| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Address all correspondence and requests for reprints to: D. Langin, Laboratoire de Recherches sur les Obésités, Institut National de la Santé et de la Recherche Médicale Unit 858-I2MR, Equipe 4, IFR31, Institut Louis Bugnard, 1 Avenue Jean Poulhès, BP 84225, 31432 Toulouse cedex 4, France. E-mail: dominique.langin{at}inserm.fr.
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Objective: We investigated adipose tissue gene expression from human obese women according to energy deficit and the fat and carbohydrate content of the diet.
Design and Setting: Obese subjects recruited among eight European clinical centers were followed up 10 wk of either a low-fat (high carbohydrate) or a moderate-fat (low carbohydrate) hypoenergetic diet.
Subjects: Two sets of 47 women in each dietary arm were selected among 648 subjects matched for anthropometric and biological parameters.
Main Outcome Measure: We measured adipose tissue gene expression changes in one set using a candidate gene approach. The other set was used to survey 24,469 transcripts using DNA microarrays. Results were analyzed using dedicated statistical methods. Diet-sensitive regulations were confirmed on the other set of subjects.
Results: The two diets induced similar weight loss and similar changes for most of the biological variables except for components of the blood lipid profile. One thousand genes were regulated by energy restriction. We validated an effect of the fat to carbohydrate ratio for five genes (FABP4, NR3C1, SIRT3, FNTA, and GABARAPL2) with increased expression during the moderate-fat diet.
Conclusions: Energy restriction had a more pronounced impact on variations in human adipose tissue gene expression than macronutrient composition. The macronutrient-sensitive regulation of a subset of genes may influence adipose tissue function and metabolic response.
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Proper glucose and lipid homeostasis depends on adipose tissue metabolism, which is altered in response to obesity-induced changes in adipose tissue gene expression. Modification of gene expression toward a nonobese profile could contribute to the beneficial effects of weight loss (5). In vitro studies identified several classes of transcription factors mediating the impact of fat and glucose on the regulation of gene expression (6, 7). Moreover, much progress has been achieved in the understanding of the metabolic effects of calorie restriction (8). However, the in vivo molecular adaptations of adipose tissue during energy deprivation and in response to macronutrients are mostly unknown. The characterization of dietary signatures that influence the risk of diet-related diseases may help in identifying new potential risk factors and indicators of effects of dietary components (9). Such information is lacking for dietary treatment of obesity. The complexity of the response of obese subjects suggests that large cohorts and specific technical and statistical tools should be used.
The Nutrient-Gene Interaction in Human Obesity (NUGENOB) study trial, a large randomized intervention carried out in 771 obese subjects to compare the effects of a low-fat/high-carbohydrate and a moderate-fat/low-carbohydrate hypoenergetic diet, is ideally suited to address the respective influence of calorie restriction and macronutrient composition on adipose tissue gene expression (10). Two gene expression studies have already been performed in the NUGENOB program. On a subset of the cohort and without a whole genome approach, a limited number of genes were shown regulated by calorie restriction, and none could be identified as diet-sensitive genes (11, 12). Using the two dietary arms of the NUGENOB cohort and a pangenomic DNA microarray approach combined with specific statistical and bioinformatic tools, we investigated the regulation of transcript levels according to energy deficit and the fat to carbohydrate ratio. The work was completed by a candidate gene approach and cross-validation on an independent set of subjects using RT-quantitative PCR (RT-qPCR).
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Subjects were participants in the European multicenter NUGENOB study (www.nugenob.org). Subjects were randomly assigned to 10-wk dietary intervention of two hypoenergetic diets: a low-fat, high-carbohydrate diet (LF) or a moderate-fat, low-carbohydrate diet (MF). Both diets were designed to provide 600 kcal/d less than the individual’s estimated energy requirement. Fat provided 20–25 and 40–45%, carbohydrates provided 60–65 and 40–45% of total energy in the LF and the MF diet, respectively. Both diets derived 15% of total energy from protein. Details on volunteer recruitment, dietary intervention, and compliance as well as procedures about the measurements of anthropometric and blood parameters were previously described (10).
From a total of 771 participants, 648 completed the intervention, among whom 47 women per diet were selected, matched for high quality of adipose tissue RNA, weight, height, body mass index (BMI), waist to hip ratio, energy intake, fat, carbohydrate, protein, and alcohol energy intakes before the intervention and similar changes in weight, BMI, percent of body fat mass, and energy intake during the intervention. Each dietary group was randomly divided into two similar sets of subjects (Fig. 1
). Set 1 was used for a candidate gene approach using RT-qPCR, which was further confirmed on set 2 [experiments performed by the Institut National de la Santé et de la Recherche Médicale (Inserm) Lyon laboratory].
|
Subcutaneous adipose tissue biopsy and total RNA preparation
An abdominal sc fat specimen (
1 g) was obtained by needle aspiration under local anesthesia after an overnight fast before and after the dietary intervention. Biopsies were washed and stored in RNA Later preservative solution (QIAGEN, Courtaboeuf, France) at –80 C until analysis. Total RNA was extracted using the RNeasy total RNA minikit (QIAGEN) in the Inserm Toulouse laboratory. Total RNA concentration and integrity were estimated using Agilent 2100 bioanalyzer (Agilent Technologies, Massy, France).
Microarray experiments
One µg of total RNA from each sample from set 2 was amplified with the MessageAmp RNA kit (Ambion, Austin, TX). Quantity and quality of amplified RNA (aRNA) were systematically checked using an Agilent 2100 bioanalyzer. Fluorescent targets were synthesized from 1 µg of aRNA with cyanine dyes (Cy) using the CyScribe first-strand cDNA postlabeling kit (Amersham Biosciences, Orsay, France). Fluorescent probes obtained before and after the dietary intervention for a same subject were cohybridized on cDNA pangenomic microarrays containing 41,760 spots representing 24,469 Unigene clusters (http://genome-www5.stanford.edu/). Dye switch of aRNA labeling was done to remove dye-bias effect. For half of the subjects, aRNA from the fat biopsy obtained before dietary intervention was labeled with Cy3, whereas Cy5 was used for aRNA from the biopsy obtained after the intervention. For the second half of subjects, the labeling was reversed.
Hybridization was performed according to the standard protocol described elsewhere (http://cmgm.stanford.edu/pbrown/protocols/index) with slight modifications. Arrays were scanned using a GenePix 4000A scanner (Axon Instruments-Molecular Devices, Sunnyvale, CA).
Data analysis
Images were analyzed using Genepix pro 5.0 software (Axon Instruments-Molecular Devices). The loess normalization smoothing procedure was applied for intensity normalization using the Limma package with R software. Intensity ratios between Cy5 and Cy3 were filtered using an average intensity 2.5 times above the background signal. The SD of the log2 ratio was set to 1 after data normalization. Clones with more than five missing values within each dietary group were excluded from the analysis. Thus, the log2 Cy5 to Cy3 signal ratios were retrieved for 18,442 spots. When labeling was reversed, values were multiplied by –1. Analysis of the function of regulated genes was done using the Protein Analysis Through Evolutionary Relationships (PANTHER) database with focus on the biological process ontology. For genes down-regulated by calorie restriction, biological processes overrepresented among the list of differentially expressed genes were considered significant at P < 0.05, determined by the binomial statistic with Bonferroni correction for multiple testing. We performed further analysis using the Ingenuity Pathways Analysis software 6.0 (Ingenuity Systems, Redwood City, CA) to identify networks between these genes.
mRNA level measurements by RT-qPCR
First-strand cDNA was synthesized from 1 µg of adipose tissue total RNA after deoxyribonuclease treatment using random hexamers and poly(dT) according to the SuperScript II reverse transcriptase kit (Invitrogen, Cergy-Pointoise, France). All measurements were performed in duplicate using the Sybergreen method and Applied Biosystems 7500 real-time PCR (Applied Biosystems, Foster City, CA) system (Inserm Toulouse) or a Roche Light Cycler (Roche Diagnostics, Meylan, France) system (Inserm Lyon). mRNA expression level was normalized by 18S rRNA expression.
Statistical analysis
Biological and anthropometric parameters were compared between sets and dietary groups by ANOVA. All changes (anthropometric, biological parameters, mRNA levels) during energy restriction were analyzed by paired-sample t test. Diet effect on the change in variables after dietary intervention was analyzed by a univariate general linear model on the measurement after energy restriction with an adjustment for measurement before energy restriction (13).
To analyze microarray experiments, the significance analysis of microarray (SAM) procedure was used to compare the two dietary groups and analyze gene expression changes induced by energy restriction (14). Next, two distinct statistical methods were used to identify the genes differentially regulated in response to the two diets. Random forests (15, 16) and nearest shrunken centroids (17) are two classifiers, already used in the context of microarray data analysis, including a feature selection mechanism that enable to extract the most discriminative genes. In the random forest classifier, genes can be ordered according to their importance based on the change in classification accuracy when expression ratio value is randomly permuted in the other dietary group. This method was applied on normalized after-to-before expression ratios from microarray features without missing data among the 48 microarray experiments. The nearest shrunken centroids method (included in the Prediction Analysis of Microarray software) improve the more usual nearest centroid method, which computes an average gene expression (centroid) for each gene in each group divided by the within-class SD for that gene. This method was used on normalized expression ratios of the 100 most over up- and down-regulated transcripts by each dietary program as determined with SAM. To select transcripts for RT-qPCR validation analysis, the effect of the diets on after-to-before dietary intervention gene expression ratio was analyzed by Student t test. Pearson correlation coefficient was determined on log-transformed data with normal distribution to study relationship of gene expression and plasma parameters. Spearman rho correlation coefficient was used on data without normal distribution.
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Comparison of the two dietary groups is presented on Table 1. Averaged fat and carbohydrate intakes before the dietary program (35 and 46% of total energy intake, respectively) were similar between the LF and MF groups. During energy restriction, in the LF group, about 23 and 59% of total energy supply was derived from fat and carbohydrates, respectively. In the MF group, about 42 and 41% of total energy was derived from fat and carbohydrates, respectively. The differences in fat or carbohydrate intakes between the two dietary groups were highly significant. There was no difference in energy intake derived from proteins.
|
). Energy restriction induced a significant and similar weight loss (–6.8 ± 0.2 kg), fat mass (–5.2 ± 0.2 kg), fat-free mass (–1.6 ± 0.2 kg), and BMI (–2.5 ± 0.1 kg . m–2) decrease in the two dietary groups. Mean values for plasma cortisol (+ 58 ± 20 nmol/liter), insulin (–1.41 ± 0.46 µU/ml), and leptin (–12.0 ± 0.8 ng/ml) levels and the insulin resistance index homeostasis model assessment (HOMA) (–0.34 ± 0.12) were also significantly and similarly modulated by the LF and MF diet. As observed in the whole population of the NUGENOB program (10), changes in plasma levels of triacylglycerols and cholesterol fractions differed between the two dietary groups.
|
). There was no difference between sets 1 and 2 regarding food intake, anthropometric and biological parameters. Changes induced by the diets were similar in the two sets (data not shown). Changes in candidate gene expression in adipose tissue during low-calorie diets
Information related to the selected genes is presented in supplementary Table 1, published as supplemental data on The Endocrine Society’s Journals Online Web site at http://jcem.endojournals.org. Selection was based on experimental evidence of in vitro or in vivo regulation of mRNA levels by fatty acid or glucose in adipocytes or potential activation during energy restriction. Nuclear or membrane receptors, transcriptional modulators, and secreted factors that may be activated by a hypoenergetic diet were also considered.
The effects of the two hypoenergetic diets were determined on sc adipose tissue mRNA levels by RT-qPCR in set 1. The mRNA levels of stearoyl-CoA desaturase (SCD), acetyl-coenzyme A carboxylase-
(ACACA), fatty acid synthase (FASN), and adiponutrin (ADPN) were significantly reduced after a LF or a MF diet (Fig. 2). Five genes, fatty acid binding protein 4 (FABP4), liver X receptor- (LXR), adipose triglyceride lipase (ATGL), and sirtuin 2 and 3 (SIRT2, SIRT3), showed a differential expression between the diets. When the two sets of subjects were combined (n = 43–47 for each diet), a diet effect was observed for the five genes (P < 0.05).
|
To gain a global insight on changes in gene expression, pangenomic microarray analysis was performed on adipose tissue samples from set 2 (Fig. 1). As a preliminary approach, the SAM two-class unpaired procedure was used to compare changes induced between the LF and MF diets and showed no significant difference in gene expression. We next performed the analysis of the effect of energy restriction on the entire set combining the MF and LF groups. With a false discovery rate of 5 and 10%, one and 81 cDNA clones were up-regulated, and 925 and 1283 clones were down-regulated during energy restriction, respectively (supplementary Table 2).
To validate these data, we quantified adipose tissue mRNA levels by RT-qPCR before and after each diet in set 2 for 10 genes with similar expression changes in both diets. We selected six down-regulated genes: glycerol-3-phosphate acyltransferase mitochondrial (GPAM), hydroxysteroid 17β-dehydrogenase 12 (HSD17 β12), aldolase C fructose-bisphosphate (ALDOC), cyclin D1 (CCND1), NAD(P)H dehydrogenase quinone 1 (NQO1), and enoyl coenzyme A hydratase domain-containing 1 (ECHD). The four remaining genes were up-regulated genes: proline-rich nuclear receptor coactivator 1 (PNRC1), glutathione peroxidase 3 (GPX3), alcohol dehydrogenase 1C (ADH1C), and N-deacetylase/N-sulfotransferase 1 (NDST1). Except for NDST1, results were in complete agreement with microarray data (Table 3). Moreover, the down-regulation of SCD and ACACA mRNA levels detected with microarray experiments was in accordance with the decrease observed in the two dietary groups of set 1 by RT-qPCR (Fig. 2). Energy, lipid, and carbohydrate metabolisms are the main biological processes involving genes down-regulated by calorie restriction (Fig. 3). The most relevant and dense networks assembled by Ingenuity Pathways Analysis with down-regulated genes (see supplementary Figure 1, published as supplemental data on The Endocrine Society’s Journals Online Web site at http://jcem.endojournals.org, and supplementary Table 3) are related to lipid metabolism, cellular assembly, organization, and small molecule biochemistry. Network 1 illustrates possible coregulation of electron transport chain complex I and acyl CoA synthase. Network 2 shows that TNF could play a central role in the interaction between enzymes involved in β-oxidation and fat storage. The third network shows linkage between electron transport chain complex IV, heat shock proteins, and the release of cytochrome c.
|
|
The candidate gene analysis revealed genes differentially regulated in the two diets. We extended our study using whole-genome DNA microarray analysis of set 2 to analyze whether more transcripts could be detected. We used two statistical methods dedicated to classification and selection of genes that are more relevant than SAM.
In the random forest classification analysis, genes with the highest importance in the classification accuracy (significance value under 0.05) were selected. A list of 64 cDNA clones as potential markers of macronutrient composition of the diet was obtained (supplementary Table 4). Independently, Prediction Analysis of Microarray identified the most specific changes in gene expression induced by each diet. Different values of threshold shrinkage were tested and the best accuracy of class prediction by cross validation (77%) was obtained with a list of 57 cDNA clones (supplementary Table 5).
The transcripts identified by the two methods map to 84 genes. Combining the output from microarray and candidate gene approaches 89 genes could be responsive to changes of the fat to carbohydrate ratio of energy-restricted diet. Ontological classification identified eight genes involved in oncogenesis, 12 in apoptosis, 24 in signal transduction, seven in cell cycle control, three in angiogenesis, and nine in lipid and steroid metabolism.
Of the 84 genes identified by the microarray approach as differentially expressed according to the macronutrient content of the diet, we restricted the list excluding clones not linked to known genes in the GenBank database. Moreover, only clones presenting a significant difference in after-to-before energy restriction expression ratio between the MF and LF groups were selected. A list of 12 genes was obtained (supplementary Table 6).
Confirmatory analyses of biomarkers
To validate the accuracy of the selection methods, adipose tissue mRNA levels of the 12 selected genes were quantified before and after dietary intervention by RT-qPCR in samples from set 2 (Fig. 1). According to the general linear model analysis, a differential effect of diet (P < 0.1) tended to be confirmed for nuclear receptor subfamily 3, group C, member 1 (NR3C1), GABA A receptor-associated protein-like 2 (GABARAPL2), ornithine decarboxylase antizyme 2 (OAZ2), phosphatase and tensin homolog (PTEN), farnesyltransferase CAAX box (FNTA), and amyloid-βA4 precursor protein (APP). A similar trend was observed on set 1 for FNTA (P = 0.06), NR3C1 (P = 0.08), and GABARAPL2 (P < 0.05). If the two sets of subjects were combined (n = 43–47 for each diet), we detected a significant effect of the macronutrient composition of the diet on NR3C1, GABARAPL2 (both at P < 0.01), and FNTA (P < 0.05) expression change. After an adjustment for BMI and other matching criteria, similar results were obtained (data not shown).
Using candidate gene and pangenomic approaches, cross-validation on two independent sets of subjects and stringent statistical procedures, a differential expression of five genes, FABP4, SIRT3, NR3C1, GABARAPL2, and FNTA, was observed between LF and MF diets (Fig. 4). Of note, a strong correlation was found between NR3C1 and FABP4 mRNA levels (rho = 0.41, P < 0.0001, n = 89). We studied the relationship in diet-induced changes of plasma lipid (Table 2) and mRNA levels (Fig. 4) of parameters influenced by the composition of the diet on the entire group of subjects (n = 91). NR3C1 expression correlated with total and low-density lipoprotein-cholesterol (r = 0.25, P = 0.02 and r = 0.24, P = 0.02, respectively) whereas GABARAPL2 expression correlated with total and high-density lipoprotein-cholesterol (r = 0.22, P = 0.04 and r = 0.23, P = 0.03, respectively).
|
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
As reported for variations in clinical parameters (10), we found a predominant impact of energy restriction and/or weight loss over macronutrient composition on adipose tissue gene expression. Compared with the previous studies on the NUGENOB cohort restricted to a subset of subjects and limited number of transcripts, the list of genes regulated by energy restriction is considerably expanded by our microarray analysis (11, 12). In agreement with these studies, the vast majority of the genes are down-regulated during the diets. Among the most down-regulated transcripts are genes involved in lipid, carbohydrate, and energy metabolism, which represented the majority of genes regulated during the hypocaloric diets. The genes included rate-limiting enzymes of de novo lipogenesis such as FASN and ACACA involved in saturated fatty acid synthesis; SCD, fatty acid desaturase 1 and 2 (FADS1, FADS2) involved in unsaturated fatty acid synthesis; and ELOVL family member 5, elongation of long chain fatty acids (ELOVL5), a long-chain fatty acid elongase. A decrease in triacylglycerol storage capacity was also suggested by the down-regulation of ADPN, GPAM, and lipoprotein lipase (LPL).
Another important result of this study is the differential expression of transcripts according to the dietary fat to carbohydrate ratio. This was achieved because of our unique experimental design. Five transcripts, SIRT3, FABP4, NR3C1, GABARAPL2, and FNTA, showed a reproducible differential expression between the two diets. Of note, the genes showed a similar pattern of regulation, i.e. up-regulation or no change during the diet with an elevated proportion of lipids and down-regulation or no change during the low-fat diet. Expression of these genes was 15–65% higher in the MF than the LF diet.
The regulation of SIRT3 is of interest in the context of metabolic adaptations to various fat and carbohydrate intakes. SIRT3 is found in the matrix of the mitochondrion, the importance of which in the adaptation to energy restriction is increasingly documented (19, 20). SIRT3 deacetylates the mitochondrial enzyme, acetyl-coenzyme A synthetase 2, thereby allowing the entry of carbon from acetate, generated following a diet enriched with fat, into the tricarboxylic cycle (21, 22). This step might be important during times of food limitation to both harvest dietary acetate and make use of acetate generated by the liver during ketogenesis (23).
We also showed a differential regulation of FABP4 by the two diets. FABP4 is involved in the intracellular traffic of fatty acids in the adipocyte and could be a critical mediator of insulin sensitivity (24). An increased FABP4 expression may allow proper handling of the flux of fatty acids. The gene is induced by fatty acids in adipocytes, an effect likely to be mediated by peroxisomal proliferator-activated receptors (25, 26).
Increased expression of the glucocorticoid receptor NR3C1 could increase the cortisol-mediated transcriptional activity, contributing to the regulation of FABP4 and potentially other glucocorticoid receptor targets (27). The very high correlation between NR3C1 and FABP4 gene expression supports this notion. These transcriptional effects could modify the metabolic profile induced by weight loss as suggested by the correlation between NR3C1 adipose tissue gene expression and plasma cholesterol parameters.
The relationship between macronutrients and GABARAPL2 and FNTA has never been documented. GABARAPL2 modulates intra Golgi transport through an interaction with soluble N-ethylmaleimide-sensitive factor attachment protein receptor complexes (28), which was described to control adipocyte GLUT4 translocation in response to insulin (29, 30). The role of the farnesyltransferase FNTA in adipose tissue remains elusive. Of note, it has been suggested that farnesylation of proteins may influence adipose tissue development (31).
To increase the relevance of this work, highly stringent criteria were applied to transcripts identified as sensitive to dietary intake. It is highly likely that other genes selected by the classification and prediction methods show real differential expression between the MF and LF diets.
To conclude, during hypoenergetic diets, the primary determinant of changes in adipose tissue gene expression is energy restriction rather than the composition in fat and carbohydrate. The regulation in energy metabolism-related processes and coregulatory pathways may explain the variations in anthropometric and biological parameters. However, the macronutrient content of the diets influences expression of a subset of genes, which may contribute to differential response in blood lipid profile.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Disclosure Statement: The authors have nothing to declare.
First Published Online September 9, 2008
1 Author Affiliations are shown at the bottom of the page 2 
Abbreviations: aRNA, Amplified RNA; BMI, body mass index; Cy, cyanine dyes; HOMA, homeostasis model assessment; Inserm, Institut National de la Santé et de la Recherche Médicale; LF, low-fat, high-carbohydrate diet; MF, moderate-fat, low-carbohydrate diet; NUGENOB, Nutrient-Gene Interaction in Human Obesity; RT-qPCR, RT-quantitative PCR; SAM, significance analysis of microarray.
Institut National de la Santé et de la Recherche Médicale, Institut de Médecine Moléculaire de Rangueil, Unit 858, Laboratoire de Recherches sur les Obésites (F.C., N.Vi., E.K., D.L.), 31432 Toulouse, France; Institut Louis Bugnard IFR31, Centre Hospitalier Universitaire de Toulouse and Université Paul Sabatier (F.C., N.Vi., E.K., D.L.), 31403 Toulouse, France; Franco-Czech Laboratory for Clinical Research on Obesity (F.C., N.Vi., E.K., D.L.), 100 00 Prague, Czech Republic; Institut National de la Santé et de la Recherche Médicale, Unité Mixte de Recherche 870, Institut National de la Recherche Agronomique Unit 1235, Faculté de Medecine Lyon-Sud and Université de Lyon 1 (N.Ve., H.V.); 69600 Oullins France; Centre de Recherche en Nutrition Humaine (N.Ve., H.V.), Hospices Civils de Lyon, 69622 Lyon, France; Institut de Mathématiques de Toulouse, Unité Mixte de Recherche 5219 (S.D.), Université Paul-Sabatier, 31403 Toulouse, France; Department of Medicine, Karolinska Institute, Karolinska University Hospital Huddinge (P.A.), M54, 141 86 Stockholm, Sweden; Department of Sports Medicine, Centre of Preventive Medicine, Third Faculty of Medicine, Charles University (E.K.), 100 10 Prague 10, Czech Republic; Department of Physiology and Nutrition, University of Navarra (J.A.M.), 31080 Pamplona, Spain; Department of Human Biology, Nutrition and Toxicology Research Institute Maastricht, Maastricht University (W.H.M.S.), NL-6200 MD Maastricht, The Netherlands; Institute of Preventive Medicine, Centre for Health and Society and Copenhagen University Hospital (C.H., T.I.A.S.), DK-1357 Copenhagen, Denmark; School of Biomedical Sciences Queen’s Medical Centre and University of Nottingham Medical School (M.T.), Nottingham NG7 2UH, United Kingdom; Institut National de la Santé et de la Recherche Médicale, Nutriomique Unit 872 (Team 7) (J.M.O., K.C.), 75004 Paris, France; Unité Mixte de Recherche-S 872, Universités Pierre et Marie Curie Paris 6 (J.M.O., K.C.), 75005, Paris, France; Cordelier Research Center and Paris Descartes (J.M.O., K.C.), 75006 Paris, France; and Department of Nutrition and Endocrinology, Assistance Publique/Hôpitaux de Paris, Pitié Salpêtrière Hospital (J.M.O., K.C.), 75013 Paris, France
Received April 15, 2008.
Accepted August 29, 2008.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  K. R. Ong, A. H. Sims, M. Harvie, M. Chapman, W. B. Dunn, D. Broadhurst, R. Goodacre, M. Wilson, N. Thomas, R. B. Clarke, et al. Biomarkers of Dietary Energy Restriction in Women at Increased Risk of Breast Cancer Cancer Prevention Research, August 1, 2009; 2(8): 720 - 731. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
, P < 0.05 between the two diets.
