ASSOCIATION BETWEEN INSULIN-LIKE GROWTH FACTOR 1 AND BIOMARKERS OF ENDOTHELIAL DYSFUNCTION AND VASCULAR INFLAMMATION IN OBESE WOMEN

Given its possible protective effect on vascular endothelial function, we hypothesized that the circulating level of insulin-like growth factor 1 (IGF-1) could be associated with biomarkers of endothelial dysfunction and vascular inflammation, i.e., soluble intercellular adhesion molecule 1 (sICAM-1), soluble E-selectin (sE-selectin) and high-sensitivity C-reactive protein (hs-CRP, in non-diabetic obese women. A total of 40 obese women (median age 35 (31-39) years; moderately obese I (n=21) body mass index (BMI) 30.0-34.9 kg/m2 and severely obese (n=19) BMI 35.0-39.9 kg/m2) and 40 controls, age-matched healthy nonobese women were included in the study. Anthropometric parameters (BMI, waist circumference, fat adipose tissue percentage − FAT% measured by bioelectrical impedance analysis), serum concentrations of biomarkers (IGF-1, sICAM-1, sE-selectin and hs-CRP) and common carotid artery intima-media thickness Z-score (CCIMT Z-score) were assessed in all subjects. Severely obese subjects had a significantly lower median concentration of IGF-1 compared to moderately obese and nonobese subjects (95 (73.75-109.25) vs. 120 (108.3-141.7) vs. 115.5 (94.1-165.5) mg/mL, P<0.05). In moderately obese subjects, the median concentration of hs-CRP (3.41 (2.3-5.1) vs. 5.9 (3.2-9.57) mg/L, P=0.01) and sICAM-1 (244.5 (192.7-262.2) vs. 325.2 (229.2-339.6) mg/L, P=0.03) were significantly lower compared to severely obese subjects. In regression analysis IGF-1 concentration was independently associated with FAT% (β=-3.14; P=0.03), hs-CRP (β=-4.24; P=0.001) and CCIMT Z-score (β=-19.2; P=0.001). The circulating level of IGF-1 is associated with hs-CRP, but not with sE-selectin and sICAM-1 in non-diabetic obese women.


INTRODUCTION
Obesity is a chronic, multifactorial and complex disease (Tsigos et al., 2008) associated with accelerated atherosclerosis and cardiovascular diseases (CVD) (Nichols et al., 2012).It is known that endothelial dysfunction, a shift from a healthy endothelium to a damaged proinflammatory, pro-coagulative, and provasoconstrictive phenotype, represents an early event in the development of atherosclerosis (Lačković et al., 2011).The prolonged exposure of the vascular endothelium to classical and obesity-associated risk factors (insulin resistance, hyperglycemia, dyslipidemia) could further promote the deterioration of endothe-lial function and progression of atherosclerosis (Lobato et al., 2012).In particular, it has been shown that obesity-associated chronic low-grade inflammation, which primarily originates from dysfunctional or hypertrophic/hyperplastic adipose tissue, plays a crucial role in the development of obesity-related CVD (Hajer et al., 2008, Gregor andal., 2011).Changes in circulating levels of inflammatory mediators (cytokines and acute phase proteins) and adhesion molecules (selectins and cell adhesion molecules) may reflect alterations in endothelial status and, as such, could represent circulating biomarkers of endothelial dysfunction and vascular inflammation (Libby et al., 2009).In addition to endothelial dysfunction, gradual remodeling and thickening of the arterial wall could reflect the presence of early vascular changes associated with subclinical atherosclerotic disease (Kardassis et al., 2014).
There is substantial interest in the role of the insulin-like growth factor system as a possible atheroprotective factor (Bach et al., 2015).Insulin-like growth factor 1 (IGF-1) is an endocrine and autocrine/paracrine growth factor that directly (through its effects during development, cell growth and differentiation) (Clemmons et al., 2012) and indirectly (through metabolic actions on fat, protein and glucose metabolism) exerts its effect on the structure and function of the vascular system (Ungvari et al., 2012).There is considerable evidence from several studies suggesting that reduced IGF-1 could promote the progression of atherosclerosis (Higashi et al., 2014, De Smedt et al., 2011).Insulin resistance, glucose intolerance (Manrique et al., 2014), oxidized low-density lipoprotein and a sedentary life (Conti et al., 2004), as known risk factors for accelerated atherosclerosis, have been associated with low IGF-1 circulating levels in some cross-sectional studies.In addition, some studies suggest that IGF-1 significantly interacts with and acts on endothelial physiology (Hirai et al., 2011;Liu et al., 2014).Furthermore, several pathophysiological mechanisms of the effects of IGF-1 on endothelial cells activation have been suggested.IGF-1 may promote endothelial nitric oxide (NO) synthase activity and reduce the expression of proinflammatory cytokines and adhesion molecules from endothelial cells (Abbas et al., 2011).
Endothelial phenotype and vascular changes mostly depend on the balance between exposure to risk factors, and the capacity of endothelial repair (Deanfield et al., 2007).Hence, given its possible protective effect on vascular endothelial function, we hypothesized that the circulating concentration of IGF-1 could be associated with biomarkers of endothelial dysfunction and vascular inflammation (i.e.serum concentrations of soluble intercellular adhesion molecule 1 (sICAM-1), soluble E selectin (sE-selectin) and high-sensitivity C reactive protein (hs-CRP)) in non-diabetic obese women.

Subjects
The study was conducted according to the Declaration of Helsinki and approved by the Ethical Committees of CCV.Informed consent was obtained from every participant of the study.This cross-sectional study was performed at the Clinical Center of Vojvodina (CCV), from February 2014 to October 2014.The study comprised 40 obese female participants (body mass index (BMI) >30 kg/m 2 ) and 40 age-matched healthy nonobese (BMI 20-24.9kg/m 2 ) controls.Healthy obese middle-aged women were included in the study, and they were consecutively recruited from the Department of Endocrinology, Diabetes and Metabolic Disorders, CCV.All participants were without known acute or chronic diseases and without recent (i.e. in the last 3 months) changes in weight and physical activity.Exclusion criteria included inflammatory (hs-CRP >10 mg/l) (Pearson et al., 2003) or metabolic diseases, diabetes mellitus, hypertension, CVD, cerebrovascular diseases, known or suspected pituitary, thyroid, adrenal or gonadal dysfunction, pregnancy and medications (lipid lowering, growth hormone, anabolic steroids, glucocorticoids, hormone replacement, and hormonal contraception).Obese patients were classified into two subgroups according to the BMI value: subgroup I (Class 1) moderate obesity, BMI 30.0-34.9 kg/m 2 , and subgroup II (Class 2) severe obesity, BMI 35.0-39.9kg/m 2 .

Study protocol
All participants simultaneously underwent the following procedures on the same day: clinical and anthropometric measurement, blood sampling and oral glucose (75 g) tolerance test.Venous blood samples were drawn from the antecubital vein after 12-h overnight fast.Analyses were performed immediately after sampling, except for serum samples for laboratory determination of sICAM-1, sE-selectin and IGF-1, which were poured into aliquots and stored at-70ºC for no longer than one month.In order to measure carotid intima media thickness (CIMT), Doppler ultrasonography of the carotid blood vessels was performed within a month.

Anthropometric and clinical measurements
All anthropometric measurements were done with subjects wearing light indoor clothes and no shoes.Body weight (BW, kg) and fat adipose tissue percentage (FAT%) were determined by the bioelectrical impedance method (Tanita TBF-310 Body Composition Analyzer; Tanita Corporation, Tokyo, Japan); body height (BH) was measured using a Harpenden anthropometer (Holtain Ltd, Croswell, United Kingdom).Waist circumference (WC) was measured using a flexible metric tape with the precision of 0.1 cm in standing position at a level midway between the lowest point of the rib and the uppermost border of the iliac crest.Blood pressure was measured three times after at least 10 min rest in a sitting position by the Riva-Rocci method using a mercury sphygmomanometer.The average of the last two readings was used for analyses.

Laboratory assay
Fasting plasma (FPG) and 2-hour plasma glucose (2hPG)) were quantified by Dialab GOD PAP (glucose oxidase-phenol + aminophenazone method); the concentration of fasting plasma insulin (FPI) and 2-hour plasma insulin (2hPI)) was measured by Chemiluminescent Immunoassay method ((CMIA), ADVIA Centaur XP, Siemens).Insulin resistance was defined as homeostatic model assessment insulin resistance (HOMA-IR) ≥2.5 (Matthews et al., 1985).Plasma glycated hemoglobin (HbA 1c ) was measured by immunoinhibitory test; high-sensitivity CRP (hs-CRP) serum concentration was measured by the immunoturbidimetric method (commercial Beckman-Coulter kits-Ireland, Olympus AU 400).Circulating concentrations of sICAM-1 and sE-selectin were determined by the commercially available enzyme-linked immunosorbent assay (ELISA) kit (R&D Systems, USA).Results were expressed as ng/ml.The limit of detection of sE-selectin was 0.04 ng/mL and for sICAM-1 0.25 ng/mL.Serum IGF-1 concentration were measured by commercially available ELISA kit (Immunodiagnostic Systems Ltd, ids IGF ELISA) and expressed as mg/mL with a limit of detection of 16 mg/mL.The reference range for circulating IGF-1 was between 49 and 147 mg/mL (age category 31-40 years).Lipid profile was determined by standard biochemical methods (commercial Siemens kits, on Olympus AU 400).

Doppler ultrasonography of the carotid blood vessels
Ultrasonographic examination of the carotid arteries was performed by high-resolution ultrasound in B mode in real-time, using a linear probe of 12 MHz on the Aloka SSD-650 US system (Aloka, Tokyo).Measurements of the CIMT were done according to proposed standard procedure (Touboul et al., 2012).From the obtained values, median values were calculated and used to define the CIMT.Based on the equations (Engelen et al., 2013), we calculated the expected "normal" mean CIMT values for each female (mean CIMT (in µm) =321.7 + 4.971 × age, and SD CIMT(in µm) =54.50 + 0.8256 × age) and calculated age-and sex-specific CIMT Z-scores as (observed CIMT -expected CIMT)/SD x expected CIMT).CIMT Z-score was based on data for the healthy population, regarding age and sex.

Statistical analysis
The normal distribution of continuous variable was assessed with the Shapiro-Wilk test.Data were presented using descriptive statistical methods, continuous variables as mean (x ± s.d.), median and interquartile range (IQR), while categorical data were summarized as a number of observations divided with the total number of subjects within the groups (Simundic et al., 2012).In order to evaluate the effect of obesity on the analyzed variables, we divided the study group according to BMI value: subgroup I (Class 1) moderate obesity, if BMI is 30.0-34.9 kg/ m 2 , and subgroup II (Class 2) severe obesity, if BMI is 35.0-39.9kg/m 2 .The differences in variables were assessed by parametric Student T-test or nonparametric Mann-Whitney U test, as appropriate.Relations among variables were assessed using Pearson's or Spearman's correlation coefficient.All correlations were weighted with BMI.Multiple regression analysis was performed to estimate the independent association between IGF-1 serum concentration and the analyzed variables that met statistical significance in univariate analysis.A 2-tailed P value of less than 0.05 was considered statistically significant.All analyses were performed using the MedCalc 12.1.4.0 statistical software (MedCalc Software, Mariakerke, Belgium).

DISCUSSION
In the present study, we simultaneously examined the association between IGF-1 circulating level, parameters of adiposity and biomarkers of endothelial dysfunction and vascular inflammation in non-diabetic obese middle-aged women.We found significantly a lower IGF-1 circulating level in severely obese compared to moderately obese and nonobese women.Also, obese compared to nonobese subjects had a significantly higher circulating level of observed biomarkers of endothelial dysfunction and vascular inflammation.In addition to significantly higher hs-CRP and sICAM-1 levels in severely compared to moderately obese subjects, the difference in sEselectin level between the obese subgroups was insignificant.Additionally, besides fat adipose tissue percentage, the circulating concentration of IGF-1 was independently associated with hs-CRP.
Obesity could be associated with abnormalities in the growth hormone/IGF-1 axis (Berryman et al., 2013), and also with changes in IGF system binding proteins (Rasmussen, 2010, Puche et al., 2012).Our results are mostly similar with several previous findings of a low normal circulating concentration of IGF-1 in obese women (Galli et al., 2012;Hjortebjerg et al., 2012).Moreover, results of this study indicate that IGF-1 circulating levels decrease with the increase in the obesity class.Some studies have shown that an increase in BMI may induce a reduction in IGF-1 production due to impaired sensitivity of the grown hormone (GH) to stimulate IGF-1 synthesis (Gómez et al., 2003).Considering that BMI, as a practical measure of weight, cannot distinguish fat mass from lean, our results may even suggest a link between body fat mass and IGF-1 circulating concentration.In this study, fat adipose tissue percentage, as opposed to BMI, had an independent association with IGF-1 circulating concentration.
Regarding the soluble form of adhesion molecules and hs-CRP in this study, only hs-CRP was independently and significantly associated with IGF-1 serum level.Evidence from other studies shows that an increase in acute-phase proteins in the liver could be a rate-limiting factor in the production of both tissue and circulating concentrations of IGF-1 (Succurro et al., 2008).Furthermore, hepatic CRP synthesis stimulated by proinflammatory cytokines expressed in dysfunctional adipose tissue (Gregor et al., 2011) might be an important mediator of the negative association observed in the present study between IGF-1 and hs-CRP.Additionally, since IGF-1 may directly oppose the effects of CRP on proinflammatory endothelial activation (Liu et al., 2014), our results suggest that decreased levels of IGF-1 and increased levels of hs-CRP in severely obese women could promote vascular endothelial perturbations.The elevated levels of sE-selectin and sICAM-1 in the studied obese subjects may reflect proinflammatory endothelial activation, since the vascular endothelium covers 98% of all vasculature and it is the most important determinant of adhesion molecule soluble form levels in circulation (Damnjanovic et al., 2009;Scalia et al., 2013).Moreover, our results suggest that severely obese women have significantly higher levels of sICAM-1, but not sE-selectin compared to moderately obese women.A possible explanation for this result may be that, besides the endothelium, there is also evidence of circulating levels of sICAM-1 being partly secreted in adipocytes and macrophages within the adipose tissue (Hajer et al., 2008).Also, we did not observe differences in lipid parameters and HOMA-IR between moderately and severely obese women, except for the triglyceride level.In previous studies, the sE-selectin level was strongly associated with insulin resistance (Adamska et al., 2014, Pontiroli et al., 2004).Some studies even suggested that a low IGF-1 level results in insulin resistance and accelerated atherosclerosis (Clemmons et al., 2011).Insulin resistance, unlike the action of the IGF system, leads to a reduced bioavailability of NO in the endothelium (Tsukahara et al. 1994).Furthermore, it was observed in experimental studies that IGF-1 may directly oppose endothelial dysfunction by promoting insulin sensitivity, potassium-channel opening and by preventing postprandial dyslipidemia (Conti et al., 2004).Moreover, in some clinical studies, low IGF-1 levels were inversely associated with endothelial dysfunction (Empen et al., 2010) and insulin resistance (Succurro et al., 2008).In the present study, despite the fact we found significant and inverse correlations between the circulating concentrations of IGF-1, sEselectin, sICAM-1 and HOMA-IR in obese subjects, they were not independently associated.The reason for this discrepancy could be related to the difference in the methods used for endothelial dysfunction assessment.Furthermore, this study included only obese women, which may account for the difference between their results and ours.
Regarding the early vascular changes in the studied subject, the increase in thickness of the intima and media layers of the arterial wall above that of the healthy population mean, i.e. higher CCIMT Z-score, was significantly higher in obese women compared to the lean controls.Tesults of this study suggest that the IGF-1 circulating level was inversely associated with CCIMT Z-scores.Even though results from other studies suggest the significant associations between IGF-1 and CCIMT (Hirai et al., 2011;Marini et al., 2007), our study was the first to evaluate values of CCIMT Z-score.The importance of this new parameter is its capability to compare CCIMT values between individuals with different cardiovascular risk profiles (Engelen et al., 2013).Hence, our results suggest that the presence of arterial wall thickening in obese women is partly associated with IGF-1 circulating levels.Also, the results from several studies suggest that women with a metabolically healthy obese (MHO) phenotype, a possible transient obesity phenotype (Blüher, 2014), have low normal IGF-1 circulating levels, which has been associated with carotid atherosclerosis (Marini et al., 2007;Abd El-Hafez et al., 2014).However, the relation between IGF system and carotid atherosclerosis is complex (Higashi et al., 2012), mostly due to the presence of other risk that favor atherosclerosis development (van den Oord et al., 2013, Succurro et al., 2010), and further investigation is needed.
The present study has some potential limitations, primarily the small number of included subjects.Although the cross-sectional design of the study does not allow us to make conclusions regarding the pathophysiological mechanisms underlying associations of IGF-1, hs-CRP and CCIMT Z-score, these findings could support the associations of IGF-1 and subclinical vascular changes in non-diabetic obese middle age women.Moreover, IGF-1 binding proteins and free IGF-1 levels were not measured, and this should be considered another limitation of the study.

CONCLUSIONS
The circulating level of IGF-1 was inversely associated with hs-CRP in non-diabetic obese women, appointing to the potential role of IGF-1 in vascular inflammatory perturbation.Additional longitudinal studies are necessary to determine if the treatment of obesity has any beneficial effects on circulating IGF-1 level and vascular endothelial function over time.

Table 2 .
Comparison of anthropometrical characteristics, metabolic parameters and circulating biomarkers between moderately and severely obese subjects

Table 3 .
Correlations between serum IGF-1 (mg/mL) and analyzed parameters in all study subjects.