Obesity is defined as an enlargement of adipose tissue mass due to both hypertrophy and hyperplasia of adipocytes . The aim of this study was to determine the gene expression profile of adipose tissue during obesity by using a focused microarray especially designed to study cell-cycle- and lipid-metabolism-related genes. We analyzed subcutaneous adipose tissue, which is the largest adipose depot in the body, accounting for approximately 80% of the total body fat. The metabolic complications associated with obesity correlate strongly with central obesity . Subcutaneous abdominal fat, as a component of central adiposity, has a strong association with insulin resistance and plays an important role in the pathophysiology of obesity [6–8].
Present microarray analysis revealed that in the subcutaneous adipose tissues of morbidly obese patients the expression of genes known to be involved in adipocyte differentiation and cell-cycle control is profoundly altered. We found that the expression of C/EBPβ and JUN, which are transcription factors that regulate the first stages of adipocyte differentiation, was increased in the adipose tissue of morbidly obese patients. Conversely, the expression of PPARγ1, a transcription factor that controls the final steps of preadipocyte conversion into mature adipocytes, was reduced. C/EBPβ is not only an important regulator of adipocyte terminal differentiation since, in addition, it is a critical regulator of body weight, adiposity and tumour growth . C/EBPβ deletion in Leprdb/db mice reduces obesity, fatty liver, and diabetes . Leptin levels are also modulated by C/EBPβ, and mice lacking C/EBPβ have severely reduced leptin levels . Consistently with this, we found an increase in both C/EBPβ and leptin mRNA levels in obese patients.
In the different study subsets performed in the present work, the expression of PPARγ1 in adipose tissue was reduced in obese subjects as compared to controls, whereas no statistically significant differences were observed for PPARγ2. In Lepob/ob mice, however, both PPARγ1 and PPARγ2 were overexpressed. Although PPARγ is a well-characterized regulator of energy metabolism, the relationship between PPARγ expression and obesity is not clear. In fact, previous studies by others have produced conflicting results regarding the association between PPARγ and obesity in both humans and animals. Some studies have shown changes in PPARγ1, PPARγ2, or total PPARγ expression in subcutaneous adipose tissue [29–33], whereas others have reported no changes [30, 34, 35]. The different characteristics of the populations and the different methodologies used to determine PPARγ expression in these studies could explain these discrepancies. Addressing this issue, Sewter et al. also found a decrease in PPARγ1 mRNA expression in subcutaneous adipocytes from morbid obese compared with lean subjects, and a strong inverse correlation between BMI and PPARγ1 mRNA levels. In contrast, they found a significant increase in PPARγ2 mRNA expression in the morbid obese group . The reduction in PPARγ1 expression in obese patients found in our study is consistent with the increase in Leptin mRNA expression also found herein, since it has been described that Leptin suppress the expression of PPARγ in adipocytes .
Cyclin-dependent kinase inhibitors (CDKIs) play an important role in cell cycle regulation and some of them are especially involved in adipocyte differentiation. In the present work we describe for the first time that the expression of three CDKIs is altered in human adipose tissue from morbid obese patients: CDKN1A (p21) was increased whereas CDKN2C (p18) and CDKN1B (p27) were decreased. Morrinson et al. were the first to demonstrate that these CDKIs are highly regulated in 3T3-L1 preadipocytes differentiation. Thus, p18 mRNA is expressed only during terminal differentiation, p27 mRNA is highly expressed during the whole differentiation process except in S phase, and p21 mRNA is expressed in G1 phase, decreases in S phase, and increases again at postmitotic growth arrest . The importance of these CDKIs has been underscored recently in knockout mice lacking p21 and p27. Loss of one or both CDKIs, results in adipocyte hyperplasia, obesity and insulin resistance . These results suggest that theses CDKIs are major regulators of adipocyte number in vivo and can have an important role in the development of adipose tissue hyperplasia during obesity. Moreover, p21 has been involved in adipocyte hypertrophy since it protected the hypertrophied adipocytes against apoptosis . However, the precise contribution of these CDKIs to obesity development in humans is unclear. In this context, it is worth to note that GADD45B, a member of the growth arrest and DNA-damage-inducible gene family, also was overexpressed in the adipose tissues of morbidly obese subjects. This gene is involved in terminal myeloid differentiation and growth suppression , but its relationship to adipocyte differentiation has not been established. It appears, thus, that the expression of CDKIs and GADD45B, all of which regulate cell cycle and differentiation, is altered in human adipose tissue from morbid obese patients, which may reflect the relative abundance of a characteristic adipocyte subtype in fat depots from obese.
Most of our knowledge of adipogenesis is based on studies in murine-derived embryonic 3T3-L1 cells and much less in known about adipocyte differentiation in humans [40, 41]. It has been established that most of the adipogenic program is similar in murine and human cells, since the expression pattern of the adipocyte differentiation-specific transcription factors C/EBPβ, C/EBPδ, PPARγ, and C/EBPα was similar in both species [15, 42]. However, in contrast to murine preadipocytes, human preadipocytes do not require clonal expansion to enter the differentiation process in vitro, and they differentiate directly in response to stimulus . It has been suggested that this phenomenon could reflect that adipocyte precursor cells from human adipose tissue have already undergone critical cell divisions and may be in a late stage of adipocyte differentiation . In agreement with this hypothesis, it has recently been established that in humans the number of adipocytes is set during childhood and adolescence and it stays constant in adulthood . However, the gene expression pattern found herein could reflect an increase in undifferentiated adipocytes as a consequence of the increased renewal rate in obese individuals. Consistently with this, Spalding et al demonstrated that obese individuals generate significantly more adipocytes per year than lean individuals . Based on the relationship between adipocyte size and total body fat, they also developed a method to quantitatively estimate adipose morphology . They describe that subjects can be categorized as having different degrees of either adipose hypertrophy or hyperplasia, and demonstrate that low generation rates of adipocytes are associated with adipose tissue hypertrophy whereas high generation rates are associated with adipose hyperplasia . In morbid obese individuals, such as those studied herein (BMI > 35 kg/m2), coexist both hypertrophy and hyperplasia, however in these severe forms of obesity hyperplasia became most predominant [9, 45]. It is important also to keep in mind that although the principal cellular component of the adipose tissue are the adipocytes other cellular components are also present, such as smooth muscle cells, endothelial cells, fibroblasts, and blood cells [46–48] and, thus, a contribution of these cells to this expression pattern cannot be ruled out.
Strong similarities in the expression pattern of adipogenic genes such as Jun, C/EBPβ, p21, and p18 were observed in subcutaneous adipose tissue from obese Lepob/ob mice as compared to human obese subjects. However, a major difference between both species occurs in PPARγ1 expression. Given that PPARγ1 is one of the master regulators of the adipocyte differentiation, differences in this gene may be involved in the differences observed between human and mouse adipogenesis and could be characteristic of adipocyte precursors of hypertrophic and hyperplasic adipose tissue of each species. The possibility exists, however, that the increase in PPARγ1 in Lepob/ob mice is just a consequence of the absence of leptin in this mouse model, given the role of this hormone in the regulation of PPARγ expression in adipocytes .
A notable finding of this work is the up regulation of ANGPTL4 expression in the adipose tissue from obese subjects, which was also found in obese mouse. This protein is a member of the angiopoietin-like family of proteins, which regulate angiogenesis  and may have a role in the stimulation of endothelial cell growth necessary for adipose tissue expansion . ANGPTL4 is predominately expressed in adipose tissue and liver and its mRNA expression increases dramatically in the early stages of adipocyte differentiation and in the adipose tissue of diabetic (Leprdb/db) and obese (Lepob/ob) mice . More recently, ANGPTL4 has been shown to be involved in the regulation of lipid and glucose metabolism, independently of its angiogenic effects. Thus, Angptl4-deficient mice are hypolipidemic and have increased lipoprotein lipase activity , whereas Angptl4 adenovirus-mediated overexpression potently increases plasma triglycerides, decreases blood glucose, and improves glucose tolerance [53, 54]. Hence, our results indicate that ANGPTL4 may be relevant to human obesity and, together with previous findings, point to this protein as a potential therapeutic target for obesity and obesity-related complications [54, 55].
Although not consistently found in every group of obese patients of this study, the expression of LMNA is also upregulated in obesity. LMNA encodes the nuclear structural proteins lamin A and C produced by alternative splicing, which are members of the intermediate filament family of the nuclear lamina. Mutations in LMNA have been associated with a number of disorders, including familial partial lipodystrophy (OMIM 151660). Sequence variations in LMNA are also associated with greater susceptibility to the development of metabolic syndrome, dyslipidemia, insulin resistance, diabetes, and obesity [56, 57]. The significance of its elevated expression in obesity remains to be established.
PCK1 gene was found to be significantly downregulated in human obese adipose tissue, whereas its expression was normal in obese mice. PCK1 encodes the cytosolic isozyme of phosphoenolpyruvate carboxykinase, which is the main enzyme controlling gluconeogenesis in the liver and kidney . However, the major role of PCK1 in white adipose tissue seems to be glyceroneogenesis for reesterification of free fatty acids [59, 60]. The reduction of PCK1 expression in obesity could be the consequence of the excess of triglyceride storage and the adaptation of adipocytes to get rid of some of these triglycerides throught lipolysis, perhaps mediated by glucocorticoids, which downregulate PCK1 in adipose tissue .
We have found sex differences in the expression of two genes involved in fatty acid utilization, such as LPL and SCD. While LPL was downregulated in female obese patients there were no differences in male obese patients. Sex differences in mRNA LPL expression and activity have been previously reported . The differences between sexes were more pronounced in SCD: while SCD expression in obese female was lower than in control females, in obese males it was clearly upregulated. SCD expression is regulated by many dietary, hormonal and environmental factors that could explain the sex differences observed in our study . Though, we cannot exclude the possibility that some of the sex differences were influenced by differences in the glucidic and triglyceride metabolism (Table 2).
We also found differences in other genes involved in lipid metabolism such as NPC2, ACSL1 and ACSL4, although we have not confirmed these results in an independent population of obese patients. Acyl-CoA synthetase (ACS) enzymes are essential for de novo lipid synthesis and fatty acid catabolism. This complex family of proteins catalyzes the activation of fatty acids necessary for their metabolism. Among these, the long-chain acyl-CoA synthetases (ACSL) activate fatty acids with chain lengths of 12 to 20 carbon atoms . ACSL1 is the major form in adipocytes and it has been proposed that it mediates free fatty acid reesterification, efflux and lipid-mediated signal transduction . ACSL4 is also expressed in adipocytes and it has been described that associates with lipid droplets after the lipolytic stimulation of 3T3-L1 adipocytes in vitro . Interestingly, these two isoforms of ACLS presented a different pattern of expression in obese patients: ACLS1 was downregulated whereas ACLS4 was found to be overexpressed.
NPC2 is a small secreted glycoprotein that binds cholesterol and plays an important role in intracellular cholesterol trafficking . The upregulation of NPC2 in obese subjects as found herein may relate with the novel role of NPC2 in adipocyte differentiation and the maintenance of the metabolic state of mature adipocytes . These novel roles of NPC2 open a new perspective in the study of the adipocyte dysfunction associated with obesity that needs to be studied in more detail.