Adipose tissue in obesity

AT comprises mature adipocytes and non-fat stromovascular cells including fibroblasts, endothelial cells, pre-adipocytes and tissue resident macrophages⁽Reference Maury and Brichard¹³⁾. Originally thought of as connective tissue where excess energies were stored as TAG, in the last decade, the endocrine function of AT has been demonstrated and it is now known to produce numerous products collectively called ‘adipokines’ involved in energy metabolism, inflammatory response and cardiovascular activity⁽Reference Fantuzzi^14, Reference Lago, Dieguez and Gomez-Reino¹⁵⁾. The two major adipokines secreted are leptin and adiponectin.

Leptin is a protein product of the ob gene initially known for its role in regulating appetite⁽Reference Friedman¹⁶⁾. In obesity, circulating levels are increased in proportion to fat mass⁽Reference Considine, Sinha and Heiman^17, Reference Maffei, Halaas and Ravussin¹⁸⁾, possibly due to increased expression of the ob gene and elevated leptin secretion⁽Reference Hamilton, Paglia and Kwan^19, Reference Lonnqvist, Arner and Nordfors²⁰⁾. Women have higher-serum leptin concentrations than men regardless of percent body fat or fat mass⁽Reference Saad, Damani and Gingerich^21, Reference Ostlund, Yang and Klein²²⁾, gender differences also seen in obese children⁽Reference Hassink, Sheslow and de Lancey²³⁾. Leptin shares homology with some pro-inflammatory cytokines and is indeed regarded as such; the leptin receptor is expressed by immune cells⁽Reference Otero, Lago and Lago^24, Reference Sanchez-Margalet, Martin-Romero and Santos-Alvarez²⁵⁾ and in human subjects circulating levels of leptin are positively associated with those of C-reactive protein, a marker of inflammation⁽Reference Sugiura, Tamakoshi and Yatsuya²⁶⁾. Adiponectin is regarded as anti-inflammatory and also regulates insulin sensitivity⁽Reference Fantuzzi¹⁴⁾. Normal-weight individuals have high circulating levels of adiponectin; however, this decreases as adiposity increases⁽Reference Arita, Kihara and Ouchi²⁷⁾ and levels also negatively correlate with insulin resistance, T2DM and CVD⁽Reference Tilg and Moschen²⁸⁾. Adiponectin synthesis is increased by weight loss and thiazolidinediones, used in the treatment of T2DM, conversely TNFα and IL-6 reduce adiponectin synthesis⁽Reference Bruun, Lihn and Verdich^29, Reference Maeda, Takahashi and Funahashi³⁰⁾. In human subjects, there is an inverse correlation between plasma C-reactive protein levels and adiponectin⁽Reference Sugiura, Tamakoshi and Yatsuya^26, Reference Ouchi, Kihara and Funahashi^31, Reference Matsushita, Yatsuya and Tamakoshi³²⁾. Furthermore, in mice, adiponectin gene knockdown results in diet-induced insulin resistance⁽Reference Maeda, Shimomura and Kishida³³⁾ and an increased inflammatory response to dextran sulfate sodium-induced colitis⁽Reference Nishihara, Matsuda and Araki³⁴⁾ and tissue ischaemia⁽Reference Shibata, Sato and Pimentel³⁵⁾.

Chronic excessive energy intake leads to AT expansion through both hypertrophy and hyperplasia, and in adults hypertrophy seems to dominate⁽Reference Spalding, Arner and Westermark^36, Reference van Harmelen, Skurk and Rohrig³⁷⁾. In some adults, however, hyperplasia can be predominate, which is thought to result in the development of a metabolically benign form of obesity⁽Reference Stefan, Kantartzis and Machann³⁸⁾. Dysregulated secretion of adipokines occurs as obesity develops and there is a general shift in adipokine and cytokine production towards a pro-inflammatory composition. This is thought to be mediated by the activation of tissue-resident macrophages in addition to a significant infiltration of other immunocytes, predominantly macrophages⁽Reference Zeyda and Stulnig³⁹⁾ although the triggering factors for this are, as yet, unclear. However, the main factors implicated in the initiation of inflammation and oxidative stress in obesity are thought to be metabolic endotoxaemia, increased plasma NEFA, hypertrophic adipocytes and increased AT hypoxia⁽Reference Maury and Brichard¹³⁾ (Fig. 1). These factors do not develop in isolation and there are most likely mechanisms involved which act to feed-forward these processes.

Fig. 1. The development of pathogenic adipose tissue in obesity is a complex interplay of many factors. LPS, lipopolysaccharide; ER, endoplasmic reticulum; TLR, Toll-like receptors.

Metabolic endotoxaemia

It is now well known that there is a post-prandial inflammatory response, especially following meals high in fat and carbohydrate⁽Reference Margioris⁴⁰⁾. Macronutrient intake has also been shown to induce pro-inflammatory changes in the immunocytes of normal-weight subjects, a response that is greater and more prolonged in the obese⁽Reference Patel, Ghanim and Ravishankar⁴¹⁾, and which is not seen following consumption of a meal rich in fruit and fibre⁽Reference Ghanim, Abuaysheh and Sia⁴²⁾. Recent research suggests that this may be due to the presence of Toll-like receptor (TLR) stimulants, such as bacterial lipopeptides and lipopolysaccharides (LPS), which have been found primarily in meat and processed food products even though they are fit for consumption⁽Reference Erridge⁴³⁾. Microbiota of the gut has recently been identified as a triggering factor of inflammation in obesity, and indeed for the development of obesity itself. Germ-free mice, which contain no gut microbiota, have 42% less total body fat than conventionally raised mice; however, re-colonisation with microbiota from normal mice, increased their total body fat content by 57%⁽Reference Backhed, Ding and Wang⁴⁴⁾. It is thought that microbiota aid the absorption of monosaccharides and induce hepatic lipogenesis, promoting energy storage with a subsequent increase in adiposity⁽Reference Backhed, Ding and Wang⁴⁴⁾.

In trying to further identify the causative factors for post-prandial inflammation, research indicates that it may be absorption of the endotoxin LPS, released during the death of Gram-negative bacteria within the gut; this is also proposed to have a role in the development of obesity and inflammation⁽Reference Cani, Amar and Iglesias⁴⁵⁾. In mice, a high-fat or high-carbohydrate meal increases plasma endotoxin levels⁽Reference Cani, Amar and Iglesias⁴⁵⁾. Similarly, in human subjects, consumption of a high-fat meal has been found to increase plasma levels of endotoxin⁽Reference Ghanim, Abuaysheh and Sia^42, Reference Erridge, Attina and Spickett⁴⁶⁾ and positive correlations have been found between energy or fat intake and plasma endotoxin levels in healthy men⁽Reference Amar, Burcelin and Ruidavets⁴⁷⁾. One study has further shown there to be a difference between fatty acids in increasing plasma LPS concentrations. Genetically obese JCR:LA-cp (James C Russell corpulent) rat fed a diet containing either 5% or 10% PUFA had significantly lower levels of LPS-binding protein than those on the isoenergetic, lipid-balanced control diets⁽Reference Hassanali, Ametaj and Field⁴⁸⁾. It would be interesting to elucidate if PUFA have similar properties in human subjects.

This significant increase in plasma LPS in response to diet composition has been termed as ‘metabolic endotoxaemia’⁽Reference Cani, Amar and Iglesias⁴⁵⁾. Plasma levels of LPS-binding protein, a marker of endotoxaemia, are significantly higher in obese human subjects and are associated with metabolic syndrome and T2DM⁽Reference Sun, Yu and Ye⁴⁹⁾. Research in mice has found that a high-fat or high-carbohydrate diet increases the Gram-negative bacteria population of the gut along with increasing plasma levels of LPS⁽Reference Cani, Amar and Iglesias⁴⁵⁾. Conversely, modulation of gut flora by antibiotics reduced metabolic endotoxaemia in diet-induced and the genetically obese ob/ob mice⁽Reference Cani, Bibiloni and Knauf⁵⁰⁾. The binding of LPS to TLR4 on enterocytes triggers its phagocytosis and it is subsequently translocated across the intestinal barrier and into the lymph; from there LPS is packaged into chylomicrons along with dietary lipids and transported around the body⁽Reference Neal, Leaphart and Levy⁵¹⁾. Indeed, one study conducted in fasting normal-weight, healthy individuals has found that while intake of water, orange juice or a glucose drink has no effect on plasma endotoxin concentrations, consumption of cream increased plasma LPS levels⁽Reference Deopurkar, Ghanim and Friedman⁵²⁾. It has also been shown in the genetically obese mouse models, ob/ob and db/db, that high-fat feeding and LPS can increase intestinal wall permeability, elevating endotoxin levels in the portal circulation⁽Reference Cani, Bibiloni and Knauf^50, Reference Brun, Castagliuolo and Leo⁵³⁾. Recent research suggests that it is gut microbiota that mediate this effect, by modulating the intestinal endocannabinoid system⁽Reference Muccioli, Naslain and Backhed⁵⁴⁾. In obese mice, blockade of the endocannabinoid receptor, CB₁ led to a reduction in gut permeability with a concomitant reduction in plasma LPS, which was also seen following prebiotic administration⁽Reference Muccioli, Naslain and Backhed⁵⁴⁾.

Although it has recently been shown that tea catechins such as epigallocatechin-3-O-gallate (EGCG)⁽Reference Korte, Dreiseitel and Schreier⁵⁵⁾, anthocyanins and anthocyanidins⁽Reference Korte, Dreiseitel and Schreier⁵⁶⁾ have a moderate affinity for endocannabinoid receptors at concentrations <50 μm, there is, as yet, no research into their ability to modulate gut permeability and thereby reduce plasma LPS levels. Furthermore, limited research has been conducted investigating their ability to limit the increases in LPS found following a high-energy meal. A small study using normal-weight, healthy subjects investigated plasma changes in endotoxin levels, inflammation and oxidative stress following consumption of a high-fat, high-carbohydrate meal with water, a glucose drink or orange juice⁽Reference Ghanim, Sia and Upadhyay⁵⁷⁾. Consumption of orange juice abrogated the increase in plasma endotoxin, inflammation and oxidative stress which was seen with glucose drink and water intake⁽Reference Ghanim, Sia and Upadhyay⁵⁷⁾. In a similar study, healthy, normal-weight individuals were given a nutritional supplement containing resveratrol (major polyphenol of red-wine grapes) and muscadine grape polyphenols along with a high-fat, high-carbohydrate meal. The supplement ameliorated the increase in plasma LPS-binding protein and endotoxin concentrations associated with intake of such a meal, along with reducing monocytic expression of TLR4⁽Reference Ghanim, Sia and Korzeniewski⁵⁸⁾. The results of these studies indicate the merit of further research into this area and due to the extended post-prandial inflammatory response seen in obesity, should be extended to include this population.

TLR are transmembrane proteins that recognise microbial components such as bacterial LPS and flagellin and also possibly endogenous ligands released during inflammation⁽Reference Shizuo, Kiyoshi and Tsuneyasu⁵⁹⁾. TLR4 is thought to mediate cellular response to LPS, and TLR2 is regarded as the main receptor for Gram-positive bacterial and fungal cell well components⁽Reference Lin, Lee and Berg⁶⁰⁾; however, LPS activation of TLR4 on 3T3-L1 murine adipocytes has been found to induce the expression of TLR2, leading to an increase in IL-6 gene expression⁽Reference Lin, Lee and Berg⁶⁰⁾. TLR expression has been observed in AT and adipocytes from normal-weight and obese human subjects and these are responsive to LPS, resulting in the activation of NF-κB signalling, which regulates immune and inflammatory responses, and subsequent cytokine release⁽Reference Vitseva, Tanriverdi and Tchkonia^61–Reference Bes-Houtmann, Roche and Hoareau⁶³⁾. In mice fed a high-fat diet, knockdown of TLR ameliorated the increase in adiposity, reducing serum insulin concentrations and circulating markers of inflammation, although the mechanisms behind this are not fully understood⁽Reference Davis, Gabler and Walker-Daniels⁶⁴⁾.

The bioactive xanthones of the mangosteen fruit, α- and γ-mangostin, have recently been found to abrogate the activation of NF-κB signalling following incubation with LPS, with a concomitant reduction in pro-inflammatory gene expression in human adipocytes, ex vivo ⁽Reference Bumrungpert, Kalpravidh and Chitchumroonchokchai⁶⁵⁾. Furthermore, α-tocopherol treatment of 3T3-L1 cells stimulated with LPS abrogated production of the cytokine, IL-6⁽Reference Lira, Rosa and Cunha⁶⁶⁾. Plant compounds have also been found to reduce LPS-activation of immunocytes. Stimulation of the human acute monocytic leukemia cell line THP-1 with LPS increased the expression and secretion of TNFα and IL-6, which was significantly reduced by treatment with the common polyphenol, cyanidin 3-O-β-glucoside⁽Reference Zhang, Lian and Zhu⁶⁷⁾. EGCG is also known to be anti-inflammatory and abrogates the macrophage inflammatory response to LPS. Recent research suggests that this is mediated by its ability to inhibit TLR signalling by down-regulating TLR4 expression⁽Reference Hong Byun, Fujimura and Yamada⁶⁸⁾.

NEFA

Elevated plasma NEFA are a feature of obesity thought to result from increased AT basal lipolytic activity⁽Reference Nielsen, Guo and Johnson⁶⁹⁾, which is in turn due to increased adipocyte size⁽Reference Arner⁷⁰⁾ and cellular hypoxia⁽Reference Yin, Gao and He⁷¹⁾. Historically, increased plasma NEFA have been connected with obesity⁽Reference Opie and Walfish⁷²⁾ and T2DM⁽Reference Fraze, Donner and Swislocki⁷³⁾ although further investigations have failed to find any significant associations⁽Reference Charles, Fontbonne and Thibult⁷⁴⁾. Regardless, there is much research to indicate the ability of NEFA to impair the insulin responsiveness of tissues including the muscle and liver, which is thought to result from the inhibition of GLUT translocation caused by the intracellular accumulation of fatty acid metabolites⁽Reference Samuel, Petersen and Shulman⁷⁵⁾. The polyphenolic compounds, EGCG⁽Reference Zhang, Li and Liang⁷⁶⁾ and naringenin⁽Reference Zygmunt, Faubert and MacNeil⁷⁷⁾, have been shown to increase GLUT translocation in rat L6 skeletal muscle cells, thereby enhancing glucose uptake; this has also been found in 3T3-L1 adipocytes treated with the phenolic acid, gallic acid, derived from seabuckthorn⁽Reference Vishnu Prasad, Anjana and Banerji⁷⁸⁾. Conversely, in isolated rat adipocytes, the polyphenols catechin-gallate, myricetin and quercetin, widely present in fruits and vegetables, have been shown to directly interact with GLUT4, reducing glucose transport⁽Reference Strobel, Allard and Perez-Acle⁷⁹⁾; however, this has not been replicated in vivo. In a fructose-fed rat model of insulin resistance, AT concentrations of GLUT4 were reduced by 58% compared to the control group; however, in those supplemented with green-tea powder, GLUT4 levels were not significantly different from the controls (19%) thereby increasing insulin sensitivity⁽Reference Wu, Juan and Hwang⁸⁰⁾. Further to this, the induction of insulin resistance in rats by 48 h intravenous infusion of a TAG emulsion, inhibited the translocation of GLUT4 in both adipose and skeletal muscle tissue. However, this was abrogated in both tissues with the co-injection of 10 mg/kg EGCG; additionally there was a concomitant improvement in insulin-induced glucose uptake⁽Reference Li, Zhao and Zhang⁸¹⁾.

Recently it has been found that normal fasting NEFA levels are maintained in abdominally obese men through a reduction in lipolytic enzyme expression and therefore reduced fasting state NEFA release⁽Reference McQuaid, Hodson and Neville⁸²⁾. However, it was further found that post-prandial fatty acid uptake by AT was significantly reduced, which is attributed to a down-regulation of fat storage mechanisms and resulted in a substantial increase in plasma NEFA following a meal⁽Reference McQuaid, Hodson and Neville⁸²⁾. This is proposed to lead to an increase in ectopic fat deposition; however, this requires further research⁽Reference McQuaid, Hodson and Neville⁸²⁾. Phytochemicals such as luteolin, a flavonoid found in olive oil and carrots⁽Reference Liu, Ma and Wang⁸³⁾, resveratrol⁽Reference Shang, Chen and Xiao⁸⁴⁾ and theaflavins found in black tea⁽Reference Lin, Huang and Lin⁸⁵⁾ have been found to limit lipid accumulation in human liver HepG2 cells. This has also been found with curcumin treatment of human Hep3B cells⁽Reference Kim, Davis and Zhang⁸⁶⁾ and EGCG treatment of primary mouse hepatocytes⁽Reference Collins, Liu and Pi⁸⁷⁾. The proposed mechanism of action is thought to be activation of AMP-activated protein kinase α signalling pathway, which subsequently inhibits fatty acid and cholesterol synthesis and up-regulates fatty acid oxidation and glycolysis, thereby regulating cellular energy balance⁽Reference Carling⁸⁸⁾.

There are a number of animal studies indicating the benefits of phyotchemicals in reducing liver lipotoxicity. Supplementation of the genetically obese KK/A^y mice with rhaponticin, extracted from rhubarb, reduced plasma NEFA and TAG levels in addition to preventing the development of liver steatosis⁽Reference Chen, Ma and Lu⁸⁹⁾. Theaflavin treatment of rats fed a high-fat diet ameliorated the development of liver steatosis with a concomitant reduction in plasma NEFA and TAG, which was shown to be mediated through activation of AMP-activated protein kinase signalling⁽Reference Lin, Huang and Lin⁸⁵⁾. In a rat model of liver steatosis, where treatment with carbon tetrachloride induces liver fat accumulation and hepatic cell death, prior feeding with an apricot extract significantly attenuated this⁽Reference Ozturk, Gul and Ates⁹⁰⁾. Olive leaf extract has also been shown to attenuate the development of liver steatosis in a high-fat fed rat model of obesity, with a concomitant reduction in plasma TAG; however, circulating NEFA levels were not affected⁽Reference Poudyal, Campbell and Brown⁹¹⁾. Resveratrol supplementation of rats with diet-induced liver steatosis significantly reduced fat deposition in the liver⁽Reference Bujanda, Hijona and Larzabal⁹²⁾. None of these studies investigated alterations in adipose or muscle tissue gene expression of lipid-processing mechanisms, which given the recent findings in obese human subjects⁽Reference McQuaid, Hodson and Neville⁸²⁾ would be pertinent. Furthermore, there are, as yet, few studies investigating their effects in human subjects.

Characterisation of fasting plasma NEFA in healthy, normal-weight individuals has found that SFA comprise just over half (56%) of the total NEFA concentration, and of this fraction palmitate predominates⁽Reference Staiger, Staiger and Stefan⁹³⁾. Further research has found that consumption of a saturated NEFA (sNEFA)-rich diet by abdominally obese participants for 8 weeks resulted in an increase in AT inflammatory gene expression compared to those on a MUFA-rich diet⁽Reference van Dijk, Feskens and Bos⁹⁴⁾. In cell culture, sNEFA have been shown to induce release of pro-inflammatory cytokines from macrophages⁽Reference Laine, Schwartz and Wang^95, Reference Haversen, Danielsson and Fogelstrand⁹⁶⁾ and in adipocytes, increase reactive oxygen species generation and TNFα, IL-6 and monocyte chemotactic protein-1 gene expression along with reducing adiponectin gene expression⁽Reference Ajuwon and Spurlock^97–Reference Takahashi, Yamaguchi and Shimoyama⁹⁹⁾. TNFα has been found to decrease adipocyte insulin sensitivity, which increases lipolysis and therefore plasma levels of NEFA⁽Reference Zhang, Halbleib and Ahmad^100, Reference Permana, Menge and Reaven¹⁰¹⁾. Even short-term exposure of adipocytes to TNFα stimulates lipolysis through increased inducible nitric oxide synthase expression, which in turn up-regulates nitric oxide production and subsequently phosphorylates and activates hormone-sensitive lipase⁽Reference Lien, Au and Tsai¹⁰²⁾. Further research has found that immunocyte response to LPS, mediated by TLR4 activation, is approximately three-fold higher in those pre-treated with sNEFA than those treated with either sNEFA or LPS alone⁽Reference Schwartz, Zhang and Karnik¹⁰³⁾. A positive feedback loop is proposed whereby sNEFA released from adipocytes activate macrophages, stimulating further release of TNFα and IL-6 and these in turn cause inflammatory changes in adipocytes⁽Reference Suganami, Nishida and Ogawa¹⁰⁴⁾.

Studies in vitro have shown that sNEFA activate TLR2 and TLR4 on adipocytes and macrophages triggering c-Jun N-terminal kinase and NF-κB pro-inflammatory signalling cascades resulting in pro-inflammatory cytokine production⁽Reference Shi, Kokoeva and Inouye^105–Reference Han, Kargi and Omer¹⁰⁸⁾. However, recent research suggests that it may be LPS contamination of the fatty acid-free bovine serum albumin, a reagent commonly used to complex with sNEFA in in vitro studies, which stimulates TLR4 as uncomplexed sNEFA had no effect on TLR-dependent signalling⁽Reference Erridge and Samani¹⁰⁹⁾. Although this is a matter that requires further clarification, numerous cell and animal studies do show that sNEFA are able to elicit a pro-inflammatory response; furthermore, these low levels of LPS may better represent the in vivo milieu⁽Reference Kennedy, Martinez and Chuang¹¹⁰⁾. There are, as yet, limited studies investigating the inhibitory effects of phytochemicals on the pro-inflammatory response of macrophages or adipocytes to sNEFA. However, as mentioned previously, studies in both adipocytes and macrophages have shown that phytochemicals, such as α- and γ-mangostin⁽Reference Bumrungpert, Kalpravidh and Chitchumroonchokchai⁶⁵⁾, α-tocopherol⁽Reference Lira, Rosa and Cunha⁶⁶⁾ and EGCG⁽Reference Hong Byun, Fujimura and Yamada⁶⁸⁾, are able to inhibit signalling pathways downstream of LPS-mediated TLR activation, ameliorating pro-inflammatory gene expression.

Adipocyte hypertrophy

Hypertrophic adipocytes have been associated with hypertension⁽Reference Berglund, Ljungman and Hartford¹¹¹⁾ and an increased risk of CVD⁽Reference Lonn, Mehlig and Bengtsson¹¹²⁾, and further research in human subjects and mice has found that adipocyte size positively correlates with degree of AT inflammation⁽Reference Cinti, Mitchell and Barbatelli^113–Reference Lee, Nair and Rousseau¹¹⁶⁾. Hypertrophic adipocytes isolated from human subjects have been found to have altered gene expression⁽Reference Jernas, Palming and Sjoholm¹¹⁷⁾ leading to increased secretion of pro-inflammatory factors and altered adipokine secretion⁽Reference Skurk, Alberti-Huber and Herder^118–Reference Gustafson, Hammarstedt and Andersson¹²⁰⁾. Anti-obesity research has investigated the ability of nutraceuticals to promote adipocyte apoptosis and inhibit adipocyte differentiation, thereby reducing AT accumulation. The two most well-studied transcription factors regulating adipogenesis are PPARγ and cytidine-cytidine-adenosine-adenosine-thymidine-enhancer-binding proteins (C/EBP)⁽Reference Rosen and MacDougald¹²¹⁾. PPARγ is considered to be essential and adequate for the induction of adipocyte differentiation and has two isoforms, PPARγ1 and PPARγ2, whose distinct roles are yet to be elucidated. Members of the C/EBP family are either anti- or pro-adipogenic and include C/EBPα, C/EBPβ, C/EBPγ and C/EBPδ. Their expression is necessary for the development of adipocyte insulin sensitivity⁽Reference Wu, Rosen and Brun¹²²⁾ and while they are important for adipogenesis, the presence of PPARγ is still required.

The anti-adipogenic potential of various natural compounds has been investigated and many have been found to inhibit adipocyte differentiation. The addition of curcumin, the major polyphenol found in turmeric, to rodent models of obesity has been shown to reduce body weight gain and ameliorate the development of diabetes⁽Reference Weisberg, Leibel and Tortoriello^123, Reference Ejaz, Wu and Kwan¹²⁴⁾. This is thought to be due to its ability to suppress adipocyte differentiation and indeed this has been found in studies using the murine 3T3-L1 cell line, with a concomitant down-regulation of PPARγ⁽Reference Zhao, Sun and Ye^125, Reference Lee, Lee and Hwang¹²⁶⁾; however, in another study utilising 3T3-L1 cells, these transcriptional markers were not affected⁽Reference Ejaz, Wu and Kwan¹²⁴⁾. A recent comprehensive investigation into the ligand activity of curcumin found it to have no binding activity at the PPARγ ligand-binding site and subsequently no effect on adipocyte differentiation⁽Reference Narala, Smith and Adapala¹²⁷⁾. These disparities in curcumin activity are most likely related to the different extractions used as some studies have utilised purified curcumin, while others have prepared ethanolic extracts of turmeric which may contain other active compounds.

The green-tea polyphenol, EGCG, induces cancer cell apoptosis and has been shown to increase weight loss and reduce fat accumulation in both human⁽Reference Wolfram, Wang and Thielecke¹²⁸⁾ and rodent obesity studies⁽Reference Wolfram, Raederstorff and Wang^129, Reference Bose, Lambert and Ju¹³⁰⁾. In one study, administration of EGCG to the diets of high-fat fed rats abrogated the development of glucose intolerance with a concomitant increase in PPARγ gene expression; however, contrastingly, administration of green tea suppressed PPARγ gene expression⁽Reference Chen, Bezzina and Hinch¹³¹⁾. In 3T3-L1 cells, investigations have attributed the anti-obesity effects of EGCG to its ability to induce adipocyte apoptosis and inhibit adipocyte differentiation and proliferation, through down-regulation of PPARγ and C/EBPα⁽Reference Furuyashiki, Nagayasu and Aoki^132–Reference Lin, Della-Fera and Baile¹³⁴⁾. The concentrations used in these studies ranged from 0 to 200 μm; however, one study found that EGCG administration at 0·5–10 μm, while having no effect on cell activity, enhanced expression of genes related to adipocyte differentiation and insulin sensitivity and reduced fat accumulation⁽Reference Sakurai, Mochizuki and Kameji¹³⁵⁾. In human subjects, studies investigating the bioavailability of green-tea polyphenols following consumption of green-tea solids in water have found the maximum plasma concentration of EGCG to be <1 μg/ml⁽Reference Lee, Maliakal and Chen^136, Reference Yang, Chen and Lee¹³⁷⁾; therefore, studies carried out in cells using lower levels of EGCG may better mimic physiological levels.

However, is the inhibition of adipocyte differentiation truly beneficial? Recent research has indicated that in the moderately obese (BMI 26–36 kg/m²), an increase in the number of smaller adipocytes may also contribute to AT inflammation. In the study, abdominal AT biopsies were taken from healthy, moderately obese individuals and analysed for cell size distribution and inflammatory gene expression⁽Reference McLaughlin, Deng and Yee¹³⁸⁾. Adipocyte size was not found to be associated with inflammatory gene expression, instead an increase in the proportion of smaller adipocytes predicted the expression of inflammatory genes, which was independent from sex, insulin resistance and BMI, although this association was stronger in insulin-resistant than insulin-sensitive individuals⁽Reference McLaughlin, Deng and Yee¹³⁸⁾. A total of eight inflammatory genes were analysed and of these CD14 and CD45 are specific for monocyte lineage cells, suggesting the acquisition of a monocytic phenotype by the smaller adipocytes. Previous research has found that pre-adipocytes acquire functional properties similar to macrophages when cultured in contact with one another and their gene profile has been found to be closer to macrophages than adipocytes⁽Reference Charriere, Cousin and Arnaud¹³⁹⁾. As outlined in Fig. 2, there are many factors in pathogenic AT which may affect the normal differentiation of pre-adipocytes; this then results in the development of a pro-inflammatory, macrophage-like phenotype as indicated by the elevated secretion of cytokines and chemokines⁽Reference Isakson, Hammarstedt and Gustafson^140, Reference Lacasa, Taleb and Keophiphath¹⁴¹⁾, most likely by activation of the key transcriptional regulator of inflammation, NF-κB⁽Reference Lacasa, Taleb and Keophiphath¹⁴¹⁾. Further research in subcutaneous AT biopsies obtained from obese individuals suggests that there is a failure of pre-adipocytes to differentiate⁽Reference Isakson, Hammarstedt and Gustafson^140, Reference Lacasa, Taleb and Keophiphath¹⁴¹⁾, possibly caused by elevated levels of TNFα⁽Reference Isakson, Hammarstedt and Gustafson¹⁴⁰⁾.

Fig. 2. Factors contributing to conversion of pre-adipocyte to one with macrophage-like activity. LPS, lipopolysaccharide.

LPS, high circulating levels of which are found in the obese or those following a high-fat meal, has also been found to affect adipocyte differentiation. Pre-adipocytes are known to have a macrophage-like phenotype and have a high gene expression and secretion of pro-inflammatory cytokines⁽Reference Cousin, Andre and Casteilla^142, Reference Cousin, Munoz and Andre¹⁴³⁾ in response to TNFα and LPS⁽Reference Charriere, Cousin and Arnaud^139, Reference Chung, LaPoint and Martinez¹⁴⁴⁾. IL-6 expression levels were found to be higher in LPS-stimulated 3T3-L1 pre-adipocytes than in mature 3T3-L1 adipocytes⁽Reference Harkins, Moustaid-Moussa and Chung¹⁴⁵⁾, and continuous treatment of these cells with LPS impaired their differentiation⁽Reference Poulain-Godefroy and Froguel¹⁴⁶⁾. Furthermore, in primary cultures of human adipocytes, LPS stimulation up-regulated pro-inflammatory cytokine mRNA, predominantly in the pre-adipocyte fraction and, as differentiation into mature adipocytes occurred pro-inflammatory cytokine expression decreased⁽Reference Chung, LaPoint and Martinez¹⁴⁴⁾. Moreover, the therapeutic effects of thiazolidinediones used in the treatment of diabetes have been attributed, in part, to their ability to increase adipocyte differentiation⁽Reference Lehmann, Moore and Smith-Oliver¹⁴⁷⁾. Treatment of genetically obese Zucker rats with the thiazolidinedione, troglitazone normalised hyperglycaemia and hyperinsulinaemia in addition to reducing AT levels of TNFα and leptin⁽¹⁾. This was concomitant with an increase in the number of small adipocytes and a reduced number of large adipocytes, without alterations to the total weight of white AT⁽Reference Okuno, Tamemoto and Tobe¹⁴⁸⁾.

Supplementation of diet-induced obese rats with bitter melon extract powder or the thiazolidinedione, pioglitazone, prevented the development of hyperinsulinaemia and glucose intolerance⁽Reference Huang, Hong and Wong¹⁴⁹⁾. The number of large adipocytes (>180 μm) was significantly lower in these two groups compared with those on the high-fat diet alone. In addition, the number of smaller adipocytes (60–100 μm) in the bitter melon group was similar to those fed the low-fat diet, which was significantly higher than the high-fat and thiazolidinedione-treated groups. Bitter melon powder had the added effect of reducing AT mass and lipid content, suggesting that lipogenesis within AT was also attenuated⁽Reference Huang, Hong and Wong¹⁴⁹⁾. In vitro studies have identified further compounds able to enhance adipocyte differentiation. Phloretin, a flavonoid found in apples and strawberries, was found to increase TAG accumulation with an attendant up-regulation of PPARγ and C/EBPα, concomitantly increases in adiponectin expression and secretion were also found⁽Reference Hassan, Yazidi and Landrier¹⁵⁰⁾. Found in citrus fruits, the flavone nobiletin has been shown to enhance 3T3-L1 differentiation; however, this was not due to any activity at PPARγ. Instead, nobiletin activated C/EBPβ, which is up-stream of PPARγ and induces its expression⁽Reference Saito, Abe and Sekiya¹⁵¹⁾. Further research clearly needs to be undertaken to fully determine the impact of individual cell populations within AT on adipose inflammation; however, it can be seen that failure of pre-adipocytes to differentiate can be detrimental with respect to either overloading of mature adipocytes rendering them hypertrophic or increasing the inflammatory pre-adipocyte population.

Adipose tissue hypoxia

Secondary to the development of hypertrophic adipocytes is AT hypoxia. Adipocytes can expand up to 150–200 μm in diameter⁽Reference Skurk, Alberti-Huber and Herder¹¹⁸⁾ and the maximum diffusion distance of oxygen is 200 μm; therefore adipocyte hypertrophy can lead to AT hypoxia as their size may impair oxygen diffusion into the cell⁽Reference Trayhurn and Wood⁴⁾. A recent study in obese human subjects has shown their abdominal AT to have a lower capillary density and oxygen partial pressure than lean subjects and this correlated negatively with percentage body fat and macrophage inflammatory protein-1 secretion, which is involved in inflammatory cell recruitment and release of pro-inflammatory cytokines⁽Reference Pasarica, Sereda and Redman¹⁵²⁾. Furthermore, the increase in AT blood flow associated with the post-prandial state is not seen in the obese⁽Reference McQuaid, Hodson and Neville⁸²⁾. AT hypoxia is also found in rodent models of obesity, in addition to increased expression of hypoxia-regulated genes and dysregulated adipokine secretion⁽Reference Yin, Gao and He^71, Reference Hosogai, Fukuhara and Oshima^153, Reference Rausch, Weisberg and Vardhana¹⁵⁴⁾. Hypoxia-inducible factor (HIF)-1α and HIF-2α are transcription factors induced by hypoxia that affect angiogenesis, glycolysis, cell proliferation, apoptosis and inflammation⁽Reference Patel and Simon^155, Reference Weidemann and Johnson¹⁵⁶⁾. Gene expression of HIF-1α has been positively correlated with body mass⁽Reference Maumus, Sengenes and Decaunes¹⁵⁷⁾ and in adipocytes, HIF-1α and HIF-2α accumulation have been shown to promote the development of an insulin-resistant state by decreasing insulin receptor phosphorylation⁽Reference Regazzetti, Peraldi and Gremeaux¹⁵⁸⁾. In adipocyte cell lines, hypoxia reduces NEFA uptake and increases lipolysis as well as inducing necrosis and apoptosis⁽Reference Yin, Gao and He⁷¹⁾, pro-inflammatory cytokine production is also increased and adipokine secretion dysregulated⁽Reference Wang, Wood and Trayhurn^159, Reference Ye, Gao and Yin¹⁶⁰⁾. Furthermore, the differentiation of pre-adipocytes is inhibited⁽Reference Kim, Song and Chung¹⁶¹⁾.

Production of angiogenic factors, such as vascular endothelial growth factor and platelet-derived growth factor, are increased in response to hypoxia⁽Reference Lolmede, Dunand de Saint Front and Galitzhy^162, Reference Wang, Wood and Trayhurn¹⁶³⁾ and circulating levels of these are found to be elevated in obese human subjects and mice⁽Reference Silha, Krsek and Sucharda^164, Reference Pang, Gao and Yin¹⁶⁵⁾. A recent rodent study has shown that angiogenesis supports, and is essential for adipogenesis and that they occur together in cell clusters; however, as fat mass increases hypertrophy dominates with a reduction in the occurrence of adipogenic/angiogenic cell clusters⁽Reference Nishimura, Manabe and Nagasaki¹⁶⁶⁾. Nutraceuticals have been found to reduce the expression and activity of HIF in cancer cells⁽Reference Singh and Agarwal¹⁶⁷⁾ with the overall aim to reduce tumour angiogenesis and therefore growth; however, similar inhibition in AT may only serve to further exacerbate hypoxia. In ischaemia-reperfusion injury, the promotion of HIF and its downstream signalling targets is beneficial, limiting infarct size by stimulating angiogenesis⁽Reference Paul, Simons and Mabjeesh¹⁶⁸⁾. The subunits of HIF, HIF-1α, HIF-2α and the recently discovered HIF-3α, differentially regulate adipogenesis⁽Reference Hatanaka, Shimba and Sakaue¹⁶⁹⁾; however, their exact roles are, as yet, ill defined. The role of HIF and its subunits in AT dysfunction is a recent area of research and needs to be further understood before investigations into its regulation by nutraceuticals can get underway.

Adipocyte dysfunction

Researchers have attempted to establish what the fundamental triggering factor is that tips healthy AT towards one which is pro-inflammatory; one potential candidate is endoplasmic reticulum (ER) stress. Markers of ER stress occur in AT of both diet and genetically induced mouse models of obesity⁽Reference Ozcan, Cao and Yilmaz¹⁷⁰⁾ and more recently in obese human subjects⁽Reference Sharma, Das and Mondal^171–Reference Gregor, Yang and Fabbrini¹⁷³⁾, and have been found to be reduced by weight loss⁽Reference Gregor, Yang and Fabbrini¹⁷³⁾. Recently, ER stress has been linked with reduced secretion of adiponectin, a key adipokine in the regulation of insulin sensitivity and inflammation in obesity. Incubation of 3T3-L1 adipocytes with a protease inhibitor induced ER stress which resulted in a reduction in adiponectin synthesis through the activation of c-Jun N-terminal kinase signalling pathway and subsequent induction of activating transcription factor 3⁽Reference Koh, Park and Park¹⁷⁴⁾.

The ER is the principal site of protein synthesis within the cell, ensuring the transport and release of correctly folded proteins and within adipocytes, also facilitates lipid droplet formation⁽Reference Gregor and Hotamisligil^175–Reference Xu, Bailly-Maitre and Reed¹⁷⁷⁾. An oxidative environment is present within the ER and is critical for disulfide bond formation and correct folding of proteins⁽Reference Xu, Bailly-Maitre and Reed¹⁷⁷⁾. However, in reaction to disturbances such as nutrient deprivation, lipids and increased workload, unfolded proteins accumulate within the ER triggering the unfolded protein response (UPR), which results in reactive oxygen species accumulation and cellular oxidative stress⁽Reference Haynes, Titus and Cooper¹⁷⁸⁾. This is mediated by three protein kinases, protein kinase RNA-like ER kinase, inositol-requiring 1α and activating transcription factor 6⁽Reference Ron and Walter^176, Reference Xu, Bailly-Maitre and Reed¹⁷⁷⁾. As a transient measure, the UPR initially reduces protein synthesis and translocation into the ER, followed by a longer-term increase in the ability of the ER to handle unfolded proteins. If protein misfolding persists or is excessive then cell death is triggered, usually through apoptosis⁽Reference Ron and Walter^176, Reference Xu, Bailly-Maitre and Reed¹⁷⁷⁾.

The UPR is linked to inflammation through production and accumulation of reactive oxygen species, activation of the acute-phase response and activation of transcription factors regulating inflammatory signalling pathways such as NF-κB and c-Jun N-terminal kinase⁽Reference Gregor and Hotamisligil^175, Reference Zhang and Kaufman¹⁷⁹⁾. It is unknown as to what causes ER stress in obesity and there is little research into ER stress in the adipocyte; however, some causative factors may include nutrient deprivation, resulting from decreased vascular density, with concomitant hypoxia, and increased protein synthesis due to adipocyte hypertrophy⁽Reference Kaufman¹⁸⁰⁾. Elevated NEFA levels may also be a culprit, as these have been shown to activate the UPR in other cell types, including pancreatic β-cells and hepatocytes⁽Reference Wei, Wang and Topczewski^181–Reference Das, Chu and Mondal¹⁸³⁾. Conversely, ER stress may also result in elevated blood lipid levels, and research has found that protease inhibition in adipocytes, which induces ER stress, suppresses TAG synthesis and the transcription of lipogenic genes⁽Reference Parker, Flint and Mulvey¹⁸⁴⁾.

Although antioxidants may promote ER stress by adversely affecting the oxidising environment of the ER, it has recently been shown that antioxidants could possibly have some benefit in reducing UPR-induced oxidative stress. One study investigated the expression of coagulation factor VIII, which is deficient in haemophilia A and is prone to misfolding in the ER⁽Reference Malhotra, Miao and Zhang¹⁸⁵⁾. It was found that factor VIII misfolding in the ER resulted in oxidative and ER stress in vitro in Chinese hamster ovary-(H9) cells with eventual apoptosis, which was attenuated by the addition of butylated hydroxyanisole, a phenolic, lipid-soluble antioxidant. An effect that was also mimicked in vivo in mice. Further to this, butylated hydroxyanisole treatment reduced the intracellular accumulation of misfolded factor VIII with a concomitant increase in functional factor VIII secretion, both in vitro and in vivo. Treatment of Chinese hamster ovary-H9 cells with ascorbic acid, which has weaker antioxidant properties than butylated hydroxyanisole, produced inconsistent results of less intensity⁽Reference Malhotra, Miao and Zhang¹⁸⁵⁾.

In another study, oxidative and ER stress were induced in human umbilical vein endothelial cells by incubation in hyperglycaemic conditions⁽Reference Sheikh-Ali, Sultan and Alamir¹⁸⁶⁾. Interestingly though, while treatment with ascorbic acid and α-tocopherol eliminated oxidative stress, no effect was found on ER stress, similar to the results found by Malhotra et al. ⁽Reference Malhotra, Miao and Zhang¹⁸⁵⁾. This discrepancy may be due to the different cell lines used or, more likely, due to butylated hydroxyanisole having phenolic activity. Research in human colon cancer cell lines has shown that quercetin, a phenolic flavonoid, is able to reduce ER stress through the inhibition of the phosphoinositide 3-kinase pathway, which is not replicated by ascorbic acid or α-tocopherol⁽Reference Natsume, Ito and Satsu¹⁸⁷⁾. In a cell-free study, oxidative stress resulted in the loss of function of the ER protein-folding enzyme, protein disulfide isomerase causing an accumulation of misfolded proteins which was prevented by the addition of the polyphenolic compounds, curcumin and masoprocol⁽Reference Pal, Cristan and Schnittker¹⁸⁸⁾. While it has recently been shown that reduced adiponectin secretion from adipocytes is due, in part, to ER stress, it has yet to be investigated whether antioxidants could ameliorate ER stress in the adipocyte and, in so doing, possibly re-assert an anti-inflammatory secretory pattern. Given the evidence, it seems unlikely that a compound with purely antioxidant activity would have this effect, and therefore polyphenolic compounds may have more benefit.