as pre-DM.

Statistical analyses Statistical analyses were performed using the SAS software (SAS version 9.22; SAS Institute, Inc, Wellington, New Zealand). Data were presented as median  interquartile range (IQR). Differences between two independent samples were analysed by the Wilcoxon rank-sum test for the nonparametric data.

Table 1 Age- and gender-specific cut-off points for metabolic syndrome in children Age Components of MetS

Gender

8 years

11 years

13 years

Waist circumference (cm)

Boys

≥ 77

≥ 89

≥ 86
5

Girls

≥ 75

≥ 87

≥ 84
5

Boys

≥ 118

≥ 108

≥ 115

Girls

≥ 98

≥ 134

≥ 133

Boys

≥ 49

≥ 47

≥ 42

Girls

≥ 46

≥ 41

≥ 39

Boys

≥ 119
8

≥ 119

≥ 123

Girls

≥ 115

≥ 124

≥ 123

Boys

≥ 70
8

≥ 77

≥ 73

Girls

≥ 69

≥ 77

≥ 75

Boys

≥ 94

≥ 94

≥ 102

Triglyceride (mg/dL)

HDL cholesterol (mg/dL)

Blood pressure (mmHg) Systolic blood pressure (SBP)

Diastolic blood pressure (DBP)

Fasting plasma glucose (mg/dL)

Girls

≥ 94

≥ 97

≥ 98
5

HDL, high-density lipoprotein cholesterol; MetS, metabolic syndrome. Individuals with the presence ≥ 3 criteria listed above were classified as MetS.

386

Variables not normally distributed were log-transformed to achieve normal distribution and to allow the use of parametric tests. However, untransformed values were used for reporting the results. The association between IL-10 and parameters of MetS was assessed using Pearson’s correlation coefficient. Multivariate linear regression tests were performed to examine the relationship between IL-10 and metabolic risk profiles. P < 0
05 was considered statistically significant.

Results Participant characteristics The overall prevalence of overweight was 18
0% (63
2% for boys and 36
7% for girls) and obesity was 25
0% (52
9% for boys and 47
0% for girls). The prevalence of MetS was 3
5% (boys, 1
8% and girls, 1
6%). There were no differences in the prevalence of MetS and its individual components between boys and girls. The prevalences of individual components of MetS were as follows: low HDL cholesterol, 8
6% (boys 4
5% and girls 4
0%); high TG, 10
5% (boys 5
7% and girls 4
8%); high BP, 18
9% (boys 9
7% and girls 9
2%); high FPG, 11
2% (boys 6
4% and girls 4
8%); and abdominal obesity, 9
0% (boys 4
8% and girls 4
2%). Four children (two normal weight and two overweight/obese) were diagnosed with type 2 diabetes, and 13 children (five normal weight and eight overweight/obese) were classified as pre-DM. Table 2 shows the clinical characteristics, anthropometric and biochemical parameters of the normal weight and overweight/obese children. Ordinary P-trend analysis showed anthropometric indices (height, weight, BMI and waist), total cholesterol, fasting plasma insulin, HOMA-IR and diastolic blood pressure (DBP) were increased with increasing ages in both normal weight and overweight/obese children (Table 2; all P < 0
05). A positive trend for metabolic risk profiles [% body fat, systolic blood pressure (SBP), TG, HDL and FPG] was found only in overweight/obese children (Table 2; all P < 0
05). IL-10 and TNF-a concentrations were also positively associated with age in both normal weight and overweight/ obese children (all P < 0
0001). Serum IL-10 concentrations in 20
6% of the children were below ELISA detection limit. Further analysis showed 37% of overweight/obese children had undetectable IL-10 concentrations compared with 8% in normal weight children. This resulted in a lower serum IL-10 concentrations in overweight/obese children compared with normal weight children in the same age group (Table 2; all P < 0
01).

Association between IL-10 concentrations and clustering of metabolic syndrome We next evaluated the association between serum IL-10 and clustering of metabolic factors by IL-10 quartiles. The clinical characteristics of the children in relation to IL-10 quartiles are

ª 2014 Stichting European Society for Clinical Investigation Journal Foundation

125
50

24
30

Height (cm)

Weight (kg)

59
00

54
00

Diastolic BP (mmHg)

TG (mg/dL)

4
96

1
11

0
12

2
48

0
87

4
02

Fasting plasma insulin (lIU/mL)

HOMA-IR

IL-10 (pg/mL)

TNF-a (pg/mL)

IL1-b (pg/mL)

NO (lM)

7
07

1
55

2
69

0
89

1
07

5
15

8
00

11
00

36
00

29
00

9
00

13
50

6
00

5
00

2
37

4
70

9
50

0
53

5
02

0
91

0
39

0
00

7
22

1
24

2
69

0
12

1
47

1
92### ##

5
65

8
00

16
00

42
00

46
00

14
50

14
50

6
90

7
00

3
74

8
90

6
55

0
43

8
91###

86
00

59
00

167
00

68
00

61
50

105
00

31
40

67
00

21
45

37
40

133
11

8
10

3
77

0
80

1
02

0
35

1
46

6
76

89
50

61
00

168
50

50
50

60
00

104
50

19
40

60
2

17
20

36
30

143
85

10
88

(20/20)

3
52

1
96

4
86

0
89

1
14

5
04

6
50

13
00

45
00

27
50

11
00

13
00

8
80

5
00

3
08

8
65

9
25

0
49

IQR

7
75

1
12

1
02

0
00

##

2
63###

11
83###

89
00

55
00

10
06

1
15

3
98

0
35

2
62

11
31

7
00

12
00

41
00

44
00

74
00## 178
00

13
00

15
00

8
00

10
00

3
19

13
70

12
10

0
55

66
00

104
50

32
40

80
00

23
25

50
90

148
70

10
73

(44/35)

IQR

OW/obese* Median

6
13

0
90

16
69

9
87

2
58

11
37

90
00

58
00

158
00

62
00

62
00

109
00

19
00

64
00

18
40

44
70

155
00

13
27

(122/109)

Median

Normal

5
74

0
41

10
37

3
69

1
41

5
96

8
00

17
00

37
00

32
00

11
00

12
00

8
10

8
00

3
76

8
30

9
00

1
06

IQR

8
40

1
15

21
45

8
33

#

4
55###

19
88###

92
00#

45
50

153
50

76
50##

64
00

114
50

28
10

83
00

25
06

62
25

157
95

13
55

(53/41)

Median

5
19

0
72

11
83

3
42

4
03

16
51

10
00

12
00

43
00

46
00

10
00

12
00

8
00

9
50

3
44

12
70

8
80

1
16

IQR

OW/obese*

Junior high school

Grade 7

< 0
0001

< 0
0001

0
5876

< 0
0001

< 0
0001

< 0
0001

0
5502

0
1261

0
5877

< 0
0001

< 0
0001

0
5848

0
0003

0
0001

0
8449

0
7139

0
7853

0
9723

0
1592

0
2440

0
0169

0
5876

< 0
0001

< 0
0001

0
3832

0
3438 0
0368

0
0195

0
0317

0
001

0
0667

0
0667

0
0029

0
5953

0
0172

0
1489

0
3199

0
0034

0
3480

< 0
0001

< 0
0001

< 0
0001

< 0
0001

< 0
0001

< 0
0001

< 0
0001

0
5143

< 0
0001

< 0
0001

< 0
0001

544

Pooled

0
0081

< 0
0001

< 0
0001

0
6397

234

OW/ obese*

310

Normal

Ordinary P-trend

BMI, body mass index; HDL, high-density lipoprotein cholesterol; HOMA-IR, homoeostasis model assessment-estimated insulin resistance; IQR, interquartile range; NO, nitric oxide; TG, triglyceride. *OW/obese: overweight and obese. # P < 0
05, ##P < 0
01, and ###P < 0
001, comparing normal weight with overweight/obese in the same age group using Wilcoxon’s rank-sum test.

85
00

Fasting glucose (mg/dL)

63
00

107
00

Systolic BP (mmHg)

HDL cholesterol (mg/dL)

17
80

Body fat (%)

171
00

57
00

Waist (cm)

Total cholesterol (mg/dL)

15
67

BMI (m /kg)

2

7
67

Age (years)

(37/24)

(15/24)

Number (male/female)

Median

Normal

IQR

OW/obese* Median

Median

IQR

Normal

Elementary school

Elementary school

Variables

Grade 4

Grade 1

Table 2 Clinical and biochemical data of the 554 children and young adolescents according to nutritional status

INTERLEUKIN 10 IN OBESE CHILDREN

European Journal of Clinical Investigation Vol 44

387

J.-S. CHANG ET AL.

www.ejci-online.com

Table 3 Age- and gender-adjusted clinical and biochemical characteristics of children and young adolescents according to IL-10 quartiles (n = 544) Q1 Variables

Median 2

Q2 IQR

Median

Q3 IQR

Median

Q4 IQR

Median

IQR

Ordinary P-trend

BMI (kg/m )

22
31

4
94

19
13

5
73

20
21

5
98

19
25

4
35

0
011

Body fat (%)

30
35

10
80

23
05

12
60

22
50

10
40

19
80

9
10

0
018

Waist (cm)

73
00

14
00

66
00

14
00

69
00

16
00

68
00

13
00

0
266

Fasting glucose (mg/dL)

88
00

7
00

88
00

8
00

90
00

9
00

91
00

8
00

0
867

Fasting serum insulin (lIU/mL)

11
00

10
08

11
59

7
25

13
77

9
80

13
72

8
09

0
107

2
40

2
21

1
93

1
61

3
09

2
40

3
03

1
99

0
308

Systolic BP (mmHg)

105
00

14
00

107
00

14
00

112
00

13
00

111
00

11
50

0
539

Diastolic BP (mmHg)

62
50

12
00

63
00

13
75

64
00

11
00

62
00

9
00

0
439

HOMA-IR

Total cholesterol (mg/dL)

166
50

40
50

171
50

40
00

152
00

38
00

158
00

39
50

0
625

HDL cholesterol (mg/dL)

56
00

13
50

59
00

16
00

54
00

19
00

56
00

18
50

0
103

LDL cholesterol (mg/dL)

100
00

36
00

99
00

38
00

87
00

33
00

90
50

28
00

0
772

69
00

44
50

62
50

39
50

64
00

37
00

63
00

32
00

0
293

TNF-a (pg/mL)

0
20

2
48

5
19

13
59

19
36

12
27

16
58

8
27

0
326

IFN-c (pg/mL)

1
58

2
28

6
41

6
74

6
40

3
06

5
37

2
91

< 0
0001

NO (lM)

6
43

7
94

6
85

8
37

6
77

5
89

5
89

5
41

0
095

IL1-b (pg/mL)

0
87

0
90

1
20

1
07

1
02

0
57

0
85

0
34

0
912

Triglyceride (mg/dL)

BMI, body mass index; HDL, high-density lipoprotein cholesterol; HOMA-IR, homoeostasis model assessment-estimated insulin resistance; IQR, interquartile range; LDL, low-density lipoprotein cholesterol; NO, nitric oxide. Data are presented as median  IQR and per cent for continuous and categorical variables, respectively. IL-10 cut-off point: 0 ≤ Quartile 1 < 0
12, 0
12 ≤ Quartile 2 < 6
795, 6
795 ≤ Quartile 3 < 10
20 pg/mL, and Quartile 4 ≥ 10
20 pg/mL.

shown in Table 3. After controlling for age and gender, IL-10 quartiles were inversely correlated with BMI (P = 0
011), % body fat (P = 0
018) and positively correlated with IFNc (P < 0
0001) (Table 3). The Pearson’s correlation coefficients between log serum IL10 and selected laboratory parameters are shown in Table 4. To control for the effects of age and gender, these variables were accordingly adjusted in the partial Pearson’s coefficient correlation analysis. After adjusting for age and gender, serum IL-10 concentrations were inversely correlated with % body fat (r = 0
224; P < 0
0001) and total cholesterol (r = 0
142; P = 0
004) (Table 4; pooled, model C). A very small positive correlation between IL-10 concentrations and SBP (r = 0
100, P = 0
042) and between IL-10 and fasting plasma insulin (r = 0
095, P = 0
053) was observed after adjusting for age and gender (Table 4; pooled, model C). We next separated overweight/obese children from normal weight. In normal weight children, IL-10 concentrations were inversely correlated with total cholesterol (r = 0
120, P = 0
046) and positively correlated with waist circumference (r = 0
202, P = 0
001), fasting

388

serum insulin (r = 0
171, P = 0
004) and HOMA-IR (r = 0
147, P = 0
014) after adjusting for age and gender (Table 4; normal weight, model C). These associations between IL-10 and MetS profiles were not significant in overweight and obese children, except fasting serum insulin (r = 0
167, P < 0
05). By contrast, an inverse relationship between IL-10 concentrations and % body fat was found in overweight/obese children after adjusting for age and gender (r = 0
317, P = 0
009) (Table 4; overweight and obese, model C).

Multivariate linear regression model on metabolic risk profiles Multivariate linear regression analysis was performed to evaluate whether serum IL-10 was independently associated with metabolic risk profiles. The % body fat (b = 0
0096, P < 0
0001) and total cholesterol (b = 0
726, P = 0
003) were independently associated with IL-10 (Table 5; pooled, Multivariate). SBP (b = 0
980, P = 0
027) was also independently associated with IL-10 (Table 5; pooled, Multivariate). We next evaluated the effects of BMI on IL-10-related metabolic profiles.

ª 2014 Stichting European Society for Clinical Investigation Journal Foundation

INTERLEUKIN 10 IN OBESE CHILDREN

Table 4 Pearson’s correlation coefficient of log-transformed serum IL-10 with selected anthropometry and individual components of MetS in children Crude r

Variables

Model B†

Model A* P value

r

P value

r

Model C‡ P value

r

P value

(a) Pooled all 0
1927

< 0
0001

0
2236

< 0
0001

0
1932

< 0
0001

0
2242

< 0
0001

BMI (kg/m )

0
1045

0
0299

0
0949

0
052

0
1042

0
0305

0
0952

0
0514

Waist (cm)

0
0787

0
1254

0
0548

0
2932

0
0771

0
1337

0
0558

0
2851

Log systolic BP (mmHg)

0
2342

< 0
0001

0
1000

0
0417

0
2343

< 0
0001

0
1000

0
0420

Log diastolic BP (mmHg)

0
0280

0
5635

0
0643

0
1906

0
0284

0
5589

0
0647

0
1882

Log total cholesterol (mg/dL)

0
2660

< 0
0001

0
1421

0
0035

0
2659

< 0
0001

0
1419

0
0036

Log triglyceride (mg/dL)

0
0454

0
3467

0
0163

0
7397

0
0464

0
3371

0
0152

0
7564

Log HDL cholesterol (mg/dL)

0
1319

0
0061

0
0206

0
6739

0
1331

0
0056

0
0198

0
6864

Log LDL cholesterol (mg/dL)

0
2086

< 0
0001

0
0844

0
0842

0
2087

< 0
0001

0
0838

0
0867

Log fasting serum glucose(mg/dL)

0
1539

0
0013

0
0267

0
5854

0
1537

0
0014

0
0272

0
5786

Log fasting serum insulin (lIU/mL)

0
4155

< 0
0001

0
0934

0
0558

0
4169

< 0
0001

0
0946

0
0529

Body fat (%)

0
0223

0
708

0
0309

0
6100

0
0423

0
4787

0
0296

0
6254

Waist (cm)

0
3044

< 0
0001

0
2026

0
0013

0
3054

< 0
0001

0
2021

0
0014

Log systolic BP (mmHg)

0
1118

0
0608

0
0755

0
2138

0
1113

0
0625

0
0749

0
2183

Log diastolic BP (mmHg)

0
0611

0
3058

0
0320

0
5975

0
0636

0
2876

0
0320

0
5981

Log total cholesterol (mg/dL)

0
1960

0
0009

0
1210

0
0446

0
1922

0
0011

0
1204

0
0461

Log triglyceride (mg/dL)

0
0596

0
3158

0
0251

0
6782

0
0662

0
2663

0
0267

0
6588

Log HDL cholesterol (mg/dL)

0
0974

0
1008

0
0212

0
7261

0
1089

0
0668

0
0236

0
6963

Log LDL cholesterol (mg/dL)

0
1307

0
0274

0
0404

0
5037

0
1231

0
0382

0
0390

0
5200

Log fasting serum glucose (mg/dL)

0
0817

0
1690

0
0615

0
3090

0
0808

0
1747

0
0615

0
3095

Log HOMA-IR

0
4367

< 0
0001

0
1466

0
0148

0
4385

< 0
0001

0
1474

0
0144

Log fasting serum insulin (lIU/mL)

0
4612

< 0
0001

0
1703

0
0046

0
4633

< 0
0001

0
17127

0
0044

Body fat (%)

0
1926

0
0194

0
2455

0
0031

0
1944

0
0187

0
2433

0
0035

Waist (cm)

0
3194

0
0003

0
0122

0
8942

0
3174

0
0003

0
0088

0
9243

Log systolic BP (mmHg)

0
4069

< 0
0001

0
1429

0
0909

0
4065

< 0
0001

0
1440

0
0896

Log diastolic BP (mmHg)

0
0354

0
6722

0
0635

0
4543

0
0294

0
7261

0
0592

0
4873

Log total cholesterol (mg/dL)

0
3521

< 0
0001

0
1482

0
0773

0
3517

< 0
0001

0
1487

0
0773

Log triglyceride (mg/dL)

0
1290

0
1193

0
0009

0
9914

0
1259

0
1299

0
0073

0
9315

Log HDL cholesterol (mg/dL)

0
3566

< 0
0001

0
0591

0
4835

0
3564

< 0
0001

0
0581

0
4921

Log LDL cholesterol (mg/dL)

0
2509

0
0362

0
0867

0
3035

0
2657

0
0012

0
0883

0
2963

Log fasting serum glucose (mg/dL)

0
2860

0
0164

0
0523

0
5353

0
3088

0
0001

0
0450

0
5946

Body fat (%) 2

(b) Normal weight

(c) Overweight and obesity

European Journal of Clinical Investigation Vol 44

389

J.-S. CHANG ET AL.

www.ejci-online.com

Table 4 Continued Crude Variables

r

Model B†

Model A* P value

r

P value

r

Model C‡ P value

r

P value

Log HOMA-IR

0
5667

< 0
0001

0
1539

0
0665

0
5410

< 0
0001

0
1638

0
0514

Log fasting serum insulin (lIU/mL)

0
5650

< 0
0001

0
1554

0
0638

0
5385

< 0
0001

0
1673

0
0466

BMI, body mass index; HDL, high-density lipoprotein cholesterol; HOMA-IR, homoeostasis model assessment-estimated insulin resistance; LDL, low-density lipoprotein cholesterol. *Adjusted for age. † Adjusted for gender. ‡ Adjusted for age and gender.

In normal weight children, waist circumference (b = 0
0069, P = 0
02) and fasting plasma insulin (b = 0
2912, P = 0
001) were independently associated with IL-10 concentrations (Table 5; normal weight, Multivariate). In overweight and obese children, IL-10 concentrations were independently associated with % body fat (b = 0
0164, P = 0
0009) (Table 5; overweight and obese, Multivariate). Association between IL-10 and fasting plasma insulin concentrations was weaker in overweight/obese children compared with normal weight (Table 5; b = 0
283, P = 0
011 vs. b = 0
291, P = 0
001, Multivariate).

Discussion To our knowledge, this study is the first to provide evidence regarding associations between IL-10 and metabolic risk profiles in children. We demonstrated that children who are overweight or obese had low circulating IL-10 concentrations compared with normal weight children in the same age group. IL-10 concentrations were inversely correlated with % body fat. An independent inverse relationship between IL-10 and total cholesterol and a small positive association between IL-10 and SBP and fasting plasma insulin were observed. However, relationship between IL-10 and metabolic risk profiles were weaker in overweight and obese children. Our data indicate that physiological IL-10 concentrations are involved in lipid and glucose homoeostasis, and decreased endogenous IL-10 in obese children may alter this relationship contributing to the risk of MetS. Our study also raises several questions: Why circulating IL10 is decreased in overweight and obese children? What is the clinical significance? Our study showed a strong inverse relationship between IL-10 and body fat in overweight/obese children but not normal weight. These data suggest that accumulation of body fat is likely to contribute to the low IL-10 concentrations in overweight/obese children. IL-10 is mainly secreted by macrophages or regulatory T cells and not by adipocytes. However, adipose tissue is known to secret many bioactive substances which may regulate IL-10 homoeostasis such as adiponectin and pro-inflammatory cytokines [10]. An in

390

vitro study showed that the addition of adiponectin induced IL-10 mRNA expression in human macrophages [14]. Chu et al. studied 1248 children in the Taipei Children Heart Study and reported that overweight children had lower adiponectin concentrations compared with normal weight [22]. Adiponectin is regarded as anti-inflammatory, anti-atherogenic and antihyperglycaemic agents. Therefore, decreased adiponectin levels may result in lower circulating IL-10 concentrations in obese children. However, Calcaterra et al. examined the relationship between adiponectin, IL-10 and MetS in 70 obese children and reported no correlation between adiponectin and IL-10 [9,11]. The authors suggested this may be due to the hypo-adiponectin and hyper-IL-10 concentrations in obese children. Animal study showed that lack of adiponectin expression did not lead to the development of insulin resistance nor alter progression of spontaneous colitis in adiponectin knockout, or IL-10 and adiponectin double knockout mice [15]. Taken together, these data imply that beneficial effects of adiponectin are probably mediated through the induction of IL-10. Alternatively, elevated blood glucose or lipid levels may contribute to the low circulating IL-10. In the Leiden 85-Plus Study, low IL-10 production capacity in peripheral blood mononuclear cell (PBMC) was inversely correlated with serum lipid and glucose concentrations [4]. Reinhold and colleagues cultured PBMC with anti-CD3 for 24 or 48 h and reported high glucose concentrations down-regulate IL-10, IL-2 and IL-6 secretions [23]. A recent report showed that oxidized LDL inhibits Toll-like receptor 2- and 4-induced IL-10 secretion in monocyte [24]. Our study did not find association between IL10 and FPG, and LDL in children. This may be explained by the relative low rate of type 2 DM and hypercholesterolaemia in the children studied. For example, only four children were diagnosed with type 2 DM in our study. Taken together, we speculate that low circulating IL-10 in overweight/obese children is likely to be regulated by the adiposity-related factors and not blood glucose and lipid levels. The role of endogenous IL-10 has been clearly demonstrated in human and mouse models of atherosclerosis. IL-10 expression has been identified in human atherosclerotic plaques [25],

ª 2014 Stichting European Society for Clinical Investigation Journal Foundation

INTERLEUKIN 10 IN OBESE CHILDREN

Table 5 Multivariate regression coefficients for serum IL-10 in predicting metabolic risk profiles Crude b

Variables

Multivariate A†

Model A* P

b

P

b

P

(a) Pooled 0
0148

< 0
0001

0
0106

< 0
0001

BMI (kg/m )

0
0164

0
0299

0
0094

0
0514

Waist (cm)

0
0035

0
1254

0
0017

0
2851

Log systolic BP (mmHg)

3
4666

< 0
0001

0
9304

0
0420

Log diastolic BP (mmHg)

0
2932

0
5635

0
4192

0
1882

Log total cholesterol (mg/dL)

2
1655

< 0
0001

0
7362

0
0036

Log HDL cholesterol (mg/dL)

0
8244

0
0061

0
0788

0
6864

Log LDL cholesterol (mg/dL)

1
1034

< 0
0001

0
2807

0
0867

Log triglyceride (mg/dL)

0
1506

0
3467

0
0315

0
7564

Log fasting serum glucose (mg/dL)

2
5051

0
0013

0
2770

0
5786

Log fasting serum insulin (lIU/mL)

0
8262

< 0
0001

0
1303

0
0529

Log HOMA-IR

0
7609

< 0
0001

0
1094

0
0823

Body fat (%) 2

0
0096

< 0
0001

0
9801

0
0274

0
7255

0
0033

0
0069

0
0219

0
3797

0
1242

0
2912

0
0011

0
0165

0
0009

(b) Normal weight Body fat (%)

0
0023

0
708

0
0021

0
6254

Waist (cm)

0
0204

< 0
0001

0
0096

0
0014

Log systolic BP (mmHg)

1
7301

0
0608

0
7483

0
2183

Log diastolic BP (mmHg)

0
5827

0
3058

0
1976

0
5981

Log total cholesterol (mg/dL)

1
5031

0
0009

0
6085

0
0461

Log HDL cholesterol (mg/dL)

0
6132

0
3725

0
1018

0
6963

Log LDL cholesterol (mg/dL)

0
6490

0
0274

0
1277

0
5200

Log triglyceride (mg/dL)

0
1916

0
3158

0
0554

0
6588

Log fasting serum glucose (mg/dL)

1
1812

0
1690

0
5737

0
3095

Log fasting serum insulin (lIU/mL)

0
9572

< 0
0001

0
2588

0
0044

Log HOMA-IR

0
8372

< 0
0001

0
2036

0
0144

Body fat (%)

0
0204

0
0194

0
0146

0
0035

Waist (cm)

0
0194

0
0003

0
0004

0
9243

Log systolic BP (mmHg)

5
5265

< 0
0001

1
2189

0
0896

Log diastolic BP (mmHg)

0
4136

0
6722

0
3999

0
4873

Log total cholesterol (mg/dL)

2
9589

< 0
0001

0
7591

0
0773

Log HDL cholesterol (mg/dL)

2
3687

< 0
0001

0
2369

0
4921

Log LDL cholesterol (mg/dL)

1
4886

0
0011

0
2934

0
2963

Log triglyceride (mg/dL)

0
4453

0
1193

0
0149

0
9315

Log fasting serum glucose (mg/dL)

5
6451

0
0002

0
5087

0
5946

(c) Overweight and obesity

European Journal of Clinical Investigation Vol 44

391

J.-S. CHANG ET AL.

www.ejci-online.com

Table 5 Continued Crude Variables

b

Multivariate A†

Model A* P

b

P

b

Log fasting serum insulin (lIU/mL)

1
0469

< 0
0001

0
2271

0
0466

Log HOMA-IR

1
0006

< 0
0001

0
2109

0
0514

P 0
2832

0
0114

BMI, body mass index; HDL, high-density lipoprotein cholesterol; HOMA-IR, homoeostasis model assessment-estimated insulin resistance; LDL, low-density lipoprotein cholesterol. *Model A: adjusted age and gender. † Multivariate A: adjusted for the factors according to Model A by P < 0
05.

and circulating IL-10 levels are associated with a more favourable prognosis in patients with acute coronary syndromes [5]. Narverud et al., [26] studied children with familial hypercholesterolaemia (FH) and reported decreased circulating IL-10 and increased TNF-a concentrations in FH children compared with healthy children. IL-10 modulates lipid metabolism by enhancing both uptake and efflux of cholesterol in macrophages [27], and IL-10 over-expression macrophage inhibits atherosclerosis in LDLR / mice [28]. These data indicate that endogenous IL-10 is important in cholesterol metabolism and defect in circulating IL-10 may contribute to the development of atherosclerosis in overweight or obese subjects. A small positive relationship between IL-10 and fasting plasma insulin was found in children. Straczkowski et al. [6] showed plasma IL-10 concentrations were negatively correlated with fasting plasma insulin in young adults (r = 0
31, P = 0
0026). Our study observed that both fasting plasma insulin and IL-10 concentrations were closely associated with normal physical development of school age children. However, this relationship differs between overweight/obese and normal weight, indicating that decreased endogenous IL-10 in obese children may interfere with the action of insulin. Indeed, Bhargava et al. [29] showed administration of lacto-N-fucopentaose III (LNFPIII) improved insulin sensitivity in dietinduced obese mice and this effect was mediated partly through increased IL-10 production by LNFPIII-activated macrophages. There are several limitations in our study, which need to be taken into account when interpretation of the results. Firstly, the prevalence of MetS is largely depending on the diagnostic methods used, which may vary between 2% and 5% in the same study population [30]. Secondly, the cross-sectional nature and the effect of pubertal growth may dilute the relationship between IL-10 concentrations and the lipid and glucose profiles. This study did not measure the pubertal stage, and therefore, we cannot discriminate the potential influence of sexual hormones on IL-10 concentrations. Decreased insulin sensitivity has been shown in adolescents [31].

392

In conclusion, our study in children was in agreement with previous studies in adults linking low circulating IL-10 with obesity. Accumulation of body fat may contribute to the decreased IL-10 and the consequent relationship between IL-10 and metabolic risk profiles. Overall, our data indicate that changes in circulating IL-10 are associated with metabolic risk in children.

Acknowledgements We express our sincere appreciation to the study participants. We also wish to thank staff from the Taipei Medical Hospital and University.

Sources of funding Dr. Jung-Su Chang was supported by grant 101TMU-TMUH-04 and NSC102-2320-B038-013.

Conflict of interest The authors have declared that no competing interest exists.

Authors’ contributions Jung-Su Chang: Dr. Chang conceptualized and designed the study, drafted the initial manuscript, and approved the final manuscript as submitted. Chyi-Huey Bai: Dr.Bai contributed to the initiation of the project and supervision of data analysis. ZuChieh Huang: Ms.Huang carried out the initial data analyses. Eddy Owaga: Mr. Owaga was responsible for sample collection at four sites and conducted cytokine analysis. Kuo-Ching Chao: Dr. Chao responsible for the blood biochemistry examination. Chun-Chao Chang: Dr. Chang contributed to the optimization of protocols and preparation and examination of IL-10 measurements. Hung-Yi Chiou: Dr. Chiou critically reviewed the manuscript and approved the final manuscript as submitted.

Address School of Nutrition and Health Sciences, College of Public Health and Nutrition, Taipei Medical University, 250 Wu-Hsing Street, Taipei City, 110 Taiwan (J.-S. Chang, Z.-C. Huang, E. Owaga); Department of Public Health, College of

ª 2014 Stichting European Society for Clinical Investigation Journal Foundation

INTERLEUKIN 10 IN OBESE CHILDREN

Medicine, Taipei Medical University, 250 Wu-Hsing Street, Taipei City, 110 Taiwan (C.-H. Bai); Division of Gastroenterology and Hepatology, Department of Internal Medicine, Taipei Medical University Hospital, No. 252, Wu-Hsing Street, Taipei City, 110 Taiwan (K.-C. Chao, C.-C. Chang); Department of Internal Medicine, School of Medicine, College of Medicine, Taipei Medical University, No. 252, Wu-Hsing Street, Taipei City, 110 Taiwan (K.-C. Chao, C.-C. Chang); School of Public Health, Taipei Medical University, No. 252, Wu-Hsing Street, Taipei City, 110 Taiwan (H.-Y. Chiou). Correspondence to: Dr Hung-Yi Chiou, School of Public Health, College of Public Health and Nutrition, Taipei Medical University, 250 Wu-Hsing Street, Taipei City, 110 Taiwan. Tel.: +886 (2)27361661#6512; fax: +886 (2)2738 4831; e-mail: [email protected] Received 11 July 2013; accepted 23 January 2014 References 1 Saraiva M, O’Garra A. The regulation of IL-10 production by immune cells. Nat Rev Immunol 2010;10:170–81. 2 Mocellin S, Panelli MC, Wang E, Nagorsen D, Marincola FM. The dual role of IL-10. Trends Immunol 2003;24:36–43. 3 Kuhn R, Lohler J, Rennick D, Rajewsky K, Muller W. Interleukin10-deficient mice develop chronic enterocolitis. Cell 1993;75:263–74. 4 van Exel E, Gussekloo J, de Craen AJ, Frolich M, Bootsma-Van Der Wiel A, Westendorp RG. Low production capacity of interleukin-10 associates with the metabolic syndrome and type 2 diabetes: the Leiden 85-Plus Study. Diabetes 2002;51:1088–92. 5 Heeschen C, Dimmeler S, Hamm CW, Fichtlscherer S, Boersma E, Simoons ML et al. Serum level of the antiinflammatory cytokine interleukin-10 is an important prognostic determinant in patients with acute coronary syndromes. Circulation 2003;107:2109–14. 6 Straczkowski M, Kowalska I, Nikolajuk A, Krukowska A, Gorska M. Plasma interleukin-10 concentration is positively related to insulin sensitivity in young healthy individuals. Diabetes Care 2005;28: 2036–7. 7 Smith DA, Irving SD, Sheldon J, Cole D, Kaski JC. Serum levels of the antiinflammatory cytokine interleukin-10 are decreased in patients with unstable angina. Circulation 2001;104:746–9. 8 Gozal D, Serpero LD, Sans Capdevila O, Kheirandish-Gozal L. Systemic inflammation in non-obese children with obstructive sleep apnea. Sleep Med 2008;9:254–9. 9 Calcaterra V, De Amici M, Klersy C, Torre C, Brizzi V, Scaglia F et al. Adiponectin, IL-10 and metabolic syndrome in obese children and adolescents. Acta Biomed 2009;80:117–23. 10 Arslan N, Erdur B, Aydin A. Hormones and cytokines in childhood obesity. Indian Pediatr 2010;47:829–39. 11 Tam CS, Garnett SP, Cowell CT, Heilbronn LK, Lee JW, Wong M et al. IL-6, IL-8 and IL-10 levels in healthy weight and overweight children. Horm Res Paediatr 2010;73:128–34. 12 Arrigo T, Chirico V, Salpietro V, Munafo C, Ferrau V, Gitto E et al. High-mobility group protein B1: a new biomarker of metabolic syndrome in obese children. Eur J Endocrinol 2013;168:631–8.

13 Beauloye V, Zech F, Tran HT, Clapuyt P, Maes M, Brichard SM. Determinants of early atherosclerosis in obese children and adolescents. J Clin Endocrinol Metab 2007;92:3025–32. 14 Kumada M, Kihara S, Ouchi N, Kobayashi H, Okamoto Y, Ohashi K et al. Adiponectin specifically increased tissue inhibitor of metalloproteinase-1 through interleukin-10 expression in human macrophages. Circulation 2004;109:2046–9. 15 Pini M, Gove ME, Fayad R, Cabay RJ, Fantuzzi G. Adiponectin deficiency does not affect development and progression of spontaneous colitis in IL-10 knockout mice. Am J Physiol Gastrointest Liver Physiol 2009;296:G382–7. 16 Guidelines for reporting health research: How to promote their use in your journal. Written by the EQUATOR Network Group; December 2011. 17 Choy CS, Chan WY, Chen TL, Shih CC, Wu LC, Liao CC. Waist circumference and risk of elevated blood pressure in children: a cross-sectional study. BMC Public Health 2011;11:613. 18 Pan WH, Lee MS, Chuang SY, Lin YC, Fu ML. Obesity pandemic, correlated factors and guidelines to define, screen and manage obesity in Taiwan. Obes Rev 2008;9(Suppl 1):22–31. 19 Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF, Turner RC. Homeostasis model assessment: insulin resistance and beta-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia 1985;28:412–9. 20 Exceutive Summary of The 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). JAMA 2001;285:2486–97. 21 Tan CE, Ma S, Wai D, Chew SK, Tai ES. Can we apply the National Cholesterol Education Program Adult Treatment Panel definition of the metabolic syndrome to Asians? Diabetes Care 2004;27:1182–6. 22 Chu NF, Shen MH, Wu DM, Lai CJ. Relationship between plasma adiponectin levels and metabolic risk profiles in Taiwanese children. Obes Res 2005;13:2014–20. 23 Reinhold D, Ansorge S, Schleicher ED. Elevated glucose levels stimulate transforming growth factor-beta 1 (TGF-beta 1), suppress interleukin IL-2, IL-6 and IL-10 production and DNA synthesis in peripheral blood mononuclear cells. Horm Metab Res 1996;28:267–70. 24 Bzowska M, Nogiec A, Skrzeczynska-Moncznik J, Mickowska B, Guzik K, Pryjma J. Oxidized LDLs inhibit TLR-induced IL-10 production by monocytes: a new aspect of pathogen-accelerated atherosclerosis. Inflammation 2012;35:1567–84. 25 Uyemura K, Demer LL, Castle SC, Jullien D, Berliner JA, Gately MK et al. Cross-regulatory roles of interleukin (IL)-12 and IL-10 in atherosclerosis. J Clin Invest 1996;97:2130–8. 26 Narverud I, Ueland T, Nenseter MS, Retterstol K, Telle-Hansen VH, Halvorsen B et al. Children with familial hypercholesterolemia are characterized by an inflammatory imbalance between the tumor necrosis factor alpha system and interleukin-10. Atherosclerosis 2011;214:163–8. 27 Han X, Kitamoto S, Lian Q, Boisvert WA. Interleukin-10 facilitates both cholesterol uptake and efflux in macrophages. J Biol Chem 2009;284:32950–8. 28 Han X, Kitamoto S, Wang H, Boisvert WA. Interleukin-10 overexpression in macrophages suppresses atherosclerosis in hyperlipidemic mice. FASEB J 2010;24:2869–80.

European Journal of Clinical Investigation Vol 44

393

J.-S. CHANG ET AL.

29 Bhargava P, Li C, Stanya KJ, Jacobi D, Dai L, Liu S et al. Immunomodulatory glycan LNFPIII alleviates hepatosteatosis and insulin resistance through direct and indirect control of metabolic pathways. Nat Med 2012;18:1665–72. 30 Lin SY, Su CT, Hsieh YC, Li YL, Chen YR, Cheng SY et al. Risk Factors Correlated With Risk of Insulin Resistance Using

394

www.ejci-online.com

Homeostasis Model Assessment in Adolescents in Taiwan. Asia Pac J Public Health 2013;doi: 10.1177/1010539512471075 31 Bottner A, Kratzsch J, Muller G, Kapellen TM, Bluher S, Keller E et al. Gender differences of adiponectin levels develop during the progression of puberty and are related to serum androgen levels. J Clin Endocrinol Metab 2004;89:4053–61.

ª 2014 Stichting European Society for Clinical Investigation Journal Foundation