A Dysregulation in CES1, APOE and Other Lipid Metabolism-Related Genes Is Associated to Cardiovascular Risk Factors Linked to Obesity

Abstract

Objective

The aim of the present study was to investigate the relationship between the differential expression of genes related to lipid metabolism in subcutaneous adipose tissue and metabolic syndrome features in lean and obese subjects with habitual high fat intake.

Methods

Microarray and RT-PCR analysis were used to analyze and validate differential gene expression in subcutaneous abdominal adipose tissue samples from lean and obese phenotype subjects.

Results

Several genes and transcripts involved in lipolysis were down-regulated, such as AKAP1, PRKAR2B, Gi and CIDEA, whereas NPY1R and CES1 were up-regulated, when comparing obese to lean subjects. Similarly, transcripts associated with cholesterol and lipoprotein metabolism showed a differential expression, with APOE and ABCA being decreased and VLDLR being increased in obese versus lean subjects. In addition, positive correlations were found between different markers of the metabolic syndrome and CES1 and NPY1R mRNA expressions, while APOE showed an inverse association with some of them.

Conclusion

Different expression patterns in transcripts encoding for proteins involved in lipolysis and lipoprotein metabolism were found between lean and obese subjects. Moreover, the dysregulation of genes such as CES1 and APOE seems to be associated with some physiopathological markers of insulin resistance and cardiovascular risk factors in obesity.

Introduction

Adipose tissue plays a critical role in energy homeostasis by hydrolyzing triacylglycerol reserves to provide fatty acids, which are important oxidative fuels for other tissues in situations of energy demand such as fasting and exercise [1]. Thus, a fine regulation of lipolysis is crucial for the maintenance of body energy homeostasis as well as for the prevention of metabolic diseases. Indeed, excess fat accumulation is associated with a constellation of metabolic risk factors related to cardiovascular disease and type 2 diabetes, mainly due to a disproportionate release of fatty acids and an impairment in the secretion of adipokines [2]. On the other hand, lipodystrophy may also lead to physiopathological alterations often accompanying the metabolic syndrome [3]. The common link between both diseases may be a defective storage capacity in adipose tissue depots: in the case of lipodystrophy due to lack of proper amount of adipose tissue and in the case of obesity due to saturation of storage capacity [4]. Thus, the functional failure of the adipose tissue to buffer postprandial lipids due to a chronically disturbed energy balance could result in changes in adipokine secretion and vascular effects that influences the body homeostasis system, which has been proposed as an explanation of the metabolic syndrome [5]. Furthermore, variations in the nature and magnitude of the appearance of the metabolic syndrome manifestations are attributable to the interaction of genetic factors with environmental influences, most notably diet and physical activity [6, 7].

In this context, we investigated the relationship between the expression of several genes related to lipid metabolism and metabolic syndrome features in two phenotypically well characterized groups of subjects [8] that significantly differed with respect to body weight despite reporting similar daily fat intake and physical activity patterns; one group of volunteers remained lean and with no features of metabolic syndrome and the other group was obese. Indeed, elucidating differences in gene expression in the subcutaneous abdominal adipose tissue (SCAAT) from these two groups of volunteers could help to develop therapies to prevent metabolic syndrome.

Subjects and Methods

Study Design and Subjects

Nine lean (22–33 years old) and 9 age-matched obese high-fat consumers (21–35 years old) were recruited as previously described, using a validated questionnaire based on self-reported questions about lifestyle and food frequency consumption [8]. All subjects were healthy, non-diabetic, non-hyperlipidemic males taking no oral medications and showing a stable body weight during at least the previous 3 months. In order to confirm that the amount and composition of energy ingested was >40% from fat, each subject completed a 3-day weighed food record for 2 weekdays and 1 weekend day. The food records were analyzed by a trained nutritionist using a computerized program (Medisystem, SanoCare, Madrid, Spain). To quantify the level of physical activity, each participant completed a validated questionnaire [9] based on self-reported questions about their leisure time and their work time physical activities on a typical workday and on a typical weekend day. In addition, sedentary lifestyle was assessed through the number of hours per week spent sitting down (watching TV or videos, reading or listening to music, etc.) on a typical workday and on a typical weekend day. All volunteers remained living in the same conditions during at least the last 3 years.

Anthropometrical Measurements and Adipose Tissue Biopsy

On the experimental day, volunteers arrived at the Clinica Universidad de Navarra after 12 h of overnight fast. Anthropometrical measurements were made using standard procedures as previously described [8]. Then, biopsies of SCAAT (1–2 g) were performed by liposuction under local anesthesia. The samples were washed and soaked in RNA-later (Qiagen, Valencia, CA, USA) to avoid RNA degradation and then stored at –80 °C until utilization. The protocol was approved by the Ethical Committee of the University of Navarra meeting the standards of the Declaration of Helsinki [10], and all subjects gave their written informed consent before participating in the study.

Blood Pressure and Measurements

Blood pressure (systolic and diastolic) was measured with a standard mercury sphygmomanometer (Minimus II, Riester, Jungingen, Germany) after the subject was quietly sitting for 5 min following WHO criteria. Fasting blood measurements were made by standard procedures as previously described [8]. The quantitative insulin sensitivity check index (QUICKI) was determined using the inverse of the sum of the logarithms of the fasting insulin (µU/ml) and fasting glucose (mg/dl).

RNA Preparation

Total RNA was isolated from each human SCAAT sample (including adipocytes and stroma-vascular fraction) by homogenization with an ultra-turrax^® T 25 basic using TRIzol (Life Technologies, Gaithersburg, MD, USA) according to the manufacturer’s instructions and incubated with RNase-free DNase (Ambion, Austin, TX, USA) for 30 min at 37 °C. RNA concentration was measured spectrophotometrically, and its quality was verified by ethidium bromide staining after agarose gel electrophoresis.

Microarray Experiments and Analysis

RNA was pooled to minimize the biological variation between the individual lean and obese subjects. Thus, 15 µg of total RNA from two pools of 3 lean subjects each (L1 and L2) and two other from 3 obese individuals each (O1 and O2) were used in the standard protocol from Affymetrix to label targets. These targets (biotinylated complementary RNA) were hybridized to the Human HG-U133 A GeneChip arrays (Affymetrix; Santa Clara, CA, USA) at Progenika Biopharma Inc. (Derio-Vizcaya, Spain), using tools obtained from Affymetrix and according to the manufacturer’s protocol (Affymetrix). Thus, a total of 4 array hybridizations were performed. The obtained signal values were further analyzed with the Affymetrix microarray suite 5.0. software (MAS 5.0) system to examine obesity-dependent increases and decreases in gene expression. After global scaling, the signal detection, the signal log ratio (SLR) and the different call change were calculated and compared between GeneChips. The comparison was performed for all four combinations: obese versus lean samples; O1 versus L1, O1 versus L2; O2 versus L1; O2 versus L2 in a double-cross analysis (concordance analysis). The alteration ratios of the gene expression were represented as means of SLR of the four quotients [11].

GARBAN software [12, 13] was used to classify the differentially expressed genes according to Gene Ontology biological process criteria.

Real-Time PCR Analysis

Differential gene expression was further confirmed by real-time PCR of a subset of genes in individual samples (n = 9 in each group). Reagents for real-time PCR analysis of AKAP1, NPY1R, CES1, APOE, VLDLR and 18S (Assays-on-Demand, TaqMan Reverse Transcriptase reagents, and TaqMan Universal PCR Master mix) were purchased from Applied Biosystems (Foster City, CA, USA) and used according to the manufacturer’s protocol. Amplification and detection of specific products were performed with the ABI PRISM 7000HT system (Applied Biosystems). Human 18S was used as reference to normalize the expression levels between samples, allowing data to be expressed relative to 18S rRNA and thus compensating for any differences in reverse transcriptase efficacy, as previously described [11].

Statistical Analysis

Data are expressed as means ± SE (n = 9 in each group). Differences between the lean and obese groups were analyzed by the unpaired Student’s t test or Mann-Whitney U test after testing the normality with the Kolmogorov-Smirnoff and Shapiro-Wilk tests, and Pearson and Spearman correlation coefficients were used to identify related variables. A p value < 0.05 was considered statistically significant. The SPSS 14.0 version for Windows (SPSS, Chicago, IL, USA) was used for the statistical analysis.

Results

Baseline Characteristics of Lean and Obese Subjects

Not surprisingly with respect to the study design, BMI, waist circumference and waist-to-height ratio were significantly higher in obese than in lean subjects. Insulin sensitivity revealed by QUICKI was significantly lower in obese compared to lean subjects. The fasting lipid profile, including triglycerides, total cholesterol and total / high-density lipoprotein (HDL) cholesterol ratio, was significantly higher in obese than in lean participants. The systolic and diastolic blood pressure values were significantly elevated in obese when compared to lean subjects. Moreover, the amount of leptin in blood was higher in obese subjects, while adiponectin was lower (table 1). Even though high-fat diet is associated with presence of the metabolic syndrome, lean volunteers showed no features of metabolic syndrome. According to the National

Table 1.

Anthropometrical and clinical parameters of volunteers^a

	Lean (n = 9)	Obese (n = 9)	p
Energy intake, kcal/day	2,766.7 ± 258.7	2,799.1 ± 171.4	0.918
BMI, kg/m2	23.1 ± 0.4	34.7 ± 1.2	0.000
Waist circumference, cm	78.7 ± 1.2	105.7 ± 2.6	0.000
Waist/height ratio	0.45 ± 0.01	0.60 ± 0.01	0.000
Leptin, ng/ml	8.3 ± 2.6	33.3 ± 4.6	0.000
Adiponectin, µg/ml	20.1 ± 3.3	10.0 ± 1.7	0.037
Glucose, mg/dl	90.3 ± 3.9	92.7 ± 2.3	0.070
QUICKI	0.40 ± 0.00	0.35 ± 0.01	0.004
Triglycerides, mg/dl	85.0 ± 6.7	142.2 ± 10.6	0.001
Total cholesterol, mg/dl	167.4 ± 17.7	188.5 ± 6.3	0.008
HDL cholesterol, mg/dl	43.3 ± 1.7	40.0 ± 2.4	0.059
Total/HDL cholesterol ratio	3.5 ± 0.2	5.0 ± 0.3	0.003
SBP, mm Hg	122.5 ± 3.6	139.1 ± 2.8	0.002
DBP, mm Hg	74.0 ± 2.2	82.8 ± 3.0	0.002

HDL = high-density lipoprotein; LDL = low-density lipoprotein;

QUICKI = quantitative insulin-sensitivity check index; SBP = systolic blood pressure; DBP = diastolic blood pressure.

^aValues are means ± S.E.M. Independent Student’s t-test or Mann-Whitney U-test was performed, as appropriate, depending on the results of Kolmogorov-Smirnoff and Shapiro-Wilk normality tests.

Cholesterol Education Program, 5 of the obese volunteers were considered obese with metabolic syndrome based on the presence of three or more of the following characteristics: waist circumference greater than 102 cm, blood pressure of at least 130/85 mm Hg, serum glucose level of at least 110 mg/dl, serum triacylglycerol level of at least 150 mg/dl and HDL cholesterol level of less than 40 mg/dl.

Microarray Data

The analysis of the microarray data revealed that several relevant transcripts involved in lipid metabolism were differentially expressed in SCAAT of obese and lean subjects (table 2). Some genes involved in the hormonally regulated lipolysis – AKAP1, PRKAR2B – were down-regulated, whereas NPY1R was up-regulated (table 2). As expected, several transcripts related to TNF-α-induced lipolysis were down-regulated including Gi and CIDEA. In contrast, CES1, an adipocyte lipase involved in basal lipolysis, was up-regulated (table 2).

Table 2.

Differentially expressed genes involved in lipid and cholesterol metabolism in SCAAT of obese vs lean subjects

	Gene symbol	SLR ± SDa	RT-PCRb
			lean	obese
Basal and hormonally stimulated lipolysis
Insulin receptor substrate 2	IRS2	–0.58 ± 0.05
Protein kinase, cAMP-dependent, regulatory, type II, β	PRKAR2B	–0.50 ± 0.18
A kinase (PRKA) anchor protein 1	AKAP1	–0.85 ± 0.37	1.0 ± 0.42	0.24 ± 0.03*
Gprotein, a inhibiting activity polypeptide 1	Gi	–0.95 ± 0.31
Phosphodiesterase 3 B	PDE3B	–0.85 ± 0.55
NeuropeptideYreceptor Y1	NPY1R	0.93 ± 0.57	1.0 ± 0.11	3.31 ± 0.22*
Carboxylesterase 1	CES1	0.80 ± 0.50	1.0 ± 0.14	1.86 ± 0.17*
Cell death-inducing DFFA-like effector A	CIDEA	–1.13 ± 0.38
Lipoprotein and cholesterol metabolism
Apolipoprotein E	APOE	–1.15 ± 0.65	1.0 ± 0.24	0.39 ± 0.07**
ATP-binding cassette, sub-family A (ABC1)	ABCA	–0.63 ± 0.21
Very low density lipoprotein receptor	VLDLR	0.65 ± 0.39	1.0 ± 0.18	2.47 ± 0.25*
Lysosomal acid, cholesterol esterase	LIPA	0.80 ± 0.63

Significantly different between groups

^*p &lt; 0.05,

^**p &lt; 0.01.

^aThe alteration ratios of the gene expression are represented as means of signal log ratio (SLR) ± standard deviation (SD).

^bDifferential gene expression was further confirmed by RT-PCR of a selected subset of genes. Data expressed as means ± SE (n = 9 in each group) were calculated by the 2^–ΔΔCt method (mean value for lean subjects was set at 1). Human18S rRNA was used as reference to normalize the expression levels.

In addition, several transcripts involved in cholesterol and lipoprotein metabolism, e.g. APOE and ABCA, a cellular cholesterol transporter, were down-regulated (table 2) while VLDLR was up-regulated in SCAAT in the obese subjects (table 2).

mRNA Expression of Selected Genes Correlated with Several Features of the Metabolic Syndrome

Gene expression of CES1, NPY1R, AKAP1, VLDLR and APOE was analyzed for potential association with different cardiovascular risk markers (table 3). A positive correlation was found between CES1 mRNA expression in SCAAT and waist circumference (r = 0.55; p = 0.032) and the waist-to-height ratio (r = 0.53; p = 0.044). Also, CES1 mRNA levels correlated with plasma glucose (r = 0.55; p = 0.035) and triglyceride levels (r = 0.73; p = 0.002) and with the cardiovascular risk factor total/HDL cholesterol (r = 0.65; p = 0.009). A positive correlation was found between NPY1R mRNA expression in SCAAT and waist circumference (r = 0.65; p = 0.009) and the waist-to-height ratio (r = 0.55; p = 0.035). Moreover, NPY1R mRNA levels were associated with the cardiovascular risk index (r = 0.52; p = 0.048).

Table 3.

Correlation between gene expression and metabolic syndrome features^a

	APOE		CES1		NPY1R
	r	p	r	p	r	p
BMI, kg/m2	–0.68	0.006	0.43	0.114	0.30	0.272
Waist circumference, cm	–0.61	0.015	0.55	0.032	0.65	0.009
Waist/height ratio	–0.58	0.023	0.53	0.044	0.55	0.035
Leptin, ng/ml	–0.56	0.030	0.31	0.262	0.36	0.191
Adiponectin, µg/ml	0.52	0.049	–0.16	0.568	0.30	0.282
Glucose, mg/dl	–0.46	0.088	0.55	0.035	0.30	0.275
Triglycerides, mg/dl	–0.40	0.139	0.73	0.002	0.34	0.218
Total/HDL cholesterol ratio	–0.65	0.009	0.65	0.009	0.52	0.048

^aPearson and Spearman correlations were performed between gene expression (arbitary units 2^–ΔΔCt) and other parameters.

Negative correlations were found between APOE mRNA expression in SCAAT and BMI (r = –0.68; p = 0.006), waist circumference (r = –0.61; p = 0.015) as well as waist-to-height ratio (r = –0.58; p = 0.023). APOE was negatively correlated with leptin (r = –0.56; p = 0.030) and positively correlated with adiponectin (r = 0.52; p = 0.049). Moreover, APOE mRNA levels were inversely correlated with the cardiovascular risk index (r = –0.65; p = 0.009) (table 3) and also with diastolic blood pressure (r = –0.62; p = 0.014).

AKAP1 mRNA levels were only significantly correlated with the waist-to-height ratio (–0.59; p = 0.016) while no significant associations were found between VLDLR mRNA levels and metabolic syndrome features.

Discussion

In the present study, mRNA expression differences of genes related to lipid and cholesterol metabolism between obese and lean subjects with habitual high-fat diet and physical activity, otherwise phenotypically well characterized [8], were investigated. The analysis of anthropometrical and clinical variables of the obese subjects suggest that they have a higher risk for developing insulin resistance, hyperlipidemia and cardiovascular disease, which are typical features of the metabolic syndrome. Indeed, 5 of the obese subjects investigated could be diagnosed with metabolic syndrome according to Adult Treatment Panel III and World Health Organization criteria [6]. However, new criteria are under discussion [14] and may change the assignment. The mechanisms underlying the metabolic syndrome are not well understood. Current evidence supports that the accumulation of abdominal fat is a major determinant of the metabolic syndrome, while gluteofemoral body fat seems to exert a protective role [15]. Upper-body subcutaneous fat is the dominant contributor to circulating free fatty acids and the main source of excess fatty acid release in upper-body obesity and of abnormalities in subcutaneous lipolysis, which appears to be an important cause of peripheral insulin resistance [16], possibly through altering fat storage and the expression of TNF-α by adipocytes in visceral fat [17]. We investigated the differential expression of genes encoding proteins participating in lipid and lipoprotein metabolism in adipose tissue. The microarray analysis revealed that several transcripts involved in the stimulation of lipolytic pathway were down-regulated in obese but not in lean subjects. Among them is the regulatory subunit type II β of PKA (PRKAR2B) which plays a crucial and central role in the regulation of energy expenditure and glucose and lipid metabolism [18, 19]. Its activity is known to be modulated by its specific location in the cell, a process mediated by A-kinase anchoring proteins (AKAPs) [20]. In fact, AKAP1, identified as a major adipocyte protein kinase Abinding protein [21], was down-regulated in obese subjects. We also found that the transcription of NPY1R, a powerful antilipolytic agent [22], was up-regulated in SCAAT of obese patients, which is in accordance with a previous study where a similar up-regulation was found in omental adipose tissue [23]. These data agree with those of several studies, which have shown that obesity is associated with an impaired catecholamine-induced lipolysis and a reduced HSL and ATGL expression in adipocytes [24, 25]. However, elevated basal lipolysis has been observed in obese individuals which is related to the bigger fat cell size and to increased levels of proinflammatory cytokines in adipose tissue [26, 27].

Moreover, in our study a down-regulation of Giα and CIDEA in obese subjects was found. CIDEA has been assumed to have a specific role in lipolysis in white human fat cells mediated by cross-talk with TNF-α [28]. Thus, the elevated basal lipolysis observed in obesity implies that other TG lipases may be up-regulated. Our study revealed that the adipocyte lipase carboxylesterase 1 (CES1) an orthologue of carboxylesterase 3 (also named TGH2), which was first described by Soni et al [29], was up-regulated in obese subjects, which is in agreement with other studies [30, 31]. This lipase is highly expressed in adipocytes and is transcriptionally regulated in response to changes in nutritional conditions, suggesting a relevant role in lipolytic response to extended fasting in adipocytes [32]. In our study, the positive association between CES1 and obesity was further supported by the fact that CES1 mRNA expression was positively correlated with waist circumference, a recognized marker of adiposity and dyslipidemia [33], and with the waist-to-height ratio, a useful predictor of coronary heart disease risk factor [34]. In addition, the positive correlation between CES1 mRNA expression and total/HDL cholesterol ratio, a known factor for cardiovascular risk, suggests that CES1 could play a role in the impaired basal lipolysis in SCAAT from obese individuals. However, a recent publication has found no correlation between CES1 expression and lipolytic activity in isolated human adipocytes [31].

On the other hand, the impaired capacity for the sensing of the replenishment of their lipid stores by adipocytes which results in alterations in the buffering capacity of adipose tissue has been suggested as the cause of metabolic syndrome [5, 35]. Based on the fact that triacylglycerol and cholesterol storage are closely linked in adipocytes, it has been postulated that cholesterol might participate in the intracellular sensing for fat cell size and triacylglycerol content [36, 37]. Our microarray and RT-PCR analysis showed a down-regulation of the APOE gene in obese subjects. This is in accordance with the marked suppression of adipose APOE observed in both diet-induced and genetic (ob/ob mice) models of obesity [38]. A recent study evidenced that excess fat accumulation via an APOE-dependent pathway might play a role in development of the metabolic syndrome [39]. APOE is secreted by cells with high lipid efflux, and it has been shown to be expressed highly in adipose tissue [40, 41]. In these cells, APOE modulates cellular cholesterol metabolism by facilitating cholesterol transfer. In fact, the study of Huang et al. [38] suggested that these changes in adipose tissue APOE expression have significant impact on adipose tissue lipid flux and lipoprotein metabolism [38]. These data and our present observations in human adipose tissue suggest that adipose APOE may participate in defending adipose tissue and organismal energy homeostasis in response to high fat feeding. In contrast to APOE, VLDLR, another gene related with cholesterol transport, was up-regulated. It is well known that high levels of plasma VLDL are associated with obesity, type 2 diabetes [42, 43] and cardiovascular disease [44]. Moreover, VLDLR can play a role in the delivery of VLDL-TG-derived free fatty acids to adipose tissue, modulating circulating levels of TG and VLDL [45, 46]. Indeed, VLDLR–/– mice were protected from high-fat diet-induced obesity showing a decreased adipocyte size [47] and triglyceride storage in the adipocyte compared to wild-type counterparts [46, 48]. Interestingly, a positive correlation between CES1 and VLDLR mRNA levels was found (data not shown).

In the present study, whole adipose tissue was used for gene expression profile analysis, even if this tissue only provides a mixture of cells (stroma-vascular cells and adipocytes). An interesting question to be answered is which of these cell fractions are mainly contributing to the observed gene expression changes in adipose tissue. Regarding CES1, several previous studies have observed that CES1 expression is higher in isolated adipocytes compared to intact adipose tissue [31]. Previous studies have also shown that APOE mRNA is found in differentiated 3T3-L1 adipocytes as well as in biopsies of human adipose tissue, but not in undifferentitated 3T3-L1 adipocytes [49], suggesting that the main contribution in APOE gene expression observed in adipose tissue from obese subjects could come from the adipocytes. However, another study described that APOE is also down-regulated in adipose tissue-recruited macrophages during high-fat diet when compared with resident macrophages [50]. In any case, it is important to emphasize that information obtained about differential gene expression from the tissue as a whole is relevant. Indeed, total adipose tissue has been widely used to identify genes associated with obesity using DNA microarrays [51, 52].

During the last years many efforts have been made to define and search for genes associated with obesity and metabolic syndrome. In this context, several studies have aimed to look for genes differentially expressed in adipose tissue from lean and obese subjects [23, 53]. The main novelty of the present study arise from the fact that it was designed to look for adipose tissue genes conferring susceptibility or resistance to develop obesity and metabolic syndrome in two well characterized groups of subjects. Despite living in same environmental conditions (habitual high fat dietary intake and similar moderate physical activity), one group was resistant to gain weight and to develop metabolic syndrome features, while the other was susceptible to develop obesity and its associated metabolic complications.

In summary, our data show a different expression of adipose transcripts encoding for proteins involved in lipolysis and lipoprotein metabolism in subjects ‘resistant’ or ‘susceptible’ to develop high-fat diet-induced obesity. These results, taken as a whole, suggest that the changes in the expression levels of these adipose tissue genes may have significant impact on adipose tissue lipolysis rate as well as on lipid flux and lipoprotein metabolism. Nevertheless, the molecular mechanisms underlying the control of these key lipid metabolism processes are still not fully understood. Our data also suggest that CES1, a recently discovered lipase, and APOE could play a role in obesity-associated cardiovascular risk factors and might be promising targets in metabolic syndrome treatment.

Disclosure

The authors declared no conflict of interest.