Funding agencies: This research was supported by Basic Science Research Program through the National Research Foundation of Korea (NRF), funded by the Ministry of Education, Science and Technology (2010-0021599) and a grant from the Asan Institute for Life Sciences (2012-106) Seoul, Korea.
Objective: Secreted protein acidic and rich in cysteine (SPARC) is expressed in most tissues and is also secreted by adipocytes. The associations of SPARC mRNA expression in visceral adipose tissue (VAT), subcutaneous abdominal adipose tissue (SAT), serum SPARC concentration, and metabolic parameters in Korean women are investigated.
Design and Methods: This is a cross-sectional study. Fifty-eight women were recruited, of whom 15 women who underwent bariatric surgery for morbid obesity (BMI mean ± SD: 40.2±5.7 kg/m2), 16 who underwent metabolic surgery for type 2 diabetes (BMI: 28.9±4.5 kg/m2), and, as a control group, 27 who underwent gynecological surgery (BMI: 22.7±2.4 kg/m2). Anthropometric variables, metabolic parameters, SPARC mRNA expression in adipose tissue, and serum SPARC concentration were measured.
Results: In all subjects, SPARC mRNA expression was significantly higher in SAT than in VAT. Serum SPARC concentrations (mean ± SE) in morbidly obese subjects, subjects with type 2 diabetes, and normal weight subjects were 267.3±40.2 ng/mL, 130.4±33.0 ng/mL, and 53.1±2.8 ng/mL, respectively. SPARC mRNA in SAT was significantly correlated with BMI, whereas SPARC mRNA in VAT was significantly correlated with BMI and VAT area. Serum SPARC concentration was significantly correlated with BMI, waist circumference, total adipose tissue area, and SAT area. After BMI adjustment, serum SPARC concentration was significantly correlated with fasting insulin concentration and HOMA-IR score. Multivariate regression analysis showed that BMI and HOMA-IR were independently associated with serum SPARC concentration.
Conclusions: Serum SPARC concentration is significantly correlated with obesity indices and might be influenced by insulin resistance. These findings suggest that SPARC may contribute to the metabolic dysregulation associated with obesity in humans.
Secreted protein acidic and rich in cysteine (SPARC), also known as osteonectin or BM-40, was initially found to be secreted from bone . Later, however, a variety of cell types, including osteoblasts, macrophages, fibroblasts, smooth muscle cells, and endothelial cells, were reported to express SPARC mRNA [2, 3]. SPARC is involved in osteogenesis, angiogenesis, wound healing, inflammation, tumorigenesis, and the pathogenesis of fibrosis [4-7].
Although several tissues secrete SPARC, adipose tissue is likely the major contributor to elevated SPARC observed in obesity [8, 9]. For example, SPARC mRNA expression in adipose tissues was upregulated in ob/ob mice  and db/db mice . The correlation between SPARC mRNA expression and adipose tissue hyperplasia and adipogenesis suggests that SPARC is involved in the growth and differentiation of adipose tissue . Circulating SPARC concentrations were found to positively correlate with BMI in humans , suggesting that the secretion of SPARC from adipose tissue may account for the majority of circulating SPARC. In addition, plasma SPARC concentrations were significantly elevated in age- and BMI-matched subjects with coronary artery disease .
The profibrotic qualities of SPARC have been found to reduce the ability of fat expansion  and fibrosis in adipose tissue impairs metabolic function . Preadipocytes secrete SPARC during differentiation, but little is known about the contribution of stromal cells to the maintenance of the extracellular matrix during adipose tissue fibrosis. Since subcutaneous adipose tissue (SAT) is more fibrotic in individuals with severe obesity than that in lean controls , SPARC may contribute to the pathogenesis of obesity-associated diseases, including insulin resistance . Indeed, SPARC mRNA expression in adipose tissue was influenced by fat mass, leptin, insulin and glucose , and plasma concentrations of SPARC were elevated in patients newly diagnosed with type 2 diabetes mellitus .
Although SPARC is a newly identified autocrine and/or paracrine factor that could affect key functions of adipose tissue , few studies have examined the pathophysiological role of SPARC protein in human obesity and metabolic dysfunction. Furthermore, little is known about the correlations of serum SPARC concentration and SPARC mRNA expression in adipose tissue with abdominal fat distribution. We therefore measured serum SPARC concentration, depot-specific expression of SPARC mRNA, and metabolic parameters in Korean women with subcutaneous and visceral fat types of obesity, as well as investigating the associations among SPARC concentration, expression in adipose tissue, and obesity-related metabolic diseases.
We recruited 15 subjects who underwent bariatric surgery for morbid obesity (typically body weight > 100 kg; predominantly subcutaneous fat type obesity), and 16 who underwent metabolic surgery for type 2 diabetes mellitus (DM) (predominantly visceral fat type obesity) in the Department of General Surgery at Inha University, Incheon, Korea. As a control group, we enrolled 27 subjects of normal weight who underwent elective gynecological surgery for benign diseases (uterine myoma, adenomyosis, endometriosis, or cystadenoma) in the Department of Gynecology at Asan Medical Center, Seoul, Korea. Women with secondary causes of obesity, those who were pregnant or lactating, and those with evidence of malignancy or severe hepatic or renal diseases were excluded. The study protocol was approved by the institutional review boards of Asan Medical Center and Inha University, and all participants provided written informed consent upon enrollment. We certify that all applicable institutional regulations regarding the ethical use of human volunteers were followed during this study.
Blood pressure and anthropometric measurements
Blood pressure was measured in the morning. Each patient was seated in a quiet room for 10 min, after which blood pressure was measured using a standard mercury sphygmomanometer on the patient's arm. A day before surgery, anthropometric measurements were taken while the subjects were dressed in light clothing, but without shoes. Height to the nearest 0.1 cm and weight to the nearest 0.1 kg were measured using an automatic height–weight scale. Body mass index (BMI) was calculated as weight (in kilograms) divided by the square of the height (in meters). Waist circumference (WC) was measured at the midpoint between the lower border of the ribcage and the iliac crest.
Estimation of abdominal fat distribution
The distribution of abdominal fat in each study subject was assessed by computed tomography (CT) using a Siemens Somatom Scanner (Erlangen, Germany), as described . Each subject was placed in a supine position, and a cross-sectional scan of 10 mm thickness, centered at the L4-L5 vertebral disc space, was obtained using a skeletal radiograph to establish the position of the scans to the nearest millimeter. The area of total adipose tissue (TAT) in the abdomen was measured by computation using an attenuation range of −190 to 30 Hounsfield units (Syngo, Siemens, Erlangen, Germany). The area of visceral adipose tissue (VAT) was measured within the muscle wall surrounding the abdominal cavity, and the area of SAT was calculated by subtracting the VAT area from the TAT area. In addition, the ratio of VAT area to SAT area (VSR) was calculated.
Measurements of metabolic variables and serum SPARC concentration
Blood samples were obtained in the morning after a 12-hour overnight fast. In patients who had taken medications for glucose or lipid control (e.g. hypoglycemic agents, insulin, or hypolipidemic agents), blood samples were obtained after a 3-day cessation of medications. Lipid profiles were measured by enzymatic procedures using an autoanalyzer (Hitachi-747, Tokyo, Japan). Plasma glucose concentrations were measured by the glucose oxidase method, hemoglobin A1c (HbA1c) by HPLC (Tosoh Bioscience, Tokyo, Japan), and insulin by human insulin radioimmunoassay (TFB, Tokyo, Japan). Insulin resistance was determined by calculating the homeostasis model assessment for insulin resistance (HOMA-IR) score, using the formula: fasting insulin (μU/mL) × fasting plasma glucose (mg/dL)/405 . Concentration of serum SPARC (Adipo Bioscience, CA) was measured using enzyme-linked immunosorbent assay kits, according to the manufacturer's protocol. The assay sensitivity was 0.64-1.2 ng/ml, and the intra- and interassay coefficients of variance were 4-6 and 8-10%, respectively.
Sampling of adipose tissue and fractionation of adipocytes and stroma/vascular (SV) cells
Samples of VAT and SAT were removed during surgery. SAT was obtained from the site of surgical incision, and VAT was removed from the distal portion of the greater omentum (i.e., the epiploon). The samples were immediately immersed in ice-cold 0.9% saline, carried to the laboratory, frozen in liquid nitrogen, and stored at –80°C. Portions of fresh VAT samples from 10 subjects were separated into adipocyte and stroma/vascular (SV) cell fractions. Each sample was washed with Krebs Ringer Henseleit (KRH) buffer to remove blood and digested with 1 mg/mL collagenase (Worthington, Freehold, NJ) in KRH buffer containing 1% BSA for 40-60 minutes at 37°C. The collagenase digest was separated from undigested tissues by filtration through 100 μm nylon mesh (Falcon, Franklin Lakes, NJ). The floating adipocyte fraction was collected and washed three times with KRH buffer. Non-floating cells isolated from the collagenase digest were centrifuged for 15 min at 1000 rpm, and the pellet (i.e., the SV cell fraction) was collected. The adipocyte and SV cell fractions were snap-frozen in liquid nitrogen and stored at –80°C for later RNA extraction.
Measurement of depot-specific expression of SPARC
We measured depot-specific expression of SPARC using real-time quantitative RT-PCR. Total RNA was extracted from adipose tissue samples using TRIzol (Invitrogen Carlsbad, CA, 1 mL/100 mg tissue), according to the manufacturer's instructions. The purity of the extracted RNA was assessed using a NanoDrop spectrophotometer. The RNA was reverse transcribed into cDNA using Superscript III reverse transcriptase (Invitrogen) and oligo-dT primers. Amplified mRNA was quantified using a Roche Light Cycler system (Roche Molecular Biochemicals, Mannheim, Germany). Each reverse transcriptase reaction was amplified in a 25-μL PCR mixture using SYBR Green QPCR master mix (Bio-Rad, Hercules, CA). The primers for human SPARC mRNA were 5'-acatcgggccttgcaaatac-3' (forward) and 5'-cagtcagaaggttgatgtcctcat-3' (reverse); and the primers for human beta-actin were 5'-gacggggtcacccacac-3' (forward) and 5'-gtggtggtgaagctgtagcc-3' (reverse). Expression of human SPARC and beta-actin mRNA was quantified by the second derivative maximum method, which determines the crossing points of individual samples using an algorithm that identifies the first turning point of the fluorescence curve. SPARC mRNA expression was calculated relative to the expression of beta-actin using the ΔΔCT method, normalizing the Ct values of the SPARC mRNA to the Ct values of β-actin relative to a control sample in arbitrary units. Amplification of specific transcripts was confirmed via melting curve profiles.
Data are presented as means ± SEs. The ANOVA test was used to compare anthropometric variables, metabolic parameters, and serum SPARC concentration among the three groups (morbid obesity, type 2 diabetes, and normal weight). SPARC mRNA expression in SAT and VAT of each group was compared by paired t-test. Spearman's test was employed to assess the correlations among SPARC mRNA expression, serum SPARC concentration, and metabolic parameters. Partial correlation analysis was used to examine their correlations adjusted with BMI. Multiple linear regression analysis was used to assess the relationship between serum SPARC concentration as a dependent variable and anthropometric and metabolic parameters as independent variables. Skewed data were log transformed and analyzed parametrically, or else nonparametric methods were used. For all tests, a P-value < 0.05 was considered statistically significant. All statistical analyses were performed using SPSS (version 19.0 for Windows; SPSS, Chicago, IL).
Clinical characteristics of the study subjects
The basic characteristics of the study subjects, including their anthropometric variables, CT measurements of abdominal fat, and metabolic parameters, are shown in Table 1. Morbidly obese patients showed the highest BMI (mean±SD: 40.18±5.66 kg/m2), WC, TAT area, SAT area, systolic and diastolic blood pressure, and fasting insulin levels, although they were younger than the other two groups. Type 2 DM patients showed BMI of 28.85±4.46 kg/m2 and the highest VSR, fasting plasma glucose, and triglyceride concentrations. Normal weight group showed the lowest BMI (22.74±2.41 kg/m2), WC, VAT area, and HOMA-IR. Serum SPARC concentration (mean±SE) was higher in morbidly obese (267.25±40.2 ng/mL) than in type 2 DM (126.51±32.0 ng/mL) patients and was higher in both than in the normal weight group (53.12±2.82 ng/mL).
Table 1. Baseline characteristics of the study subjects
Type 2 DM
(n = 15)
(n = 16)
(n = 27)
Mean ± SD
Mean ± SD
Mean ± SD
P value by ANOVA test among the three groups.
aP < 0.05 vs. normal weight;
bP < 0.05 vs. morbidly obese (by multiple comparisons)
Abbreviations: SD, standard deviation; SE, standard error; BMI, body mass index; WC, waist circumference; SBP, systolic blood pressure; DBP, diastolic blood pressure; FPG, fasting plasma glucose; TAT, total adipose tissue; VAT, visceral adipose tissue; SAT, subcutaneous adipose tissue; VSR, visceral/subcutaneous adipose tissue ratio; HOMA-IR, homeostasis model assessment for insulin resistance; SPARC, secreted protein, acidic and rich in cysteine.
SPARC mRNA expression in adipose tissues and serum SPARC concentration
We found that SPARC mRNA expression was significantly greater in adipocytes than in SV cells (Figure 1). Moreover, in each of the three groups, SPARC mRNA expression was significantly higher in SAT than in VAT (Figure 2). SPARC mRNA expression in SAT and VAT was significantly correlated (r = 0.438, P = 0.001). In addition, serum SPARC concentration was significantly correlated with SPARC mRNA expression in SAT (r = 0.269, P = 0.043) and in VAT (r = 0.323, P = 0.014) (Table 2).
Table 2. Correlations between SPARC mRNA expression in adipose tissue, serum SPARC concentration, anthropometric variables, and metabolic parameters of the study subjects
SPARC mRNA expression in SAT
SPARC mRNA expression in VAT
Serum SPARC concentration
aP < 0.005;
bP < 0.05 by Spearman correlation analysis or partial correlation analysis.
Abbreviations: BMI, body mass index; WC, waist circumference; SBP, systolic blood pressure; DBP, diastolic blood pressure; FPG, fasting plasma glucose; TAT, total adipose tissue; VAT, visceral adipose tissue; SAT, subcutaneous adipose tissue; VSR, visceral/subcutaneous adipose tissue ratio; HOMA-IR, homeostasis model assessment for insulin resistance; SPARC, secreted protein, acidic and rich in cysteine.
Correlations among SPARC mRNA expression in adipose tissue, serum SPARC concentration, and metabolic parameters
Table 2 shows the correlations among SPARC mRNA expression in adipose tissues, serum SPARC concentration, and metabolic parameters in the study subjects. SPARC mRNA expression in SAT was significantly correlated with BMI, whereas SPARC mRNA expression in VAT was significantly correlated with BMI, VAT area, triglyceride concentration, fasting insulin concentration, and HOMA-IR score. However, after BMI adjustment, their correlations with metabolic parameters were disappeared. Serum SPARC concentration showed significant positive correlations with BMI, WC, TAT area, SAT area, systolic blood pressure, triglyceride, fasting insulin concentration, and HOMA-IR score. After BMI adjustment, serum SPARC concentration was significantly correlated with fasting insulin concentration and HOMA-IR score.
Multiple linear regression analyses relating serum SPARC concentration with anthropometric and metabolic parameters
Table 3 shows multiple linear regression analyses using serum SPARC concentration as a dependent variable and age, BMI, HbA1c, and HOMA-IR as independent variables. We found that BMI and HOMA-IR were independent predictors of serum SPARC concentration.
Table 3. Multiple linear regression analyses using serum SPARC concentration as a dependent variable and anthropometric and metabolic parameters as independent variables
P value by multiple linear regression analyses (n = 58).
Abbreviations: BMI, body mass index; HOMA-IR, homeostasis model assessment for insulin resistance; SPARC, secreted protein, acidic and rich in cysteine.
We have demonstrated that SPARC mRNA expression in adipose tissues and serum SPARC concentrations were associated with some anthropometric and metabolic parameters. Serum SPARC concentration was highly correlated with obesity indices and abdominal fat distribution (BMI, WC, TAT area, and SAT area) and metabolic parameters (fasting insulin concentrations and HOMA-IR score) in this group of individuals. Consistent with our results, a previous human study reported that plasma concentration of SPARC was significantly correlated with BMI and was closely correlated with abdominal SAT .
Plasma concentrations of SPARC were found to be higher in patients newly diagnosed with type 2 diabetes than those in individuals with normal glucose tolerance . Moreover, plasma SPARC concentrations were reported to be elevated in patients with coronary artery disease . Our study demonstrated that serum SPARC concentration was highly correlated with components of the metabolic syndrome. Additionally, we found that HOMA-IR, an indicator of insulin resistance, was independently associated with serum SPARC concentration. The mechanisms involving SPARC in the pathogenesis of insulin resistance remain poorly understood. However, some data suggest that SPARC may be involved in inflammatory process, causing insulin resistance [19, 20]. A previous study demonstrated that high-sensitivity C-reactive protein correlated significantly with SPARC expression both in VAT and SAT. Moreover, macrophage migration inhibitory factor, a proinflammatory adipokine, had a positive correlation with SPARC expression in VAT .
The associations of fasting plasma glucose concentration with SPARC mRNA expression and serum SPARC concentration are inconsistent. Increased glucose concentration decreased SPARC mRNA expression in cultures of visceral fat explants . However, SPARC mRNA expression in VAT but not in SAT positively correlated with fasting glucose in vivo . We did not observe any significant correlation between SPARC mRNA expression or serum SPARC concentration and fasting plasma glucose, in agreement with a recent human study . One of the possible reasons for the lack of association with plasma glucose concentration might be that diabetic patients included in our study were under treatment for glucose control. These differences may be related to the stage of glucose tolerance or insulin resistance in study subjects. A recent study reported that HbA1c, chronic glycemic indicator, was more sensitive marker than fasting glucose for the association with circulating SPARC levels . However, we did not find any association between SPARC expression in adipose tissue or serum SPARC concentration with the HbA1c level.
We found that SPARC expression was greater in adipocytes than in SV cells, in agreement with a previous study . In addition, similar to a recent study , we found that SPARC mRNA expression was higher in SAT than in VAT in all three groups. In human SAT, adipocytes are twice as numerous as SV cells ; in comparison to SAT, VAT compared with SAT is more vascular, innervated, and contains a larger number of inflammatory and immune cells and lesser preadipocyte differentiating capacity . Thus, adipocyte-dominant expression of SPARC is consistent with our finding that SPARC expression was higher in SAT than in VAT. Our results comparing fat tissue and serum SPARC seemed likely that the SPARC production from adipose tissue, particularly adipocytes, will cause the surplus of SPARC in obese and possibly type 2 diabetic subjects. It could be better if we sampled adipose tissue cumulatively from the whole fat rather than the biopsy of a small piece of adipose tissue. When we assessed correlations with depot-specific SPARC mRNA expression, we found that SPARC mRNA expression in VAT was significantly correlated with BMI and VAT area. However, SPARC expression in SAT was only significantly correlated with BMI, but not with SAT area. Some study showed that culture of VAT with insulin dose-dependently increased SPARC mRNA expression .
This study had some limitations. First, the numbers of study subjects were small, and especially there was a lack of subjects with intermediate range BMI levels; therefore, our study might not be powerful enough to get strongly positive outcomes. Second, sampling sites of SAT were different between subjects undergoing bariatric or metabolic surgery and those undergoing gynecological surgery. SAT of morbidly obese subjects and type 2 diabetic subjects (n = 31) was collected around umbilical area, and that of control subjects was collected in lower abdomen (n = 27). Third, subjects in our study were limited to Korean women; therefore, our findings may not be directly applicable to men or to other ethnic groups. However, a previous study reported that they did not find a significant difference in SPARC expression of adipose tissue between BMI-matched men and women . Fourth, although SPARC has profibrotic qualities, we could not measure the degree of fibrosis of adipose tissue. Therefore, we did not assess the associations of SPARC mRNA expression or concentration with fibrosis of SAT and metabolic derangement in the study subjects.
Despite these limitations, we found that serum SPARC concentration, rather than SPARC mRNA expression in adipose tissue, was highly correlated with obesity indices and insulin resistance and that BMI and HOMA-IR were independently associated with serum SPARC concentration. These results suggest that SPARC may play a pathophysiological role in human obesity and metabolic diseases.