Adipose tissue-derived cytokines and their correlations with clinical characteristics in Vietnamese patients with type 2 diabetes mellitus

Background Adipokines are involved in the pathogenesis of metabolic disorders including obesity and type 2 diabetes mellitus (T2DM). This study investigates the levels of leptin, resistin, visfatin, secreted frizzled-related protein 5 (SFRP5), monocyte chemoattractant protein-1 (MCP-1) and retinol-binding protein 4 (RBP4) and their correlations with clinical parameters of overweight and T2DM. Methods We recruited overweight 50 patients with T2DM, 88 non-overweight patients with T2DM, 29 overweight and 100 non-overweight individuals devoid of T2DM for this study. The levels of studied adipokines were measured by enzyme-linked immunosorbent assay and correlated with clinical parameters. Results The levels of MCP-1 and SFRP5 were decreased while visfatin and RBP4 levels were increased in patients with T2DM compared to those in the control individuals (P < 0.01). Among patients with T2DM, leptin and resistin levels were higher while RBP4 levels were lower in patients with overweight T2DM compared to those in patients with non-overweight T2DM (P < 0.0001, 0.019 and 0.05, respectively). Leptin and MCP-1 levels were correlated with HOMA-IR, QUICKI and HOMA-β. Leptin/MCP-1 ratio was correlated with insulin levels, HOMA-IR and HOMA-β indexes. Resistin/RBP4, visfatin/MCP-1 and MCP-1/RBP4 ratios were strongly correlated with the levels of fasting glucose, HbA1c and HOMA-β. In addition, ROC curve analyses indicated a diagnostic potential of resistin/RBP4 and MCP-1/RBP4 indexes for T2DM (AUC = 0.81 and 0.83, respectively) and β-cell function (AUC = 0.76 and 0.74, respectively). Conclusions Adipokines (leptin, resistin, visfatin, SFRP5, MCP-1, and RBP4) are associated with overweight and T2DM and may serve as a potential prognostic marker and therapeutic intervention for overweight-related T2DM.

Background Type 2 diabetes mellitus (T2DM) is a chronic metabolic disorder and is exponentially increasing in developing countries including Vietnam due to an increased consumption of energy-rich food, sedentary lifestyle, and urbanization [1]. According to the International Diabetes Federation (IDF), there were 382 million diabetic cases in 2013 and the number of diabetic cases is predicted to be 592 million by the year 2035 [2,3]. Approximately 80% of diabetic cases are in developing countries and there are 5.1 million deaths worldwide annually due to this disease [2,3]. The prevalence of T2DM is 1.2% in developed countries while there is four times higher in developing countries [3]. In Vietnam, the prevalence of T2DM is increasing rapidly and approximately 3.3 million diabetic cases were reported in 2014. The prevalence of diabetes in the age group of 30-69 years is estimated to be 5.7% across the country and 7% in the urban areas [4,5]. However, a recent study conducted with a large sample size in the Red River Delta of Vietnam showed that the general population aged 40-64 years had a low level of diabetes knowledge, especially in the rural areas [6]. This might be one of the challenges for public health to control the disease in Vietnam.
T2DM constitutes up to 90% of all diabetes and it is characterized by chronic hyperglycaemia resulting from defects in insulin secretion and/or insulin action and metabolic disorders of protein and lipids [7,8]. The main pathogenesis of T2DM includes insulin resistance and insufficient insulin production of pancreatic β cells, which lead to the disability to control glucose level in the circulation [9]. Additionally, the increased production of pro-inflammatory cytokines due to obesity contributes to the increased risk of T2DM development [7,10]. Insulin resistance is significantly related to obesity and develops years before the clinical manifestations of T2DM [11]. Adipose tissue is considered as an endocrine organ that produces and metabolites numerous proteins, hormones, cytokines, and sex steroids, namely adipocytokine (or adipokines) [12,13]. Adipocytes are also involved in various biological processes such as energy metabolism, neuroendocrine function and immune response through the activities of their receptors for hormones of the endocrine and central nervous system [13]. The increased adipose tissue is associated with insulin resistance, elevated blood glucose levels, lipid metabolic disorders, hypertension and inflammation [14].
Adipose tissue-derived proteins with hormone activities regulate metabolic functionalities and inflammation [14]. The adipose tissue-derived adipokines such as adiponectin, leptin, visfatin, resistin and adipsin are associated with obesity and obesity-related metabolic disorders including T2DM [15]. Other adipose tissue-derived proteins also play important roles in metabolism, immune response, and metabolic disorders. Particularly, secreted frizzled-related protein 5 (SFRP5) is a member of the Sfrp family that is involved in the modulation of the Wnt signaling [16]. SFRP5 is highly expressed in adipose tissue and has been associated with β-cell function, glucose metabolism, obesity and T2DM [17][18][19][20]. Monocyte chemoattractant protein-1 (MCP-1), known as chemokine (C-C motif ) ligand 2 (CCL2), is a chemoattractant protein for several immune cells including monocytes/macrophages, T cells and natural killer (NK) cells, and this protein has been implicated in adipose tissue inflammation [21]. In addition, retinolbinding protein 4 (RBP4) is an adipose tissue-derived cytokine that transports retinol from the liver to various tissues and it is related to insulin resistance, visceral fat distribution, and dyslipidemia [22,23].
Our previous study showed that the levels of adiponectin and pro-inflammatory cytokines including tumor necrosis factor (TNF)-α, IL-1β and IL-10 are modulated and correlated with clinical parameters of overweight and T2DM in Vietnamese patients with T2DM and in overweight individuals [24]. The present study aims to investigate the levels of leptin, resistin, visfatin, SFRP5, MCP-1/CCL2, and RBP4 as well as their correlations with insulin resistance and clinical parameters of overweight and T2DM in a Vietnamese study group.

Study population
The study was designed as a case-controlled study combined with clinical observations and experimental analyses. One hundred and thirty-eight (n = 138) Vietnamese patients with T2DM and one hundred and twenty-nine (n = 129) control individuals were recruited during 2013 and 2014 for this study (Table 1).
Patients with T2DM were diagnosed based on the standard criteria reported by the World Health Organization (WHO) and by the International Diabetes Federation (IDF) [25,26]. The inclusion and exclusion criteria for the patient group as well as the selection criteria for the control groups were described in our previous study [24]. At the time of sampling, most of the recruited patients with T2DM were newly diagnosed and had not been treated with any anti-diabetic drug. A small number of patients were immediately treated with the diabetes drug Diamicron MR and/or a low dose of insulin injection. Clinical examinations and biochemical tests for the function of the heart, lung, liver and kidney were performed as the routine. History of diabetes-associated symptoms, personal and family history of diabetes, family history of premature cardiovascular diseases, smoking, alcohol consumption, and lifestyle were obtained from all participants using the standard study questionnaires ( Table 1). The anthropometric data such as height, weight, waist and hip circumferences were measured using standard procedures for all study participants. Body mass index (BMI) and waist-hip ratio (WHR) were calculated based on their anthropometric measurements. Based on their BMI and T2DM status, the patient group was divided into two subgroups: overweight (BMI ≥ 25) with T2DM (n = 50) and non-overweight (BMI < 25) with T2DM (n = 88). Similarly, the control group was also divided into two subgroups including overweight control individuals (BMI ≥ 25; n = 29) and non-overweight control individuals (BMI ranging from 16 to 25) without T2DM (n = 100) ( Table 1).

Measurement of leptin, resistin, visfatin, SFRP5, MCP-1, and RBP4 levels
The levels of leptin, resistin, visfatin, SFRP5, MCP-1 and RBP4 were measured in the respective serum samples of the study participants by using commercially available ELISA kits following the manufacturer's instruction. No.: RAB0414-1KT) was purchased from Sigma-Aldrich, Missouri, the United States. The detection limit of the ELISA kits is 0.2 ng/ml for leptin, 0.012 ng/ml for resistin, 0.125 ng/ml for visfatin, 15.6 pg/ml for SFRP5, 0 pg/ ml for MCP-1, and 0.1 pg/ml for RBP4.

Statistical analysis
Clinical and demographic data were presented in median values with range for continuous variables. Chi square or Fisher's exact tests were used to compare categorical variables. Multi-factor ANOVA was performed for comparing means between groups and to adjust for the confounding effect of age and sex on the clinical parameters and investigated adipokines. Kruskal-Wallis, Mann-Whitney U test was used to analyze the serum levels of leptin, resistin, visfatin, SFRP5, MCP-1, and RBP4 in patients with T2DM and in controls wherever appropriate. Spearman's rank correlation coefficient was used to analyze the correlation between two studied variables. The diagnostic potential of investigated adipose tissuederived cytokines and their relative ratios in T2DM, insulin resistance/sensitivity and β-cell function was evaluated by receiver operating characteristic (ROC) analysis and the area under the curve (AUC) was also calculated. All statistical analyses were performed using IBM Statistics SPSS v.19 (IBM Corp, Armonk, NY. the USA). P values of less than 0.05 were considered significant.

Demographic, anthropometric, biochemical and clinical characteristics of the study subjects
The demographic, anthropometric, biochemical and clinical characteristics of the patients with T2DM and control individuals are summarized in Table 1. No statistical differences between the mean age of study group (with T2DM, 58.6 ± 7.2 years) and control group (without T2DM; 56.8 ± 8.2 years), between patients with overweight T2DM and those with non-overweight T2DM, as well as between control individuals with and without overweight (P > 0.05) were observed. Regarding gender, there was a higher number of male participants in the group with T2DM compared to that in control group (P = 0.0001). The proportions of smokers and alcohol users in the patient group were also higher compared to that in control group (P = 0.001 and < 0.0001, respectively). In addition, the proportion of patients with T2DM whose at least one of their direct relatives (parents, siblings, children) had T2DM was higher compared to that of the control group (P = 0.009). These results demonstrate that the smoking, alcohol usage and family history of T2DM are associated with the prevalence of T2DM (Table 1).
The levels of fasting glucose, triglycerides and glycosylated hemoglobin (HbA1c) were higher in patients with T2DM compared to those in control individuals (P < 0.0001, 0.023, and < 0.0001, respectively). Glucose was detected more frequently in urine samples of patients with T2DM compared to that in the control group (P = 0.0026). The levels of high-density lipoproteincholesterol (HDL-C) were lower in patients with T2DM compared to those in control individuals (P = 0.002). However, no significant difference in the levels of lowdensity lipoprotein-cholesterol (LDL-C), insulin and total cholesterol between the two groups was observed (P > 0.05). HOMA-IR and HOMA-β indexes were higher while QUICKI index was lower in patients with T2DM compared to those in the control group (P < 0.0001) ( Table 1).
Within the group of patients with T2DM, the levels of fasting glucose and triglycerides were higher while HDL-C levels were decreased in the patients with overweight T2DM compared to those with non-overweight T2DM (P = 0.002, 0.014, 0.042, respectively). Insulin levels and HOMA-IR index were higher while QUICKI index was lower in patients with overweight T2DM compared to those with non-overweight T2DM (P = 0.04, 0.028 and 0.016, respectively). However, within the control individuals, insulin levels, HOMA-IR, HOMA-β and QUICKI indexes were not significantly different between control individuals with and without overweight (P > 0.05). In addition, no difference in the levels of fasting glucose and triglycerides was observed between the two control subgroups (P > 0.05) ( Table 1).

Levels of leptin, resistin, visfatin, SFRP5, MCP-1, and RBP4
The levels of MCP-1 and SFRP5 were significantly decreased (P < 0.0001 and P = 0.003, respectively), while the levels of visfatin and RBP4 were significantly increased in patients with T2DM compared to those in control individuals (P = 0.005 and P = 0.021, respectively). However, no difference in the levels of leptin and resistin between the two groups was observed (P > 0.05) (Fig. 1). Among the patients with T2DM, the levels of leptin and resistin were higher while the RBP4 levels were lower in patients with overweight T2DM compared to those in patients with non-overweight T2DM (P = 0.0001, P = 0.019, and P = 0.05, respectively). Among the control group, the levels of visfatin, MCP-1 and RBP4 were higher in overweight individuals compared to those in non-overweight controls (P = 0.022, 0.048 and 0.05, respectively). We also observed a similar trend for leptin, resistin, and SFRP5 between overweight and nonoverweight control subgroups, but the difference did not reach the statistical significance. In addition, we observed a difference of MCP-1, SFRP5 and RBP4 levels between overweight with T2DM and overweight control groups as well as between non-overweight with T2DM and nonoverweight control groups (P < 0.05) (Fig. 1).

Correlation of leptin, resistin, visfatin, SFRP5, MCP-1, and RBP4 levels with metabolic parameters
We analyzed a correlation of the levels of the studied adipokines with the metabolic parameters of overweight and T2DM. We observed that the leptin levels were positively correlated with insulin levels (Spearman's rho = 0. 43 Table 3).

Discussion
Adipose tissue-derived cytokines play a significant role in the pathogenesis of inflammation and metabolic disorders such as obesity and T2DM [21,30,31]. Our previous study has shown that the levels of adiponectin and several pro-inflammatory cytokines were modulated in patients with T2DM compared to individuals without T2DM [24]. Similarly, we show in the present study that the levels of RBP4 and visfatin are increased while the MCP-1 and SFRP5 levels are decreased in patients with T2DM compared to control group. Although not all the comparisons reached the statistical significance, the levels of studied adipokines are modulated in individuals with overweight compared to those without. Our results revealed that the levels of studied adipokines including leptin, resistin, visfatin, SFRP5, MCP-1, and RBP4 are correlated with clinical parameters of overweight and T2DM. In addition, the relative ratios of studied adipokines are correlated with fasting glucose and HbA1c levels as well as with insulin resistance/sensitivity indexes such as HOMA-IR, QUICKI and β-cell function index (HOMA-β). These results indicate that adipose-derived cytokines are involved in the pathogenesis of obesity and T2DM and are associated with clinical outcomes of overweight and T2DM, and thus may be used as indicators of insulin resistance and β-cell function.
Leptin has been shown to play an important role in the pathogenesis of atherosclerosis, cardiovascular disease, inflammation, obesity and T2DM [32]. The biological

Table 2 Correlation between pair of studied adipokines
The correlations between pair of studied adipokines were calculated by using the Spearman's rank correlation coefficient. Spearman's rho (ρ) and P values are presented

Table 3 Correlation between studied adipokines and clinical parameters of obesity and type 2 diabetes mellitus
The correlations between studied adipokines and clinical parameters of overweight and T2DM were calculated by using Spearman's rank correlation coefficient. role of leptin is to regulate appetite, hunger, body temperature and expenditure of energy through the hypothalamus [30,31]. A recent study showed that lower leptin is associated with T2DM in obese adolescents [33]. However, leptin levels are not significantly different between elderly patients with T2DM and age-matched individuals without T2DM as shown in this study. This contradictory result might be due to the higher age of study subjects in our study. In addition, although leptin levels are not significantly correlated with metabolic parameters of obesity and T2DM, strong correlations of leptin levels with insulin levels and with indexes of insulin resistance/sensitivity and β cell function. These results suggest that leptin can be a promising indicator of insulin production and its resistance.
In line with a previous finding that resistin serum levels were significantly correlated with the inflammatory chemokines such as MCP-1 and RBP4, which are major players in the pathogenesis of metabolic syndromes [34]. A decreased resistin concentration by knocking out the resistin encoding gene or blocking resistin by antibody could protect against obesity-associated hyperglycemia, which is mainly due to recovering the response of the liver to insulin [35]. However, visfatin levels are not affected by the treatment of T2DM with basal insulin [36]. Although the difference is not significant between patients with and without T2DM in our study, resistin levels are increased in patients with overweight T2DM compared to those with non-overweight. These results are similar to other studies indicating that resistin levels are higher in patients with overweight T2DM compared to controls [37,38]. Similarly, our results of visfatin levels in individuals with and without T2DM are consistent with several previous studies showing that visfatin levels are increased in individuals with overweight and T2DM compared to controls [39,40]. Visfatin serum levels are significantly correlated with the accumulation of white adipose tissue (WAT), and visfatin expression was increased during the differentiation of adipocytes and according to the destruction of β cells [41,42]. The negative correlation between the levels of visfatin and glucose indicates that visfatin is an important indicator for the development of obesity and related T2DM.
Similar to leptin, MCP-1 is significantly involved in the pathogenesis of several metabolic disorders including inflammation, obesity and diabetes [43]. MCP-1 also has chemotaxis activity that can promote leukocytes to leave the circulation and form the foam cells that subsequently lead to arterial plaque. Therefore, elevated level of MCP-1 production is a risk factor for the development of cardiovascular disease in patients with obesity-related T2DM [44]. Our study showed that the MCP-1 levels are decreased in patients with T2DM compared to healthy individuals, but increased in overweight healthy individuals compared to those with non-overweight. The strong correlations of MCP-1 levels with the levels of fasting glucose and HbA1c support the fact that MCP-1 contributes to increasing obesity-induced insulin resistance, the effect of macrophage infiltration into adipose tissue and elevated hepatic triacylglycerol [44]. These results suggested that increased MCP-1 production in adipose tissue caused by macrophage infiltration into adipose tissue leads to an increase of insulin resistance and consequently leads to T2DM [43]. Additionally, MCP-1 levels Fig. 3 Diagnostic performance of adipose tissue-derived cytokines and their relative ratios in insulin resistance/sensitivity and β-cell function. ROC curves of adipose tissue-derived cytokines and their relative ratios based on insulin resistance/sensitivity and β-cell function. HOMA-IR, QUICKI, and HOMA-β were used as the predictor variables and the cut-off values were defined as 25% percentile, median and 75% percentile of the control group. Different cut-off values were used to analyze and the best diagnostic performance was presented. a diagnostic performances of resistin/ RBP4 ratio, resistin and RBP4 in β-cell function with HOMA-β of 65.6 (25% percentile) as cut-off value. b diagnostic performances of MCP-1/RBP4 ratio, MCP-1 and RBP4 in β-cell function with HOMA-β of 65.6 (25% percentile) as cut-off value. Diagnostic performances of Leptin/MCP-1, Leptin and MCP-1 in insulin resistance/sensitivity with HOMA-IR of 1.99 (median) as cut-off value (c) and QUICKI of 0.93 (median) as cut-off value (d) Toan et al. Diabetol Metab Syndr (2018) 10:41 may be considered as a promising indicator for β-cell function and insulin sensitivity.
SFRP5 serum levels are correlated with markers of obesity (e.g. BMI, waist-hip ratio, percentage of body fat), and T2DM (e.g. insulin resistance and disorders of lipid metabolism) [45]. Our study showed that the SFRP5 levels are significantly increased in control group with overweight compared to those without overweight and are correlated with levels of fasting glucose, LDL-C and HOMA-β. This observation supports a recent study showing the association between Sfrp5 gene expression and fat deposition in healthy adipose tissue [18]. Lower SFRP5 levels observed in patients with T2DM compared to controls in the current study is in accordance with a previous study showing the SFRP5 levels in Chinese patients [46]. However, these results are contradictory to a previous study showing that SFRP5 levels are elevated in Spanish patients with T2DM compared to prediabetic subjects and controls [19]. The association between SFRP5 levels and HOMA-β index clearly indicates the association of SFRP5 with β-cell function [17]. In addition, SFRP5 is sensitive to nutritional therapy suggesting the use of bioactive molecule as a biomarker for antiinflammatory effects of diet [47].
Corroborating with previous studies [48,49], our results also indicate that the RBP4 levels are significantly increased in patients with T2DM compared to controls. RBP4 has been shown to be associated with insulin resistance, visceral fat distribution, dyslipidemia and diabetes [22,50]. The strong correlations of RBP4 levels with the levels of fasting glucose and HbA1c and HOMA-β index as well as the correlation of relative ratios based on RBP4 levels with QUICKI and HOMA-β indexes suggest that RBP4 plays an important role in β-cell dysfunction and has a potential diagnostic value for T2DM. RBP4 serum levels have a potential to be an indicator of insulin resistance and decreasing the RBP4 serum levels may be considered as one of the strategies for anti-diabetic therapies of overweight-related T2DM.
In addition, a previous study has proposed a novel index based on the relative proportion of adiponectinto-resistin to predict insulin resistance [29]. The relative ratios of adiponectin/TNF-α, adiponectin/IL-1β, adiponectin/IL-10, TNF-α/IL-10 and IL-1β/IL-10 have been shown to be strongly correlated with the insulin resistance/sensitivity indexes (HOMA-IR and QUICKI) [24]. In this study, only the relative ratios of leptin/MCP-1 and MCP-1/RBP4 are correlated with resistance/sensitivity indexes. Nevertheless, the relative ratios of resistin/RBP4, visfatin/MCP-1 and MCP-1/RBP4 could be used as indicators for T2DM as these relative ratios are strongly correlated with the levels of fasting glucose, HbA1c and HOMA-β index. Furthermore, ROC curve analyses demonstrate the diagnostic potential of resistin/ RBP4 and MCP-1/RBP4 indexes for T2DM and β-cell function while leptin/MCP-1 index is an additional indicator for insulin resistance/sensitivity. However, a panel comprised of adipose tissue-derived cytokines should be systematically developed and optimized before using as a diagnostic indicator for screening and predicting the development of T2DM.
Although our data indicate a significant association of adipokine levels with overweight and T2DM, the study remains several limitations. It was designed as a crosssectional study, which cannot determine causal relationships between adipokine levels and the development of overweight and T2DM. Another limitation could be the limited number of data points of visfatin, SFRP5 and RBP4 levels in overweight controls. This may lead to an unsuccessful detection of the modulation of these adipokines in individuals with overweight and may weaken the findings that adipose tissue-derived cytokines play a vital role during the development of obesity and T2DM as well as reduce their diagnostic value for insulin resistance/sensitivity and β-cell function.

Conclusions
Our study showed that the levels of leptin, resistin, visfatin, SFRP5, MCP-1, and RBP4 are significantly modulated during the development of overweight and T2DM and are correlated with clinical parameters of overweight and T2DM. The relative ratios of studied adipokines (e.g. resistin/RBP4, leptin/MCP-1 and MCP-1/RBP4) are correlated with fasting glucose and HbA1c levels as well as with insulin resistance/sensitivity and β-cell function indexes. Adipose-derived cytokines play important roles in the pathogenesis of obesity and T2DM and may serve as a prognostic marker to predict overweight-related T2DM.