The FASEB Journal • Research Communication

Weight loss regulates inflammation-related genes in white adipose tissue of obese subjects KARINE CLE´MENT,*,1 NATHALIE VIGUERIE,†,‡ CHRISTINE POITOU,* CLAIRE CARETTE,* VE´RONIQUE PELLOUX,* CYRILE A. CURAT,§ AUDREY SICARD,†,‡ SOPHIE ROME,¶ ARRIEL BENIS,*,
JEAN-DANIEL ZUCKER,*,
HUBERT VIDAL,¶ MARTINE LAVILLE,¶ GREGORY S. BARSH,# ARNAUD BASDEVANT,* VLADIMIR STICH,‡,** RAFFAELLA CANCELLO,* AND DOMINIQUE LANGIN†,‡,1 *INSERM "Avenir" Paris 6 University, EA 3502, Nutrition Department, AP/HP, Hoˆtel-Dieu, Paris, France; †Unite´ de Recherches sur les obe´site´s Inserm UPS U586, Institut Louis Bugnard, Centre Hospitalier Universitaire de Toulouse, Universite´ Paul Sabatier, Toulouse, France; ‡Franco-Czech Laboratory for Clinical Research on Obesity, French Institute of Health and Medical Research (Inserm U586) and 3rd Faculty of Medicine, Charles University, Prague, Czech Republic; §Institut fu¨r Kardiovasculare Physiologie, J. W. Goethe-Universita¨t, Frankfurt am Main, Germany; ¶INSERM U449 and Human Nutrition Research Center of Lyon, Faculty of Medicine R. Laennec, Lyon, France;
LIM & BIO, Paris Nord University, France; #Department of Pediatrics and Genetics, Howard Hughes Medical Institute, Beckman Center, Stanford University School of Medicine, California, USA; and **Department of Sports Medicine and Obesity Unit, 3rd Faculty of Medicine, Charles University, Prague, Czech Republic Adipose tissue produces inflammation and immunity molecules suspected to be involved in obesity-related complications. The pattern of expression and the nutritional regulation of these molecules in humans are poorly understood. We analyzed the gene expression profiles of subcutaneous white adipose tissue from 29 obese subjects during very low calorie diet (VLCD) using cDNA microarray and reverse transcription quantitative PCR. The patterns of expression were compared with that of 17 nonobese subjects. We determined whether the regulated genes were expressed in adipocytes or stromavascular fraction cells. Gene expression profiling identified 100 inflammation-related transcripts that are regulated in obese individuals when eating a 28 day VLCD but not a 2 day VLCD. Cluster analysis showed that the pattern of gene expression in obese subjects after 28 day VLCD was closer to the profile of lean subjects than to the pattern of obese subjects before VLCD. Weight loss improves the inflammatory profile of obese subjects through a decrease of proinflammatory factors and an increase of anti-inflammatory molecules. The genes are expressed mostly in the stromavascular fraction of adipose tissue, which is shown to contain numerous macrophages. The beneficial effect of weight loss on obesity-related complications may be associated with the modification of the inflammatory profile in adipose tissue.— Cle´ment, K., Viguerie, N., Poitou, C., Carette, C., Pelloux, V., Curat, C. A., Sicard, A., Rome, S., Benis, A., Zucker, J.-D., Vidal, H., Laville, M., Barsh, G. S., Basdevant, A., Stich, V., Cancello R., Langin, D. Weight loss regulates inflammation-related genes in white adipose tissue of obese subjects. FASEB J. 18, 1657–1669 (2004)

ABSTRACT

0892-6638/04/0018-1657 © FASEB

Key Words: obesity 䡠 gene profiling 䡠 calorie restriction 䡠 microarray

Adipose tissue has a high capacity to produce and secrete molecules involved in numerous functions including metabolism, insulin secretion, reproduction, as well as immunity and inflammation (1). This latter class of factors is of utmost interest since inflammationrelated molecules released from adipose tissue could contribute to the development of complications associated with obesity such as diabetes and cardiovascular diseases (2– 4). The contribution of inflammation to the development and maintenance of metabolic diseases in humans mostly stems from the measurement of circulating inflammatory molecules in obese and nonobese cohorts. Adult and younger obese subjects show increased circulating levels of acute-phase proteins, proinflammatory cytokines, and chemokines (5–9). Some circulating cytokines are correlated with increased fat mass, such as interleukin (IL)-6, IL-8, and tumor necrosis factor ␣ (TNF-␣) (2–5). A modest elevation of these inflammation-related molecules in the circulation may contribute to a substantial increased risk of cardiovascular stroke and mortality (6). Obesity could then be viewed as a metabolic as well as a low-grade inflammatory disease like atherosclerosis (7, 8). Weight loss is usually associated with decreased 1

Correspondence: K. C., Department of Nutrition, Hoˆtel-Dieu, 75181 Paris cedex 04, France. E-mail: [email protected] or D.L., Inserm UPS U586, Institut Louis Bugnard, Baˆtiment L3, CHU Rangueil, 31403 Toulouse Cedex 4, France. E-mail: [email protected] doi: 10.1096/fj.04-2204com 1657

concentrations of inflammation-related products in the circulation (9 –13). Adipose tissue contains mature adipocytes and cell types of various lineages in the stromavascular fraction (SVF) that include preadipocytes, endothelial cells, and macrophages. The cellular origin of inflammatory markers is not known and the nutritional regulation in humans of adipose tissueproduced factors is poorly understood. Our objective was to study the consequence of negative energy balance on the inflammation-related genes in adipose tissue. We determined gene expression in subcutaneous white adipose tissue of obese subjects after different nutritional challenges using cDNA microarrays and reverse transcription-quantitative PCR (RT-qPCR). Gene expression profiles of these obese subjects were compared with those of non-obese subjects. Expression of regulated genes was measured in adipocytes and SVF cells. Our results show the expression of a large panel of inflammation-related proteins in human adipose tissue. Weight loss significantly improves the inflammatory profile of obese subjects through a decrease in proinflammatory factors and an increase in anti-inflammatory molecules. The larger fraction of these genes is expressed in the SVF of adipose tissue.

MATERIALS AND METHODS Subjects and clinical investigation protocols This study included Caucasian obese women involved in two nutritional challenges and a group of non-obese subjects. All women were premenopausal. None of the subjects studied had a familial or personal history of diabetes or was taking medication. The subjects were at their maximal peak weight and were weight stable at the time of the study. None were involved in an exercise program. All clinical investigations were performed according to the Declaration of Helsinki and approved by the Ethical Committees of Hoˆtel-Dieu (Paris) and the Third Faculty of Medicine (Prague). Informed consent was obtained from obese and lean subjects. The clinical characteristics of the patients involved in all the clinical investigation procedures are presented in Table 1. In the first nutritional challenge, we assessed the changes in gene expression after a very low calorie diet (VLCD 800 kcal/day) of 28 days in obese women (28 day VLCD, subgroup 1). Microarray analysis was performed in 10 of 21 obese women (Table 1). In the second nutritional challenge, we analyzed changes in gene expression after a 2 day 650 kcal caloric restriction in eight severely obese women (2 day VLCD, subgroup 2). This corresponded to a 70% decrease of their usual daily energy intake. We studied the profile of adipose tissue gene expression in 6 healthy weight-stable women from a group of 17 individuals (Table 1). In all subjects, subcutaneous abdominal adipose tissue samples were obtained by needle biopsy from the periumbilical area under local anesthesia (1% xylocaine) at day 1 (baseline) and day 28 for the obese women subjected to the 28 day VLCD (subgroup 1) and on days 1 (baseline) and 3 for subjects involved in the 2 day VLCD (subgroup 2), as well as in the weight-stable control subjects at baseline. Adipose tissue specimens, obtained after an overnight fast following the same standard operating procedures in all the groups, 1658

Vol. 18

November 2004

were immediately frozen in liquid nitrogen and stored at – 80°C until analysis. Blood samples were obtained for biochemical and hormonal evaluation. Body composition was assessed by dual-energy X-ray absorptiometry performed with a total body scanner. Resting metabolic rate was evaluated after a 1 h rest in supine position. Oxygen consumption (VO2) and carbon dioxide production (VCO2) were monitored over 30 min using an open-circuit ventilated canopy system (Deltatrac II monitor, Datex Instrumentarium Corp., Helsinki, Finland) calibrated with a reference gas. Resting metabolic rate was derived from VO2 and VCO2 by using indirect calorimetry. Plasma glucose was determined with the glucose oxidase technique (Biotrol kit, Merck-Clevenot, Nogent-sur-Marne, France). Plasma insulin concentrations were measured using RIA kits from Sanofi Diagnostics Pasteur (Marnes la Coquette, France). Leptin plasma levels were determined using radioimmunoassay kits from Linco Research (Saint Louis, MO, USA) according to the manufacturer’s recommendations. Plasma triglyceridemia, total and HDL cholesterolemia were measured in the hospital biochemistry departments using clinically validated protocols. Isolation of human white adipocytes and cells from the stromavascular fraction Subcutaneous abdominal white adipose tissue was obtained from nine healthy female subjects (mean body mass index, BMI 28⫾7 kg/m2) undergoing plastic surgery in agreement with French laws on biomedical research. This group of subjects was different from the subjects who participated in the clinical investigation protocols. The tissue sample was cut into small pieces and processed as described previously (14). Briefly, mature adipocytes were collected after collagenase (type II, Gibco, Cergy pontoise, France) cell dissociation at 37°C, filtration, decantation, and centrifugation. The floating cellular layer was kept free of any detectable stromavascular element and contained only mature adipocytes filled with triglyceride droplets (tested by light microscopy). SVF cells were recovered at the bottom of the tubes after centrifugation. For measurements of mRNA levels, adipocytes and SVF cells were lysed with denaturing buffer from RNeasy kit (Qiagen, Courtaboeuf, France), then stored at – 80°C until RNA preparation. Isolation of macrophages from the stromavascular fraction The SVF, homogenized at the maximal concentration of 40 million cells/mL of PBS/2% fetal calf serum (FCS), was incubated at room temperature for 15 min with 100 ␮L of Easysep positive selection cocktail (Stem Cell Technologies, Meylan, France)/mL of cell suspension, followed by a 10 min incubation with 50 ␮L of Easysep magnetic nanoparticles/mL of cell suspension. After successive magnetic sorting steps and washes with PBS/2%FCS, the CD34 negative subset, collected after each sorting step was centrifuged at 200 ⫻ g and homogenized in 1 mL PBS/2%FCS. The CD34 negative cell fraction was used to isolate the resident macrophages with 50 ␮L of CD14-coupled magnetic microbeads (Dynal Biotech, Hamburg, Germany)/mL of CD34-negative cell fraction. After 20 min incubation at 4°C under constant shaking, the cell suspension containing the beads was briefly homogenized in 10 mL PBS/2%FCS and placed in front of the magnet for 1 min. After two washes in PBS/0.1% bovine serum albumin, the magnetic beads-coupled fraction, which corresponds to the CD34-/CD14⫹ cell population, was collected and centrifuged for 10 min at 200 ⫻ g.

The FASEB Journal

CLE´MENT ET AL.

Immunomorphological analysis of adipose tissue in obese subjects The cellular origin of gene products expressed in macrophages or isolated adipocytes was studied on samples collected during surgical biopsies of subcutaneous abdominal white adipose tissue from morbidly obese women (n⫽10, BMI 48⫾4 kg/m2, age 41⫾14 years) undergoing bariatric surgery. The samples were fixed in 4% paraformaldehyde, dehydrated, paraffin embedded, then sectioned (thin sections 5 ␮m thick). Sections were stained with hematoxylin and eosin and/or processed for immunohistochemical detection of macrophage markers according to standard immunoperoxidase procedure. Immunohistochemical detection of HAM56 (Dako Cytomation, Trappes, France), CD68 (Santa Cruz Biotechnologies, Heidelberg, Germany), and serum amyloid (SAA) (Dako) was performed with the avidin-biotin peroxidase (ABC) method (15). De-waxed sections were processed through incubation steps: 1) antigen unmasking by 750W microwave washing three times in a solution of citrate buffer 10 mM pH6; 2) hydrogen peroxide 3% in water for 15 min to block endogenous peroxidase; 3) Tris-buffered saline/ Tween20-casein 0.02 M solution (TBS-TC) for 10 min; 4) monoclonal mouse antibodies diluted 1:100 for SAA antibody, 1:300 for CD68 and 1:500 for HAM56 were incubated for 1 h in TBS-TC at room temperature; 5) multilink antimouse biotinylated immunoglobulins (Dako) diluted 1:200 in TBS-TC for 20 min; 6) standard streptavidin– biotin–peroxidase complex (SABC) method was applied using a commercially available kit (ABCYS GMR4-61; Biospa, Milano, Italy); 7) the staining was visualized using diaminobenzidine and slides were counterstained with Mayer’s hematoxylin. Methods specificity tests were performed: omission of primary antibodies in the staining and use of preimmune serum instead of the first antiserum. Total RNA extraction and mRNA amplification Total RNA was prepared using the RNeasy total RNA Mini kit (Qiagen). RNA concentration and integrity were assessed using the Agilent 2100 Bioanalyzer (Agilent Technologies, Massy, France). For microarray experiments, 1 ␮g of total RNA from each total RNA sample preparation was amplified by MessageAmp RNA Kit (Ambion, Austin, TX, USA) and 3 ␮g of amplified RNA was labeled with cyanin dyes (Cy) using the CyScribe First-Strand cDNA labeling kit (Amersham Biosciences, Orsay, France). We (16) and others (17) have checked that mRNA amplification did not induce a distortion in mRNA representation. To compare microarrray experiments, we used a common reference pool generated by mixing equal amounts of total RNA extracted from adipose tissue samples of patients undergoing plastic surgery. Amplified RNA from the reference pool was labeled with Cy3 and aRNA from the testing samples were labeled with Cy5. A total of 42 individual arrays for each condition were performed. We compared total RNA isolated from adipocytes and from SVF cells. Here, the microarray experiments were performed after pooling an equal amount of total RNA from adipocyte and from SVF cell preparations and repeated six times. Amplified RNA from SVF cells was labeled with Cy3 whereas aRNA from adipocytes was labeled with Cy5. The labeled cDNA mixtures were hybridized according to the protocol described at http://cmgm.stanford.edu/pbrown/protocols/ index. mRNA quantification by reverse transcription-real-time PCR Reverse transcription and real-time PCR were performed as described (18). We used 18S rRNA (rRNA Control TaqMan WEIGHT LOSS AND INFLAMMATION IN ADIPOSE TISSUE

Assay kit; Applied Biosystems, Foster City, CA, USA) as normalization control for gene expression; primers and TaqMan probes for mRNA were also obtained from Applied Biosystems. These probes were labeled with a reporter dye (FAM) on the 5⬘ end. The probe for 18S rRNA was labeled with the reporter dyes VIC and TAMRA on the 5⬘ end and the 3⬘ end, respectively. For each primer and probe pair, a standard curve was obtained using serial dilutions of human adipose tissue cDNA prior to mRNA quantitation. Data analysis of microarray The cDNA microarrays produced at Stanford University consisted of PCR-amplified cDNAs printed on glass slides with 42,786 spots representing 29,308 UniGene clusters. The clones and description are available online at the Stanford Microarray Database website (http://genome-www5.stanford.edu/). The detailed procedure has been described (16, 18, 19). After scanning the arrays, the images were analyzed and ratio intensities were filtered and normalized using Stanford Microarray Database Online resources (http://genome-ww5.stanford.edu/ MicroArray/SMD/). We performed a filtering procedure omitting manually flagged elements (i.e., bad quality spots). A uniform scale factor was then applied to all measured intensities to normalize signal intensities between both images. The log2 Cy5/Cy3 ratios for all conditions were extracted and spots with an average intensity ⬍2.5-fold above the median average intensity of the local background were eliminated. Using this procedure, 39,691 spots were recovered in VLCD experiments. Before calculations, the data from all experiments were normalized using a lowess procedure (http://www.tigr.org/software/tm4/ midas.html) (20). Unless otherwise indicated, differential gene expression was assessed using the significant analysis of microarray (SAM) procedure (http://www-stat.stanford.edu/⬃tibs/ SAM/index.htlm), which provides a list of significant genes and an estimate of the false discovery rate representing the percentage of genes that could be identified by chance (21). To analyze the function of the known genes, we used Gene Ontology annotations (22) (www.geneontology.org). We focused on genes whose ontological criteria shared the following syntax: immune or inflammatory response, acute-phase response, cellular defense, and response to stress. For the selected genes, we identified the other clones representing on the array the same gene (that is, the same UniGene number) as the one sorted using SAM. Several cDNA clones (i.e., different Clone ID) may correspond to the same gene (i.e., same UniGene number). To add an additional step of selection in our procedure, we tested whether transcripts hybridizing to the different cDNAs derived from the same gene are similarly regulated. For each UniGene number corresponding to a cDNA selected by SAM, values for all corresponding clone IDs were tested using the paired Student t test. Only genes with a common pattern of expression among the replicates were selected. We then examined the pattern of expression of the selected inflammatory-related genes in all patient groups (obese patients of subgroups 1 and 2 at baseline and after calorie restriction and in lean subjects). We performed an agglomerative ascending hierarchical clustering to both mean gene ratios and experiments using the nonparametric Pearson rank correlation coefficient as a measure of similarities and average linkage clustering (23). The resulting dendrograms were visualized by the Tree View Software (http:// www.microarrays.org/software/). The significance of a cluster is related to the distance measured between two profiles used by the clustering algorithm and the normalization methods. To further assess the significance of the hierarchy proposed by the agglomerative ascending hierarchical clustering algorithm, we computed intercluster dis1659

tances of alternate hierarchies. This distance computation provides a quantitative measure of the quality of the clustering of conditions. Different distances (Euclidian, Manhattan) were used to evaluate the dependency of the gene ranking with respect to the bias introduced by the choice of the distance. C0, C1, C2, C3, and C4 represented lean subjects, obese subjects before 28 day VLCD, obese subjects before 2 day VLCD, obese subjects after 2 day VLCD, and obese subjects after 28 day VLCD, respectively. Mean values of expression are used for each partition. Cluster K1 gathered C0 and C4 whereas cluster K2 gathered C1, C2, and C3. The quality of the partition of the five experimental conditions into the two clusters K1 and K2 was measured by computing the distance between these two clusters. The fact that the clustering algorithm outputs this partition suggests that the distance between K1 and K2 is maximal. To assess the significance of this partition, we first computed the distance between K and K’ of all partitions of the five conditions into two clusters of at least two elements. We used different distances to control the importance of distance itself in the computation. Statistical analysis of target genes Data are expressed as means ⫾ se. For microarray experiments on samples collected in obese subjects before and after VLCD, SAM was applied with a false discovery rate of 0.025 (see above). For comparison between adipocytes and SVF cells, SAM was used with a false discovery rate of 0.05. Comparison of inflammatory gene expression between obese and lean subjects was performed with T permutation statistics implemented in MeV (MultiExperimentViewerTIGR), a Java application designed for analysis of microarray data, with P ⬍ 0.1. Age adjustments were made when necessary. Conventional statistical analysis was performed with JMP statistics software (SAS Institute Inc., Cary, NC, USA). For genes analyzed by RT-qPCR, significant differences were determined by Wilcoxon nonparametric paired test (before or after VLCD; adipocyte vs. SVF) and Student t test with unequal variance (obese vs. lean subjects). The correlations between mRNA levels of the different transcripts were examined by the nonparametric Spearman’s rank correlation test. P ⬍ 0.05 was the threshold of significance.

RESULTS Clinical and metabolic characteristics of obese subjects before and after 2 or 28 day VLCD Table 1 shows the characteristics of non-obese women and of obese subjects at various energy intakes. VLCD during 28 day resulted in a significant reduction of weight, BMI, fat mass (P⬍0.001), and a decrease in plasma concentration (P⬍0.0001) and adipose tissue mRNA levels (data not shown) of leptin as well as in an improvement of insulin sensitivity, as expected. An improvement of triglyceride and cholesterol levels was noted. The short-term reduction of food intake (2 day VLCD) performed in obese subjects with higher BMI led to a small decrease in body weight. Simultaneous decrease in plasma leptin levels and an improvement in insulin sensitivity were observed (Table 1). The subgroups of subjects used for cDNA chips and RT-qPCR 1660

Vol. 18

November 2004

experiments were not different from the whole groups for any of the tested parameters (data not shown). Global gene expression profiling in adipose tissue before and after 28 day VLCD Control for multiple testing of cDNA microarray experiments was performed using the SAM procedure (21). On cDNAs with signals recovered in at least 8/10 experiments, using an estimated false discovery rate of 0.025, we detected 1923 genes with Gene Ontology annotations (http://source.stanford.edu). The four main functional classes were metabolism (58%), cell growth and/or maintenance (33%), response to external stimulus (9.6%), and response to stress (7.5%). From Gene Ontology annotations, 171 genes encoded proteins linked to the inflammatory process. Seventyseven genes were represented by a unique clone on the microarray whereas other genes had at least two cDNA replicates. We tested whether the ratios of the different replicates for each UniGene number were significant using the paired Student’s t test (P⬍0.05). The final list of genes comprised 100 differentially expressed genes. Forty-one were overexpressed after VLCD (mean fold change: 1.59 range 1.2–2.9) whereas 59 were downregulated (mean fold change: 0.62 range 0.8 – 0.4). It must be stressed that this selection procedure is highly stringent because each mRNA variant of a gene produced by alternative promoters, splicing, or polyadenylation sites has the same UniGene number but is not necessarily identically regulated. As shown in Table 2, data obtained from microarrays were validated using RT-qPCR on 10 genes, including up-, down-, and unregulated genes. As shown on supplementary Table 1 (http://www. fasebj.org) and Fig. 1, the 100 genes were more precisely classified in 12 functional categories. The list included cytokines, interleukins, complement-related factors and acute-phase proteins but also molecules involved in cell-cell and cell-matrix adhesion as well as factors involved in cell proliferation, growth, and differentiation. Genes encoding factors involved in acutephase response, growth, and differentiation were mostly down-regulated after caloric restriction whereas other functional categories were equally represented in up and down-regulated genes (Fig. 1). Expression patterns of inflammation-related genes in the different nutritional conditions We further analyzed the pattern of expression of the set of genes differentially expressed after 28 day VLCD in obese subjects before and after 2 or 28 day calorie restriction and in non-obese women. Nonparametric hierarchical clustering was performed with the mean ratio of the abundance of the transcripts of each gene to the median abundance across the experimental conditions. As shown on Fig. 2, gene expression clustering led to two major groups. Cluster 1 gathered transcripts repressed by 28 day VLCD in obese subjects

The FASEB Journal

CLE´MENT ET AL.

TABLE 1. Clinical characteristics of the subjects Non obese

Obese Subgroup 1

Number

Baseline

Baseline

17

21 (10) 39 ⫾ 2b (41.3 ⫾ 2.2) 94 ⫾ 3.0b (98.0 ⫾ 4.0) 35.0 ⫾ 0.9b (36.0 ⫾ 0.9) 46.0 ⫾ 2.0b (48.0 ⫾ 2.5) 35.5 ⫾ 0.84 (35.0 ⫾ 1.3) 4.8 ⫾ 0.13 (5.04 ⫾ 0.2) 5.9 ⫾ 0.5 (5.4 ⫾ 0.9) 0.37 ⫾ 0.006 (0.38 ⫾ 0.01) 1.4 ⫾ 0.1b (1.6 ⫾ 0.3) 5.2 ⫾ 0.2a (5.7 ⫾ 0.4) 1.05 ⫾ 0.04a (1.1 ⫾ 0.1) 30.8 ⫾ 2.7 (32.0 ⫾ 4.5)

Age (years)

30.0 ⫾ 1.8

Body weight (kg)

65.7 ⫾ 1.7

BMI (kg/m2)

23.0 ⫾ 0.4

Body fat mass (kg)

19.2 ⫾ 1.9

RMR (kcal/kg FFM/day) Glucose (mmol/L)

ND 5.0 ⫾ 0.13

Insulin (␮U/mL)

5.5 ⫾ 0.7

QUICKI

0.37 ⫾ 0.009

Triglycerides (mmol/L)

0.48 ⫾ 0.05

Total cholesterol (mmol/L)

4.3 ⫾ 0.4 1.44 ⫾ 0.08

HDL-C (mmol/L) Leptin (ng/mL)

ND

Subgroup 2

28 day VLCD

Baseline

2 day VLCD

8 41 ⫾ 2.3a 88 ⫾ 3.0** (91.0 ⫾ 3.8)* 32.0 ⫾ 0.98** (34 ⫾ 0.8)* 40.0 ⫾ 2.0** (43 ⫾ 3)* 32.5 ⫾ 0.96** (32.0 ⫾ 1.5)* 4.5 ⫾ 0.15 (4.6 ⫾ 0.2) 4.2 ⫾ 0.6* (3.7 ⫾ 0.8) 0.41 ⫾ 0.013** (0.42 ⫾ 0.02)* 1.1 ⫾ 0.07* (1.3 ⫾ 0.2) 4.6 ⫾ 0.2** (5.1 ⫾ 0.4)* 0.9 ⫾ 0.03* (1 ⫾ 0.1) 14.5 ⫾ 2.6** (16.0 ⫾ 4)*

120.0 ⫾ 7.7b 46.6 ⫾ 2.5b 61 ⫾ 7.0b 40.0 ⫾ 1.6

119.0 ⫾ 8† 46.5 ⫾ 2.5 ND ND

5.1 ⫾ 0.16 18.4 ⫾ 2.4

5.1 ⫾ 0.3 14 ⫾ 2.9†

0.31 ⫾ 0.004a

0.34 ⫾ 0.01†

1.24 ⫾ 0.09b

ND

5.1 ⫾ 0.2a

ND

1.4 ⫾ 0.2

ND

57.5 ⫾ 6

51.0 ⫾ 5†

BMI, body mass index; FFM, fat free mass; HDL-C, high density lipoprotein-cholesterol; RMR, resting metabolic rate; QUICKI, quantitative insulin resistance index; ND: not determined. Values in italics and parentheses represent clinical and biological data from the individuals used a in the pangenomic microarray study in subgroup 1 (see Materials and Methods). P ⬍ 0.05 and b P ⬍ 0.001 clinical data compared with baseline values of nonobese subjects. * P ⬍ 0.05 and ** P ⬍ 0.001 comparison between baseline value (day 1) and 1 month VLCD (day † 28) in obese subgroup 1. P ⬍ 0.01 comparison between baseline values (day 1) and 2 day VLCD (day 3) in obese subgroup 2.

whereas cluster 2 gathered genes induced by 28 day VLCD. For a majority of genes, expression patterns were similar in obese subjects before short- and longterm VLCD and after short-term VLCD. RT-qPCR was used to validate clustering findings in 18 obese and 16

lean subjects (Table 3). In accordance with cluster 1 data, SAA4 and carboxylesterase 1 (CES1) were overexpressed in obese subjects; in cluster 2, we validated the transcriptional changes for integrin ␣ L. The pertinence of the first cluster analysis was further supported

TABLE 2. Validation of gene expression changes in response to 28 day VLCDa Response to VLCD

Up

Down Nonsignificant

UG cluster

Name

Hs.673 Hs.174103 Hs.1511739 Hs.50002 Hs.1955 Hs.76688 Hs.170917 Hs.173894 Hs.234726

Interleukin 12A Integrin alpha L Matrix metalloproteinase 9 Chemokine (COC) motif ligand 19 Serum amyloid A4 Carboxylesterase 1 Prostaglandin E receptor 3 Colony-stimulating factor 1 Serine (or cysteine) proteinase inhibitor, clade A, member 3 H factor 1 (complement)

Hs.250651

Array ratio (n⫽10)

RTqPCR ratio (n⫽21)

P value RTqPCR

1.46 1.72 1.50 1.82 0.58 0.60 0.71 1.13 1.14

1.84 1.60 1.80 1.20 0.55 0.57 0.78 1.4 1.20

0.04 0.0025 0.03 NS 0.05 0.01 0.09 NS NS

0.70

1.20

NS

a VLCD, very low calorie diet, NS, nonsignificant. To validate microarray data, reverse transcription quantitative PCR (RT-qPCR) was performed on subcutaneous adipose tissue total RNA of 21 obese subjects before and after VLCD. Statistical tests of RT-qPCR data were performed using nonparametric Wilcoxon signed rank tests. Nonsignificant genes were chosen among genes related to inflammation that were not statistically different after our filtering and analysis procedures.

WEIGHT LOSS AND INFLAMMATION IN ADIPOSE TISSUE

1661

TABLE 3. Validation of differences in the expression of 3 genes between obese and non-obese subjectsa

Cluster

1 1 2

UG Cluster

Name

Hs.1955 Serum amyloid A4 Hs.76688 Carboxylesterase 1 Hs.174103 Integrin alpha L

Obese/lean ratio

P value

4.80 3.30 0.18

0.01 0.001 0.0007

a Cluster 1 and 2 gather respectively down- and up-regulated by the VLCD. To validate microarray data, reverse transcription quantitative PCR (RT-qPCR) was performed on subcutaneous adipose tissue total RNA of 18 obese and 16 nonobese subjects. Statistical tests of RT-qPCR data were performed using Student’s t test with unequal variance.

Figure 1. Histogram for functional classes of inflammationrelated genes regulated after 28 day of very low calorie diet (VLCD) in obese subjects. Open and filled bars represent down and overexpressed genes, respectively. Chi2 ⫽ 16.4, P ⫽ 0.1 when comparing number of genes in all functional classes between up- and down-regulated genes. Chi2 ⫽ 6.4, P ⫽ 0.04 when comparing only the number of genes in acute-phase response, growth, and differentiation, and complement-related factor classes between up and down-regulated genes.

by a second, independent analysis. Three experimental conditions (obese subjects before 28 day VLCD, obese subjects before 2 day VLCD, obese subjects after 2 day VLCD) clustered together in a cluster called K1 (Table 4). The clustering revealed similarities between the global gene expression pattern of obese patients after weight loss induced by 28 day VLCD and that of non-obese subjects in a cluster K2. Comparing distances between possible clusters, the distance between cluster K2 and K1 was ranked the longest. Increased expression of genes encoding anti-inflammatory proteins after VLCD We hypothesized that the increased expression of inflammation-related genes (cluster 2) could be associ-

Figure 2. Cluster of inflammation-related genes in non-obese subjects and obese patients after different nutritional challenges. Each row represents a single gene. For each gene are shown the mean average log2 ratios of 6, 10, and 8 hybridizations, respectively, performed in non-obese subjects and obese subjects before and after 2 or 28 day VLCD. Thus, each column represents the mean ratios of an experimental condition in several subjects. The ratio of the abundance of the transcripts of each gene to the mean abundance across the experimental conditions is represented by the color of the corresponding cell in the matrix file. Green, red, and black lines represent transcripts respectively below, above, or equal to the median. The upper dendrogram shows similarities in the expression pattern between each group of subjects. The left dendrogram represents the clustering of genes in two groups.

1662

Vol. 18

November 2004

The FASEB Journal

CLE´MENT ET AL.

TABLE 4. Representation of the distances between metabolic conditions using Euclidian and Manhattan distancesa First partition

C4 C3 C3 C4 C4 C0 C2 C0 C0 C4

C0 C2 C1 C1 C2 C3 C1 C2 C1 C3

Second partition

C3 C4 C4 C0 C0 C4 C4 C4 C4 C0

C2 C0 C0 C3 C3 C2 C0 C3 C3 C2

C1 C1 C2 C2 C1 C1 C3 C1 C2 C1

Euclidian

Manhattan

4.1 3.8 3.2 3.2 3.0 3.0 2.7 2.5 2.4 2.4

30.9 29.3 22.4 25.6 20.4 19.6 18.4 18.7 16.4 17.7

a

C0, C1, C2, C3, and C4 represent lean subjects, obese subjects before 28d VLCD, obese subjects before 2d VLCD, obese subjects after 2d VLCD, and obese subjects after 28d VLCD, respectively. Mean values of expression are used for each partition.

ated with the regulation of molecules with anti-inflammatory properties. We analyzed the individual variation in expression of IL-10 and IL-1 receptor antagonist (IL-1ra), two factors with recognized high anti-inflammatory properties. The two genes were not represented on our microarray. As shown on Fig. 3, we observed a significant increase in IL-10 and IL-1ra expression after the VLCD in obese subjects (P⬍0.001). Moreover, there was a strong positive correlation between the changes in expression of the two transcripts (Rho⫽ 0.83,P⫽0.0006). IL-1ra mRNA levels were 2-fold higher in obese than in lean subjects (P⬍0.05). Origin of the differentially expressed genes related to inflammation We examined whether the 100 inflammation-related genes regulated by caloric restriction were expressed in adipocytes or SVF. Mature adipocytes were therefore separated from cells of the SVF. We checked that there was a selective enrichment of the adipocyte fraction in fat cell markers and of the SVF in macrophage markers (Table 5). CD14⫹ cells isolated from the SVF using CD14-coupled magnetic microbeads express characteristic markers of the monocyte/macrophage lineage (Fig. 4). The data indicate the presence of resident macrophages in the SVF of human subcutaneous adipose tissue. We could determine that there were 4042 ⫾ 440 macrophages (CD14⫹ cells) per gram of adipose tissue and that the cells represented 8.8 ⫾ 1.1% of SVF cells. The presence of macrophages in human adipose tissue was also shown by immunohistochemistry. The experiments were performed on surgical biopsies of subcutaneous abdominal white adipose tissue from morbidly obese patients (Fig. 5). The immunopositivity for HAM56 confirmed that the inflammatory cells infiltrating white adipose tissue of obese patients are mature macrophages. These cells stained positive for CD68, another macrophage marker whose mRNA level has been determined (Fig. 4) in isolated cells. On the WEIGHT LOSS AND INFLAMMATION IN ADIPOSE TISSUE

contrary, only mature adipocytes tested positive for SAA. The intensity of SAA staining was evident in the perinuclear area where the cytoplasm is thicker and rich in the Golgi complex and endoplastic reticulum. All the other cell types (fibroblasts, endothelial cells, smooth muscle cells of the vessel walls) tested negative, indicating that SAA protein production derives from fully mature adipocytes. As shown on supplementary Table 1, less than onefourth of the genes regulated during 28 day VLCD was significantly overexpressed in adipose cells whereas the remaining were expressed in adipocytes and SVF or mainly in SVF. Increased expression in isolated adipocytes was validated by RT-qPCR for CES1 and prostaglandin E receptor 3 as well as for SAA4, in agreement with immunohistochemistry data (Table 6). Several factors contributing to cellular adhesion and extracellular matrix remodeling were mostly expressed in SVF (e.g., matrix metalloproteinase 9). Different chemokines and interleukins were expressed in SVF (e.g., CCL19 and IL-12A). Expression of the anti-inflammatory IL-10 and IL-1ra mRNAs was higher in macrophages and SVF than in mature adipocytes (Fig. 4).

DISCUSSION Using gene profiling analysis, we describe here that caloric restriction-induced weight loss leads to the regulation of a wide variety of inflammation-related molecules in human adipose tissue. Weight loss decreases the expression of inflammatory markers in white adipose tissue of obese subjects and leads to the concomitant increased expression of molecules with anti-inflammatory properties. The vast majority of the gene transcripts were expressed in cells from the SVF of adipose tissue. The presence of macrophages in human adipose tissue may contribute to the increased expres-

Figure 3. Gene expression changes of interleukin-10 and -1 receptor antagonist mRNA in obese subjects before and after 28 day VLCD. Open and dark bars show mRNA levels of IL-10 and IL-1Ra in obese subjects before and after 28 day VLCD, respectively. ***P ⬍ 0.001 after VLCD vs. before VLCD. mRNA levels were normalized using 18S rRNA (⫻10⫺5 and 10⫺6 for IL-10 and IL-1Ra, respectively). 1663

TABLE 5. Expression of mature fat cell and macrophage markers in the adipocyte fraction and stromavascular fraction (SVF) of human subcutaneous adipose tissuea UG cluster

Marker

Name

Adipocyte/SVF ratio

Hs.103253 Hs.334305 Hs.80485 Hs.180878 Hs.241570 Hs.174142 Hs.246381 Hs.75627

Adipocyte Adipocyte Adipocyte Adipocyte Macrophage Macrophage Macrophage Macrophage

Perilipin Diacylglycerol O-acyltransferase 2 Adiponectin Lipoprotein lipase Tumor necrosis alpha Colony-stimulating factor 1 receptor CD68 antigen or macrosialin CD14 antigen or lipopolysaccharide receptor

23.9 12.8 5.2 3.9 0.19 0.29 0.37 0.38

a

Microarray experiments (n⫽6) were performed on amplified RNA from the adipocyte fraction and SVF of white adipose tissue.

sion of a large fraction of inflammatory genes in obese subjects. Previous analysis using large-scale gene expression studies performed in adipose tissue of lean and obese mice have shown the regulation of a repertoire of inflammatory genes (24 –26). Our study confirms that this feature is found in human subcutaneous adipose tissue and reveals the presence of new unexpected factors in adipose tissue. Furthermore, the group of transcripts that exhibited a decreased expression after 28 day VLCD was often expressed at higher levels in weight-stable obese subjects than in lean controls (cluster 1 in Fig. 2 and supplementary Table 1). This agrees with the presence of a proinflammatory state in obesity. Most of the transcriptional change of this set of genes after 28 day VLCD was not observed after an acute variation in energy balance (2 day VLCD). The level of obesity was significantly higher in the 2 day VLCD group than in the 28 day VLCD group. Differences between 2 and 28 days of caloric restriction may be

attributed to the length of time of caloric restriction and weight loss or to differences in both groups at baseline. However, the pattern of gene expression in these two groups clusters together before the diet despite the difference in the level of obesity. In addition, the energy deficit was more important in the 2 day VLCD group with the higher BMI. Despite this very high energy restriction that could have led to more pronounced changes in the 2 day VLCD group than in the 28 day VLCD group, there was no significant variation in gene expression for most of the genes. This supports the conclusion that the changes are associated with the length of caloric restriction and weight loss. Further experiments comparing the kinetic of inflammatory gene changes over the course of weight loss in the same individuals are necessary to confirm this finding. Our gene expression study was performed in subcutaneous adipose tissue of obese women. Differences in fat cell metabolism and secretory capacity among adi-

Figure 4. Gene expression of monocyte/macrophage and antiinflammatory markers in different cellular fractions of human subcutaneous adipose tissue. mRNA levels were normalized using 18S rRNA (⫻10⫺3). For all transcripts, mRNA levels were significantly different (P⬍0.05) between macrophages and adipocytes and between adipocytes and the stromavascular fraction.

1664

Vol. 18

November 2004

The FASEB Journal

CLE´MENT ET AL.

Figure 5. Immunostaining of inflammatory markers in macrophages and mature adipocytes of human subcutaneous adipose tissue. Subcutaneous abdominal white adipose tissues from obese subjects were tested by immunohistochemistry to detect HAM56, CD68, and serum amyloid A (SAA) proteins. Several cells were positive for HAM56, showing the presence of fully mature infiltrating macrophages. Accordingly, strong positivity was observed for another macrophage marker, CD68. SAA protein was detected only in fully mature adipocytes. The stromavascular fraction tested negative for this marker, indicated by arrows (SAA, 100⫻). These images are representative of immunostaining in subcutaneous white adipose tissues from 10 obese subjects.

pose tissue depots are well known. Several genes, including inflammatory or immunity mediators identified from candidate gene or larger scale approaches, appear to have depot-related variations (27–30). Secretory factors produced by visceral adipose tissue have been proposed to mediate the relationships between obesity, insulin resistance, and cardiovascular compli-

cations, although the mass of the visceral fat represents a small part of the total body fat mass. We observe here that many inflammatory genes expressed in visceral adipose tissue are expressed in subcutaneous adipose tissue of obese women (31). Our study thus suggests a relevant contribution of subcutaneous adipose tissue, especially in women in whom adipose tissue from the subcutaneous depot is abundant. This agrees with studies showing that at any level of body fat mass, visceral and subcutaneous adipose tissue are both good predictors of metabolic risk and insulin resistance (32, 33). It will be necessary to analyze in the future whether the secretion rates of the inflammatory markers are different between the fat depots in order to define new molecular links between adipose tissue distribution and the metabolic derangements associated with obesity. The regulated transcripts encoded proteins with a wide variety of functions on different cell targets. TNF receptors, ligands, and associated proteins were grouped in cluster 1. These changes in the TNF family are in agreement with the role of the TNF-␣ pathway in obesity-induced insulin resistance (34). The adipose tissue expression of TNF receptors (TNFR1 and TNFR2) is increased in rodent and human obesity, and TNFR2 expression levels in adipose tissue are correlated with BMI (35). There was, however, no change in TNF-␣ mRNA levels (data not shown) as previously reported during VLCD (3). The fact that numerous TNF receptors, ligands, and associated proteins were regulated suggests a coordinating response of the TNF-␣ pathway to nutritional changes. Another group of genes represented in cluster 1 was acute-phase reactants (36). These secreted factors constitute important markers of inflammation, infection, and malignancy (37). They are up-regulated in conditions with low level of systemic inflammation such as Alzheimer’s disease, amyloidosis, atherosclerosis, and insulin resistance (36, 38). Haptoglobin, which is modulated by TNF-␣, is overexpressed in different rodent models of obesity (39, 40). ␤2 Microglobulin, ␣2 macroglobulin, and serum amyloids have been involved in the pathophysiology of diseases associated with the depot of amyloid components (41). In agreement with a role of acute-phase reactants in the adaptations occurring dur-

TABLE 6. Gene expression in isolated adipocytes and in the stroma vascular fraction of human adipose tissuea

UG cluster

Hs.673 Hs.193717 Hs.81134 Hs.1511739 Hs.50002 Hs.1955 Hs.76688 Hs.170917

Name

Array ratio (n⫽6)

Adipocyte/SVF ratio (n⫽7)

P value

Interleukin 12A Interleukin 10 Interleukin 1 receptor antagonist Matrix metalloproteinase 9 Chemokine (COC) motif ligand 19 Serum amyloid A4 Carboxylesterase 1 Prostaglandin E receptor 3

0.66 ND ND 0.27 0.76 5.0 4.9 1.9

0.39 ⫾ 0.14 0.09 ⫾ 0.03 0.10 ⫾ 0.03 0.06 ⫾ 0.01 0.20 ⫾ 0.06 22 ⫾ 7 26 ⫾ 5 11 ⫾ 3

0.03 0.01 0.01 0.01 0.02 0.001 0.01 0.04

a

ND, not determined. To validate microarray data, reverse transcription quantitative PCR (RT-qPCR) was performed on total RNA from adipocyte and cells of the stromavascular fraction (SVF). Statistical tests of RT-qPCR data were performed using nonparametric Wilcoxon signed rank test.

WEIGHT LOSS AND INFLAMMATION IN ADIPOSE TISSUE

1665

ing VLCD, plasma SAA levels are decreased during VLCD (unpublished data). Soluble factors, membrane linked-proteins, and chemokines, chemotactic cytokines often induced by proinflammatory stimuli, are similarly suspected to be involved in chronic inflammatory pathologies, like atherosclerosis, possibly via their contribution on tissue immunity cell recruitment and differentiation (8, 42). One of the chemokines downregulated in VLCD, the monocyte chemoattractant protein 1 (C-C motif ligand 2), highly expressed in rodent adipose tissue, may induce adipocyte dedifferentiation and contribute to the development of insulin resistance (43). Due to the large amount of body fat in humans, adipose tissue might contribute to the elevated secretion of these molecules in obesity and may explain the decreased systemic production after weight loss. Weight loss induced by medical treatment or surgery in obese patients is indeed associated with decreased circulating concentrations of acute-phase response and immunological markers (9, 10). Two methods of clustering show that a significant proportion of the genes have similar patterns of expression in obese patients after 28 day VLCD and in lean subjects, although the obese subjects are still overweight after the nutritional challenge. It is well known that obese subjects benefit from moderate weight loss. The improvement of the inflammatory profile of adipose tissue may contribute to the overall improvement of obesity-related complications. In follow-up studies, improvement in insulin resistance indexes (9, 11, 12) and an amelioration of endothelial function in obese subjects (13) have been associated with decreased inflammatory factor concentrations. Together with a decreased expression of proinflammatory factors, we observed a concomitant increase of other inflammation-related molecules during weight loss in obese subjects. Cluster 2 contains genes whose products are involved in extracellular matrix remodeling such as the matrix metalloproteinase 9, an enzyme known to enhance the degradation of the basement membrane and extracellular matrix allowing endothelial cell migration. Other transcripts encoding cell-cell and cell-matrix adhesion factors such as integrins and cadherins were regulated. These changes may participate in the morphological modifications observed in adipose tissue during weight loss. Several interleukins and related factors involved in interferon ␥ activities were located in cluster 2. IL-12 and IL-18 favor T helper 1 cell response. Up-regulation of the receptor for the Fc portion of immunoglobulin G, CD32, is also indicative of such a response. The significance of these regulations in the inflammatory or immune adaptations during VLCD needs further cellular work to determine their biological effects. We investigated whether there was an up-regulation of anti-inflammatory molecules after weight loss. Among these molecules, IL-10, which has a major role in the regulatory network of cytokines is well recognized to have anti-inflammatory properties. It inhibits the macrophage production of IL-1, IL-6, and TNF-␣ (44, 45). IL-6 protein is decreased in 1666

Vol. 18

November 2004

adipose tissue of obese subjects after calorie restriction (3). IL-1ra, an antagonist of IL-1 receptor, is an acutephase factor that can be secreted by a large variety of cells. IL-1ra, which is highly expressed in adipose tissue (46), is the major modulator of the IL-1 proinflammatory pathway in vivo (47). VLCD was associated with increased expression of both IL-10 and IL-1ra in obese subjects and changes in IL-10 and IL-1ra expression were highly correlated. Activation of the major histocompatibility complex molecule CD1d induced during VLCD, has been shown to result in rapid and sustained production of IL-10 (48). It is noteworthy that IL receptor-like 1, a member of IL-1 receptor family, was down-regulated during calorie restriction. These data suggest a coordinated role for these factors in the improvement of the inflammatory profile. Our microarray and RT-qPCR experiments performed on human isolated adipocytes and SVF cells show the importance of the SVF cells in the regulation induced by calorie restriction. Indeed, the factors preferentially expressed in adipocytes represented ⬍25% of the inflammation-related mRNAs regulated during VLCD. We observed a predominant expression in the adipose cell of several inflammatory factors such as acute-phase proteins (e.g., haptoglobin, SAA4) and other molecules (e.g., monocyte to macrophage differentiation-associated protein, prostaglandin E receptor 3, and CES1). Using immunohistochemistry experiments in obese subjects, we confirmed that SAA protein production derives from fully mature adipocytes. The data agree with accumulating evidence that adipocytes and immune cells share many features with expression in both cell types of transcripts previously considered to be specific of one lineage. However, ⬃45% of the transcripts were more abundant in SVF cells than in adipocytes. SVF is constituted of several cell types including preadipocytes not yet filled with lipids, fibroblasts, histiocytes, endothelial cells, and monocyte/ macrophages. A large fraction of the regulated mRNAs is known to be expressed in the latter cell type. Our characterization of CD14⫹ cells shows that the population of resident macrophages constitutes an important component of the SVF in human subcutaneous adipose tissue. Immunohistochemistry experiments show that macrophage cells are infiltrating the adipose tissue of obese subjects. The number of macrophages is elevated in rodent models of obesity (49, 50). Moreover, a recent paper revealed that the number of macrophages was positively correlated to the BMI in humans (51). Preadipose cells possess phagocytic and microbicidal activities and can be converted into cells expressing markers characteristic of macrophages (52, 53). Alternatively, accumulation of macrophages may be due to an influx of bone marrow-derived precursors into adipose tissue and their subsequent differentiation into mature macrophages (49, 50). The lower expression of IL-1ra mRNA in lean subjects than in obese subjects as recently reported (46) might be in apparent contradiction with the increased expression observed during calorie restriction. The

The FASEB Journal

CLE´MENT ET AL.

tier University, Toulouse, France) for critical reading of the manuscript and Dr. Anne Bouloumie´ (Institut fu¨r Kardiovasculare Physiologie, J.W. Goethe-Universita¨t, Frankfurt am Main, Germany, and Inserm U586, Toulouse, France) for advice on macrophage isolation and characterization. This work was supported by INSERM (ATC-nutrition no. 4NU10G), the Direction de la Recherche Clinique/Assistance Publique-Hopitaux de Paris (AOR 02076), the Contrat Hospitalier de Recherche Clinique (CRC No.97123), CRC/Alfediam, the Claude Bernard association (Formation Associe´e), and the Servier Research Institute (IRIS). K.C. received a grant from the French Association for Research on Obesity (AFERO) and Institut Benjamin Delessert. C.P. received a grant from Alfediam/Merck-Lipha.

REFERENCES 1. 2.

3.

Figure 6. Schematic diagram of the changes of inflammationrelated gene expression in adipose tissue during the development of obesity and calorie restriction-induced weight loss. Hypothetical adaptations of macrophages deduced from gene expression profiles are in italics.

complexity of macrophage biology may explain this discrepancy. Adipose tissue accumulates macrophages in the obese state (49, 50). IL-1 ra mRNA is predominantly expressed in macrophages vs. adipocytes. Hence, the higher expression of IL-1 ra mRNA seen in obesity may merely be the reflection of the higher content in macrophages at basal state. It is known, however, that macrophage cells express different functional programs in response to environmental changes. During calorie restriction, the nature of resident macrophages may be modified with shift from M1 macrophages to M2 macrophages producers of IL-10 and IL-1 ra (54). The adaptation is certainly complex as the markers of M1 macrophages, IL-12 and monocyte chemoattractant protein 1, are respectively increased and decreased during VLCD. Isolation and characterization of the different populations of adipose tissue macrophages is warranted to fully understand the nutritional regulation of gene expression in these cells. Taken together, these findings show that weight loss globally improves the inflammatory profile of white adipose tissue in obese subjects (Fig. 6). The beneficial effect of weight loss is associated with modification of the inflammation-related gene expression in adipocytes and macrophages. This work paves the way for future clinical and cellular studies aimed at determining the impact of these molecular adaptations on the development of insulin resistance. We thank Dr. Bernard Pipy (Macrophages, inflammatory mediators and cellular interactions, UPS EA-2405, Paul SabaWEIGHT LOSS AND INFLAMMATION IN ADIPOSE TISSUE

4.

5. 6. 7.

8. 9.

10.

11.

12.

13.

14.

Mora, S., and Pessin, J. E. (2002) An adipocentric view of signaling and intracellular trafficking. Diabetes Metab. Res. Rev. 18, 345–356 Bastard, J. P., Jardel, C., Delattre, J., Hainque, B., Bruckert, E., and Oberlin, F. (1999) Evidence for a link between adipose tissue interleukin-6 content and serum C-reactive protein concentrations in obese subjects. Circulation 99, 2221–2222 Bastard, J. P., Jardel, C., Bruckert, E., Blondy, P., Capeau, J., Laville, M., Vidal, H., and Hainque, B. (2000) Elevated levels of interleukin 6 are reduced in serum and subcutaneous adipose tissue of obese women after weight loss. J. Clin. Endocrinol. Metab. 85, 3338 –3342 Mohamed-Ali, V., Flower, L., Sethi, J., Hotamisligil, G., Gray, R., Humphries, S. E., York, D. A., and Pinkney, J. (2001) betaAdrenergic regulation of IL-6 release from adipose tissue: in vivo and in vitro studies. J. Clin. Endocrinol. Metab. 86, 5864 –5869 Bruunsgaard, H., and Pedersen, B. K. (2003) Age-related inflammatory cytokines and disease. Immunol. Allergy Clin. N. Am. 23, 15–39 Rosenson, R. S., and Koenig, W. (2003) Utility of inflammatory markers in the management of coronary artery disease. Am. J. Cardiol. 92, 10i–18i Festa, A., D'Agostino, R., Jr., Howard, G., Mykkanen, L., Tracy, R. P., and Haffner, S. M. (2000) Chronic subclinical inflammation as part of the insulin resistance syndrome: the Insulin Resistance Atherosclerosis Study (IRAS). Circulation 102, 42– 47 Grimble, R. F. (2002) Inflammatory status and insulin resistance. Curr. Opin. Clin. Nutr. Metab. Care 5, 551–559 Hanusch-Enserer, U., Cauza, E., Spak, M., Dunky, A., Rosen, H. R., Wolf, H., Prager, R., and Eibl, M. M. (2003) Acute-phase response and immunological markers in morbid obese patients and patients following adjustable gastric banding. Int. J. Obes. Relat. Metab. Disord. 27, 355–361 Laimer, M., Ebenbichler, C. F., Kaser, S., Sandhofer, A., Weiss, H., Nehoda, H., Aigner, F., and Patsch, J. R. (2002) Markers of chronic inflammation and obesity: a prospective study on the reversibility of this association in middle-aged women undergoing weight loss by surgical intervention. Int. J. Obes. Relat. Metab. Disord. 26, 659 – 662 Kopp, H. P., Kopp, C. W., Festa, A., Krzyzanowska, K., Kriwanek, S., Minar, E., Roka, R., and Schernthaner, G. (2003) Impact of weight loss on inflammatory proteins and their association with the insulin resistance syndrome in morbidly obese patients. Arterioscler. Thromb. Vasc. Biol. 23, 1042–1047 Esposito, K., Pontillo, A., Di Palo, C., Giugliano, G., Masella, M., Marfella, R., and Giugliano, D. (2003) Effect of weight loss and lifestyle changes on vascular inflammatory markers in obese women: a randomized trial. J. Am. Med. Assoc. 289, 1799 –1804 Ziccardi, P., Nappo, F., Giugliano, G., Esposito, K., Marfella, R., Cioffi, M., D'Andrea, F., Molinari, A. M., and Giugliano, D. (2002) Reduction of inflammatory cytokine concentrations and improvement of endothelial functions in obese women after weight loss over one year. Circulation 105, 804 – 809 Gerhardt, C. C., Romero, I. A., Cancello, R., Camoin, L., and Strosberg, A. D. (2001) Chemokines control fat accumulation

1667

15.

16.

17.

18.

19.

20. 21. 22.

23. 24. 25.

26.

27.

28.

29.

30.

31.

1668

and leptin secretion by cultured human adipocytes. Mol. Cell. Endocrinol. 175, 81–92 Hsu, S. M., Raine, L., and Fanger, H. (1981) A comparative study of the peroxidase-antiperoxidase method and an avidinbiotin complex method for studying polypeptide hormones with radioimmunoassay antibodies. Am. J. Clin. Pathol. 75, 734 –738 Clement, K., Viguerie, N., Diehn, M., Alizadeh, A., Barbe, P., Thalamas, C., Storey, J. D., Brown, P. O., Barsh, G. S., and Langin, D. (2002) In vivo regulation of human skeletal muscle gene expression by thyroid hormone. Genome Res. 12, 281–291 Feldman, A. L., Costouros, N. G., Wang, E., Qian, M., Marincola, F. M., Alexander, H. R., and Libutti, S. K. (2002) Advantages of mRNA amplification for microarray analysis. Biotechniques 33, 906 –912 Viguerie, N., Clement, K., Barbe, P., Courtine, M., Benis, A., Larrouy, D., Hanczar, B., Pelloux, V., Poitou, C., Khalfallah, Y., et al. (2004) In vivo epinephrine-mediated regulation of gene expression in human skeletal muscle. J. Clin. Endocrinol. Metab. 89, 2000 –2014 Rome, S., Clement, K., Rabasa-Lhoret, R., Loizon, E., Poitou, C., Barsh, G. S., Riou, J. P., Laville, M., and Vidal, H. (2003) Microarray profiling of human skeletal muscle reveals that insulin regulates approximately 800 genes during a hyperinsulinemic clamp. J. Biol. Chem. 278, 18063–18068 Quackenbush, J. (2002) Microarray data normalization and transformation. Nat. Genet. 32, Suppl., 496 –501 Tusher, V. G., Tibshirani, R., and Chu, G. (2001) Significance analysis of microarrays applied to the ionizing radiation response. Proc. Natl. Acad. Sci. USA 98, 5116 –5121 Ashburner, M., Ball, C. A., Blake, J. A., Botstein, D., Butler, H., Cherry, J. M., Davis, A. P., Dolinski, K., Dwight, S. S., Eppig, J. T., et al. (2000) Gene ontology: tool for the unification of biology. The Gene Ontology Consortium. Nat. Genet. 25, 25–29 Eisen, M. B., Spellman, P. T., Brown, P. O., and Botstein, D. (1998) Cluster analysis and display of genome-wide expression patterns. Proc. Natl. Acad. Sci. USA 95, 14863–14868 Soukas, A., Cohen, P., Socci, N. D., and Friedman, J. M. (2000) Leptin-specific patterns of gene expression in white adipose tissue. Genes Dev. 14, 963–980 Moraes, R. C., Blondet, A., Birkenkamp-Demtroeder, K., Tirard, J., Orntoft, T. F., Gertler, A., Durand, P., Naville, D., and Begeot, M. (2003) Study of the alteration of gene expression in adipose tissue of diet-induced obese mice by microarray and reverse transcription-polymerase chain reaction analyses. Endocrinology 144, 4773– 4782 Takahashi, K., Mizuarai, S., Araki, H., Mashiko, S., Ishihara, A., Kanatani, A., Itadani, H., and Kotani, H. (2003) Adiposity elevates plasma MCP-1 levels leading to the increased CD11bpositive monocytes in mice. J. Biol. Chem. 278, 46654 – 46660 Dusserre, E., Moulin, P., and Vidal, H. (2000) Differences in mRNA expression of the proteins secreted by the adipocytes in human subcutaneous and visceral adipose tissues. Biochim. Biophys. Acta 1500, 88 –96 Eriksson, P., van Harmelen, V., Hoffstedt, J., Lundquist, P., Vidal, H., Stemme, V., Hamsten, A., Arner, P., and Reynisdottir, S. (2000) Regional variation in plasminogen activator inhibitor-1 expression in adipose tissue from obese individuals. Thromb. Haemost. 83, 545–548 Gabrielsson, B. G., Johansson, J. M., Lonn, M., Jernas, M., Olbers, T., Peltonen, M., Larsson, I., Lonn, L., Sjostrom, L., Carlsson, B., et al. (2003) High expression of complement components in omental adipose tissue in obese men. Obes. Res. 11, 699 –708 Linder, K., Arner, P., Flores-Morales, A., Tollet-Egnell, P., and Norstedt, G. (2004) Differentially expressed genes in visceral or subcutaneous adipose tissue of obese men and women. J. Lipid Res. 45, 148 –154 Gomez-Ambrosi, J., Catalan, V., Diez-Caballero, A., MartinezCruz, L. A., Gil, M. J., Garcia-Foncillas, J., Cienfuegos, J. A., Salvador, J., Mato, J. M., and Fruhbeck, G. (2004) Gene expression profile of omental adipose tissue in human obesity. FASEB J. 18, 215–217

Vol. 18

November 2004

32. 33.

34. 35.

36. 37. 38. 39.

40.

41. 42. 43. 44. 45. 46.

47.

48.

49.

50.

51.

Albu, J. B., Kovera, A. J., and Johnson, J. A. (2000) Fat distribution and health in obesity. Ann. N.Y. Acad. Sci. 904, 491–501 Smith, S. R., Lovejoy, J. C., Greenway, F., Ryan, D., deJonge, L., de la Bretonne, J., Volafova, J., and Bray, G. A. (2001) Contributions of total body fat, abdominal subcutaneous adipose tissue compartments, and visceral adipose tissue to the metabolic complications of obesity. Metabolism 50, 425– 435 Hotamisligil, G. S., Shargill, N. S., and Spiegelman, B. M. (1993) Adipose expression of tumor necrosis factor-alpha: direct role in obesity-linked insulin resistance. Science 259, 87–91 Hotamisligil, G. S., Arner, P., Atkinson, R. L., and Spiegelman, B. M. (1997) Differential regulation of the p80 tumor necrosis factor receptor in human obesity and insulin resistance. Diabetes 46, 451– 455 Friedland, R. P. (2002) Lipid metabolism, epidemiology, and the mechanisms of Alzheimer's disease. Ann. N.Y. Acad. Sci. 977, 387–390 Argiles, J. M., and Lopez-Soriano, F. J. (1998) Catabolic proinflammatory cytokines. Curr. Opin. Clin. Nutr. Metab. Care 1, 245–251 Marette, A. (2002) Mediators of cytokine-induced insulin resistance in obesity and other inflammatory settings. Curr. Opin. Clin. Nutr. Metab. Care 5, 377–383 Friedrichs, W. E., Navarijo-Ashbaugh, A. L., Bowman, B. H., and Yang, F. (1995) Expression and inflammatory regulation of haptoglobin gene in adipocytes. Biochem. Biophys. Res. Commun. 209, 250 –256 Chiellini, C., Bertacca, A., Novelli, S. E., Gorgun, C. Z., Ciccarone, A., Giordano, A., Xu, H., Soukas, A., Costa, M., Gandini, D., et al. (2002) Obesity modulates the expression of haptoglobin in the white adipose tissue via TNFalpha. J. Cell. Physiol. 190, 251–258 Merlini, G., and Bellotti, V. (2003) Molecular mechanisms of amyloidosis. N. Engl. J. Med. 349, 583–596 Shin, W. S., Szuba, A., and Rockson, S. G. (2002) The role of chemokines in human cardiovascular pathology: enhanced biological insights. Atherosclerosis 160, 91–102 Sartipy, P., and Loskutoff, D. J. (2003) Monocyte chemoattractant protein 1 in obesity and insulin resistance. Proc. Natl. Acad. Sci. USA 100, 7265–7270 Braat, H., Peppelenbosch, M. P., and Hommes, D. W. (2003) Interleukin-10-based therapy for inflammatory bowel disease. Expert Opin. Biol. Ther. 3, 725–731 Asadullah, K., Sterry, W., and Volk, H. D. (2003) Interleukin-10 therapy–review of a new approach. Pharmacol. Rev. 55, 241–269 Juge-Aubry, C. E., Somm, E., Giusti, V., Pernin, A., Chicheportiche, R., Verdumo, C., Rohner-Jeanrenaud, F., Burger, D., Dayer, J. M., and Meier, C. A. (2003) Adipose tissue is a major source of interleukin-1 receptor antagonist: upregulation in obesity and inflammation. Diabetes 52, 1104 –1110 Marculescu, R., Endler, G., Schillinger, M., Iordanova, N., Exner, M., Hayden, E., Huber, K., Wagner, O., and Mannhalter, C. (2002) Interleukin-1 receptor antagonist genotype is associated with coronary atherosclerosis in patients with type 2 diabetes. Diabetes 51, 3582–3585 Colgan, S. P., Hershberg, R. M., Furuta, G. T., and Blumberg, R. S. (1999) Ligation of intestinal epithelial CD1d induces bioactive IL-10: critical role of the cytoplasmic tail in autocrine signaling. Proc. Natl. Acad. Sci. USA 96, 13938 –13943 Xu, H., Barnes, G. T., Yang, Q., Tan, G., Yang, D., Chou, C. J., Sole, J., Nichols, A., Ross, J. S., Tartaglia, L. A., et al. (2003) Chronic inflammation in fat plays a crucial role in the development of obesity-related insulin resistance. J. Clin. Invest. 112, 1821–1830 Weisberg, S. P., McCann, D., Desai, M., Rosenbaum, M., Leibel, R. L., and Ferrante, A. W., Jr. (2003) Obesity is associated with macrophage accumulation in adipose tissue. J. Clin. Invest. 112, 1796 –1808 Curat, C. A., Miranville, A., Sengenes, C., Diehl, M., Tonus, C., Busse, R., and Bouloumie, A. (2004) From blood monocytes to adipose tissue-resident macrophages: induction of diapedesis by human mature adipocytes. Diabetes 53, 1285–1292

The FASEB Journal

CLE´MENT ET AL.

52.

Charriere, G., Cousin, B., Arnaud, E., Andre, M., Bacou, F., Penicaud, L., and Casteilla, L. (2003) Preadipocyte conversion to macrophage. Evidence of plasticity. J. Biol. Chem. 278, 9850 – 9855 53. Saillan-Barreau, C., Cousin, B., Andre, M., Villena, P., Casteilla, L., and Penicaud, L. (2003) Human adipose cells as candidates in defense and tissue remodeling phenomena. Biochem. Biophys. Res. Commun. 309, 502–505

WEIGHT LOSS AND INFLAMMATION IN ADIPOSE TISSUE

54.

Mantovani, A., Sozzani, S., Locati, M., Allavena, P., and Sica, A. (2002) Macrophage polarization: tumor-associated macrophages as a paradigm for polarized M2 mononuclear phagocytes. Trends Immunol. 23, 549 –555 Received for publication May 5, 2004. Accepted for publication July 7, 2004.

1669