n-3 PUFA present in fish oil are known to have numerous beneficial effects on health (Connor, Reference Connor2000; Ruxton, Reference Ruxton2004), suggesting important anti-adipogenic properties of EPA-enriched diets (Raclot et al. Reference Raclot, Groscolas, Langin and Ferre1997). However, other trials have not observed any effects on body-weight gain and visceral fat after EPA treatment (Kusunoki et al. Reference Kusunoki, Tsutsumi and Hara2003). In this context, previous studies by our group in primary cultured rat adipocytes (Perez-Matute et al. Reference Perez-Matute, Marti, Martinez, Fernandez-Otero, Stanhope, Havel and Moreno-Aliaga2005) and by others in 3T3-L1 cells (Murata et al. Reference Murata, Kaji, Takahashi, Iida, Mizuno, Okimura, Abe and Chihara2000) demonstrated that EPA was able to stimulate leptin gene expression and protein secretion. These data suggest potential anti-obesity effects of EPA in vivo since leptin is a hormone primarily secreted by white adipose tissue (WAT) involved in the regulation of body weight and body fat stores via its central actions on food intake and energy expenditure (Marti & Martinez, Reference Marti and Martinez1999).
In order to better elucidate the potential EPA anti-obesity actions, we analysed the effects of its administration in rats fed on a standard diet or a high-fat (cafeteria) diet (Moreno-Aliaga et al. Reference Moreno-Aliaga, Lamas, Marti and Martinez2002; Perez-Echarri et al. Reference Perez-Echarri, Perez-Matute, Martinez, Marti and Moreno-Aliaga2005). Our cafeteria-fed rats constitute a useful model for obesity studies because a high fat intake is a key factor in the development of obesity (Lopez et al. Reference Lopez, Marti, Milagro, Zulet Md Mde, Moreno-Aliaga, Martinez and De Miguel2003).
The mechanisms (alterations of adipogenesis, lipolysis, apoptosis) involved in these potential anti-obesity actions of EPA remain unclear. Some studies have suggested that an inhibition of PPARγ gene expression, a transcription factor with a key role as master regulator of adipogenesis, could be involved in the potential anti-obesity actions of EPA (Reseland et al. Reference Reseland, Haugen, Hollung, Solvoll, Halvorsen, Brude, Nenseter, Christiansen and Drevon2001). However, controversial data are described (Chambrier et al. Reference Chambrier, Bastard, Rieusset, Chevillotte, Bonnefont-Rousselot, Therond, Hainque, Riou, Laville and Vidal2002). Furthermore, contradictory results regarding the lipolytic action of n-3 fatty acids have also been shown (Rustan et al. Reference Rustan, Hustvedt and Drevon1993; Raclot et al. Reference Raclot, Groscolas, Langin and Ferre1997). On the other hand, several authors suggested that EPA limits fat cell trophic growth without modifying the cell number (Parrish et al. Reference Parrish, Pathy and Angel1990). However, others have observed a reduction in cellularity of mice epididymal adipose tissue after treatment with n-3 PUFA-enriched diets (Ruzickova et al. Reference Ruzickova, Rossmeisl, Prazak, Flachs, Sponarova, Veck, Tvrzicka, Bryhn and Kopecky2004), suggesting a possible apoptotic effect of EPA, although this fact has not been yet addressed. Therefore, the aim of the present study was to assess for the first time the direct potential apoptotic effects of EPA in WAT. In addition, we also tested EPA effects on WAT lipolysis ex vivo and on adipogenic PPARγ gene expression.
EPA treatment has also been suggested to improve insulin resistance in several models of obesity and diabetes in rodents (Mori et al. Reference Mori, Murakawa, Katoh, Hata, Yokoyama, Tajima, Ikeda, Nobukata, Ishikawa and Shibutani1997, Reference Mori, Murakawa, Yokoyama, Tajima, Ikeda, Nobukata, Ishikawa and Shibutani1999; Nobukata et al. Reference Nobukata, Ishikawa, Obata and Shibutani2000). Obesity has been associated with an inflammatory state which has been proposed to be the link with several obesity-associated disorders such as insulin resistance, CVD and metabolic syndrome (Dandona et al. Reference Dandona, Aljada and Bandyopadhyay2004; Moreno-Aliaga et al. Reference Moreno-Aliaga, Campion, Milagro, Berjón and Martínez2005). This inflammatory state associated with obesity appears to be predominantly triggered and to reside in adipose tissue (Wellen & Hotamisligil, Reference Wellen and Hotamisligil2003). The expression and release of inflammation-related adipokines generally rises as adipose tissue expands, except for adiponectin, which has an anti-inflammatory action (Trayhurn & Wood, Reference Trayhurn and Wood2005). Thus, plasma adiponectin concentrations are decreased in obesity and are positively associated with whole-body insulin sensitivity (Stefan et al. Reference Stefan, Bunt, Salbe, Funahashi, Matsuzawa and Tataranni2002). EPA and DHA are well known to have anti-inflammatory properties (Mori & Beilin, Reference Mori and Beilin2004), and the ability of fish oil to alter cytokine production from several cell types and reduce pro-inflammatory cytokines has been described (Calder, Reference Calder2001; Browning, Reference Browning2003). However, to our knowledge there are no data available regarding the potential anti-inflammatory actions of n-3 fatty acids in adipose tissue. In order to address this question, in the present study we analysed the effects of EPA supplementation on the pro-inflammatory cytokine TNFα gene expression in WAT. In addition, adiponectin gene expression and protein production were also studied.
In summary, the present trial provides novel insights into the mechanisms involved in the anti-adipogenic and insulin-sensitising actions of EPA on WAT, especially regarding its effects on apoptosis, TNFα and adiponectin production.
Material and methods
Animals and treatment
Twenty-nine male Wistar rats weighing about 180 g (6 weeks old) were obtained from the Centre of Applied Pharmacology (CIFA, Pamplona, Spain). Animals were housed (two or three per cage) in polycarbonate cages in temperature-controlled rooms (22 ± 2°C) with a 12 h light–dark cycle, fed a pelleted chow diet (Rodent Toxicology Diet; B&K Universal, Aldbrough, Hull, UK) and given deionised water ad libitum for an adaptation period of 4 d. Rats were then assigned to two dietary groups for 5 weeks. The control group (n 15) was fed a standard pelleted chow laboratory diet containing 18 % of energy as protein, 76 % of energy as carbohydrate and 6 % of energy as lipid by dry weight. The overweight group (n 14) was fed on a fat-rich hyperenergetic ‘cafeteria’ diet composed of paté, chips, chocolate, bacon, biscuits and pelleted diet in a proportion of 2 : 1 : 1 : 1 : 1 : 1 as previously described (Moreno-Aliaga et al. Reference Moreno-Aliaga, Lamas, Marti and Martinez2002; Perez-Echarri et al. Reference Perez-Echarri, Perez-Matute, Martinez, Marti and Moreno-Aliaga2005). The composition was 9 % of energy as protein, 29 % as carbohydrate and 62 % as lipid by dry weight. The fatty acid composition of both diets (control and cafeteria) is showed in Table 1. Both dietary groups were divided into two new subgroups. One subgroup from each dietary group received by oral administration EPA ethyl ester (1 g/kg per d) daily for 5 weeks (control EPA and overweight EPA (CEPA and OEPA) groups). The other subgroups of rats (control and overweight) received by oral administration the same volume of water as previously shown in other studies (Mori et al. Reference Mori, Murakawa, Katoh, Hata, Yokoyama, Tajima, Ikeda, Nobukata, Ishikawa and Shibutani1997; Kusunoki et al. Reference Kusunoki, Tsutsumi and Hara2003). These groups of rats without treatment with any other type of fatty acid (such as saturated fatty acids with equal chain length as EPA or oleic acid) are more likely to be considered as control groups in our study design. This is because it has been demonstrated that supplementation with some other fatty acids is able to modify adiposity and the circulating levels of biochemical and hormonal markers planned to be determined in the present study (Mori et al. Reference Mori, Murakawa, Katoh, Hata, Yokoyama, Tajima, Ikeda, Nobukata, Ishikawa and Shibutani1997).
Table 1 Fatty acid composition (%) of control and cafeteria diets
Body weight and food intake were daily recorded. At the end of the experimental period and after an overnight fast, rats were killed by decapitation and blood and tissue samples were collected and kept at − 80 °C for subsequent analysis. All procedures were conducted under protocols approved by the University Ethics Committee for the use of laboratory animals.
Biochemical measurements
Serum glucose was measured using an autoanalyser (Cobas Roche Diagnostic, Basel, Switzerland). Plasma insulin and leptin were assayed by RIA (Linco Research, St Charles, MO, USA). Adiponectin concentrations in serum were determined by a Mouse/Rat Adiponectin ELISA kit (B-Bridge International, Inc., Sunnyvale, CA, USA).
Adipocyte isolation and ex vivo lipolysis
Adipocytes were isolated from retroperitoneal fat pads of male Wistar rats. The fat pads were minced into pieces in Krebs–Ringer HEPES buffer (pH 7·4; containing 5 mm-d-glucose, 2 % bovine serum albumin, 135 mm-NaCl, 2·2 mm-CaCl2.2H2O, 1·25 mm-MgSO4.7H2O, 0·45 mm-KH2PO4, 2·17 mm-Na2HPO4 and 10 mm-HEPES). Adipose tissue fragments were digested in the same buffer with type I collagenase (1·25 mg/ml per 0·5 g tissue; Worthington, Lakewood, NJ, USA) at 37 °C with gentle shaking at sixty cycles/min for 30 min. The resulting cell suspension was diluted in HEPES-phosphate buffer and the isolated adipocytes were then separated from the undigested tissue by filtration through a 400 μm nylon mesh and washed three times. Ex vivo lipolysis was performed in isolated adipocytes according to Campion & Martínez (Reference Campion and Martinez2004). Results were expressed as μmol glycerol released per 100 mg total lipids after 90 min of incubation in Krebs–Ringer buffer.
Detection of white adipose tissue apoptosis
Apoptotic cell death was measured by using the Cell Death Detection ELISAPLUS kit (Boehringer Mannheim Biochemicals, Indianapolis, IN, USA). This is a photometric enzyme-immunoassay for quantitative in vitro determination of cytoplasmic histone-associated DNA fragments, which are generated during apoptosis, using monoclonal antibodies directed against DNA and histones, respectively. All experimental procedures were carried out on ice. The assay was performed according to the manufacturer's instructions, reading absorbance at 405 nm and subtracting absorbance at 490 nm for each well (A405–A490). Data represent average values from double absorbance measurements of each sample and were calculated subtracting the background value of the immunoassay from each value (Margareto et al. Reference Margareto, Aguado, Oses-Prieto, Rivero, Monge, Aldana, Marti and Martinez2000).
Gene expression levels in white adipose tissue
Epididymal adipose tissue mRNA expression levels were determined by Northern blotting as described by Pérez-Matute et al. (Reference Perez-Matute, Marti, Martinez and Moreno-Aliaga2003). The leptin cDNA probe was a 388 bp fragment whereas adiponectin cDNA probe was a 497 bp fragment. They were kindly provided by Dr Charles Mobbs (Mount Sinai School of Medicine, New York), which is gratefully acknowledged. The PPARγ2 cDNA probe (Genbank accession number Y12882) was a 570 bp fragment of rat cDNA, and the primers used were 5′-TCTGATTATGGGTGAAACTC-3′ (sense 43–62) and 5′-TTTCTACTCTTTTTGTGGATC-3′ (antisense 592–612). A 320 bp TNFα DNA fragment was prepared by the PCR with primers corresponding to the nucleotides 923–942 (5′-CCCCATTACTCTGACCCCTT-3′) and 1223–1242 (5′-AGGCCTGAGACATCTTCAGC-3′) of rat TNFα cDNA (Genbank accession number AF329987) (Kawaguchi et al. Reference Kawaguchi, Sugiyama, Hibasami and Toyoda2003). The 18S ribosomal probe was obtained from Ambion (Austin, TX, USA). mRNA expression levels were normalised with respect to this 18S ribosomal signal. Signals were quantified using a Sharp JX-325 densitometric scanner (Sharp Electric, Tokyo, Japan) and the analysis was carried out with the software ImageMaster™ (Pharmacia Biotech, Uppsala, Sweden).
Statistical analysis
Results are given as mean values with their standard errors. Statistical differences and interactions were evaluated with a factorial two-way ANOVA (diet × EPA treatment). When statistically significant differences resulted at the interaction level, a Student's t test was carried out to compare the effects of each treatment (GraphPad Prism; GraphPad Software Inc., San Diego, CA, USA). Differences were considered as statistically significant at P < 0·05.
Results
Effects on body-weight gain and food intake
High-fat-fed rats showed a significant increase in both final body weight (P < 0·001) and body-weight gain (P < 0·001) at the end of the experimental trial as compared with control animals (Fig. 1). The administration of EPA (1 g/kg) did not significantly alter body-weight gain, although a marginal decrease (P = 0·09) could be observed after EPA treatment in the group of rats fed with the cafeteria diet (OEPA).
Fig. 1 Effects of EPA (35 d treatment) on the growth curve in rats: control rats (n 8; –●–), control rats treated with EPA (1 g/kg; n 7; –○–), overweight rats (high-fat-diet; n 7; –▾–) and overweight rats treated with EPA (1 g/kg; n 7; –▿–). Body weight was daily recorded. Data are means, with standard errors represented by vertical bars. Statistical analysis was by two-way ANOVA; the effect of diet was significant (P < 0·001).
Energy intake was measured in order to find out if the differences showed in body-weight evolution could be related to differences in food intake. Overweight rats (fed with a high-fat diet) showed a significant increase in their food intake (P < 0·001), which was reduced in part by EPA treatment (P < 0·01). Feed efficiency, the weight gain:energy consumed ratio, was significantly decreased (P < 0·001) in all groups receiving the high-fat diet as compared with the control group. However, EPA treatment did not modify this parameter (Table 2).
Table 2 Total food intake, feed efficiency and white adipose tissue weights in control and eicosapentaenoic acid-treated lean and overweight rats* (Mean values with their standard errors)
CEPA, control EPA; OEPA, overweight EPA.
* Data were analysed by two-way ANOVA. When an interaction was found, comparisons between groups were analysed by a Student's t test.
† Mean value was significantly different from that of the control group (P < 0·001).
‡ Mean value was significantly different from that of the overweight group (P < 0·01).
§ Energy supplied by EPA was taken into account when total food intake was measured.
Effects on white adipose tissue weights
Wet weights (g) of WAT from overweight rats were significantly higher than in lean animals (P < 0·001). Daily administration of EPA produced a significant reduction (P < 0·05) in retroperitonal adipose tissue weight. Indeed, EPA treatment showed a tendency (non-significant) to decrease the weight of all the fat pads, but only in the group of animals fed with a standard diet (Table 2).
Regarding brown adipose tissue, animals fed with the cafeteria diet showed an increase in brown fat (P < 0·001) as compared with control rats, while the administration of EPA did not modify the size of this fat depot (Table 2).
Effects on plasma biochemical parameters
Neither the cafeteria diet nor EPA treatment modified glucose plasma levels; however, a slight increase was observed after the consumption of the high-fat diet (P = 0·08). EPA treatment induced a significant reduction (P < 0·05) in both insulin plasma levels and the homeostasis model assessment (HOMA) index, an index of insulin resistance, suggesting that EPA is able to partially prevent the development of insulin resistance induced in a non-significant manner by the cafeteria diet (Table 3).
Table 3 Serum glucose, plasma insulin and homeostasis model assessment (HOMA) index in control and eicosapentaenoic acid-treated lean and overweight rats* (Mean values with their standard errors)
CEPA, control EPA; OEPA, overweight EPA.
* Data were analysed by two-way ANOVA.
† HOMA = insulin (μUnits) × glucose (mmol/l)/22·5.
Effects on retroperitoneal white adipose tissue apoptosis
No differences in the relative apoptosis rate were observed in overweight rats as compared with the basal apoptosis of the control rats. EPA treatment induced a marked increase (P < 0·05) in the levels of histone-associated DNA oligonucleosomal fragments, a characteristic marker of apoptosis in the group of animals fed with the standard diet but not in the high-fat-fed rats (Fig. 2).
Fig. 2 Effects of EPA on apoptosis in white adipose tissue of control and overweight rats. Data are means for at least seven independent animals per group, with standard errors represented by vertical bars. Statistical analysis was by two-way ANOVA; the diet × EPA interaction was significant (P < 0·01). When an interaction was found, comparisons between groups were analysed by a Student's t test. *Mean value was significantly different from that of the control group (P < 0·05). CEPA, control EPA; OEPA, overweight EPA.
Effects on ex vivo lipolysis
High-fat-diet feeding caused a significant decrease (P < 0·05) in isoproterenol-stimulated lipolytic response in both EPA-treated and untreated rats (Fig. 3). EPA treatment, however, did not affect either basal lipolysis (data not shown) or isoproterenol-induced lipolysis (Fig. 3).
Fig. 3 Effects of EPA on isoproterenol-induced lipolysis in white adipose tissue from control and overweight-cafeteria rats. Data are means for at least four independent animals per group, with standard errors represented by vertical bars. Statistical analysis was by two-way ANOVA; the effect of diet was significant (P < 0·05). CEPA, control EPA; OEPA, overweight EPA.
Effects on leptin circulating levels and gene expression
Regarding leptin circulating levels, a significant interaction (P < 0·01) was obtained between diet and EPA treatment. As expected, leptin levels were significantly higher (P < 0·05) in the cafeteria-fed group as compared with control animals. EPA treatment significantly decreased leptin circulating levels (P < 0·01) in the group of animals fed on a standard diet while a significant increase (P < 0·05) was observed in overweight rats treated with the fatty acid (Fig. 4 (A)). When leptin serum concentrations were corrected for adiposity (leptin per g WAT), which was estimated by the sum of all fat depots collected (see Table 2), a significant decrease (P < 0·05) was observed in the group of control animals treated with EPA. On the contrary, EPA caused a significant elevation of leptin production levels (P < 0·01) in the high-fat-fed rats (Fig. 4 (B)). Similarly, leptin gene expression in epididymal fat showed the same pattern as circulating levels (Fig. 4 (C)).
Fig. 4 Effects of EPA on leptin circulating levels (A), leptin concentrations expressed per g white adipose tissue (WAT) (B) and mRNA expression in epididymal fat (C) in control and overweight-cafeteria rats. The expression level of 18S ribosomal RNA was determined and used as an internal control to correct for minor variation in total RNA amount. Densitometric scanning was used to determine the relative amount of leptin mRNA and 18S RNA. Data are means for at least seven independent animals per group (n 4 in Northern blot), with standard errors represented by vertical bars. Statistical analysis was by two-way ANOVA; the diet × EPA interaction was significant for leptin (P < 0·01) and for leptin/WAT (P < 0·001). When an interaction was found, comparisons between groups were analysed by Student's t test. Mean value was significantly different from that of the control group: *P < 0·05, **P < 0·01. Mean value was significantly different from that of the overweight group: †P < 0·05, ††P < 0·01. CEPA, control EPA; OEPA, overweight EPA.
Effects on adiponectin circulating levels and gene expression
Regarding adiponectin, cafeteria-diet feeding significantly increased adiponectin circulating levels (P < 0·001); however, EPA treatment did not affect the circulating levels of this adipokine (Fig. 5 (A)). When adiponectin concentrations were corrected for adiposity (adiponectin per g WAT), a marginally non-significant decrease (P = 0·07) was observed in the cafeteria-fed group. Furthermore, EPA treatment induced a significant increase (P < 0·05) in adiponectin concentrations per g WAT (Fig. 5 (B)).
Fig. 5 Effects of EPA on adiponectin circulating levels (A), adiponectin concentrations expressed per g white adipose tissue (WAT) (B) and mRNA expression in epididymal fat (C) in control and overweight-cafeteria rats. The expression level of 18S ribosomal RNA was determined and used as an internal control to correct for minor variation in total RNA amount. Densitometric scanning was used to determine the relative amount of adiponectin mRNA and 18S RNA. Data are means for at least seven independent animals per group (n 4 in Northern blot), with standard errors represented by vertical bars. Statistical analysis was by two-way ANOVA; the effect of diet was significant for adiponectin (P < 0·001), the effect of EPA was significant for adiponectin/WAT (P < 0·05) and the effect of diet was marginally significant for adiponectin/WAT (P = 0·07). CEPA, control EPA; OEPA, overweight EPA.
Northern blot analysis revealed three mRNA bands for adiponectin in rats (Hu et al. Reference Hu, Liang and Spiegelman1996). In agreement with the observed levels of serum adiponectin, an increase in adiponectin gene expression in WAT was also observed after EPA treatment (Fig. 5 (C)).
Several studies have suggested that there is a strong relationship between low adiponectin levels and insulin resistance (Hotta et al. Reference Hotta, Funahashi, Bodkin, Ortmeyer, Arita, Hansen and Matsuzawa2001). In fact, our data showed an inverse relationship between serum adiponectin levels corrected for adiposity and HOMA index (r − 0·43; P < 0·05).
Effects on white adipose tissue peroxisome proliferator-activated receptor-γ and tumour necrosis factor-α gene expression
The expression of the adipogenic factor PPARγ was significantly (P < 0·001) decreased by EPA treatment in both control and cafeteria-fed animals (Figs. 6 (A) and (B)).
Fig. 6 Effects of EPA on PPARγ mRNA (A and B) and TNFα mRNA (C and D) expression in epididymal fat in control and overweight-cafeteria rats. The expression level of 18S ribosomal RNA was determined and used as an internal control to correct for minor variation in total RNA amount. Densitometric scanning was used to determine the relative amount of PPARγ, TNFα mRNA and 18S RNA. Data are means for at least four independent animals per group, with standard errors represented by vertical bars. Statistical analysis was by two-way ANOVA; the effect of EPA was significant for PPARγ mRNA (P < 0·001) and the diet × EPA interaction was significant for TNFα mRNA (P < 0·001). When an interaction was found, comparisons between groups were analysed by a Student's t test. Mean value was significantly different from that of the control group: *P < 0·05, **P < 0·01. ††Mean value was significantly different from that of the overweight group (P < 0·01). CEPA, control EPA; OEPA, overweight EPA.
As is observed in Figs. 6 (C) and (D), TNFα gene expression was significantly stimulated (P < 0·05) by cafeteria-diet feeding. EPA treatment was able to prevent (P < 0·01) the rise in the expression of this inflammatory cytokine induced by the high-fat diet.
Discussion
Animals fed with a high-energy-yielding diet during 35 d exhibited a significant increase in body-weight gain. EPA treatment, however, was not able to prevent this increase induced by the high-fat diet, although a moderate lower body-weight gain was observed in the group of animals fed with the high-fat diet and orally treated with EPA. These results are in agreement with those from Kusunoki et al. (Reference Kusunoki, Tsutsumi and Hara2003) who did not observe any significant effect of EPA treatment (concentrations from 0·3 to 1 g/kg) on body-weight gain in Otsuka Long-Evans Tokushima Fatty (OLETF) rats. Energy intake was measured in order to find out if the moderate differences showed in body-weight evolution could be related to differences in food intake. As in previous studies, animals fed with the high-fat diet showed an increase in food intake in comparison with the control group, probably because palatability is the predominant force driving the hyperphagia in cafeteria-fed animals (Berraondo et al. Reference Berraondo, Marti, Duncan, Trayhurn and Martinez2000; Margareto et al. Reference Margareto, Gomez-Ambrosi, Marti and Martinez2001). Interestingly, EPA treatment was able to decrease food intake, but only in the group of animals fed on the cafeteria diet (OEPA). It is well known that rising concentrations of leptin generate a signal to reduce food intake and increase energy expenditure (Havel, Reference Havel2004). Taking into account this role of leptin in body-weight regulation, it could be suggested that the EPA-induced increase in leptin production specifically observed in the cafeteria-fed rats could explain the decrease in food intake observed in this group of animals. Peyron-Caso et al. (Reference Peyron-Caso, Taverna, Guerre-Millo, Veronese, Pacher, Slama and Rizkalla2002) also described an increase in leptin concentrations in rats fed an n-3 PUFA-enriched diet. In addition, our previous studies in primary cultured rat adipocytes also demonstrated a stimulatory action of EPA on leptin production (Perez-Matute et al. Reference Perez-Matute, Marti, Martinez, Fernandez-Otero, Stanhope, Havel and Moreno-Aliaga2005). However, other studies showed an inhibitory action of this fatty acid on leptin (Raclot et al. Reference Raclot, Groscolas, Langin and Ferre1997; Reseland et al. Reference Reseland, Haugen, Hollung, Solvoll, Halvorsen, Brude, Nenseter, Christiansen and Drevon2001). These conflicting results could be due to the observations that the effects of n-3 PUFA on leptin are dependent on a number of factors such as diet composition and energy balance. In this way, the present results showed opposite effects of EPA treatment in control rats in comparison with the overweight group. Other studies have also observed differential actions of some anti-obesity treatments (such as β-adrenergic agonist) on genes related to energy homeostasis between control and obese animals (Berraondo et al. Reference Berraondo, Marti, Duncan, Trayhurn and Martinez2000).
With regard to adiposity, EPA treatment induced a depot-specific reduction in the weight of retroperitoneal fat, as previously observed by Raclot et al. (Reference Raclot, Groscolas, Langin and Ferre1997). Furthermore, EPA treatment also induced a moderate reduction in the weight of all fat depots but only in control animals. Hill et al. (Reference Hill, Peters, Lin, Yakubu, Greene and Swift1993) observed a decrease in all fat depots of adult male Wistar rats after 6 months of treatment with dietary n-3 PUFA, suggesting that the potential anti-adipogenic actions of n-3 PUFA are dependent on treatment duration and metabolic status. In this sense, if our EPA treatment lasted more than 5 weeks, it could be possible to observe a significant decrease in the weight of all fat depots and not only in the retroperitoneal fat pad. Paradoxically, the lowering effects of EPA on adiposity contrast with the lack of effects on body-weight gain observed in control rats. Previous studies have found that a fish oil-enriched diet induces greater gain in lean body mass and a lower gain in fat in comparison with animals fed with other type of enriched diets (Su & Jones, Reference Su and Jones1993). Although we have not performed any body composition measurements, our data also suggest that EPA treatment could induce a body composition redistribution, which could explain that a reduction in adipose tissue weight did not always mean a decrease in total body weight, as observed in the present study.
An increase of lipolysis could be a potential mechanism responsible for the limitation of retroperitoneal fat weight after EPA treatment. The increase in fat pad weights in high-fat-fed animals was accompanied by a decrease in the isoproterenol-induced lipolytic response, as previously shown (Portillo et al. Reference Portillo, Simon, Garcia-Calonge and Del Barrio1999). The present study demonstrated a lack of effect of EPA on both basal and isoproterenol-stimulated lipolysis, suggesting that the potential anti-adipogenic properties of this fatty acid are not mediated by a lipolytic action in WAT. In addition, our in vitro study in cultured adipocytes also argues against a direct lipolytic action of this fatty acid (Perez-Matute et al. Reference Perez-Matute, Marti, Martinez, Fernandez-Otero, Stanhope, Havel and Moreno-Aliaga2005).
PPARγ, a member of the nuclear hormone receptor superfamily, is an important adipogenic transcription factor. Several articles have described an increase in the expression of this factor in WAT of high-fat-fed obese animals (Margareto et al. Reference Margareto, Rivero, Monge, Aldana, Marti and Martinez2002; Lopez et al. Reference Lopez, Marti, Milagro, Zulet Md Mde, Moreno-Aliaga, Martinez and De Miguel2003). In the present study, we have also observed a slight non-significant increase in the expression of this factor after the cafeteria-diet consumption without statistical significance. Several studies have reported the ability of EPA to modulate the expression and activity of this adipogenic factor, although the results are controversial. Thus, some studies have observed an inhibitory action of EPA on PPARγ (Reseland et al. Reference Reseland, Haugen, Hollung, Solvoll, Halvorsen, Brude, Nenseter, Christiansen and Drevon2001) while others have observed an stimulatory effect (Chambrier et al. Reference Chambrier, Bastard, Rieusset, Chevillotte, Bonnefont-Rousselot, Therond, Hainque, Riou, Laville and Vidal2002). Our data revealed that EPA-treated animals exhibited a decrease in the adipogenic PPARγ gene expression levels, which could impair the adipogenic process, contributing to the decrease in WAT weight observed in these animals.
Targeting apoptotic pathways in adipocytes has been proposed as a potential novel treatment for obesity (Nelson-Dooley et al. Reference Nelson-Dooley, Della-Fera, Hamrick and Baile2005). We tested if an increase in apoptosis could be contributing to the reduced WAT size. Ruzickova et al. (Reference Ruzickova, Rossmeisl, Prazak, Flachs, Sponarova, Veck, Tvrzicka, Bryhn and Kopecky2004) demonstrated that n-3 PUFA limit diet-induced obesity in mice by reducing cellularity of adipose tissue and, although they did not directly measure apoptosis, their data suggest a potential apoptotic action of EPA. Our data showed for the first time that an increase in apoptosis is involved in the more efficient action of EPA on the reduction of WAT size observed in control rats. Thus, the rise in the apoptotic cell death rate would decrease adipocyte number and consequently adipose tissue weight. Finally, we also observed in the present study that the cafeteria diet strongly impaired this apoptotic action induced by EPA, suggesting that EPA-induced apoptosis depends on the nutritional and metabolic status of animals. Furthermore, the opposite effects of EPA observed in control and cafeteria-fed rats could be also due to the different percentage of EPA in relation to total lipids supplied by both types of diets. In the present study, the EPA supplemented by oral administration in rats fed with the cafeteria diet (62 % of energy as fat) accounts for 2·1 % of the total lipids. However, EPA accounts for a 30·2 % of the total lipids ingested by the rats fed with the control diet (6 % of energy as fat) and orally treated with EPA. In this context, recent evidence suggests that the ratio between different types of fatty acids such as the n-6 and n-3 fatty acids is an important fact in determining their effects on health (Simopoulos, Reference Simopoulos2002; Weiss et al. Reference Weiss, Barrett-Connor and Von Muhlen2005).
Several studies have suggested that obesity could be an inflammatory disease (Das, Reference Das2001; Ramos et al. Reference Ramos, Xu, Romanova, Middleton, Chen, Quinn, Inui, Das and Meguid2003). In this context, several markers of chronic inflammation, such as TNFα, have been found to be associated with obesity and insulin resistance (Ramos et al. Reference Ramos, Xu, Romanova, Middleton, Chen, Quinn, Inui, Das and Meguid2003). In fact, similarly to our present data, an induction of TNFα mRNA expression and activity was observed in adipose tissue from different rodent models of obesity (Hotamisligil et al. Reference Hotamisligil, Shargill and Spiegelman1993; Morin et al. Reference Morin, Eckel, Marcel and Pagliassotti1997). Here, we have demonstrated for the first time that EPA treatment was able to down regulate the increased adipose TNFα gene expression induced by a high-fat diet. Although the mechanisms involved in this inhibitory action have not been yet elucidated, it has been suggested that EPA could prevent the NF-κB activation (Zhao et al. Reference Zhao, Joshi-Barve, Barve and Chen2004). The inflammatory state associated with obesity seems to be predominantly initiated and to reside in adipose tissue, although other metabolically key organs may also be involved during the course of the disease (Wellen & Hotamisligil, Reference Wellen and Hotamisligil2003). In addition, this inflammatory state associated with obesity has been proposed to be the link with several obesity-associated disorders such as insulin resistance (Dandona et al. Reference Dandona, Aljada and Bandyopadhyay2004). Indeed, increased TNFα levels have been associated with increased insulin resistance in adipose tissue, by reducing the phosphorylation of insulin receptor sustrate-1 (Hotamisligil, Reference Hotamisligil1999). Therefore, our data lead us to suggest that the improvement in insulin sensitivity induced by EPA could be mediated, at least in part, by decreasing the inflammation in the adipose tissue of the overweight rats.
Adiponectin is an anti-inflammatory adipokine with an important role in insulin sensitivity (Combs et al. Reference Combs, Pajvani and Berg2004). Thus, hypoadiponectinaemia has been related to the development of insulin resistance and cardiovascular risk associated with obesity (Hotta et al. Reference Hotta, Funahashi, Bodkin, Ortmeyer, Arita, Hansen and Matsuzawa2001; Kubota et al. Reference Kubota, Terauchi and Yamauchi2002; Kumada et al. Reference Kumada, Kihara and Sumitsuji2003). However, our data showed that levels of circulating adiponectin were increased in overweight rats fed with the high-fat diet. Several studies have also shown an increase in both adiponectin circulating levels and gene expression in early stages of obesity (11, 15 and 21 d of high-fat-diet consumption) (Li et al. Reference Li, Yu, Pan and Unger2002; Lopez et al. Reference Lopez, Milagro, Marti, Moreno-Aliaga, Martinez and De Miguel2004; Perez-Echarri et al. Reference Perez-Echarri, Perez-Matute, Martinez, Marti and Moreno-Aliaga2005) suggesting that this could be an initial mechanism in the adaptation to high energy intake. In the present study, when the levels of circulating adiponectin were corrected for adiposity, a moderate non-significant decrease was observed in the untreated cafeteria rats, accompanied by a decrease in the expression of this adipokine, suggesting an impairment in adiponectin production processes by the overweight adipocytes. These results suggest that when obesity is more established (for example, after 65 d of a high-fat diet as described by Lopez et al. (Reference Lopez, Milagro, Marti, Moreno-Aliaga, Martinez and De Miguel2005)) the ability of adipocytes to express and secrete adiponectin is more strongly impaired. Our data also suggest that this ability of adipocytes to produce adiponectin was significantly increased by the administration of EPA, which could contribute to the insulin-sensitising properties of this fatty acid. Furthermore, a negative relationship was found in the present study between adiposity-corrected adiponectin plasma levels and the HOMA index which supports the beneficial effects of EPA supplementation on insulin sensitivity. A recently published study (Flachs et al. Reference Flachs, Mohamed-Ali, Horakova, Rossmeisl, Hosseinzadeh-Attar, Hensler, Ruzickova and Kopecky2006) has also observed that feeding mice with a high-fat diet enriched with EPA and DHA leads to elevated systemic concentrations of adiponectin, and suggested that this increase could explain, to some extent, the n-3 EPA antidiabetic properties.
In summary, our data suggest that EPA supplementation could help to reduce adiposity through the down regulation of PPARγ and through the increase of apoptosis in WAT. EPA also exhibited antidiabetic properties which could be related to its reducing effects on the pro-inflammatory cytokine TNFα mRNA levels in adipose tissue and increasing the capacity of the adipose tissue to produce the insulin-sensitising adipokine, adiponectin. In addition, the present study suggests that some EPA actions seem to be dependent on diet composition and the physiological and metabolic status of animals, which could be important to take into account when considering supplementation with n-3-enriched-products.
Acknowledgements
The present study has been supported in part by the Government of Navarra (Department of Education and Culture) and Línea Especial de Investigación ‘Nutrición, Salud y Obesidad’ (University of Navarra). P. P.-M. was supported by a doctoral grant from the Danone Institute (Spain) and the Asociación de Amigos de la Universidad de Navarra (Spain). The expert technical assistance of Ana Lorente and Veronica Ciaurriz is gratefully acknowledged. EPA was kindly provided by Brudy S.L (Spain).
