Introduction

The importance of adipose tissue in maintaining metabolic equilibrium is well known from studies in both animal models and humans. Triglyceride synthesis is a determining factor for adipose tissues mass. The fatty acids (FA) from adipose tissue are primarily an important energy substrate for a number of organs. In addition, triglyceride synthesis in cells may protect them from the potentially toxic effects of excess FA. These FA alter insulin action and glucose metabolism, as well as altering gene function like transcription factors and regulating insulin production [1, 2]. The efficient energy store is influenced by the supply of fatty acids–coenzyme A (CoA), of glycerol phosphate precursors, or by the activity of enzymes at later stages of triglyceride synthesis. Availability of glycerol phosphate is a determinant of triglyceride synthesis and lipolysis. Another key enzyme in triglyceride synthesis is acyl-coenzyme A/cholesterol acyltransferase (DGAT1), which catalyzes the final step of lipogenesis. Inactivation of DGAT1 results in lean mice [3]. Although lipogenesis accounts for only a small fraction of total triglyceride flux, its importance in the determination of adipose tissue mass has not been resolved [4].

The regulation of lipolysis is also essential to ensure an adequate supply of lipids to tissues utilizing FA. During intracellular lipolysis, adipose triglyceride lipase (ATGL) is rate limiting for the first step in triglyceride hydrolysis, generating diglycerides (DG) and FA [5]. Hormone-sensitive lipase (HSL) hydrolyzes both triglycerides and DG substrate. Perilipin is essential for the recruitment of HSL to lipid droplets and full enzyme activation [6]. The function of perilipins is to prevent lipolysis in basal conditions, favoring lipid deposition. Half of the hydrolyzed FA are recycled into triglyceride synthesis [7]. The combination of an excess calorie intake, physical inactivity, and genetic factors contributes to an alteration of the function of normal adipose tissue as an energy store and lipid buffer. Lipolysis in normal subjects is extremely sensitive to the action of insulin. However, hormone regulation of lipolysis in obesity is under debate. Regarding insulin, increased, normal, or even decreased antilipolytic effects have been reported [8].

Peroxisome proliferator-activated receptor-γ (PPARγ) are ligand-activated transcription factors strongly expressed in adipocytes, and they play a critical role in glucose and lipid homeostasis [9]. The PPARγ regulate the transcription of genes involved in lipid synthesis and facilitate the entrance of free fatty acids (FFA) into adipocytes, activate triglyceride synthesis, and stimulate the esterification of FFA by acyl-CoA synthase. They also seem to stimulate the glycerol cycle [10] and to regulate the transcription of genes involved in lipolysis [11]. Several studies have helped in understanding the function of PPARγ [12].

Morbid obesity represents an extreme phenotype involving a very important increase of adipose tissue. Most morbidly obese patients have alterations in lipid and carbohydrate metabolism [13–15], which is regulated by PPARγ [16]. To better understand the mechanisms underlying the adipose tissue mass, the aim of this study was twofold: firstly, to examine in adipose tissue from morbidly obese patients with low and with high insulin resistance the expression of PPARγ and genes involved in lipogenesis and lipolysis and secondly, to explore the expression of these genes in visceral adipose tissue (VAT) and subcutaneous adipose tissue (SAT). To date, only a few studies have been undertaken simultaneously in the two tissues. We hypothesized that the expression of lipogenic and lipolytic genes varies according to the degree of insulin resistance and the particular adipose tissue studied.

Materials and Methods

Subjects

The study included 23 morbidly obese patients (BMI 57.4 ± 5.2 kg/m2), 13 with low insulin resistance (homeostasis model assessment of insulin resistance index (HOMA-IR) <4.7) and ten with high insulin resistance (HOMA-IR >8), and 15 healthy, lean persons (BMI 22.7 ± 1.5 kg/m2). The cutoff point for the HOMA-IR was taken from previous studies carried out in a healthy population with no carbohydrate metabolism disorders (HOMA-IR <4.7) [17]. Patients were excluded if they had cardiovascular disease, arthritis, acute inflammatory disease, infectious disease, or were receiving drugs that could alter the lipid profile or the metabolic parameters at the time of inclusion in the study. None of the morbidly obese patients was being treated with insulin, oral antidiabetic agents, or diet. The weight of all the persons had been stable for at least 1 month, and none had renal involvement. All participants gave their informed consent and the study was reviewed and approved by the Ethics and Research Committee of Virgen de la Victoria Clinical University Hospital, Malaga, Spain.

Laboratory Measurements

Blood samples were collected after a 12-h fast. The serum was separated and immediately frozen at −80°C. Serum biochemical parameters were measured in duplicate. Serum glucose, cholesterol, high-density lipoprotein (HDL) cholesterol, triglycerides (Randox Laboratories Ltd., Antrium, UK) and FFA (WAKO Chemicals, Richmond, VA, USA) were measured by standard enzymatic methods. Low-density lipoprotein cholesterol was calculated by the Friedewald formula. The insulin was analyzed by an immunoradiometric assay (BioSource International, Camarillo, CA, USA). Leptin was analyzed by enzyme immunoassay (ELISA) kits (DSL, Webster, TX, USA). Adiponectin was analyzed by enzyme immunoassay (ELISA) kits (DRG Diagnostics GmbH, Germany). The HOMA-IR was calculated from fasting insulin and glucose with the following equation: HOMA-IR = fasting insulin (μIU/mL) × fasting glucose (mmol/L)/22.5.

Visceral and Subcutaneous Adipose Tissue mRNA

VAT and abdominal SAT were obtained during bariatric surgery in the morbidly obese patients or during elective abdominal surgery procedures in lean subjects [18]. The biopsy samples were washed in physiological saline and immediately frozen in liquid nitrogen. Biopsy samples were maintained at −80°C until analysis. Total RNA isolation from adipose tissues was obtained using RNeasy Lipid Tissue Mini Kit (Qiagen GmbH, Germany). The purity of the RNA was determined by the absorbance260/absorbance280 ratio on the Nanodrop. The integrity of total purified RNA was checked by denaturing agarose gel electrophoresis and ethidium bromide staining. Quantitative real-time PCR was performed in a Mastercycler gradient (Eppendorf AG, Hamburg, Germany). The cDNAs obtained were used as the template for quantitative real-time PCR by a Light Cycler Instrument (Roche Diagnostics, Basel, Switzerland), using the SYBR green fluorescent method. Primers for the PCR reaction were designed based on NCBI database sequences of the human genes PPARγ, DGAT1, phosphoenolpyruvate carboxykinase (PEPCK), glycerol kinase (GK), aquaporin 7 (AQP7), acetyl-coenzyme A carboxylase alpha (ACC1), acyl-CoA synthetase short-chain family member 2 (ACSS2), fatty acid binding protein 4 (FABP4), ATGL, HSL, phosphodiesterase type 3 (PDE3), and perilipin. Adipose tissue expression levels for each gene were normalized using β-actin. Statistical analysis showed no differences in gene expression for this housekeeping gene in the groups studied.

Statistical Analysis

The statistical analysis was done with SPSS (Version 11.5 for Windows; SPSS, Chicago, IL, USA). Comparison between the results of the controls and morbidly obese patients was made with the Mann–Whitney test. Comparison between the results of the different groups of morbidly obese patients and the controls was made with the Kruskal–Wallis test. Comparison between the results of the different adipose tissues was made with the Wilcoxon test. The Spearman correlation coefficients were calculated to estimate the correlations between variables. Values were considered to be statistically significant when the p ≤ 0.05. The results are given as the mean ± SD.

Results

Anthropometric and Biochemical Characteristics

Table 1 summarizes the characteristics of the controls and the morbidly obese patients, with low and high insulin resistance. The weight and the anthropometric data were, as expected, significantly greater in the morbidly obese patients, though no significant differences were seen between the two groups of obese patients.

Table 1 Anthropometric and biochemical variables in the controls and the morbidly obese patients, classified according to their HOMA-IR
Full size table

VAT mRNA Expression

Compared with the morbidly obese patients, the mRNA expression in the VAT from the controls was significantly lower: PPARγ (1.12 ± 0.65 vs. 2.57 ± 1.52, p = 0.001), GK (0.98 ± 0.45 vs. 2.39 ± 2.36, p = 0.010), AQP7 (1.35 ± 0.65 vs. 3.10 ± 1.34, p < 0.001), DGAT1 (1.06 ± 0.37 vs. 2.00 ± 0.86, p < 0.001), ATGL (1.33 ± 0.61 vs. 5.54 ± 2.71, p < 0.001), HSL (1.58 ± 1.12 vs. 4.48 ± 3.25, p < 0.001), and perilipin (1.15 ± 0.54 vs. 1.83 ± 0.82, p = 0.006). The mRNA expression of ACC1 (1.80 ± 0.52 vs. 1.08 ± 0.78, p = 0.004) and PEPCK (3.22 ± 2.67 vs. 1.93 ± 1.16, p = 0.049), however, was significantly greater in the controls than the morbidly obese patients.

When the morbidly obese patients were classified according to their level of insulin resistance, the differences remained significant as compared with the controls for the expression of PPARγ, DGAT1, ACC1, AQP7, ATGL, and HSL (Figs. 1 and 2). The expression of PEPCK in the controls was only significantly higher than those morbidly obese patients with high insulin resistance (Fig. 1). The expression of GK in the controls was only significantly lower than those morbidly obese patients with low insulin resistance (Fig. 1). The expression of perilipin in the morbidly obese patients with low insulin resistance was significantly higher than those morbidly obese patients with high insulin resistance and the controls (Fig. 2).

Fig. 1
figure 1

mRNA expression levels of lipogenic genes in subcutaneous (black bar) and visceral adipose tissue (white bar) in the controls (n = 15) and the morbidly obese patients, classified according to their insulin resistance. MO low IR morbidly obese persons with low insulin resistance (n = 13), MO high IR morbidly obese persons with high insulin resistance (n = 10). Adipose tissue expression levels for each gene were normalized using β-actin. The results are given as the mean ± SD. Different letters indicate significant differences between the means of the three groups in VAT or in SAT (p < 0.05). *p < 0.05; † p < 0.005; ‡ p < 0.001: significant differences within the same group of patients between subcutaneous and visceral adipose tissues

Full size image
Fig. 2
figure 2

mRNA expression levels of lipolytic genes in subcutaneous (black bar) and visceral adipose tissue (white bar) in the controls (n = 15) and the morbidly obese patients, classified according to their insulin resistance. MO low IR morbidly obese persons with low insulin resistance (n = 13), MO high IR morbidly obese persons with high insulin resistance (n = 10). Adipose tissue expression levels for each gene were normalized using β-actin. The results are given as the mean ± SD. Different letters indicate significant differences between the means of the three groups in VAT or in SAT (p < 0.05). *p < 0.05; † p < 0.005; ‡ p < 0.001: significant differences within the same group of patients between subcutaneous and visceral adipose tissues

Full size image

SAT mRNA Expression

Compared with the morbidly obese patients, the mRNA expression in the SAT from the controls was significantly lower: PPARγ (1.65 ± 0.51 vs. 2.50 ± 1.50, p = 0.022), AQP7 (1.87 ± 0.57 vs. 2.92 ± 1.36, p = 0.003), DGAT1 (1.89 ± 0.53 vs. 2.81 ± 0.76, p < 0.001), and ATGL (2.18 ± 0.59 vs. 5.47 ± 2.76, p < 0.001). The mRNA expression of ACC1 (2.52 ± 1.00 vs. 1.09 ± 0.63, p < 0.001), ACSS2 (1.65 ± 0.61 vs. 1.07 ± 0.68, p = 0.012), PEPCK (3.88 ± 4.02 vs. 1.88 ± 1.19, p = 0.034), FABP4 (1.96 ± 0.80 vs. 1.14 ± 0.73, p = 0.003), and PDE3 (2.14 ± 0.61 vs. 1.57 ± 0.86, p = 0.005) was significantly greater in the controls than the morbidly obese patients.

When the morbidly obese patients were classified according to their level of insulin resistance, the differences only remained significant as compared with the controls for the expression of PPARγ, DGAT1, FABP4, ACC1, and ATGL (Figs. 1 and 2). The expression of ACSS2, PEPCK, and PDE3 in the controls was only significantly higher than those morbidly obese patients with high insulin resistance (Figs. 1 and 2), and the expression of AQP7 in the controls was only significantly higher than those morbidly obese patients with low insulin resistance (Fig. 1).

Differences Between VAT and SAT mRNA Expression

Comparison in the controls of the different expression levels in the two adipose tissues showed that the expression of PPARγ, DGAT1, AQP7, ACC1, ACSS2, FABP4, GK, ATGL, HSL, PDE3, and perilipin was significantly higher in the SAT than the VAT (Figs. 1 and 2). In the morbidly obese patients, the expression of GK in the SAT was lower than in the VAT only in those with high insulin resistance (Fig. 1). No significant differences between the VAT and the SAT were seen in the expression of the other genes studied in the morbidly obese patients.

Correlations of the mRNA Gene Expression

A positive correlation was found between the expression of a group of genes (PPARγ, DGAT1, AQP7, GK, ATGL, HSL, and perilipin) and the BMI, insulin, and HOMA-IR in all the persons studied, in SAT (Table 2), and in VAT (Table 3). However, another group of genes (ACC1, PEPCK, ACSS2, FABP4, and PDE3) correlated negatively with the BMI, insulin, and HOMA-IR (Tables 2 and 3).

Table 2 Significant correlations (p) found between the subcutaneous adipose tissue mRNA expression levels of the genes studied and the body mass index, insulin, and HOMA-IR in all the study subjects
Full size table
Table 3 Significant correlations (p) found between the visceral adipose tissue mRNA expression levels of the genes studied and the body mass index, insulin, and HOMA-IR in all the study subjects
Full size table

The expression of PPARγ correlated positively in the VAT with the expression of DGAT1 (r = 0.635, p < 0.001), AQP7 (r = 0.636, p < 0.001), GK (r = 0.449, p = 0.005), ACSS2 (r = 0.336, p = 0.039), FABP4 (r = 0.589, p < 0.001), ATGL (r = 0.849, p < 0.001), HSL (r = 0.497, p = 0.002), PDE3 (r = 0.675, p < 0.001), and perilipin (r = 0.819, p < 0.001). The expression of PPARγ correlated positively in the SAT with the expression of DGAT1 (r = 0.504, p < 0.001), AQP7 (r = 0.504, p = 0.001), FABP4 (r = 0.398, p = 0.015), ATGL (r = 0.534, p = 0.001), HSL (r = 0.405, p = 0.013), PDE3 (r = 0.541, p = 0.001), and perilipin (r = 0.535, p = 0.001).

Discussion

Our results show an upregulation of genes involved in facilitating adipocyte lipid storage and lipolysis in morbidly obese patients, though ACC1 and PEPCK were downregulated, especially in the morbidly obese patients with high insulin resistance. They also show that VAT and SAT behave differently in lean controls compared with morbidly obese patients.

PPARs are nuclear receptors that play a key role in the entraining of adipose tissue lipid metabolism to the nutritional state [19]. As we determined in our study, the expression of PPARγ, in both VAT and SAT, was significantly increased in the morbidly obese patients. Additionally, the expression of PPARγ correlated with the expression of most genes involved in the metabolism of fatty acids in adipose tissue. Our results suggest that this favors the storage of fatty acids in adipose tissue in order to ensure the storage and relocalization of the excess triglycerides [19]. However, PPARγ also correlated with the expression of genes involved in lipolysis. Previous studies have shown that PPARγ agonism also stimulates lipolysis by increasing the lipolytic potential, including the gene expression levels, an effect that is compensated by lipid-retaining pathways [11]. Our findings also suggest that increased expression may contribute to the positive action of PPARγ agonism on adipose lipolysis.

The expression of several genes involved in lipogenesis and lipolysis (DGAT1, ACC1, AQP7, and ATGL) was significantly altered in the morbidly obese patients, regardless of the state of insulin resistance. DGAT1 catalyzes the final step of lipogenesis. Transgenic mice overexpressing DGAT1 in adipose tissue have increased body weight and adiposity [20]. Our results show that morbidly obese patients have a greater expression of DGAT1, although with no correlation with PPARγ, as suggested in previous studies [21]. However, transcription factors regulating DGAT1 genes have not been studied in detail.

The expression of ACC1 is reduced in morbidly obese persons. ACC1 is highly expressed in adipose tissue [22] and plays an important role in energy balance by controlling malonyl-CoA synthesis. The ACC1 gene has an inducible expression in response to a variety of factors, including insulin [23]. Although our results show a negative correlation with plasma insulin levels and the HOMA-IR, the relation between ACC1 and insulin resistance is unclear [22, 23], with less extensive studies in adipocytes [24].

Glycerol is a key metabolite in adipose tissue. Under normal conditions, glycerol released from the triglycerides is not recycled intracellularly. AQP7 has been identified as the transporter of glycerol in the adipocytes [25] and is increased in obesity [26]. The increased expression of AQP7 in morbidly obese patients seen in our study may represent a mechanism that supplies the necessary glycerol to the adipose tissue for the increase in triglyceride synthesis. Changes in glycerol transporter expression could also be associated with alterations in the rate of glycerol efflux (as observed during lipolysis). As in other studies [26], our data show a positive association with insulin resistance in the VAT, but not in the SAT. This may agree with studies showing that AQP7 in SAT is not affected in persons with type 2 diabetes [27].

Lipolysis and FFA release from adipose tissue are commonly increased in obesity [14]. As we determined in our study, the expression of both ATGL and HSL is mainly increased in VAT of morbidly obese patients, independently of insulin resistance. However, in SAT, only ATGL is increased. The antilipolytic effect of insulin may vary between normal and impaired in the obese [28]. However, previous studies investigating this issue have produced conflicting results [29]. In states of insulin resistance, insulin cannot exert the antilipolytic effect through stimulation of PDE3. A decrease in PDE3 activity and mRNA seems to be associated with insulin resistance [30]. Our results show that the morbidly obese subjects with high insulin resistance have a lower expression of PDE3 in SAT, the adipose tissue that is more sensitive to the action of insulin. High levels of PDE3 inhibition decrease glucose uptake by adipocytes and may lead to glucose intolerance [31].

As with PDE3 and perilipin, the expression of some of the enzymes studied, such as PEPCK and GK, depends on the state of insulin resistance. These enzymes are involved in the synthesis of glycerol-3-phosphate, key for the synthesis of triglycerides. PEPCK is the enzyme that catalyzes the rate-limiting step in glyceroneogenesis. Glyceroneogenesis is about three- to fourfold more efficient than GK, as previously observed in 3T3-F442A rodent adipocytes [32]. Our results suggest something similar. PEPCK expression was greater than that of GK in the controls, though this was different in the morbidly obese patients. This change depended on the state of insulin resistance in the patients. The reduction in PEPCK was greater in the morbidly obese patients with high insulin resistance. The transcription of PEPCK in adipose tissue is known to be inhibited by insulin [33]. Something similar happens to the VAT ACSS2. The expression of GK in the VAT from the morbidly obese patients with greater insulin resistance, however, was similar to that of the controls. This suggests that these patients have a lower capacity to synthesize glycerol-3-phosphate in VAT and a lower triglyceride storage capacity in this tissue. Recent data suggest that it may not be the absolute amount of fat accumulated but the mismatch between energy surplus and storage capacity that may be an important factor linking obesity, insulin resistance, and associated morbidities [34].

As we see in our study, visceral and subcutaneous adipose tissues are metabolically different [35]. In states of low insulin resistance, as for instance in the controls, the expression of most of the enzymes studied was greater in the SAT. This tissue is more sensitive than VAT to the action of insulin and has high avidity for FFA and triglyceride uptake, preventing their deposition in nonadipose tissue [36]. Even so, the differences between the tissues disappeared almost completely in the morbidly obese patients, in whom we found a significant increase in insulin resistance. In these patients, we noted a relatively increased expression in the VAT as compared with the SAT. Our results may be in agreement with the hypothesis suggesting that the insulin-sensitizing effects of the thiazolidinediones are explained by a selective accumulation of subcutaneous adipose tissue, with concomitant lack of change or reduction in the adiposity of visceral depots [37].

The expression of several lipogenic and lipolytic genes changes simultaneously in VAT and SAT from morbidly obese patients, regardless of the insulin resistance. However, when the expression is only modified in VAT or SAT, the behavior of these tissues is different. VAT from morbidly obese patients shows a significant increase in the expression of other genes involved in lipolysis (GK, HSL, and perilipin). These findings reinforce the notion that VAT has a greater lipolytic activity than SAT [36] and that fewer fatty acids are produced from SAT [28]. Moreover, SAT shows a decrease in the expression of genes involved in de novo lipid synthesis and energy generation (ACSS2 and FABP4). The differences between VAT and SAT may partly be explained by the regional variations in the action of catecholamines and the variations in insulin sensitivity. These actions are markedly altered in the adipose tissue of the obese [38]: A higher catecholamine-induced lipolysis has been observed in VAT than in SAT [28, 38]. This suggests that altered expression of genes might play a role in adipose tissue distribution and expansion. This relative change may also be related with the concept that VAT contributes to the morbidity associated with the metabolic syndrome features. A relatively increased expression in VAT may be associated with an increased metabolic and cardiovascular risk.

This study provides new information regarding adipose tissue physiology in obesity. In the face of increased lipolysis, the net release of FFA into the bloodstream is counteracted by increased reuptake and cycling. In adipocytes, there is constant fatty acid turnover and also enormous enzyme activity to convert large quantities of free fatty acids to neutral lipids. From these results, we can suggest the presence of an association in the upregulation of PPARγ and genes facilitating adipocyte lipid storage in vivo and lipolysis in morbidly obese patients. On the other hand, the expression of enzymes synthesizing de novo fatty acids and glycerol was reduced. The expression of some enzymes studied in the morbidly obese patients depended on the state of insulin resistance. Our results also suggest that the degree of involvement of VAT and SAT in obesity and insulin resistance differs. However, a limitation of this study is that separation of the adipocytes from the stroma vascular fraction was not made. These results all facilitate our understanding of the coordinated regulation of the different genes and serve to enhance our understanding of the molecular mechanisms that underlie the regulation of fatty acid metabolism in obesity and insulin resistance.