- RA
retinoic acid
During evolution famine has been a substantial threat to survival. Adipocytes enable the storage of large amounts of fat for use when food is scarce. In modern times many individuals face a different challenge, i.e. overnutrition and a sedentary lifestyle, and the negative health consequences have become more prevalent. Excessive intake of energy-dense foods high in fat and/or sugar combined with low physical activity is associated with the development of obesity (Drevon et al. Reference Drevon, Graff-Iversen, Klepp, Nilsen, Strømme, Sundgot-Borgen, Søgaard, Tonstad, Øien and Aasen2000; Saris et al. Reference Saris, Astrup, Prentice, Zunft, Formiguera and Verboeket-van de Venne2000). The amount as well as the type of dietary fatty acids may affect lipid homeostasis and fat deposition (Hill et al. Reference Hill, Lin, Yakubu and Peters1992). Obesity is becoming an extensive public health problem in most countries, particularly in economically-advanced countries. The prevalence of obesity in Europe is alarming, as demonstrated by recent health statistics in Norway (Meyer & Tverdal, Reference Meyer and Tverdal2005) and the UK (Rennie & Jebb, Reference Rennie and Jebb2005), which show that about 15 and 25% of the adult population respectively are classified as obese (BMI >30 kg/m2). The prevalence of obesity is even higher in the USA, with about one-third of adults classified as obese (Baskin et al. Reference Baskin, Ard, Franklin and Allison2005). Furthermore, there has been a very rapid escalation in the obesity epidemic over the past two to three decades, with the incidence rising 3-fold in Norway (Drevon et al. Reference Drevon, Graff-Iversen, Klepp, Nilsen, Strømme, Sundgot-Borgen, Søgaard, Tonstad, Øien and Aasen2000) and the UK (Prentice & Jebb, Reference Prentice and Jebb1995). Obesity is associated with increased risk of conditions such as type 2 diabetes, CHD, hypertension, dyslipidaemia (which are often included in the so-called metabolic syndrome), gall stones, certain types of cancer (breast and colon), osteoarthritis, non-alcoholic steatohepatitis, sleep apnoea, infertility and many psychological conditions. Moreover, there is a marked reduction in life expectancy of the order of several years. With an increasing amount of body fat there is an increasing risk of developing type 2 diabetes in particular; the likelihood of developing this disease increases 10-fold at a BMI of 30 kg/m2.
Type 2 diabetes mellitus is diagnosed in about 20% of pubertal children referred to diabetic clinics in the USA (Arslanian, Reference Arslanian2002). The metabolic syndrome is identified in ⩽30% of obese children. Signs of the metabolic syndrome in children tend to continue into adulthood, and non-alcoholic steatohepatitis is an increasing clinical problem in obese children and adolescents (Malecka-Tendera & Mazur, Reference Malecka-Tendera and Mazur2006).
Adipose tissue structure and distribution
Adipose tissue is unique to vertebrates. It is found in most mammals, birds, reptiles and amphibians, and a variety is found in some species of fish. Furthermore, in insects the fat body found in larvae as well as in adults shares some homology with adipose tissue. In man adipose tissue deposition begins towards the end of fetal life. A very lean healthy adult may have as little as 3–4% body fat, whereas a lean adult on average has about 10–15 kg adipose tissue and an obese individual with leptin receptor deficiency can accumulate >70% body mass as fat (Clement et al. Reference Clement, Vaisse, Lahlou, Cabrol, Pelloux and Cassuto1998).
Different adipose tissue depots (Bjorneboe et al. Reference Bjorneboe, Bjorneboe, Bodd, Hagen, Kveseth and Drevon1986) probably have different functions based on their anatomical location (Table 1). For example, it is likely that the retro-orbital fat depot, located behind the eye, has a different function from the visceral adipose tissue depots (surrounding internal organs). There is fairly good evidence that there is higher risk of developing myocardial infarction associated with visceral obesity than with fat accumulated subcutaneously around the hip (Yusuf et al. Reference Yusuf, Hawken, Ounpuu, Bautista, Franzosi and Commerford2005). This difference in risk may be related to a lower expression of leptin, leptin receptor, adipsin, adiponectin, acylating-stimulating protein, cholesteryl ester transfer protein in visceral adipose tissue v. subcutaneous adipose tissue. On the other hand, higher expression of IL-6 and -8, plasminogen-activator inhibitor-1, PPARγ, resistin and 11-hydroxysteroid dehydrogenase type 1 is observed in visceral adipose tissue v. subcutaneous adipose tissue (Schaffler et al. Reference Schaffler, Scholmerich and Buchler2005).
Table 1. Adipose tissue depot locations (modified after Abate & Garg, Reference Abate and Garg1995)
* These depots are very sensitive to shifts in energy balance, and they are often enlarged in obesity.
Adipose tissue comprises several cell types
Although the distinctive feature of adipose tissue is the adipocytes, there are several types of cells present in adipose tissues, including pericytes, preadipocytes, white and brown adipocytes, fibroblasts, endothelial cells, immune cells such as macrophages, dendritic cells, mast cells, granulocytes and lymphocytes, in addition to nerve cells linked to the autonomous nervous system (Table 2).
Table 2. Different cell types in adipose tissue
Adipocytes are derived from mesenchymal stem cells (probably pericytes) and acquire intracellular lipid droplets during differentiation (Cinti, Reference Cinti1999a). Essentially, all eukaryotic cells are capable of forming small lipid droplets, although the capacity for lipid-droplet storage is greatly enhanced in adipocytes.
Mature white adipocytes comprise ⩽85% of the adipose tissue mass and can accommodate much larger lipid droplets than any other cell type (Fig. 1). In contrast to most other cell types adipocytes have a considerable capacity for expansion. When filled with lipids white adipocytes are approximately spherical. They differ in size, having diameters of ⩽100 μm, which is larger than most other cell types. The unilocular lipid droplet can represent approximately 90% of the cell volume of a white adipocyte. In a typical mature white adipocyte the nucleus and other organelles are compressed against the cell membrane (Cinti, Reference Cinti1999a).
Fig. 1. White and brown adipocytes containing unilocular and multilocular lipid droplets, respectively.
Brown adipocytes (Fig. 1) are clustered together to form what is termed brown adipose tissue, which appears brown because of the large number of mitochondria in the cytoplasm of these cells (Cinti, Reference Cinti1999a). Brown adipocytes have a unique metabolism in which the flow of protons through the inner mitochondrial membrane bypasses the ATP synthase via uncoupling protein 1, which uncouples oxidative phosphorylation and the energy produced by the protons as they are channelled through the inner mitochondrial membrane is released as heat (Cannon & Nedergaard, Reference Cannon and Nedergaard2004).
White adipose tissue depots contain variable amounts of brown adipocytes depending on age, species and environmental conditions. Recent data (Cinti, Reference Cinti2005) have demonstrated the plasticity of the adipose organ, and under particular conditions fully differentiated white adipocytes can transdifferentiate into brown adipocytes, and vice versa. All adipocytes of the adipose organ express a specific adrenoceptor termed ss3 adrenoceptor, and treatment of genetically-obese and diet-induced obese rats with ss3 adrenoceptor agonists ameliorates their pathological condition, and is accompanied by the appearance of brown adipocytes in white areas of the adipose organ. This drug-induced modification of the anatomy of the organ is also demonstrated in rats and dogs by treatment with PPARγ agonists. The transformation of white adipose tissue into brown adipose tissue in rats treated with ss3 adrenoceptor agonists is a result of direct transformation of differentiated unilocular adipocytes (Cinti, Reference Cinti2005).
Adipose tissue consists of other cells, in addition to adipocytes. For example, preadipocytes are dispersed amongst the adipocytes and are precursor cells of brown and white adipocytes. Furthermore, adipose tissue is highly vascularized and brown adipose tissue has a particularly extensive capillary network. Adipose tissue is also richly innervated with several nerves entering the various depots, and brown adipose tissue is more innervated than the white adipose tissue. Monocytes, lymphocytes, granulocytes, macrophages and dendritic cells are also present in adipose tissue, and many of the inflammatory cells can be recruited from the bone marrow with increasing size during the development of obesity (Weisberg et al. Reference Weisberg, McCann, Desai, Rosenbaum, Leibel and Ferrante2003). Communication between inflammatory cells and adipocytes, which may involve monocyte chemoattractant protein-1 (Kamei et al. Reference Kamei, Tobe, Suzuki, Ohsugi, Watanabe and Kubota2006), is being intensively investigated.
Adipose tissue functions
Adipose tissue has several functions (Table 3) and is considered to be a vital organ.
Table 3. Adipose tissue functions
From an evolutionary perspective long-term energy storage is an essential function of the adipose tissue. In a lean healthy individual the adipose tissue can provide fuel for several weeks (Table 4).
Table 4. Energy stores in man and the approximate duration of their capacity to provide energy for different activities
The adipocytes store TAG and release fatty acids by highly-regulated processes. The adipose tissue also stores cholesterol and lipid-soluble vitamins, in particular vitamins D and E (Dueland et al. Reference Dueland, Helgerud, Pedersen, Berg and Drevon1983; Bjorneboe et al. Reference Bjorneboe, Bjorneboe, Bodd, Hagen, Kveseth and Drevon1986). Adipose tissue surrounding lymph nodes may have a function in the immune system (Pond, Reference Pond2005). In addition, several non-protein hormones are produced in adipose tissue, e.g. sex steroids and glucocorticoids are produced from precursors in adipose tissue.
Brown adipose tissue is able to generate as much as 300 times more heat relative to tissue weight than other tissues. The abundance of brown adipose tissue peaks at about the time of birth and is less prominent after the postnatal period. The importance of brown adipose tissue in man is not known, but in rodents a functional brown adipose tissue is maintained throughout life (Cannon & Nedergaard, Reference Cannon and Nedergaard2004).
TAG storage
TAG are the main constituents of lipid droplets in adipocytes, and a human adipocyte typically contains about 1 μg TAG. The fatty acids originate either from plasma TAG or from de novo lipogenesis (Fig. 2). TAG circulating in plasma in lipoprotein particles are hydrolysed to fatty acids outside the adipocyte by lipoprotein lipase attached to endothelial cells in the capillaries.
Fig. 2. Metabolism of fatty acids in adipocytes. NEFA are released from chylomicrons and VLDL by the action of lipoprotein lipase and taken up into cells mainly by protein carriers in the plasma membrane and transported intracellularly via fatty acid-binding proteins (FABP). NEFA are activated (acyl-CoA) before they can be shuttled via acyl-CoA-binding protein (ACBP) to mitochondria or peroxisomes for β-oxidation (formation of energy as ATP and heat), or to endoplasmic reticulum for esterification to different lipid classes. Acyl-CoA or certain NEFA may bind to transcription factors that regulate gene expression or may be converted to signalling molecules (eicosanoids). Glucose may be transformed to fatty acids if there is a surplus of glucose or energy in the cells. TAG are stored in lipid droplets covered with lipid droplet-binding proteins (LDBP) such as perilipin. Perilipin and hormone-sensitive lipase (HSL) are activated by phosphorylation by protein kinase A. Adipose tissue TAG lipase (ATGL) hydrolyses TAG, whereas HSL hydrolyses diacylglycerols. (Modified after Rustan & Drevon, Reference Rustan and Drevon2005.)
Glucose can be utilized for de novo lipogenesis and the resulting fatty acids are stored as TAG in the lipid droplets of adipocytes after esterification with glycerol 3-phosphate. In the fed state lipogenesis is stimulated by insulin. However, lipogenesis may contribute quantitatively to TAG storage only under a shortage of fatty acids, which is not usually the case on a mixed diet with >20% energy from fat (Frayn, Reference Frayn2003).
Fatty acid release
Fatty acids are released by lipases from adipocytes into the bloodstream, e.g. during fasting. The lipases act on the surface of lipid droplets, and at least three lipases act sequentially to release three molecules fatty acids and one molecule glycerol from each molecule TAG. The first fatty acid is removed by adipose TAG lipase (Zimmermann et al. Reference Zimmermann, Strauss, Haemmerle, Schoiswohl, Birner-Gruenberger, Riederer, Lass, Neuberger, Eisenhaber, Hermetter and Zechner2004), the second by hormone-sensitive lipase and the third by monoacylglycerol lipase (Frayn, Reference Frayn2003).
When intracellular levels of cAMP increase, hormone-sensitive lipase is phosphorylated by cAMP-dependent protein kinase and translocated to the surface of the lipid droplets. Moreover, the lipases may work in concert with putative lipid droplet-coating proteins, e.g. perilipin, mannose-6-phosphate receptor-binding protein 1 and adipose differentiation-related protein. Along with hormone-sensitive lipase, perilipin is phosphorylated, thus bringing about its restructuring or relocalization (Clifford et al. Reference Clifford, Londos, Kraemer, Vernon and Yeaman2000). This process in turn may facilitate the translocation of hormone-sensitive lipase during activation and provide access to the lipid droplets (Sztalryd et al. Reference Sztalryd, Xu, Dorward, Tansey, Contreras, Kimmel and Londos2003). In the fed state insulin inhibits this process of fatty acid release and hormone-sensitive lipase is dephosphorylated.
Adipose tissue as a source of adipokines
Adipose tissue expresses several receptors that allow it to respond to signals from traditional hormones as well as the autonomous nervous system (Trayhurn et al. Reference Trayhurn, Bing and Wood2006). Moreover, the adipose tissue is the site of expression and secretion of a range of biologically-active proteins termed adipokines. Some of the most studied adipokines are leptin, adiponectin, resistin, IL and TNFα (Fig. 3).
Fig. 3. Adipokines may act locally (autocrine or paracrine) and at the systemic (endocrine) level influencing a variety of biological processes including energy metabolism with carbohydrates as well as lipids, appetite, reproduction, immune function, angiogenesis and extracellular matrix metabolism. FIAF/PGAR, fasting-induced adipose factor/PPARγ angiopoietin-related gene (FIAF, PGAR and angiopoietin like-4 are alternative terms for the same factor); IL-1Ra, IL-1 receptor antagonist; MMP-2, MMP-9, matrix metalloproteinase-2 and -9 respectively; TIMP-1, TIMP-2, tissue inhibitor of metalloproteinases-1 and -2 respectively. (Modified after Lafontan, Reference Lafontan2005.)
Thus, adipose tissue contains the biological machinery necessary for the handling of a major proportion of energy metabolism as well as communication with most other organs. The present review will focus on some aspects of resistin, adiponectin and leptin, in particular the interaction between nutrients and the expression of these adipokines.
Resistin
Steppan et al. (Reference Steppan, Bailey, Bhat, Brown, Banerjee, Wright, Patel, Ahima and Lazar2001) discovered resistin when screening for adipocyte-derived factors that may cause insulin resistance. They looked for genes in 3T3-L1 cells that were both induced during adipogenesis and down-regulated in mature adipocytes by the anti-diabetic thiazolidinedione drugs, which enhance insulin sensitivity in liver, muscle and adipose tissue, and are ligands for the PPAR nuclear receptor, which is highly expressed in adipocytes. One of the genes that matched both criteria in 3T3-L1 cells encoded a protein that was secreted by adipocytes and was found in mouse serum; it was termed resistin. The expression of resistin increases with increasing mass of adipose tissue. The resistin gene is identical to the FIZZ3 gene discovered earlier by Holcomb et al. (Reference Holcomb, Kabakoff, Chan, Baker, Gurney and Henzel2000). Although there are several observations that are compatible with a role for resistin in the development of insulin resistance in rodents as well as in man, there are data that support the notion that the major function of resistin in man may be related to inflammation (Weisberg et al. Reference Weisberg, McCann, Desai, Rosenbaum, Leibel and Ferrante2003).
Resistin like other RELM proteins comprises 105–114 amino acids and is characterized by a specific pattern made up of ten cysteine residues (Holcomb et al. Reference Holcomb, Kabakoff, Chan, Baker, Gurney and Henzel2000). It circulates in the plasma as a trimer or a hexamer (Patel et al. Reference Patel, Rajala, Rossetti, Scherer and Shapiro2004).
Only two genes of the RELM family have been identified in the human genome that share homology with their mouse and rat counterparts: resistin, which is also known as RETN; RELMβ, which is also termed RETNLB (Holcomb et al. Reference Holcomb, Kabakoff, Chan, Baker, Gurney and Henzel2000; Weisberg et al. Reference Weisberg, McCann, Desai, Rosenbaum, Leibel and Ferrante2003). In contrast to findings in mice, the human resistin gene has its highest expression in bone marrow, whereas expression in adipose tissue is much lower than that in mouse adipose tissue (Patel et al. Reference Patel, Buckels, Kinghorn, Murdock, Holbrook, Plumpton, Macphee and Smith2003). Low expression of resistin has also been observed in several tissues including lung, breast and placenta (Yura et al. Reference Yura, Sagawa, Itoh, Kakui, Nuamah, Korita, Takemura and Fujii2003; Haugen et al. Reference Haugen, Ranheim, Harsem, Lips, Staff and Drevon2006).
Obesity and resistin
There is a strong correlation between serum concentrations of resistin and resistin mRNA expression in abdominal subcutaneous adipose tissue from obese subjects (Heilbronn et al. Reference Heilbronn, Rood, Janderova, Albu, Kelley, Ravussin and Smith2004). Thus, despite the low level of expression in adipose tissue, there is a link between adipose tissue mass and circulating levels of resistin, and obesity is an important determinant of resistin levels in human subjects.
Diabetes and resistin
Serum resistin is increased in subjects with type 2 diabetes compared with subjects without diabetes (McTernan et al. Reference McTernan, Fisher, Valsamakis, Chetty, Harte, McTernan, Clark, Smith, Barnett and Kumar2003). Plasma resistin levels are also increased in patients with Cushing's syndrome (central adiposity and insulin resistance) as compared with control subjects matched for body mass (Krsek et al. Reference Krsek, Silha, Jezkova, Hana, Marek, Weiss, Stepan and Murphy2004). However, no correlation has been found between plasma resistin levels and insulin sensitivity in human subjects. Several polymorphisms in the resistin gene have been studied, some of which are located in the 5′ flanking region (G–638A, A–537C, C–420G and G–358A) and affect circulating levels of resistin (Cho et al. Reference Cho, Youn, Chung, Kim, Lee, Yu, Park, Shin and Park2004; Osawa et al. Reference Osawa, Yamada, Onuma, Murakami, Ochi and Kawata2004). The –420G genotype is specifically recognized by the transcription factors Sp1 and Sp3, leading to increased resistin promoter activity and serum resistin levels. The –420G/G genotype has been associated with type 2 diabetes (Osawa et al. Reference Osawa, Yamada, Onuma, Murakami, Ochi and Kawata2004). Plasma resistin levels of women are higher than those of men, and it is possible that the physiology of resistin differs between the genders. In a group of women who were not diabetic the –420C/C genotype was found to be associated with a higher body mass than the –420G allele carriers (Mattevi et al. Reference Mattevi, Zembrzuski and Hutz2004).
Resistin may inhibit insulin-mediated glucose uptake in human cells (Bajaj et al. Reference Bajaj, Suraamornkul, Hardies, Pratipanawatr and DeFronzo2004; McTernan et al. Reference McTernan, Fisher, Valsamakis, Chetty, Harte, McTernan, Clark, Smith, Barnett and Kumar2003). In line with animal studies treatment of subjects with type 2 diabetes with thiazolidinedione drugs causes a decrease in plasma resistin concentration, which is positively correlated with the reduction in hepatic fat content and improvement in hepatic insulin sensitivity (Bajaj et al. Reference Bajaj, Suraamornkul, Hardies, Pratipanawatr and DeFronzo2004). Furthermore, thiazolidinediones inhibit resistin expression in macrophages and adipocytes in vitro (McTernan et al. Reference McTernan, Fisher, Valsamakis, Chetty, Harte, McTernan, Clark, Smith, Barnett and Kumar2003). In summary, it is not yet clear whether resistin has a role in insulin resistance and the development of type 2 diabetes in human subjects.
Inflammation and resistin
Patients with severe inflammatory disease have elevated levels of resistin and there is a correlation between plasma resistin concentration and markers of inflammation (Stejskal et al. Reference Stejskal, Adamovska, Bartek, Jurakova and Proskova2003). Obesity is characterized by the expression of weak chronic inflammation, and this condition may cause insulin resistance. If macrophages rather than adipocytes are the main source of resistin in human subjects, resistin could be a link between insulin resistance and inflammation. Macrophages resident in adipose tissue may prove to play an important role in the development of insulin resistance (Xu et al. Reference Xu, Barnes, Yang, Tan, Yang, Chou, Sole, Nichols, Ross, Tartaglia and Chen2003). Macrophages in adipose tissue show an increase in numbers during obesity and are responsible for most of the TNFα expression in adipose tissue and substantial amounts of inducible NO synthase and IL-6 expression (Weisberg et al. Reference Weisberg, McCann, Desai, Rosenbaum, Leibel and Ferrante2003).
There have been few studies of the direct effect of fatty acids on the expression of resistin in adipocytes. As increased plasma NEFA concentrations are associated with insulin resistance (Lam et al. Reference Lam, Carpentier, Lewis, van de Werve, Fantus and Giacca2003), the effects of individual NEFA on the expression of resistin mRNA have been examined in cultured murine 3T3-L1 adipocytes (Haugen et al. Reference Haugen, Zahid, Dalen, Hollung, Nebb and Drevon2005). The NEFA tested were not found to increase resistin expression, but both arachidonic acid and EPA were shown to reduce resistin mRNA levels. Arachidonic acid was found to be by far the most potent NEFA, reducing resistin mRNA levels to approximately 20% of the control levels at concentrations of 60–250 μm. Selective inhibitors of cyclooxygenase-1 and mitogen-activated protein kinase kinase were found to counteract the arachidonic acid-induced reduction in resistin mRNA levels. Transient overexpression of sterol-regulatory element binding protein-1a was shown to activate the resistin promoter, but no reduction in the abundance of mature sterol-regulatory element binding protein-1 was found after arachidonic acid exposure. Both actinomycin D and cycloheximide were found to abolish the arachidonic acid-induced reduction of resistin mRNA levels, indicating a dependence on de novo transcription and translation. These data suggest that reductions in resistin mRNA levels involve a destabilization of the resistin mRNA molecule. An inhibitory effect of arachidonic acid and EPA on resistin expression may explain the beneficial effect on insulin sensitivity of ingesting PUFA.
Leptin
Early studies have shown that fasting and feeding enhance and reduce plasma leptin concentrations respectively (Kolaczynski et al. Reference Kolaczynski, Considine, Ohannesian, Marco, Opentanova, Nyce, Myint and Caro1996; Weigle et al. Reference Weigle, Duell, Connor, Steiner, Soules and Kuijper1997). Although it is known that plasma leptin concentrations are correlated with the amount of adipose tissue in the body, relatively little information is available on the long-term effects of diet on leptin concentrations. The Oslo Diet and Exercise Study has investigated whether changes in dietary energy sources and exercise-mediated energy expenditure, alone or in combination, affect plasma leptin concentrations (Reseland et al. Reference Reseland, Anderssen, Solvoll, Hjermann, Urdal, Holme and Drevon2001a). In a randomized, 2×2 factorial trial 186 men with enhanced risk of developing CVD were divided into four groups: diet; exercise; a combination of diet and exercise; control. Data on dietary intake, physical fitness and demographics were collected and plasma leptin concentrations were measured before and after an intervention period of 1 year. Plasma leptin concentrations, BMI and fat mass were found to be decreased in association with long-term reductions in food intake as well as increased physical activity. When values were adjusted for either BMI or fat mass a highly significant (P<0·001) reduction in plasma leptin concentration was found after both the diet intervention and the exercise intervention. No interaction was found between the interventions, suggesting a direct and additive effect of changes in diet and physical activity on plasma leptin concentrations. Thus, long-term changes in lifestyle consisting of decreased intake of dietary fat and increased physical activity reduce plasma leptin concentrations in human subjects beyond the reduction expected as a result of changes in fat mass. This finding might indicate enhanced leptin sensitivity, which might be promoted by the newly-described PTP1B (a tyrosine phosphatase) regulating body mass and adiposity primarily through actions in the brain (Bence et al. Reference Bence, Delibegovic, Xue, Gorgun, Hotamisligil, Neel and Kahn2006). Neuronal PTP1B also regulates adipocyte leptin production and is probably essential for the development of leptin resistance.
In a small dietary intervention study lasting 16 weeks (Mori et al. Reference Mori, Burke, Puddey, Shaw and Beilin2004) a daily fish meal providing 3·65 g n-3 fatty acids as part of a weight-reducing regimen was found to be more effective in reducing plasma leptin levels than either the fish meal or the weight-reducing regimen alone. Reductions in leptin were found to be related to the substantial fall in blood pressure observed with the fish meal and weight-loss intervention.
The extension of studies in the Oslo Diet and Exercise Study population to include the measurement of other adipokines and cytokines (MH Rokling-Andersen, JE Reseland, MB Veierød, SA Anderssen, DA Jacobs Jr, P Urdal, JO Jansson and CA Drevon, unpublished results) has shown that plasma adiponectin levels are higher in the subjects who have improved their diet and physical activity as compared with the controls. Plasma adiponectin levels are unaltered, whereas BMI and fat mass decrease after a reduction in energy intake and increased physical activity, whereas the adiponectin concentration is reduced in the control group. Both diet and exercise intervention result in stable plasma adiponectin levels (P<0·01 and P=0·07, respectively) and a decrease in body fat mass (P<0·001 and P<0·01, respectively), but after adjustment for changes in body fat mass no effects on adiponectin are observed (P>0·15). Thus, 1 year after implementing changes in diet and exercise plasma adiponectin levels are increased, which can be largely explained by a reduced body fat mass.
The previous studies on leptin in the Oslo Diet and Exercise Study (Reseland et al. Reference Reseland, Anderssen, Solvoll, Hjermann, Urdal, Holme and Drevon2001a) have prompted further examination of whether specific intervention with a daily supplement of 5 g marine n-3 PUFA would affect leptin levels in plasma (Reseland et al. Reference Reseland, Haugen, Hollung, Solvoll, Halvorsen, Brude, Nenseter, Christiansen and Drevon2001b). It was found that after 6 weeks plasma leptin concentrations of male smokers are unchanged. Changes in dietary intake of SFA are positively correlated with changes in plasma leptin levels, whereas for changes in the intake of PUFA the correlation is negative. Dietary intake of n-3 PUFA-enriched diet for 3 weeks, as compared with a lard-enriched diet, reduces plasma leptin concentration and leptin mRNA expression in rat epididymal adipose tissue (Reseland et al. Reference Reseland, Haugen, Hollung, Solvoll, Halvorsen, Brude, Nenseter, Christiansen and Drevon2001b). In the human trophoblast cell line (BeWo) n-3 PUFA has a dose- and time-dependent effect on leptin expression (Reseland et al. Reference Reseland, Haugen, Hollung, Solvoll, Halvorsen, Brude, Nenseter, Christiansen and Drevon2001b). Incubation with EPA and DHA (1 mm) for 72 h reduces leptin expression by 71% and 78% respectively, as compared with the control. There is no effect on the expression of the signal transducing form of the leptin receptor. In BeWo cells transfected with the human leptin promoter n-3 PUFA reduce leptin promoter activity, whereas SFA and MUFA have no effect on leptin promoter activity (Reseland et al. Reference Reseland, Haugen, Hollung, Solvoll, Halvorsen, Brude, Nenseter, Christiansen and Drevon2001b). The transcription factors PPARγ and sterol-regulatory element binding protein-1 mRNA are reduced after incubation with n-3 PUFA, whereas the expression of CCAAT/enhancer-binding protein a is unchanged. DHA-reduced leptin expression is abolished in BeWo cells grown in cholesterol-free medium. Thus, n-3 PUFA decrease leptin gene expression both in vivo and in vitro. The direct effects of PUFA on leptin promoter activity indicate a specific regulatory action of fatty acids on leptin expression.
Retinoic acid (RA) is a ligand for some nuclear receptors and its effect on the expression of leptin has been examined in adipocytes of murine and human origin (Hollung et al. Reference Hollung, Rise, Drevon and Reseland2004). After incubation of murine 3T3-L1 adipocytes with 1 and 10 mm-all-trans RA for 48 h the expression of leptin mRNA is reduced by 56% and 65% respectively, whereas the secretion of leptin to the culture medium is reduced by 38% and 77% respectively. In human adipose tissue explants incubation with 1 mm-all-trans RA for 24 h reduces leptin mRNA expression levels by 55% and leptin secretion by 25%. In 3T3-L1 cells after incubation with RA mRNA expression levels for the transcription factors PPARγ, retinoid X receptor α, and RA receptor α are increased, whereas in human adipose tissue explants mRNA levels for these transcription factors are unchanged. In two other leptin-expressing cell systems (the human placental trophoblast cell line BeWo and normal human primary osteoblasts) there is no effect of RA on leptin mRNA expression, but in BeWo cells leptin secretion is reduced by 64% after 24 h incubation with 10 mm-all-trans RA. Thus, all-trans RA reduces both the expression and secretion of leptin in human and rodent adipose tissue. In human BeWo cells or primary osteoblasts leptin mRNA expression levels are not changed by all-trans RA, indicating a tissue-specific regulation of leptin mRNA expression by all-trans RA.
Adiponectin
Structure
Adiponectin, an approximately 30 kDa polypeptide consisting of an N-terminal signal sequence followed by a variable domain, a collagen-like domain and a C-terminal globular domain, was first described by Scherer et al. (Reference Scherer, Williams, Fogliano, Baldini and Lodish1995). In the central region of the collagen-like domain there are fifteen Gly-X-Y repeats, whereas at the beginning and end of the domain there are seven Gly-X-Pro repeats. Recombinant adiponectin produced by mammalian cells has a higher biological activity than bacterially-produced adiponectin, suggesting that post-translational modifications may be important. Lysine residues in the collagen-like domain are glycosylated to yield multiple isoforms of adiponectin with relevance for biological activity (Wang et al. Reference Wang, Xu, Knight, Xu and Cooper2002).
The adiponectin superstructure resembles a bouquet of flowers, in which three protomers form a trimer that comprises a globular head domain and a collagen triple helix ‘stalk’ and six trimers form a multimer that resembles a ‘bouquet’. The smallest adiponectin complex, referred to as the low-molecular-weight form, is simply a pair of trimers. A group of larger adiponectin complexes is referred to as the high-molecular-weight form, which is a diverse group of multimers comprising two to three pairs of trimers (Cinti, Reference Cinti1999b).
Physiological conditions
Adiponectin levels display a diurnal variation, with a nocturnal decline starting in the late evening and continuing throughout the night to reach the lowest point in the early morning (Gavrila et al. Reference Gavrila, Peng, Chan, Mietus, Goldberger and Mantzoros2003). Pre- and post-menopausal women have higher plasma levels of adiponectin than men (Nishizawa et al. Reference Nishizawa, Shimomura, Kishida, Maeda, Kuriyama and Nagaretani2002). It is possible that low levels of adiponectin are related to the high risk of insulin resistance and atherosclerosis in men. A negative correlation has been observed between plasma concentrations of the adiponectin monomer and BMI in both men and women (Arita et al. Reference Arita, Kihara, Ouchi, Takahashi, Maeda and Miyagawa1999; Weyer et al. Reference Weyer, Funahashi, Tanaka, Hotta, Matsuzawa, Pratley and Tataranni2001). Plasma adiponectin levels are also reduced in adolescent obesity (Weiss et al. Reference Weiss, Dufour, Groszmann, Petersen, Dziura, Taksali, Shulman and Caprio2003). Among obese patients who have undergone gastric partition surgery body-weight reduction is accompanied by a reduction in plasma adiponectin levels (Yang et al. Reference Yang, Lee, Funahashi, Tanaka, Matsuzawa, Chao, Chen, Tai and Chuang2001). Plasma adiponectin levels are increased in anorexia nervosa (Delporte et al. Reference Delporte, Brichard, Hermans, Beguin and Lambert2003). However, in severe anorexia nervosa the adiponectin level increases gradually until BMI is about 16 kg/m2 and then decreases subsequently (Iwahashi et al. Reference Iwahashi, Funahashi, Kurokawa, Sayama, Fukuda, Okita, Imagawa, Yamagata, Shimomura, Miyagawa and Matsuzawa2003).
Subjects with type 2 diabetes have lower adiponectin levels than subjects without diabetes (Hotta et al. Reference Hotta, Funahashi, Arita, Takahashi, Matsuda and Okamoto2000; Weyer et al. Reference Weyer, Funahashi, Tanaka, Hotta, Matsuzawa, Pratley and Tataranni2001), and among subjects without diabetes there is an association between high adiponectin levels and insulin sensitivity (Tschritter et al. Reference Tschritter, Fritsche, Thamer, Haap, Shirkavand, Rahe, Staiger, Maerker, Haring and Stumvoll2003). Adiponectin mRNA levels are reduced in omental and subcutaneous adipose tissue of obese patients with type 2 diabetes compared with lean and obese subjects who are normoglycaemic (Hu et al. Reference Hu, Liang and Spiegelman1996). Also, low adiponectin levels are associated with gestational diabetes mellitus before and after it develops (Ranheim et al. Reference Ranheim, Haugen, Staff, Braekke, Harsem and Drevon2004; Retnakaran et al. Reference Retnakaran, Hanley, Raif, Connelly, Sermer and Zinman2004; Williams et al. Reference Williams, Qiu, Muy-Rivera, Vadachkoria, Song and Luthy2004).
In a study of patients with different forms of lipodystrophies (Haque et al. Reference Haque, Shimomura, Matsuzawa and Garg2002) serum adiponectin levels were found to be lower among patients with diabetes compared with subjects without diabetes. In Pima Indians plasma adiponectin is negatively associated with insulin receptor tyrosine phosphorylation in muscle biopsies (Stefan et al. Reference Stefan, Vozarova, Funahashi, Matsuzawa, Weyer, Lindsay, Youngren, Havel, Pratley, Bogardus and Tataranni2002). Prospectively, low adiponectin levels in Pima Indians are associated with a decrease in insulin sensitivity and insulin receptor tyrosine phosphorylation increases the risk of developing type 2 diabetes (Lindsay et al. Reference Lindsay, Funahashi, Hanson, Matsuzawa, Tanaka, Tataranni, Knowler and Krakoff2002; Stefan et al. Reference Stefan, Vozarova, Funahashi, Matsuzawa, Weyer, Lindsay, Youngren, Havel, Pratley, Bogardus and Tataranni2002), and this outcome is not associated with increased adiposity (Vozarova et al. Reference Vozarova, Stefan, Lindsay, Krakoff, Knowler, Funahashi, Matsuzawa, Stumvoll, Weyer and Tataranni2002). Plasma adiponectin concentrations are lower among individuals who subsequently develop type 2 diabetes than among controls (Spranger et al. Reference Spranger, Kroke, Mohlig, Bergmann, Ristow, Boeing and Pfeiffer2003).
The accumulation of lipid in muscle cells is associated with insulin resistance, and there is a negative relationship between adiponectin and intramyocellular lipid content in a paediatric population (Weiss et al. Reference Weiss, Dufour, Groszmann, Petersen, Dziura, Taksali, Shulman and Caprio2003). Adiponectin does not contribute to the exercise-related improvements in insulin sensitivity (Hulver et al. Reference Hulver, Zheng, Tanner, Houmard, Kraus, Slentz, Sinha, Pories, MacDonald and Dohm2002; Yatagai et al. Reference Yatagai, Nishida, Nagasaka, Nakamura, Tokuyama, Shindo, Tanaka and Ishibashi2003). Adiponectin is inversely correlated with abdominal visceral fat mass and insulin resistance in patients infected with HIV who are undergoing highly-active anti-retroviral therapy (Addy et al. Reference Addy, Gavrila, Tsiodras, Brodovicz, Karchmer and Mantzoros2003; Sutinen et al. Reference Sutinen, Korsheninnikova, Funahashi, Matsuzawa, Nyman and Yki-Jarvinen2003; Tong et al. Reference Tong, Sankale, Hadigan, Tan, Rosenberg, Kanki, Grinspoon and Hotamisligil2003). On the other hand, in a study of subjects with type 1 diabetes serum adiponectin levels were found to be higher than those of healthy control subjects (Imagawa et al. Reference Imagawa, Funahashi, Nakamura, Moriwaki, Tanaka and Nishizawa2002).
Subjects with coronary artery disease have lower adiponectin levels than healthy subjects (Hotta et al. Reference Hotta, Funahashi, Arita, Takahashi, Matsuda and Okamoto2000). Male patients with hypoadiponectinaemia have a 2-fold increase in prevalence of coronary artery disease, independent of well-known risk factors for coronary artery disease such as type 2 diabetes, dyslipidaemia, hypertension, smoking and increased BMI (Kumada et al. Reference Kumada, Kihara, Sumitsuji, Kawamoto, Matsumoto and Ouchi2003). Adiponectin inhibits the TNFα inflammatory response of endothelial cells by inhibiting NF-κB signalling via a cAMP-dependent pathway (Ouchi et al. Reference Ouchi, Kihara, Arita, Okamoto, Maeda and Kuriyama2000). Patients undergoing haemodialysis have elevated plasma adiponectin levels, and among these patients plasma adiponectin levels are lower in those patients who prospectively experience new cardiovascular events than in patients who are event free (Zoccali et al. Reference Zoccali, Mallamaci, Tripepi, Benedetto, Cutrupi and Parlongo2002). In women without diabetes low plasma adiponectin concentrations are associated with dyslipidaemia such as high TAG levels and low HDL-cholesterol levels in serum (Matsubara et al. Reference Matsubara, Maruoka and Katayose2002).
In post-menopausal women low levels of adiponectin are associated with higher levels of high-sensitivity C-reactive protein and IL-6, which are inflammatory mediators and markers of increased cardiovascular risk (Engeli et al. Reference Engeli, Feldpausch, Gorzelniak, Hartwig, Heintze, Janke, Mohlig, Pfeiffer, Luft and Sharma2003). Low adiponectin levels are also linked to endothelial dysfunction in human subjects (Ouchi et al. Reference Ouchi, Ohishi, Kihara, Funahashi, Nakamura and Nagaretani2003; Shimabukuro et al. Reference Shimabukuro, Higa, Asahi, Oshiro, Takasu, Tagawa, Ueda, Shimomura, Funahashi and Matsuzawa2003).
Circulating levels of adiponectin depend on synthesis in adipose tissue, as well as clearance, possibly via the kidneys. Adiponectin levels are increased in patients with nephrotic syndrome, and proteinuria is strongly related to circulating adiponectin levels in patients with nephrotic and non-nephrotic renal diseases (Zoccali et al. Reference Zoccali, Mallamaci, Panuccio, Tripepi, Cutrupi and Parlongo2003). Urinary adiponectin excretion amounts are increased in patients with nephropathy, whereas serum adiponectin levels are also elevated, perhaps as a result of compensatory enhanced synthesis (Koshimura et al. Reference Koshimura, Fujita, Narita, Shimotomai, Hosoba, Yoshioka, Kakei, Fujishima and Ito2004).
Lifestyle intervention focused on low-fat diets can reduce the risk of developing type 2 diabetes (Tuomilehto et al. Reference Tuomilehto, Lindstrom, Eriksson, Valle, Hamalainen and Ilanne-Parikka2001). Dietary composition as part of overall lifestyle may affect abnormalities in pregnancy such as gestational diabetes mellitus and pre-eclampsia. Increased fat intake during pregnancy has been associated with the development of impaired glucose tolerance and gestational diabetes mellitus in human subjects (Saldana et al. Reference Saldana, Siega-Riz and Adair2004) and in rats (Holemans et al. Reference Holemans, Caluwaerts, Poston and Van Assche2004). Observational studies of pre-eclampsia suggest that various nutrients are associated with pre-eclampsia. For example, high intake of energy, sucrose and PUFA may be involved in the development of the disorder (Clausen et al. Reference Clausen, Slott, Solvoll, Drevon, Vollset and Henriksen2001). However, in a recent review (Roberts et al. Reference Roberts, Balk, Bodnar, Belizan, Bergel and Martinez2003) it is argued that targets for nutritional investigation based on the current knowledge of pathophysiology are warranted.
Adipokines may link the adipose tissue and reproductive function (Budak et al. Reference Budak, Fernandez Sanchez, Bellver, Cervero, Simon and Pellicer2006). One possible role of adiponectin is to ensure energy supply and to regulate energy needs for normal reproduction and pregnancy. Cord blood adiponectin levels are much higher than the serum levels in children and adults, and are positively correlated with fetal birth weights (Sivan et al. Reference Sivan, Mazaki-Tovi, Pariente, Efraty, Schiff, Hemi and Kanety2003). To address this hypothesis, adiponectin and other adipokines have been studied in various pathological states during pregnancy (Ranheim et al. Reference Ranheim, Haugen, Staff, Braekke, Harsem and Drevon2004; Haugen et al. Reference Haugen, Ranheim, Harsem, Lips, Staff and Drevon2006).
A recent study (Haugen et al. Reference Haugen, Ranheim, Harsem, Lips, Staff and Drevon2006) has investigated whether adipokine levels are altered in pre-eclampsia and whether insulin sensitivity is affected. Maternal plasma concentrations of adiponectin, resistin and leptin and their mRNA expression were monitored in the abdominal adipose tissue and placenta from two groups of patients undergoing caesarean section: (1) women with pre-eclampsia; (2) healthy pregnant women (control group). Compared with the control group the women with pre-eclampsia were found to have higher concentrations of several adipokines: adiponectin, 50%; resistin, 22%; leptin, 52%. Similar mRNA levels of adiponectin, resistin and leptin were found in abdominal subcutaneous adipose tissue in the two groups. Moreover, resistin mRNA levels in the placenta were not found to be different between the groups, whereas leptin mRNA levels were found to be higher in placentas from women with pre-eclampsia compared with the controls. Thus, increased plasma concentrations of adipokines in pre-eclampsia may not relate to altered expression levels in adipose tissue. In contrast to resistin, leptin mRNA levels in the placenta were found to be increased in pre-eclampsia.
A comparison of women with gestational diabetes mellitus and healthy pregnant women (Ranheim et al. Reference Ranheim, Haugen, Staff, Braekke, Harsem and Drevon2004) has shown that plasma adiponectin concentrations among lean control subjects are 51% higher than those of corresponding individuals with gestational diabetes mellitus. In line with this observation adiponectin mRNA levels in abdominal subcutaneous adipose tissue were found to be higher in healthy pregnant women as compared with women with gestational diabetes.
Dietary factors may influence the expression of adipokines in vivo as well as in vitro. Supply of total energy, total fat, different types of fatty acids and eicosanoids, RA and nicotine linked to the adrenergic hormonal response may all influence the expression of adipokines in adipose tissue, as well as the plasma concentration of adipokines. Thereby dietary factors can change many biological systems with extensive physiological consequences.
Altered levels of, or sensitivity to, adipokines may alter appetite, glucose tolerance, fatty acid oxidation and angiogenesis (Iversen et al. Reference Iversen, Drevon and Reseland2002; Reseland et al. Reference Reseland, Syversen, Bakke, Qvigstad, Eide, Hjertner, Gordeladze and Drevon2001c).
