- Email: [email protected]
Association of serum amyloid A levels with adipocyte size and serum levels of adipokines: Differences between men and women
Cytokine 48 (2009) 260–266
Contents lists available at ScienceDirect
Cytokine journal homepage: www.elsevier.com/locate/issn/10434666
Association of serum amyloid A levels with adipocyte size and serum levels of adipokines: Differences between men and women Kajsa Sjöholm a,*, Magdalena Lundgren b, Maja Olsson a, Jan W. Eriksson b,c a
Sahlgrenska Center for Cardiovascular and Metabolic Research, Department of Molecular and Clinical Medicine, Institute of Medicine, The Sahlgrenska Academy, Göteborg University, Göteborg, Sweden b Department of Medicine, UKBF, Umeå University Hospital, Umeå, Sweden c The Lundberg Laboratory for Diabetes Research, Sahlgrenska Center for Cardiovascular and Metabolic Research, Department of Molecular and Clinical Medicine, Institute of Medicine, The Sahlgrenska Academy, Göteborg University, Göteborg, Sweden
a r t i c l e
i n f o
Article history: Received 21 November 2008 Received in revised form 17 June 2009 Accepted 18 August 2009
Keywords: Adipocyte size Serum amyloid A C-reactive protein Adipokines
a b s t r a c t The aim of this study was to characterize the association between adipocyte enlargement and circulating levels of serum amyloid A (SAA). Furthermore, we wanted to search for possible associations with measures of glycemic control and levels of circulating adipokines and/or inflammatory markers in men and women with a large range in body mass index. The study cohort consisted of 167 subjects, 114 non-diabetic and 53 with Type 2 diabetes. Adipocyte diameter as well as circulating levels of SAA, C-reactive protein (CRP), adiponectin, leptin, interleukin-6, tumor necrosis factor alpha, glucose and insulin were measured. Women had higher serum levels of SAA than men (p = 0.044). SAA levels were weakly but positively correlated with BMI (p = 0.043) and % body fat (p = 0.027) in all subjects as well as subcutaneous adipocyte diameter (p = 0.034) in women. Furthermore, in all subjects we found correlations between SAA levels and levels of CRP (p < 0.001), interleukin-6 (p < 0.001), leptin (p = 0.003), insulin (p = 0.006), HbA1c (p = 0.02) and HOMA-IR (p = 0.002). A majority of the correlations were strongest in women. In conclusion, serum levels of SAA are strongly correlated with serum levels of inflammatory markers as well as measures of glycemic control. There seems to be large sex differences in these associations suggesting that sex-specific factors need to be considered when analyzing SAA levels in relation to metabolic disease. Ó 2009 Elsevier Ltd. All rights reserved.
1. Introduction The prevalence of obesity is increasing rapidly and to the extent that the World Health Organization has denoted it a global epidemic. Obesity is associated with modest but chronically elevated levels of circulating inflammatory markers, such as proinflammatory cytokines and acute phase proteins, and has been designated a low-grade inflammatory state [1,2]. Increasing evidence suggest that factors released from the growing adipose tissue depot may contribute to the development of the metabolic complications related to obesity, such as insulin resistance, Type 2 diabetes and atherosclerosis. Fat cells (adipocytes) produce a number of factors such as cytokines and other bioactive molecules, collectively known as adipokines [3].
* Corresponding author. Address: Sahlgrenska Center for Cardiovascular and Metabolic Research, Department of Molecular and Clinical Medicine, Institute of Medicine, The Sahlgrenska Academy, Göteborg University, SOS-sekr., Vita Stråket 15, SE-413 45 Göteborg, Sweden. Tel.: +46 31 3423029; fax +46 31 418527. E-mail address: [email protected] (K. Sjöholm). 1043-4666/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.cyto.2009.08.005
Central or visceral obesity are strongly related to metabolic disease including atherosclerosis, insulin resistance and hypertension [4]. However, not only the size and location of the adipose tissue is important. Enlargement of adipocytes is also a risk factor for metabolic disease. Adipocytes can increase several thousandfold in volume and about 20-fold in diameter and this increase in size is an independent predictor of Type 2 diabetes [5]. Enlargement of adipocytes is also associated with lowered insulin sensitivity [6–8], increased expression and secretion of adipokines [8–12] as well as changes in lipid and glucose metabolism [12– 15]. It is possible that some subjects display impaired ability to recruit new adipocytes, causing a reduction of adipose tissue lipid storage capacity and detrimental enlargement of existing adipocytes [16]. The nutritional overload may cause altered expression of adipokines and/or lipid spill-over leading to ectopic fat accumulation in other tissues such as liver and skeletal muscle, thereby causing a further increase in insulin resistance [17,18]. We, and others, have previously shown that adipocytes, and in particular enlarged adipocytes, display high expression of the acute-phase protein serum amyloid A (SAA) [19–21]. SAA binds to the cell surface receptor selenoprotein S [22], and it was re-
261
K. Sjöholm et al. / Cytokine 48 (2009) 260–266
cently reported that SAA is also a functional ligand of the toll-like receptor 2 [23]. Hence, SAA is a possible link between Type 2 diabetes, atherosclerosis and inflammation. Furthermore, serum levels of SAA can be used to predict cardiovascular risk [24]. Several of the previous reports on the association between adipocyte volume and expression/circulating levels of SAA, have not considered sex-specific factors and reported associations with circulating adipokines are conflicting [20,25,26]. Therefore, in this study, the aim was to characterize the association between adipocyte enlargement and circulating SAA levels in men and women with a large range in BMI and to search for possible associations with measures of glycemic control and levels of circulating adipokines and/or inflammatory markers.
2. Subjects and methods
2.2. Assessment of cell diameter and glucose uptake in isolated adipocytes Isolation of adipocytes and assessment of cell diameter was performed as described previously [8]. In brief, adipose tissue was minced and then subjected to collagenase digestion. Obtained adipocytes were filtered through a nylon mesh, washed and cell diameter was measured using a light microscope with an integrated ruler. The cell diameter is reported as the average cell diameter (lm) of 100 cells from the subcutaneous and, in a subset (n = 43), also from the omental depot. The rate of glucose uptake was assessed in isolated adipocytes as previously reported and it was performed in a subset (n = 99) of the subjects included in the study [8]. The maximal insulin-stimulated (1000 mU/l) glucose uptake rate (MGU) was expressed as percent of the basal, non-stimulated glucose uptake to provide a measure of insulin action in vitro.
2.1. Subjects and samples The study cohort consisted of 167 subjects (87 women and 80 men), 114 non-diabetic subjects and 53 Type 2 diabetic subjects diagnosed according to the 1998 WHO criteria [27]. Anthropometric and biochemical characteristics of study participants are shown in Table 1. Data were obtained from previous study cohorts at Umeå University Hospital [8] and subjects were included if serum stored at 80 °C was available for SAA analysis. Subcutaneous adipose tissue was obtained through a needle biopsy from the lower part of the abdomen after dermal local anesthesia with lidocain (Xylocain; AstraZeneca, Södertälje, Sweden) and fasting blood samples were collected. Omental adipose tissue biopsies were obtained after induction of general anesthesia in a subset (n = 43) of subjects undergoing elective abdominal surgery. Body fat (% of body weight) was determined by the bioimpedance technique (Bodystat 1500; Bodystat, Isle of Man, Great Britain) [28]. All subjects gave informed consent and the Umeå University Ethics Committee approved the studies.
2.3. Blood chemistry and insulin sensitivity An enzyme-linked immunosorbent assay was used to analyze serum levels of serum amyloid A (Biosource International, Camarillo, CA, USA). The inter-assay coefficient of variation for the SAA analysis was 14% at 30 lg/ml. Other measurements were performed as previously described [8] using commercial enzymatic kits for adiponectin and leptin (Linco Research, St. Charles, MO, USA), interleukin-6 (IL-6) and tumor necrosis factor alpha (TNFa) (R&D Systems, Minneapolis, MN, USA). Serum insulin concentrations were measured by microparticle enzyme immunoassays (Immulite 2000, DPC, Los Angeles, CA, USA; Abbot Imx, Abbott Laboratories, Abbott Park, IL, USA). All other measurements were made according to routine methods at the Department of Clinical Chemistry, Umeå University Hospital. In a subset of the subjects included in the study (n = 58) insulin sensitivity in vivo was assessed as glucose infusion rate, i.e., the
Table 1 Characteristics of the subjects included in the study.c
N Age BMI (kg/m2) WHR Fat (%) P-Glucose (mmol/l) S-Insulin (mU/l) HbA1c (%) HOMA S-SAA (lg/ml) S-Hs CRP (mg/l) P-Leptin (ng/ml) P-IL-6 (pg/ml) P-TNF-a (pg/ml) P-Adiponectin (lg/ml) Sc adipocyte diameter (lm) Om adipocyte diameter (lm) Sc MGU (%)a M-value (mg kg 1 min 1)b
All T2D
All Non-diabetic
Women T2D
Women Non-diabetic
Men T2D
Men Non-diabetic
53 59.2 ± 8.4 28.4 ± 4.7 0.96 ± 0.08 31.8 ± 11.4 9.9 ± 3.1 11.3 ± 7.5 7.4 ± 1.8 4.9 ± 3.7 16.4 ± 10.3 2.6 ± 2.3 12.8 ± 10.0 2.5 ± 1.8 1.8 ± 1.4 7.2 ± 3.4 105.1 ± 9.4 NA 190.3 ± 57.1 5.1 ± 2.2
114 51.3 ± 15.3** 26.8 ± 4.6* 0.89 ± 0.08*** 31.4 ± 10.0 5.1 ± 0.6*** 9.1 ± 6.0* 4.3 ± 0.5*** 2.1 ± 1.6*** 15.1 ± 12.0 1.6 ± 1.6** 14.8 ± 13.8 1.8 ± 1.2** 1.6 ± 2.2* 9.9 ± 5.0*** 100.1 ± 12.3* 94.3 ± 14.2 260.1 ± 111.2*** 10.3 ± 4.0***
22 60.2 ± 9.8 28.6 ± 6.6 0.91 ± 0.06 39.7 ± 13.0 8.6 ± 2.3 12.6 ± 8.7 6.9 ± 1.5 4.8 ± 3.6 19.2 ± 12.0 2.7 ± 2.4 20.2 ± 11.0 2.8 ± 2.0 2.1 ± 1.9 7.3 ± 3.8 109.8 ± 8.4 NA 201.2 ± 72.8 5.0 ± 2.2
65 51.4 ± 16.0* 26.5 ± 5.3 0.84 ± 0.06*** 36.5 ± 9.3 5.0 ± 0.6*** 9.3 ± 7.2* 4.2 ± 0.5*** 2.2 ± 1.9*** 16.7 ± 12.9 1.8 ± 1.9* 21.4 ± 15.2 1.7 ± 1.3** 1.7 ± 2.3 11.6 ± 5.3*** 101.0 ± 13.9* 91.5 ± 15.0 263.3 ± 117.8* 10.8 ± 4.0***
31 58.5 ± 7.2 28.2 ± 2.7 1.00 ± 0.06 26.2 ± 5.6 10.9 ± 3.3 10.3 ± 6.5 7.7 ± 1.9 5.0 ± 3.9 14.4 ± 8.6 2.5 ± 2.3 7.5 ± 4.5 2.3 ± 1.8 1.7 ± 0.9 7.2 ± 3.1 101.7 ± 8.6 NA 183.4 ± 44.7 5.1 ± 2.2
49 51.2 ± 14.3* 27.2 ± 3.6 0.95 ± 0.06** 24.7 ± 6.1 5.2 ± 0.7*** 8.7 ± 4.0 4.4 ± 0.5*** 2.0 ± 1.1*** 13.0 ± 10.5 1.4 ± 1.0 6.4 ± 3.5 1.8 ± 1.1 1.6 ± 2.1* 7.8 ± 3.5 99.0 ± 10.1 97.5 ± 12.9 256.5 ± 105.5** 9.7 ± 4.0***
NA, non-applicable. Significance between diabetic and non-diabetic subjects are given below. * p 0.05. ** p 0.01. *** p 0.001. a n = 99. b n = 58. c Data are means ± SD.
262
K. Sjöholm et al. / Cytokine 48 (2009) 260–266
M-value, at steady state (60–120 min) of a 2-h euglycemic–hyperinsulinemic (56 mU m 2 body surface area min 1) clamp. This procedure was performed as previously described in detail [29,30]. 2.4. Statistical analyses Statistical analyses were performed using the SPSS package (version 16.0; SPSS, Chicago, IL, USA). Differences in measured variables between diabetics and non-diabetics, women and men, and between lean and overweight/obese were analyzed using the Mann–Whitney U test. The relationship between serum levels of SAA and other variables was analyzed using the Spearman rank correlation test. Partial correlation was used to control for age, sex and body mass index (BMI) when appropriate. Skewed variables were log transformed prior to partial correlation analysis. 3. Results 3.1. Anthropometric and biochemical characteristics Characteristics of the subjects included in this study are shown in Table 1. Women had higher circulating levels of SAA (p = 0.044; Fig. 1A), adiponectin (p < 0.001) and leptin (p < 0.001) and lower waist-to-hip ratio (WHR) (p < 0.001), HbA1c (p = 0.005) and levels of glucose (p = 0.007) than men. In all subjects, obese (BMI 30 kg/m2) subjects had significantly higher levels of SAA than nonobese (p = 0.006; Fig. 1B). Type 2 diabetic subjects were older (p = 0.001), had larger subcutaneous adipocytes (p = 0.017), higher BMI (p = 0.037), WHR (p < 0.001), glucose (p < 0.001), insulin (p = 0.014), HbA1c (p < 0.001), homeostasis model assessment of insulin resistance (HOMA-IR) (p < 0.001), C-reactive protein (CRP) (p = 0.006), IL-6 (p = 0.001), TNF-a (p = 0.011) and lower adiponectin (p < 0.001), subcutaneous MGU (p < 0.001) and M-value (p < 0.001) (Table 1) than non-diabetic subjects. There were no significant differences in SAA levels or % fat mass between non-diabetic and Type 2 diabetic subjects (Table 1). 3.2. SAA serum levels in relation to BMI and adipocyte size In all subjects, serum SAA levels were positively correlated with BMI and % body fat (Table 2), but not with subcutaneous (Table 2) or omental adipocyte diameter (data not shown). Interestingly, in women, but not in men, SAA levels were positively correlated with BMI (r = 0.27; p = 0.012), % body fat (r = 0.27; p = 0.018) and subcutaneous (r = 0.27; p = 0.034) (Table 2), but not with omental adipocyte diameter (data not shown). However, the correlation between
SAA and subcutaneous adipocyte diameter was lost when controlling for age and BMI. When dividing female subjects by BMI range (lean, n = 23, 24.99 kg/m2, overweight, n = 20, 25.00–29.99 kg/m2 and obese, n = 21, BMI 30.00 kg/m2), SAA levels in lean subjects were positively correlated with subcutaneous adipocyte diameter (r = 0.425; p = 0.043; Fig. 2A), borderline significance was found in overweight subjects (r = 0.42; p = 0.066; Fig. 2B) and no correlation was seen in obese subjects (r = 0.066; p = 0.775; Fig. 2C). After controlling for age and BMI no correlations were found between serum levels of SAA and adipocyte diameter. In men, no correlation was seen (Fig. 2D) even after dividing subjects by BMI range (r = 0.131; p = 0.297). When dividing all subjects by Type 2 diabetes status, we found that in non-diabetic subjects, serum levels of SAA were positively correlated with % body fat (r = 0.21; p = 0.034), but not with BMI (r = 0.12; p = 0.43). No correlation was found in diabetic subjects. When splitting also by sex, we found that in non-diabetic women, borderline significance was found between SAA levels and BMI and % body fat (Table 2). In non-diabetic men, diabetic men and diabetic women no correlations were found between SAA and BMI or % body fat. 3.3. SAA serum levels in relation to glycemic control and inflammatory markers When analyzing all subjects (Table 2), SAA levels positively correlated with HbA1c, HOMA-IR and insulin. SAA levels were also positively correlated with serum levels of CRP, leptin and IL-6 (Table 2), but not with adiponectin, TNF-a or glucose (data not shown). Furthermore, in all subjects, significant associations remained significant or marginally significant after controlling for sex, age and BMI except for leptin (Table 2). No correlations were found between SAA levels and insulin-stimulated glucose uptake in isolated subcutaneous adipocytes (n = 99) or whole-body insulin sensitivity (n = 58) assessed by a euglycemic–hyperinsulinemic clamp in the subsets where data were available (data not shown). When analyses were split by sex, (Table 2), SAA serum levels positively correlated with CRP (Fig. 3A and B), HbA1c and HOMAIR in both men and women (Fig. 3C and D). In women there were also positive correlations between circulating levels of SAA and insulin, leptin and IL-6 (Table 2). In men there was a borderline significant correlation between SAA and IL-6 (Table 2) but no correlations with insulin and leptin. When dividing subjects into groups of diabetics and non-diabetics, we found that, in both groups, SAA levels correlated with CRP (r = 0.45; p = 0.002 and r = 0.33; p = 0.001, respectively).
Fig. 1. Serum levels of SAA in men (n = 80) and women (n = 87) (A) and obese (n = 43) and non-obese (n = 124) subjects (B). Women display higher levels than men (p = 0.044) and obese subjects display higher levels than non-obese (p = 0.006). Values are expressed as means ± SEM.
263
K. Sjöholm et al. / Cytokine 48 (2009) 260–266 Table 2 Correlation analysis of SAA serum levels with adiposity, glycemic control variables, adipokines, CRP and adipocyte diameter. Variablea
All subjects
Women
Men
Women T2D
Non-diabetic
T2D
Non-diabetic
N BMI (kg/m2)
167 r = 0.16; p = 0.043 r = 0.18; p = 0.027 r = 0.21; p = 0.006b r = 0.19; p = 0.020b r = 0.24; p = 0.002b r = 0.39; p < 0.001b r = 0.24; p = 0.003 r = 0.28; p < 0.001b r = 0.11; p = 0.210
87 r = 0.27; p = 0.012 r = 0.27; p = 0.018 r = 0.25; p = 0.021 r = 0.24; p = 0.035 r = 0.27; p = 0.013 r = 0.42; p < 0.001 r = 0.30; p = 0.007 r = 0.37; p = 0.001 r = 0.27; p = 0.034
80 r = 0.03; p = 0.776 r = 0.05; p = 0.694 r = 0.18; p = 0.116 r = 0.27; p = 0.019 r = 0.26; p = 0.021 r = 0.38; p = 0.001 r = 0.12; p = 0.301 r = 0.21; p = 0.075 r = 0.13; p = 0.297
22 r = 0.30; p = 0.170 r = 0.34; p = 0.156 r = 0.26; p = 0.247 r = 0.03; p = 0.895 r = 0.30; p = 0.175 r = 0.53; p = 0.019 r = 0.52; p = 0.013 r = 0.64; p = 0.002 r = 0.30; p = 0.198
65 r = 0.24; p = 0.059 r = 0.25; p = 0.056 r = 0.21; p = 0.090 r = 0.24; p = 0.068 r = 0.23; p = 0.068 r = 0.40; p = 0.004 r = 0.23; p = 0.093 r = 0.23; p = 0.084 r = 0.25; p = 0.096
31 r = 0.05; p = 0.811 r = 0.13; p = 0.534 r = 0.19; p = 0.318 r = 0.34; p = 0.061 r = 0.11; p = 0.548 r = 0.45; p = 0.019 r = 0.16; p = 0.381 r = 0.28; p = 0.123 r = 0.11; p = 0.585
49 r = 0.06; p = 0.700 r = 0.05; p = 0.768 r = 0.33; p = 0.019 r = 0.17; p = 0.284 r = 0.27; p = 0.056 r = 0.28; p = 0.081 r = 0.23; p = 0.129 r = 0.16; p = 0.301 r = 0.13; p = 0.457
Fat (%) S-Insulin (mU/l) HbA1c (%) HOMA S-Hs CRP (mg/l) P-Leptin (ng/ml) P-IL-6 (pg/ml) Sc adipocyte diameter (lm)
Men
a Data are means ± SD. Correlations with TNF-a, adiponectin, glucose, omental adipocyte diameter, MGU and M-value were omitted as correlations were not significant. Spearman correlation coefficient Rho, r; significant correlations in bold. b Correlations adjusted for age, sex and BMI significant for HOMA (p = 0.012), S-Hs CRP (p < 0.001), HbA1c (p = 0.025) and P-IL-6 (p = 0.003) and marginally significant for SInsulin (p = 0.065). Skewed variables were log transformed prior to partial correlation analysis.
For IL-6 a significant correlation was found in the diabetic group (r = 0.44; p = 0.001) but only borderline in the non-diabetics (r = 0.18; p = 0.078). In the non-diabetics, SAA levels also correlated with insulin (r = 0.24; p = 0.01), HOMA-IR (r = 0.22; p = 0.02) and leptin (r = 0.28; p = 0.005) but no association was found in the diabetic group. When splitting also by sex, we found that in non-diabetic men (n = 49), SAA levels were positively correlated with
insulin but in non-diabetic women (n = 65) only a borderline significant correlation was found (Table 2). Furthermore, in non-diabetic women there were borderline significant positive correlations between SAA levels and HbA1c, HOMA-IR, leptin and IL-6. In non-diabetic men, there was a borderline significant positive correlation between SAA levels and HOMA-IR. In diabetic women significant correlations were found between SAA and leptin
Fig. 2. Scatter plots of serum levels of SAA and adipocyte diameter in lean (n = 23) (A), overweight (n = 20) (B) and obese (n = 21) (C) women and all men included in the study (n = 65) (D) where cell diameter was available. Rho; Spearman’s correlation coefficient. SAA was log transformed for presentation due to a skewed distribution. A regression line was included for illustration.
264
K. Sjöholm et al. / Cytokine 48 (2009) 260–266
Fig. 3. Scatter plots of serum levels of SAA and CRP in women (n = 71) (A) and men (n = 67) (B) and HOMA-IR index in women (n = 86) (C) and men (n = 79) (D). Rho; Spearman’s correlation coefficient. All variables were log transformed for presentation due to skewed distributions. A regression line was included for illustration.
and IL-6 (Table 2). In diabetic men, a significant correlation was found with CRP and borderline significant correlation was found with Hba1c.
4. Discussion In this study we wanted to examine the possible association between serum levels of SAA and adipocyte size as well as with circulating adipokines and CRP in lean and obese, female and male subjects. SAA is an acute phase protein, which was previously believed only to be expressed in, and released from, the liver in response to infection or inflammation. However, we [19] and others [21], have recently discovered that adipocytes are a major site of expression during non-acute phase and that SAA levels in serum are highly correlated with SAA expression in adipose tissue [19]. Furthermore, enlarged adipocytes have increased expression [20] suggesting that, in an obese individual, the contribution of adipose SAA to serum levels may be substantial, especially in hypertrophic obesity, i.e., individuals with enlarged adipocytes. In this study, we report a correlation between adipocyte size and serum levels of SAA in non-obese women and show that there are sex differences, not only in adipose tissue expression of SAA [19] which we have reported previously, but also in associations with adipocyte size and serum levels of other adipokines. In line with previous reports [25,26,31,32], we found that SAA serum levels are increased in the obese, that SAA serum levels are positively correlated with BMI and that women have higher levels of SAA than men. Furthermore, we found that in women, serum levels of SAA were correlated with adipocyte size. We have also extended this finding by showing that the highest correlation is found in non-obese women. We have previously reported that women have higher mRNA levels of SAA in subcutaneous adipose tissue than men [19]. However, in that study, this difference was not reflected in SAA serum levels, which probably can be explained by the limited sample size (n = 6 women), and that all subjects included in that study were
morbidly obese. The findings in the current study support the fact that obesity-induced enlargement of adipocytes may cause an increase in circulating levels of SAA. However, we find no evidence that adipocyte size can be used to predict serum levels of SAA. Instead, a general increase in adiposity may cause an increase in levels of circulating SAA that, in turn, may induce other inflammatory mediators [25]. We have previously shown that subcutaneous adipose tissue displays higher levels of SAA expression than omental adipose tissue and that hypertrophic adipocytes have increased levels of SAA both when analyzing gene expression and protein levels [19,20]. Furthermore, we, and others have previously shown an association between adipocyte volume and SAA mRNA levels [20,25,26]. In the study by Yang et al. (n = 19), it is not clear if the subjects were obese or not, nor whether they were male or female. In the study by Lappalainen et al. (n = 37 men/38 women), SAA mRNA levels correlated with adipocyte size but no correlation was found between adipocyte size and serum levels of SAA [32]. The only study where an association between circulating levels of SAA and adipocyte size has previously been found is a study by Poiteu et al. where morbidly obese subjects (n = 10 men/50 women) where studied [26]. In the present study, 87 women and 80 men with a large range in BMI (16– 49 kg/m2) were included. We found no association between adipocyte diameter and SAA serum levels in all subjects, men or obese women. In contrast, we found a weak but significant association between adipocyte diameter and SAA levels in lean women and borderline significance in overweight women. Considering the higher expression of SAA in subcutaneous adipose tissue and a gynoid adipose tissue distribution with larger subcutaneous depots in women, the finding of increased levels of serum SAA is not surprising. However, further studies are needed to elucidate whether there actually is a lack of association in the higher range of BMI. The reason for the higher SAA expression in women is unknown but one possible explanation is that it could be related to hormonal status as estrogen replacement in postmenopausal women increases inflammatory factors like IL-6, CRP and SAA [33]. Another possible explanation could be direct
K. Sjöholm et al. / Cytokine 48 (2009) 260–266
or indirect interaction with leptin [32]. We found no association between omental adipocyte diameter and serum SAA in this study. This could be due to the limited number of samples available for that analysis. In T2D, hyperglycemia could lead to disease progression partly via an increased release of adipokines into the circulation. It has previously been shown that poor glycemic control in T2D patients is associated with insulin resistance but also with elevated levels of CRP and TNF-a [29]. The present study shows that levels of SAA are associated with HOMA-IR index in non-diabetics suggesting that inflammatory activity is associated with degree of insulin resistance in the pre-diabetic state. The lack of association in diabetic subjects, in concert with the lack of association between SAA and adiposity, as well as SAA and insulin levels in diabetic subjects suggest that the regulation of these relationships may be disturbed by disease progression. In a previous report consisting of mainly male T2D subjects, it was shown that HbA1c and HOMA-IR correlates with serum levels of SAA [34]. We also found a borderline significant correlation between SAA and HbA1c in male diabetic subjects. However, in women, we find a borderline significant correlation between SAA and HbA1c in non-diabetic, but not in diabetic women. This may explain the fact that we find no correlation between SAA and HbA1c in the whole group of diabetic subjects. We found no correlation between SAA and insulin-stimulated glucose uptake in isolated subcutaneous adipocytes (n = 99) or subcutaneous whole-body insulin sensitivity assessed by a euglycemic–hyperinsulinemic clamp (n = 58) in subsets of the subjects included in the study. We also performed correlation analyses between SAA and measures of glycemic control, insulin and HOMA-IR, in the two subsets but no correlations were found. Therefore, due to limited sample sizes it is quite possible that there are associations between SAA and functional parameters of glucose homeostasis that were not detected. A previous study, on a cohort mainly consisting of female subjects, has shown that serum levels of SAA are associated with IL6 levels [26], which might be expected as IL-6 plays a critical role in the synergistic induction of human SAA [35]. In line with that report, we found positive associations between serum levels of SAA with levels of IL-6, leptin and insulin only in women. Our results indicate that serum SAA levels are correlated with several serum parameters in women but not in men. In two studies [26,31], no association between serum levels of SAA and leptin was found in morbidly obese subjects whereas a correlation was found in a leaner cohort [32], further supporting that SAA may best represent adiposity in the lower range of BMI. Several studies have shown a high correlation between levels of CRP and SAA [1,25,26], which was confirmed in this study. In line with previous results [26], we found no association between serum levels of SAA and TNF-a or adiponectin. Leptin levels have previously been shown to be higher in women than in men [32,36], which may depend on differences in adipose tissue distribution, increased expression in enlarged adipocytes and effects by reproductive hormones. It has previously been shown that women have larger adipocytes than men and that subcutaneous adipocyte size is an independent predictor of plasma leptin levels [8]. It is plausible that the differences in adipocyte size, size-related expression and secretion of adipokines, adipose tissue distribution and sex hormones all affect the associations between SAA, leptin and other parameters making it even more important to analyze men and women separately. Different hypotheses exist on the association between nutritional overload and metabolic disease. One explanation is that increased adipose tissue mass causes increased inflammation, both locally in adipose tissue and systemically. Another explanation is that the adipocyte responds to increased energy availability by lipid spill-over resulting in ectopic fat storage and insulin resistance.
265
We have previously shown that enlarged adipocytes display increased expression of certain genes, including higher levels of SAA and leptin [20]. Considering the role of SAA as a proinflammatory cytokine [25] it is possible that SAA is an important link between adipocyte size, inflammation in adipose tissue and obesity-related disease, especially in hypertrophic obesity. Hence, we hypothesized that hypertrophic obesity leads to higher levels of SAA, which may lead to increased risk for cardiovascular events. However, we find no evidence that SAA is independently associated with adipocyte size. Instead, SAA levels seem correlated with a general increase in adiposity and inflammation [26]. In conclusion, we have found that, in women, adipocyte size is associated with serum levels of SAA and this association is most pronounced in lean/overweight, non-diabetic individuals. We find a strong association between serum levels of SAA and adipokines like leptin and IL-6 in women but not in men. This study suggests that it is extremely important to consider sex-specific factors when analyzing SAA in relation to metabolic disease. Acknowledgments We thank associate professor Bob Olsson, professor Björn Carlsson and professor Lena Carlsson for their helpful discussions on the study and Camilla Glad for technical assistance. This work was supported by grants from the Swedish Foundation for Strategic Research to Sahlgrenska Center for Cardiovascular and Metabolic Research, the Swedish Research Council (529-2002-6671, 5212005-6736), The Swedish federal government under the LUA/ALF agreement, the Sahlgrenska University Hospital Foundation, the Swedish Knowledge Foundation through the Industrial Ph.D. programme in Medical Bioinformatics at Corporate Alliances, Karolinska Institute, The Foundations of the National Board of Health and Welfare, The Royal Physiographic Society in Lund (NilssonEhle Foundation), the Fredrik and Ingrid Thuring Foundation, the Jeansson Foundations and the Åke Wiberg Foundation. References [1] Yudkin JS, Stehouwer CD, Emeis JJ, Coppack SW. C-reactive protein in healthy subjects: associations with obesity, insulin resistance, and endothelial dysfunction: a potential role for cytokines originating from adipose tissue? Arterioscler Thromb Vasc Biol 1999;19:972–8. [2] Das UN. Is obesity an inflammatory condition? Nutrition 2001;17:953–66. [3] Rajala MW, Scherer PE. Minireview: the adipocyte – at the crossroads of energy homeostasis, inflammation, and atherosclerosis. Endocrinology 2003;144:3765–73. [4] Montague CT, O’Rahilly S. The perils of portliness: causes and consequences of visceral adiposity. Diabetes 2000;49:883–8. [5] Weyer C, Foley JE, Bogardus C, Tataranni PA, Pratley RE. Enlarged subcutaneous abdominal adipocyte size, but not obesity itself, predicts type II diabetes independent of insulin resistance. Diabetologia 2000;43:1498–506. [6] Salans LB, Knittle JL, Hirsch J. The role of adipose cell size and adipose tissue insulin sensitivity in the carbohydrate intolerance of human obesity. J Clin Invest 1968;47:153–65. [7] Salans LB, Dougherty JW. The effect of insulin upon glucose metabolism by adipose cells of different size. Influence of cell lipid and protein content, age, and nutritional state. J Clin Invest 1971;50:1399–410. [8] Lundgren M, Svensson M, Lindmark S, Renström F, Ruge T, Eriksson JW. Fat cell enlargement is an independent marker of insulin resistance and ‘hyperleptinaemia’. Diabetologia 2007;50:625–33. [9] Winkler G, Kiss S, Keszthelyi L, Sapi Z, Ory I, Salamon F, et al. Expression of tumor necrosis factor (TNF)-alpha protein in the subcutaneous and visceral adipose tissue in correlation with adipocyte cell volume, serum TNF-alpha, soluble serum TNF-receptor-2 concentrations and C-peptide level. Eur J Endocrinol 2003;149:129–35. [10] Van Harmelen V, Reynisdottir S, Eriksson P, Thorne A, Hoffstedt J, Lönnqvist F, et al. Leptin secretion from subcutaneous and visceral adipose tissue in women. Diabetes 1998;47:913–7. [11] Sopasakis VR, Sandqvist M, Gustafson B, Hammarstedt A, Schmelz M, Yang X, et al. High local concentrations and effects on differentiation implicate interleukin-6 as a paracrine regulator. Obes Res 2004;12:454–60. [12] Skurk T, Alberti-Huber C, Herder C, Hauner H. Relationship between adipocyte size and adipokine expression and secretion. J Clin Endocrinol Metab 2007;92:1023–33.
266
K. Sjöholm et al. / Cytokine 48 (2009) 260–266
[13] Smith U. Studies of human adipose tissue in culture. I. Incorporation of glucose and release of glycerol. Anat Rec 1972;172:597–602. [14] Engfeldt P, Arner P. Lipolysis in human adipocytes, effects of cell size, age and of regional differences. Horm Metab Res Suppl 1988;19:26–9. [15] Despres JP, Fong BS, Julien P, Jimenez J, Angel A. Regional variation in HDL metabolism in human fat cells: effect of cell size. Am J Physiol 1987;252:E654–9. [16] McLaughlin T, Sherman A, Tsao P, Gonzalez O, Yee G, Lamendola C, et al. Enhanced proportion of small adipose cells in insulin-resistant vs insulinsensitive obese individuals implicates impaired adipogenesis. Diabetologia 2007;50:1707–15. [17] Sethi JK, Vidal-Puig AJ. Thematic review series: adipocyte biology. Adipose tissue function and plasticity orchestrate nutritional adaptation. J Lipid Res 2007;48:1253–62. [18] Lelliott C, Vidal-Puig AJ. Lipotoxicity, an imbalance between lipogenesis de novo and fatty acid oxidation. Int J Obes Relat Metab Disord 2004;28(Suppl. 4):S22–8. [19] Sjöholm K, Palming J, Olofsson LE, Gummesson A, Svensson PA, Lystig TC, et al. A microarray search for genes predominantly expressed in human omental adipocytes: adipose tissue as a major production site of serum amyloid A. J Clin Endocrinol Metab 2005;90:2233–9. [20] Jernås M, Palming J, Sjöholm K, Jennische E, Svensson PA, Gabrielsson BG, et al. Separation of human adipocytes by size: hypertrophic fat cells display distinct gene expression. FASEB J 2006;20:1540–2. [21] Clement K, Viguerie N, Poitou C, Carette C, Pelloux V, Curat CA, et al. Weight loss regulates inflammation-related genes in white adipose tissue of obese subjects. FASEB J 2004;18:1657–69. [22] Walder K, Kantham L, McMillan JS, Trevaskis J, Kerr L, De Silva A, et al. Tanis: a link between type 2 diabetes and inflammation? Diabetes 2002;51:1859–66. [23] Cheng N, He R, Tian J, Ye PP, Ye RD. Cutting edge: TLR2 is a functional receptor for acute-phase serum amyloid A. J Immunol 2008;181:22–6. [24] Johnson BD, Kip KE, Marroquin OC, Ridker PM, Kelsey SF, Shaw LJ, et al. Serum amyloid A as a predictor of coronary artery disease and cardiovascular outcome in women: the National Heart, Lung, and Blood Institute-Sponsored Women’s Ischemia Syndrome Evaluation (WISE). Circulation 2004;109:726–32. [25] Yang RZ, Lee MJ, Hu H, Pollin TI, Ryan AS, Nicklas BJ, et al. Acute-phase serum amyloid A: an inflammatory adipokine, potential link between obesity, its metabolic complications. PLoS Med 2006;3:e287. [26] Poitou C, Coussieu C, Rouault C, Coupaye M, Cancello R, Bedel JF, et al. Serum amyloid A: a marker of adiposity-induced low-grade inflammation but not of metabolic status. Obesity (Silver Spring) 2006;14:309–18.
[27] Alberti KG, Zimmet PZ. 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 1998;15:539–53. [28] Lukaski HC, Bolonchuk WW, Hall CB, Siders WA. Validation of tetrapolar bioelectrical impedance method to assess human body composition. J Appl Physiol 1986;60:1327–32. [29] Lindmark S, Buren J, Eriksson JW. Insulin resistance, endocrine function and adipokines in type 2 diabetes patients at different glycaemic levels: potential impact for glucotoxicity in vivo. Clin Endocrinol (Oxf) 2006;65:301–9. [30] Lindmark S, Wiklund U, Bjerle P, Eriksson JW. Does the autonomic nervous system play a role in the development of insulin resistance? A study on heart rate variability in first-degree relatives of Type 2 diabetes patients and control subjects. Diabet Med 2003;20:399–405. [31] van Dielen FM, van’t Veer C, Schols AM, Soeters PB, Buurman WA, Greve JW. Increased leptin concentrations correlate with increased concentrations of inflammatory markers in morbidly obese individuals. Int J Obes Relat Metab Disord 2001;25:1759–66. [32] Lappalainen T, Kolehmainen M, Schwab U, Pulkkinen L, Laaksonen DE, Rauramaa R, et al. Serum concentrations and expressions of serum amyloid A and leptin in adipose tissue are interrelated: the Genobin Study. Eur J Endocrinol 2008;158:333–41. [33] Wakatsuki A, Ikenoue N, Shinohara K, Watanabe K, Fukaya T. Effect of lower dosage of oral conjugated equine estrogen on inflammatory markers and endothelial function in healthy postmenopausal women. Arterioscler Thromb Vasc Biol 2004;24:571–6. [34] Leinonen E, Hurt-Camejo E, Wiklund O, Hulten LM, Hiukka A, Taskinen MR. Insulin resistance and adiposity correlate with acute-phase reaction and soluble cell adhesion molecules in type 2 diabetes. Atherosclerosis 2003;166:387–94. [35] Hagihara K, Nishikawa T, Isobe T, Song J, Sugamata Y, Yoshizaki K. IL-6 plays a critical role in the synergistic induction of human serum amyloid A (SAA) gene when stimulated with proinflammatory cytokines as analyzed with an SAA isoform real-time quantitative RT-PCR assay system. Biochem Biophys Res Commun 2004;314:363–9. [36] Hickey MS, Israel RG, Gardiner SN, Considine RV, McCammon MR, Tyndall GL, et al. Gender differences in serum leptin levels in humans. Biochem Mol Med 1996;59:1–6.
Contents lists available at ScienceDirect
Cytokine journal homepage: www.elsevier.com/locate/issn/10434666
Association of serum amyloid A levels with adipocyte size and serum levels of adipokines: Differences between men and women Kajsa Sjöholm a,*, Magdalena Lundgren b, Maja Olsson a, Jan W. Eriksson b,c a
Sahlgrenska Center for Cardiovascular and Metabolic Research, Department of Molecular and Clinical Medicine, Institute of Medicine, The Sahlgrenska Academy, Göteborg University, Göteborg, Sweden b Department of Medicine, UKBF, Umeå University Hospital, Umeå, Sweden c The Lundberg Laboratory for Diabetes Research, Sahlgrenska Center for Cardiovascular and Metabolic Research, Department of Molecular and Clinical Medicine, Institute of Medicine, The Sahlgrenska Academy, Göteborg University, Göteborg, Sweden
a r t i c l e
i n f o
Article history: Received 21 November 2008 Received in revised form 17 June 2009 Accepted 18 August 2009
Keywords: Adipocyte size Serum amyloid A C-reactive protein Adipokines
a b s t r a c t The aim of this study was to characterize the association between adipocyte enlargement and circulating levels of serum amyloid A (SAA). Furthermore, we wanted to search for possible associations with measures of glycemic control and levels of circulating adipokines and/or inflammatory markers in men and women with a large range in body mass index. The study cohort consisted of 167 subjects, 114 non-diabetic and 53 with Type 2 diabetes. Adipocyte diameter as well as circulating levels of SAA, C-reactive protein (CRP), adiponectin, leptin, interleukin-6, tumor necrosis factor alpha, glucose and insulin were measured. Women had higher serum levels of SAA than men (p = 0.044). SAA levels were weakly but positively correlated with BMI (p = 0.043) and % body fat (p = 0.027) in all subjects as well as subcutaneous adipocyte diameter (p = 0.034) in women. Furthermore, in all subjects we found correlations between SAA levels and levels of CRP (p < 0.001), interleukin-6 (p < 0.001), leptin (p = 0.003), insulin (p = 0.006), HbA1c (p = 0.02) and HOMA-IR (p = 0.002). A majority of the correlations were strongest in women. In conclusion, serum levels of SAA are strongly correlated with serum levels of inflammatory markers as well as measures of glycemic control. There seems to be large sex differences in these associations suggesting that sex-specific factors need to be considered when analyzing SAA levels in relation to metabolic disease. Ó 2009 Elsevier Ltd. All rights reserved.
1. Introduction The prevalence of obesity is increasing rapidly and to the extent that the World Health Organization has denoted it a global epidemic. Obesity is associated with modest but chronically elevated levels of circulating inflammatory markers, such as proinflammatory cytokines and acute phase proteins, and has been designated a low-grade inflammatory state [1,2]. Increasing evidence suggest that factors released from the growing adipose tissue depot may contribute to the development of the metabolic complications related to obesity, such as insulin resistance, Type 2 diabetes and atherosclerosis. Fat cells (adipocytes) produce a number of factors such as cytokines and other bioactive molecules, collectively known as adipokines [3].
* Corresponding author. Address: Sahlgrenska Center for Cardiovascular and Metabolic Research, Department of Molecular and Clinical Medicine, Institute of Medicine, The Sahlgrenska Academy, Göteborg University, SOS-sekr., Vita Stråket 15, SE-413 45 Göteborg, Sweden. Tel.: +46 31 3423029; fax +46 31 418527. E-mail address: [email protected] (K. Sjöholm). 1043-4666/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.cyto.2009.08.005
Central or visceral obesity are strongly related to metabolic disease including atherosclerosis, insulin resistance and hypertension [4]. However, not only the size and location of the adipose tissue is important. Enlargement of adipocytes is also a risk factor for metabolic disease. Adipocytes can increase several thousandfold in volume and about 20-fold in diameter and this increase in size is an independent predictor of Type 2 diabetes [5]. Enlargement of adipocytes is also associated with lowered insulin sensitivity [6–8], increased expression and secretion of adipokines [8–12] as well as changes in lipid and glucose metabolism [12– 15]. It is possible that some subjects display impaired ability to recruit new adipocytes, causing a reduction of adipose tissue lipid storage capacity and detrimental enlargement of existing adipocytes [16]. The nutritional overload may cause altered expression of adipokines and/or lipid spill-over leading to ectopic fat accumulation in other tissues such as liver and skeletal muscle, thereby causing a further increase in insulin resistance [17,18]. We, and others, have previously shown that adipocytes, and in particular enlarged adipocytes, display high expression of the acute-phase protein serum amyloid A (SAA) [19–21]. SAA binds to the cell surface receptor selenoprotein S [22], and it was re-
261
K. Sjöholm et al. / Cytokine 48 (2009) 260–266
cently reported that SAA is also a functional ligand of the toll-like receptor 2 [23]. Hence, SAA is a possible link between Type 2 diabetes, atherosclerosis and inflammation. Furthermore, serum levels of SAA can be used to predict cardiovascular risk [24]. Several of the previous reports on the association between adipocyte volume and expression/circulating levels of SAA, have not considered sex-specific factors and reported associations with circulating adipokines are conflicting [20,25,26]. Therefore, in this study, the aim was to characterize the association between adipocyte enlargement and circulating SAA levels in men and women with a large range in BMI and to search for possible associations with measures of glycemic control and levels of circulating adipokines and/or inflammatory markers.
2. Subjects and methods
2.2. Assessment of cell diameter and glucose uptake in isolated adipocytes Isolation of adipocytes and assessment of cell diameter was performed as described previously [8]. In brief, adipose tissue was minced and then subjected to collagenase digestion. Obtained adipocytes were filtered through a nylon mesh, washed and cell diameter was measured using a light microscope with an integrated ruler. The cell diameter is reported as the average cell diameter (lm) of 100 cells from the subcutaneous and, in a subset (n = 43), also from the omental depot. The rate of glucose uptake was assessed in isolated adipocytes as previously reported and it was performed in a subset (n = 99) of the subjects included in the study [8]. The maximal insulin-stimulated (1000 mU/l) glucose uptake rate (MGU) was expressed as percent of the basal, non-stimulated glucose uptake to provide a measure of insulin action in vitro.
2.1. Subjects and samples The study cohort consisted of 167 subjects (87 women and 80 men), 114 non-diabetic subjects and 53 Type 2 diabetic subjects diagnosed according to the 1998 WHO criteria [27]. Anthropometric and biochemical characteristics of study participants are shown in Table 1. Data were obtained from previous study cohorts at Umeå University Hospital [8] and subjects were included if serum stored at 80 °C was available for SAA analysis. Subcutaneous adipose tissue was obtained through a needle biopsy from the lower part of the abdomen after dermal local anesthesia with lidocain (Xylocain; AstraZeneca, Södertälje, Sweden) and fasting blood samples were collected. Omental adipose tissue biopsies were obtained after induction of general anesthesia in a subset (n = 43) of subjects undergoing elective abdominal surgery. Body fat (% of body weight) was determined by the bioimpedance technique (Bodystat 1500; Bodystat, Isle of Man, Great Britain) [28]. All subjects gave informed consent and the Umeå University Ethics Committee approved the studies.
2.3. Blood chemistry and insulin sensitivity An enzyme-linked immunosorbent assay was used to analyze serum levels of serum amyloid A (Biosource International, Camarillo, CA, USA). The inter-assay coefficient of variation for the SAA analysis was 14% at 30 lg/ml. Other measurements were performed as previously described [8] using commercial enzymatic kits for adiponectin and leptin (Linco Research, St. Charles, MO, USA), interleukin-6 (IL-6) and tumor necrosis factor alpha (TNFa) (R&D Systems, Minneapolis, MN, USA). Serum insulin concentrations were measured by microparticle enzyme immunoassays (Immulite 2000, DPC, Los Angeles, CA, USA; Abbot Imx, Abbott Laboratories, Abbott Park, IL, USA). All other measurements were made according to routine methods at the Department of Clinical Chemistry, Umeå University Hospital. In a subset of the subjects included in the study (n = 58) insulin sensitivity in vivo was assessed as glucose infusion rate, i.e., the
Table 1 Characteristics of the subjects included in the study.c
N Age BMI (kg/m2) WHR Fat (%) P-Glucose (mmol/l) S-Insulin (mU/l) HbA1c (%) HOMA S-SAA (lg/ml) S-Hs CRP (mg/l) P-Leptin (ng/ml) P-IL-6 (pg/ml) P-TNF-a (pg/ml) P-Adiponectin (lg/ml) Sc adipocyte diameter (lm) Om adipocyte diameter (lm) Sc MGU (%)a M-value (mg kg 1 min 1)b
All T2D
All Non-diabetic
Women T2D
Women Non-diabetic
Men T2D
Men Non-diabetic
53 59.2 ± 8.4 28.4 ± 4.7 0.96 ± 0.08 31.8 ± 11.4 9.9 ± 3.1 11.3 ± 7.5 7.4 ± 1.8 4.9 ± 3.7 16.4 ± 10.3 2.6 ± 2.3 12.8 ± 10.0 2.5 ± 1.8 1.8 ± 1.4 7.2 ± 3.4 105.1 ± 9.4 NA 190.3 ± 57.1 5.1 ± 2.2
114 51.3 ± 15.3** 26.8 ± 4.6* 0.89 ± 0.08*** 31.4 ± 10.0 5.1 ± 0.6*** 9.1 ± 6.0* 4.3 ± 0.5*** 2.1 ± 1.6*** 15.1 ± 12.0 1.6 ± 1.6** 14.8 ± 13.8 1.8 ± 1.2** 1.6 ± 2.2* 9.9 ± 5.0*** 100.1 ± 12.3* 94.3 ± 14.2 260.1 ± 111.2*** 10.3 ± 4.0***
22 60.2 ± 9.8 28.6 ± 6.6 0.91 ± 0.06 39.7 ± 13.0 8.6 ± 2.3 12.6 ± 8.7 6.9 ± 1.5 4.8 ± 3.6 19.2 ± 12.0 2.7 ± 2.4 20.2 ± 11.0 2.8 ± 2.0 2.1 ± 1.9 7.3 ± 3.8 109.8 ± 8.4 NA 201.2 ± 72.8 5.0 ± 2.2
65 51.4 ± 16.0* 26.5 ± 5.3 0.84 ± 0.06*** 36.5 ± 9.3 5.0 ± 0.6*** 9.3 ± 7.2* 4.2 ± 0.5*** 2.2 ± 1.9*** 16.7 ± 12.9 1.8 ± 1.9* 21.4 ± 15.2 1.7 ± 1.3** 1.7 ± 2.3 11.6 ± 5.3*** 101.0 ± 13.9* 91.5 ± 15.0 263.3 ± 117.8* 10.8 ± 4.0***
31 58.5 ± 7.2 28.2 ± 2.7 1.00 ± 0.06 26.2 ± 5.6 10.9 ± 3.3 10.3 ± 6.5 7.7 ± 1.9 5.0 ± 3.9 14.4 ± 8.6 2.5 ± 2.3 7.5 ± 4.5 2.3 ± 1.8 1.7 ± 0.9 7.2 ± 3.1 101.7 ± 8.6 NA 183.4 ± 44.7 5.1 ± 2.2
49 51.2 ± 14.3* 27.2 ± 3.6 0.95 ± 0.06** 24.7 ± 6.1 5.2 ± 0.7*** 8.7 ± 4.0 4.4 ± 0.5*** 2.0 ± 1.1*** 13.0 ± 10.5 1.4 ± 1.0 6.4 ± 3.5 1.8 ± 1.1 1.6 ± 2.1* 7.8 ± 3.5 99.0 ± 10.1 97.5 ± 12.9 256.5 ± 105.5** 9.7 ± 4.0***
NA, non-applicable. Significance between diabetic and non-diabetic subjects are given below. * p 0.05. ** p 0.01. *** p 0.001. a n = 99. b n = 58. c Data are means ± SD.
262
K. Sjöholm et al. / Cytokine 48 (2009) 260–266
M-value, at steady state (60–120 min) of a 2-h euglycemic–hyperinsulinemic (56 mU m 2 body surface area min 1) clamp. This procedure was performed as previously described in detail [29,30]. 2.4. Statistical analyses Statistical analyses were performed using the SPSS package (version 16.0; SPSS, Chicago, IL, USA). Differences in measured variables between diabetics and non-diabetics, women and men, and between lean and overweight/obese were analyzed using the Mann–Whitney U test. The relationship between serum levels of SAA and other variables was analyzed using the Spearman rank correlation test. Partial correlation was used to control for age, sex and body mass index (BMI) when appropriate. Skewed variables were log transformed prior to partial correlation analysis. 3. Results 3.1. Anthropometric and biochemical characteristics Characteristics of the subjects included in this study are shown in Table 1. Women had higher circulating levels of SAA (p = 0.044; Fig. 1A), adiponectin (p < 0.001) and leptin (p < 0.001) and lower waist-to-hip ratio (WHR) (p < 0.001), HbA1c (p = 0.005) and levels of glucose (p = 0.007) than men. In all subjects, obese (BMI 30 kg/m2) subjects had significantly higher levels of SAA than nonobese (p = 0.006; Fig. 1B). Type 2 diabetic subjects were older (p = 0.001), had larger subcutaneous adipocytes (p = 0.017), higher BMI (p = 0.037), WHR (p < 0.001), glucose (p < 0.001), insulin (p = 0.014), HbA1c (p < 0.001), homeostasis model assessment of insulin resistance (HOMA-IR) (p < 0.001), C-reactive protein (CRP) (p = 0.006), IL-6 (p = 0.001), TNF-a (p = 0.011) and lower adiponectin (p < 0.001), subcutaneous MGU (p < 0.001) and M-value (p < 0.001) (Table 1) than non-diabetic subjects. There were no significant differences in SAA levels or % fat mass between non-diabetic and Type 2 diabetic subjects (Table 1). 3.2. SAA serum levels in relation to BMI and adipocyte size In all subjects, serum SAA levels were positively correlated with BMI and % body fat (Table 2), but not with subcutaneous (Table 2) or omental adipocyte diameter (data not shown). Interestingly, in women, but not in men, SAA levels were positively correlated with BMI (r = 0.27; p = 0.012), % body fat (r = 0.27; p = 0.018) and subcutaneous (r = 0.27; p = 0.034) (Table 2), but not with omental adipocyte diameter (data not shown). However, the correlation between
SAA and subcutaneous adipocyte diameter was lost when controlling for age and BMI. When dividing female subjects by BMI range (lean, n = 23, 24.99 kg/m2, overweight, n = 20, 25.00–29.99 kg/m2 and obese, n = 21, BMI 30.00 kg/m2), SAA levels in lean subjects were positively correlated with subcutaneous adipocyte diameter (r = 0.425; p = 0.043; Fig. 2A), borderline significance was found in overweight subjects (r = 0.42; p = 0.066; Fig. 2B) and no correlation was seen in obese subjects (r = 0.066; p = 0.775; Fig. 2C). After controlling for age and BMI no correlations were found between serum levels of SAA and adipocyte diameter. In men, no correlation was seen (Fig. 2D) even after dividing subjects by BMI range (r = 0.131; p = 0.297). When dividing all subjects by Type 2 diabetes status, we found that in non-diabetic subjects, serum levels of SAA were positively correlated with % body fat (r = 0.21; p = 0.034), but not with BMI (r = 0.12; p = 0.43). No correlation was found in diabetic subjects. When splitting also by sex, we found that in non-diabetic women, borderline significance was found between SAA levels and BMI and % body fat (Table 2). In non-diabetic men, diabetic men and diabetic women no correlations were found between SAA and BMI or % body fat. 3.3. SAA serum levels in relation to glycemic control and inflammatory markers When analyzing all subjects (Table 2), SAA levels positively correlated with HbA1c, HOMA-IR and insulin. SAA levels were also positively correlated with serum levels of CRP, leptin and IL-6 (Table 2), but not with adiponectin, TNF-a or glucose (data not shown). Furthermore, in all subjects, significant associations remained significant or marginally significant after controlling for sex, age and BMI except for leptin (Table 2). No correlations were found between SAA levels and insulin-stimulated glucose uptake in isolated subcutaneous adipocytes (n = 99) or whole-body insulin sensitivity (n = 58) assessed by a euglycemic–hyperinsulinemic clamp in the subsets where data were available (data not shown). When analyses were split by sex, (Table 2), SAA serum levels positively correlated with CRP (Fig. 3A and B), HbA1c and HOMAIR in both men and women (Fig. 3C and D). In women there were also positive correlations between circulating levels of SAA and insulin, leptin and IL-6 (Table 2). In men there was a borderline significant correlation between SAA and IL-6 (Table 2) but no correlations with insulin and leptin. When dividing subjects into groups of diabetics and non-diabetics, we found that, in both groups, SAA levels correlated with CRP (r = 0.45; p = 0.002 and r = 0.33; p = 0.001, respectively).
Fig. 1. Serum levels of SAA in men (n = 80) and women (n = 87) (A) and obese (n = 43) and non-obese (n = 124) subjects (B). Women display higher levels than men (p = 0.044) and obese subjects display higher levels than non-obese (p = 0.006). Values are expressed as means ± SEM.
263
K. Sjöholm et al. / Cytokine 48 (2009) 260–266 Table 2 Correlation analysis of SAA serum levels with adiposity, glycemic control variables, adipokines, CRP and adipocyte diameter. Variablea
All subjects
Women
Men
Women T2D
Non-diabetic
T2D
Non-diabetic
N BMI (kg/m2)
167 r = 0.16; p = 0.043 r = 0.18; p = 0.027 r = 0.21; p = 0.006b r = 0.19; p = 0.020b r = 0.24; p = 0.002b r = 0.39; p < 0.001b r = 0.24; p = 0.003 r = 0.28; p < 0.001b r = 0.11; p = 0.210
87 r = 0.27; p = 0.012 r = 0.27; p = 0.018 r = 0.25; p = 0.021 r = 0.24; p = 0.035 r = 0.27; p = 0.013 r = 0.42; p < 0.001 r = 0.30; p = 0.007 r = 0.37; p = 0.001 r = 0.27; p = 0.034
80 r = 0.03; p = 0.776 r = 0.05; p = 0.694 r = 0.18; p = 0.116 r = 0.27; p = 0.019 r = 0.26; p = 0.021 r = 0.38; p = 0.001 r = 0.12; p = 0.301 r = 0.21; p = 0.075 r = 0.13; p = 0.297
22 r = 0.30; p = 0.170 r = 0.34; p = 0.156 r = 0.26; p = 0.247 r = 0.03; p = 0.895 r = 0.30; p = 0.175 r = 0.53; p = 0.019 r = 0.52; p = 0.013 r = 0.64; p = 0.002 r = 0.30; p = 0.198
65 r = 0.24; p = 0.059 r = 0.25; p = 0.056 r = 0.21; p = 0.090 r = 0.24; p = 0.068 r = 0.23; p = 0.068 r = 0.40; p = 0.004 r = 0.23; p = 0.093 r = 0.23; p = 0.084 r = 0.25; p = 0.096
31 r = 0.05; p = 0.811 r = 0.13; p = 0.534 r = 0.19; p = 0.318 r = 0.34; p = 0.061 r = 0.11; p = 0.548 r = 0.45; p = 0.019 r = 0.16; p = 0.381 r = 0.28; p = 0.123 r = 0.11; p = 0.585
49 r = 0.06; p = 0.700 r = 0.05; p = 0.768 r = 0.33; p = 0.019 r = 0.17; p = 0.284 r = 0.27; p = 0.056 r = 0.28; p = 0.081 r = 0.23; p = 0.129 r = 0.16; p = 0.301 r = 0.13; p = 0.457
Fat (%) S-Insulin (mU/l) HbA1c (%) HOMA S-Hs CRP (mg/l) P-Leptin (ng/ml) P-IL-6 (pg/ml) Sc adipocyte diameter (lm)
Men
a Data are means ± SD. Correlations with TNF-a, adiponectin, glucose, omental adipocyte diameter, MGU and M-value were omitted as correlations were not significant. Spearman correlation coefficient Rho, r; significant correlations in bold. b Correlations adjusted for age, sex and BMI significant for HOMA (p = 0.012), S-Hs CRP (p < 0.001), HbA1c (p = 0.025) and P-IL-6 (p = 0.003) and marginally significant for SInsulin (p = 0.065). Skewed variables were log transformed prior to partial correlation analysis.
For IL-6 a significant correlation was found in the diabetic group (r = 0.44; p = 0.001) but only borderline in the non-diabetics (r = 0.18; p = 0.078). In the non-diabetics, SAA levels also correlated with insulin (r = 0.24; p = 0.01), HOMA-IR (r = 0.22; p = 0.02) and leptin (r = 0.28; p = 0.005) but no association was found in the diabetic group. When splitting also by sex, we found that in non-diabetic men (n = 49), SAA levels were positively correlated with
insulin but in non-diabetic women (n = 65) only a borderline significant correlation was found (Table 2). Furthermore, in non-diabetic women there were borderline significant positive correlations between SAA levels and HbA1c, HOMA-IR, leptin and IL-6. In non-diabetic men, there was a borderline significant positive correlation between SAA levels and HOMA-IR. In diabetic women significant correlations were found between SAA and leptin
Fig. 2. Scatter plots of serum levels of SAA and adipocyte diameter in lean (n = 23) (A), overweight (n = 20) (B) and obese (n = 21) (C) women and all men included in the study (n = 65) (D) where cell diameter was available. Rho; Spearman’s correlation coefficient. SAA was log transformed for presentation due to a skewed distribution. A regression line was included for illustration.
264
K. Sjöholm et al. / Cytokine 48 (2009) 260–266
Fig. 3. Scatter plots of serum levels of SAA and CRP in women (n = 71) (A) and men (n = 67) (B) and HOMA-IR index in women (n = 86) (C) and men (n = 79) (D). Rho; Spearman’s correlation coefficient. All variables were log transformed for presentation due to skewed distributions. A regression line was included for illustration.
and IL-6 (Table 2). In diabetic men, a significant correlation was found with CRP and borderline significant correlation was found with Hba1c.
4. Discussion In this study we wanted to examine the possible association between serum levels of SAA and adipocyte size as well as with circulating adipokines and CRP in lean and obese, female and male subjects. SAA is an acute phase protein, which was previously believed only to be expressed in, and released from, the liver in response to infection or inflammation. However, we [19] and others [21], have recently discovered that adipocytes are a major site of expression during non-acute phase and that SAA levels in serum are highly correlated with SAA expression in adipose tissue [19]. Furthermore, enlarged adipocytes have increased expression [20] suggesting that, in an obese individual, the contribution of adipose SAA to serum levels may be substantial, especially in hypertrophic obesity, i.e., individuals with enlarged adipocytes. In this study, we report a correlation between adipocyte size and serum levels of SAA in non-obese women and show that there are sex differences, not only in adipose tissue expression of SAA [19] which we have reported previously, but also in associations with adipocyte size and serum levels of other adipokines. In line with previous reports [25,26,31,32], we found that SAA serum levels are increased in the obese, that SAA serum levels are positively correlated with BMI and that women have higher levels of SAA than men. Furthermore, we found that in women, serum levels of SAA were correlated with adipocyte size. We have also extended this finding by showing that the highest correlation is found in non-obese women. We have previously reported that women have higher mRNA levels of SAA in subcutaneous adipose tissue than men [19]. However, in that study, this difference was not reflected in SAA serum levels, which probably can be explained by the limited sample size (n = 6 women), and that all subjects included in that study were
morbidly obese. The findings in the current study support the fact that obesity-induced enlargement of adipocytes may cause an increase in circulating levels of SAA. However, we find no evidence that adipocyte size can be used to predict serum levels of SAA. Instead, a general increase in adiposity may cause an increase in levels of circulating SAA that, in turn, may induce other inflammatory mediators [25]. We have previously shown that subcutaneous adipose tissue displays higher levels of SAA expression than omental adipose tissue and that hypertrophic adipocytes have increased levels of SAA both when analyzing gene expression and protein levels [19,20]. Furthermore, we, and others have previously shown an association between adipocyte volume and SAA mRNA levels [20,25,26]. In the study by Yang et al. (n = 19), it is not clear if the subjects were obese or not, nor whether they were male or female. In the study by Lappalainen et al. (n = 37 men/38 women), SAA mRNA levels correlated with adipocyte size but no correlation was found between adipocyte size and serum levels of SAA [32]. The only study where an association between circulating levels of SAA and adipocyte size has previously been found is a study by Poiteu et al. where morbidly obese subjects (n = 10 men/50 women) where studied [26]. In the present study, 87 women and 80 men with a large range in BMI (16– 49 kg/m2) were included. We found no association between adipocyte diameter and SAA serum levels in all subjects, men or obese women. In contrast, we found a weak but significant association between adipocyte diameter and SAA levels in lean women and borderline significance in overweight women. Considering the higher expression of SAA in subcutaneous adipose tissue and a gynoid adipose tissue distribution with larger subcutaneous depots in women, the finding of increased levels of serum SAA is not surprising. However, further studies are needed to elucidate whether there actually is a lack of association in the higher range of BMI. The reason for the higher SAA expression in women is unknown but one possible explanation is that it could be related to hormonal status as estrogen replacement in postmenopausal women increases inflammatory factors like IL-6, CRP and SAA [33]. Another possible explanation could be direct
K. Sjöholm et al. / Cytokine 48 (2009) 260–266
or indirect interaction with leptin [32]. We found no association between omental adipocyte diameter and serum SAA in this study. This could be due to the limited number of samples available for that analysis. In T2D, hyperglycemia could lead to disease progression partly via an increased release of adipokines into the circulation. It has previously been shown that poor glycemic control in T2D patients is associated with insulin resistance but also with elevated levels of CRP and TNF-a [29]. The present study shows that levels of SAA are associated with HOMA-IR index in non-diabetics suggesting that inflammatory activity is associated with degree of insulin resistance in the pre-diabetic state. The lack of association in diabetic subjects, in concert with the lack of association between SAA and adiposity, as well as SAA and insulin levels in diabetic subjects suggest that the regulation of these relationships may be disturbed by disease progression. In a previous report consisting of mainly male T2D subjects, it was shown that HbA1c and HOMA-IR correlates with serum levels of SAA [34]. We also found a borderline significant correlation between SAA and HbA1c in male diabetic subjects. However, in women, we find a borderline significant correlation between SAA and HbA1c in non-diabetic, but not in diabetic women. This may explain the fact that we find no correlation between SAA and HbA1c in the whole group of diabetic subjects. We found no correlation between SAA and insulin-stimulated glucose uptake in isolated subcutaneous adipocytes (n = 99) or subcutaneous whole-body insulin sensitivity assessed by a euglycemic–hyperinsulinemic clamp (n = 58) in subsets of the subjects included in the study. We also performed correlation analyses between SAA and measures of glycemic control, insulin and HOMA-IR, in the two subsets but no correlations were found. Therefore, due to limited sample sizes it is quite possible that there are associations between SAA and functional parameters of glucose homeostasis that were not detected. A previous study, on a cohort mainly consisting of female subjects, has shown that serum levels of SAA are associated with IL6 levels [26], which might be expected as IL-6 plays a critical role in the synergistic induction of human SAA [35]. In line with that report, we found positive associations between serum levels of SAA with levels of IL-6, leptin and insulin only in women. Our results indicate that serum SAA levels are correlated with several serum parameters in women but not in men. In two studies [26,31], no association between serum levels of SAA and leptin was found in morbidly obese subjects whereas a correlation was found in a leaner cohort [32], further supporting that SAA may best represent adiposity in the lower range of BMI. Several studies have shown a high correlation between levels of CRP and SAA [1,25,26], which was confirmed in this study. In line with previous results [26], we found no association between serum levels of SAA and TNF-a or adiponectin. Leptin levels have previously been shown to be higher in women than in men [32,36], which may depend on differences in adipose tissue distribution, increased expression in enlarged adipocytes and effects by reproductive hormones. It has previously been shown that women have larger adipocytes than men and that subcutaneous adipocyte size is an independent predictor of plasma leptin levels [8]. It is plausible that the differences in adipocyte size, size-related expression and secretion of adipokines, adipose tissue distribution and sex hormones all affect the associations between SAA, leptin and other parameters making it even more important to analyze men and women separately. Different hypotheses exist on the association between nutritional overload and metabolic disease. One explanation is that increased adipose tissue mass causes increased inflammation, both locally in adipose tissue and systemically. Another explanation is that the adipocyte responds to increased energy availability by lipid spill-over resulting in ectopic fat storage and insulin resistance.
265
We have previously shown that enlarged adipocytes display increased expression of certain genes, including higher levels of SAA and leptin [20]. Considering the role of SAA as a proinflammatory cytokine [25] it is possible that SAA is an important link between adipocyte size, inflammation in adipose tissue and obesity-related disease, especially in hypertrophic obesity. Hence, we hypothesized that hypertrophic obesity leads to higher levels of SAA, which may lead to increased risk for cardiovascular events. However, we find no evidence that SAA is independently associated with adipocyte size. Instead, SAA levels seem correlated with a general increase in adiposity and inflammation [26]. In conclusion, we have found that, in women, adipocyte size is associated with serum levels of SAA and this association is most pronounced in lean/overweight, non-diabetic individuals. We find a strong association between serum levels of SAA and adipokines like leptin and IL-6 in women but not in men. This study suggests that it is extremely important to consider sex-specific factors when analyzing SAA in relation to metabolic disease. Acknowledgments We thank associate professor Bob Olsson, professor Björn Carlsson and professor Lena Carlsson for their helpful discussions on the study and Camilla Glad for technical assistance. This work was supported by grants from the Swedish Foundation for Strategic Research to Sahlgrenska Center for Cardiovascular and Metabolic Research, the Swedish Research Council (529-2002-6671, 5212005-6736), The Swedish federal government under the LUA/ALF agreement, the Sahlgrenska University Hospital Foundation, the Swedish Knowledge Foundation through the Industrial Ph.D. programme in Medical Bioinformatics at Corporate Alliances, Karolinska Institute, The Foundations of the National Board of Health and Welfare, The Royal Physiographic Society in Lund (NilssonEhle Foundation), the Fredrik and Ingrid Thuring Foundation, the Jeansson Foundations and the Åke Wiberg Foundation. References [1] Yudkin JS, Stehouwer CD, Emeis JJ, Coppack SW. C-reactive protein in healthy subjects: associations with obesity, insulin resistance, and endothelial dysfunction: a potential role for cytokines originating from adipose tissue? Arterioscler Thromb Vasc Biol 1999;19:972–8. [2] Das UN. Is obesity an inflammatory condition? Nutrition 2001;17:953–66. [3] Rajala MW, Scherer PE. Minireview: the adipocyte – at the crossroads of energy homeostasis, inflammation, and atherosclerosis. Endocrinology 2003;144:3765–73. [4] Montague CT, O’Rahilly S. The perils of portliness: causes and consequences of visceral adiposity. Diabetes 2000;49:883–8. [5] Weyer C, Foley JE, Bogardus C, Tataranni PA, Pratley RE. Enlarged subcutaneous abdominal adipocyte size, but not obesity itself, predicts type II diabetes independent of insulin resistance. Diabetologia 2000;43:1498–506. [6] Salans LB, Knittle JL, Hirsch J. The role of adipose cell size and adipose tissue insulin sensitivity in the carbohydrate intolerance of human obesity. J Clin Invest 1968;47:153–65. [7] Salans LB, Dougherty JW. The effect of insulin upon glucose metabolism by adipose cells of different size. Influence of cell lipid and protein content, age, and nutritional state. J Clin Invest 1971;50:1399–410. [8] Lundgren M, Svensson M, Lindmark S, Renström F, Ruge T, Eriksson JW. Fat cell enlargement is an independent marker of insulin resistance and ‘hyperleptinaemia’. Diabetologia 2007;50:625–33. [9] Winkler G, Kiss S, Keszthelyi L, Sapi Z, Ory I, Salamon F, et al. Expression of tumor necrosis factor (TNF)-alpha protein in the subcutaneous and visceral adipose tissue in correlation with adipocyte cell volume, serum TNF-alpha, soluble serum TNF-receptor-2 concentrations and C-peptide level. Eur J Endocrinol 2003;149:129–35. [10] Van Harmelen V, Reynisdottir S, Eriksson P, Thorne A, Hoffstedt J, Lönnqvist F, et al. Leptin secretion from subcutaneous and visceral adipose tissue in women. Diabetes 1998;47:913–7. [11] Sopasakis VR, Sandqvist M, Gustafson B, Hammarstedt A, Schmelz M, Yang X, et al. High local concentrations and effects on differentiation implicate interleukin-6 as a paracrine regulator. Obes Res 2004;12:454–60. [12] Skurk T, Alberti-Huber C, Herder C, Hauner H. Relationship between adipocyte size and adipokine expression and secretion. J Clin Endocrinol Metab 2007;92:1023–33.
266
K. Sjöholm et al. / Cytokine 48 (2009) 260–266
[13] Smith U. Studies of human adipose tissue in culture. I. Incorporation of glucose and release of glycerol. Anat Rec 1972;172:597–602. [14] Engfeldt P, Arner P. Lipolysis in human adipocytes, effects of cell size, age and of regional differences. Horm Metab Res Suppl 1988;19:26–9. [15] Despres JP, Fong BS, Julien P, Jimenez J, Angel A. Regional variation in HDL metabolism in human fat cells: effect of cell size. Am J Physiol 1987;252:E654–9. [16] McLaughlin T, Sherman A, Tsao P, Gonzalez O, Yee G, Lamendola C, et al. Enhanced proportion of small adipose cells in insulin-resistant vs insulinsensitive obese individuals implicates impaired adipogenesis. Diabetologia 2007;50:1707–15. [17] Sethi JK, Vidal-Puig AJ. Thematic review series: adipocyte biology. Adipose tissue function and plasticity orchestrate nutritional adaptation. J Lipid Res 2007;48:1253–62. [18] Lelliott C, Vidal-Puig AJ. Lipotoxicity, an imbalance between lipogenesis de novo and fatty acid oxidation. Int J Obes Relat Metab Disord 2004;28(Suppl. 4):S22–8. [19] Sjöholm K, Palming J, Olofsson LE, Gummesson A, Svensson PA, Lystig TC, et al. A microarray search for genes predominantly expressed in human omental adipocytes: adipose tissue as a major production site of serum amyloid A. J Clin Endocrinol Metab 2005;90:2233–9. [20] Jernås M, Palming J, Sjöholm K, Jennische E, Svensson PA, Gabrielsson BG, et al. Separation of human adipocytes by size: hypertrophic fat cells display distinct gene expression. FASEB J 2006;20:1540–2. [21] Clement K, Viguerie N, Poitou C, Carette C, Pelloux V, Curat CA, et al. Weight loss regulates inflammation-related genes in white adipose tissue of obese subjects. FASEB J 2004;18:1657–69. [22] Walder K, Kantham L, McMillan JS, Trevaskis J, Kerr L, De Silva A, et al. Tanis: a link between type 2 diabetes and inflammation? Diabetes 2002;51:1859–66. [23] Cheng N, He R, Tian J, Ye PP, Ye RD. Cutting edge: TLR2 is a functional receptor for acute-phase serum amyloid A. J Immunol 2008;181:22–6. [24] Johnson BD, Kip KE, Marroquin OC, Ridker PM, Kelsey SF, Shaw LJ, et al. Serum amyloid A as a predictor of coronary artery disease and cardiovascular outcome in women: the National Heart, Lung, and Blood Institute-Sponsored Women’s Ischemia Syndrome Evaluation (WISE). Circulation 2004;109:726–32. [25] Yang RZ, Lee MJ, Hu H, Pollin TI, Ryan AS, Nicklas BJ, et al. Acute-phase serum amyloid A: an inflammatory adipokine, potential link between obesity, its metabolic complications. PLoS Med 2006;3:e287. [26] Poitou C, Coussieu C, Rouault C, Coupaye M, Cancello R, Bedel JF, et al. Serum amyloid A: a marker of adiposity-induced low-grade inflammation but not of metabolic status. Obesity (Silver Spring) 2006;14:309–18.
[27] Alberti KG, Zimmet PZ. 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 1998;15:539–53. [28] Lukaski HC, Bolonchuk WW, Hall CB, Siders WA. Validation of tetrapolar bioelectrical impedance method to assess human body composition. J Appl Physiol 1986;60:1327–32. [29] Lindmark S, Buren J, Eriksson JW. Insulin resistance, endocrine function and adipokines in type 2 diabetes patients at different glycaemic levels: potential impact for glucotoxicity in vivo. Clin Endocrinol (Oxf) 2006;65:301–9. [30] Lindmark S, Wiklund U, Bjerle P, Eriksson JW. Does the autonomic nervous system play a role in the development of insulin resistance? A study on heart rate variability in first-degree relatives of Type 2 diabetes patients and control subjects. Diabet Med 2003;20:399–405. [31] van Dielen FM, van’t Veer C, Schols AM, Soeters PB, Buurman WA, Greve JW. Increased leptin concentrations correlate with increased concentrations of inflammatory markers in morbidly obese individuals. Int J Obes Relat Metab Disord 2001;25:1759–66. [32] Lappalainen T, Kolehmainen M, Schwab U, Pulkkinen L, Laaksonen DE, Rauramaa R, et al. Serum concentrations and expressions of serum amyloid A and leptin in adipose tissue are interrelated: the Genobin Study. Eur J Endocrinol 2008;158:333–41. [33] Wakatsuki A, Ikenoue N, Shinohara K, Watanabe K, Fukaya T. Effect of lower dosage of oral conjugated equine estrogen on inflammatory markers and endothelial function in healthy postmenopausal women. Arterioscler Thromb Vasc Biol 2004;24:571–6. [34] Leinonen E, Hurt-Camejo E, Wiklund O, Hulten LM, Hiukka A, Taskinen MR. Insulin resistance and adiposity correlate with acute-phase reaction and soluble cell adhesion molecules in type 2 diabetes. Atherosclerosis 2003;166:387–94. [35] Hagihara K, Nishikawa T, Isobe T, Song J, Sugamata Y, Yoshizaki K. IL-6 plays a critical role in the synergistic induction of human serum amyloid A (SAA) gene when stimulated with proinflammatory cytokines as analyzed with an SAA isoform real-time quantitative RT-PCR assay system. Biochem Biophys Res Commun 2004;314:363–9. [36] Hickey MS, Israel RG, Gardiner SN, Considine RV, McCammon MR, Tyndall GL, et al. Gender differences in serum leptin levels in humans. Biochem Mol Med 1996;59:1–6.