NOV/CCN3: a New Adipocytokine Involved in Obesity-Associated Insulin Resistance

2502 Diabetes Volume 65, September 2016

Cécile Martinerie,1,2 Marie Garcia,1,2 Thi Thu Huong Do,1,2 Bénédicte Antoine,1,2 Marthe Moldes,1,2 Guillaume Dorothee,1 Chantal Kazazian,1,2 Martine Auclair,1,2 Marion Buyse,1,2,3,4 Tatiana Ledent,1 Pierre-Olivier Marchal,1,2 Maria Fesatidou,1,2 Adrien Beisseiche,5 Haruhiko Koseki,6 Shuichi Hiraoka,7 Christos Evangelos Chadjichristos,8 Bertrand Blondeau,5 Raphael Georges Denis,9 Serge Luquet,9 and Bruno Fève1,2,10

NOV/CCN3: A New Adipocytokine Involved in Obesity-Associated Insulin Resistance

Diabetes 2016;65:2502–2515 | DOI: 10.2337/db15-0617

Identification of new adipokines that potentially link obesity More than 300 million people are currently obese world- to insulin resistance represents a major challenge. We wide, and this noninfectious epidemic affects both Western recently showed that NOV/CCN3, a multifunctional matri- and emerging countries (1–3). In the absence of adequate cellular protein, is synthesized and secreted by adipose care, obesity induces many comorbidities and reduces tissue, with plasma levels highly correlated with BMI. NOV life expectancy (4–7). Thus, there is a crucial need to involvement in tissue repair, fibrotic and inflammatory better understand the pathophysiology of obesity, insulin- diseases, and cancer has been previously reported. How- resistance, and type 2 diabetes. In this context, the discovery ever, its role in energy homeostasis remains unknown. We 2/2 of new adipocytokines involved in the control of energy investigated the metabolic phenotype of NOV mice fed balance is not only a challenge for understanding the mech- a standard or high-fat diet (HFD). Strikingly, the weight of 2 2 anisms responsible for the abnormal development of obesity METABOLISM NOV / mice was markedly lower than that of wild-type or insulin resistance but also offers new opportunities to mice but only on an HFD. This was related to a significant manage this disease and its complications. decrease in fat mass associated with an increased proportion of smaller adipocytes and to a higher expression The NOV/CCN3 gene (nephroblastoma overexpressed of genes involved in energy expenditure. NOV2/2 mice fed gene) (8) is a founder of the CCN (Cyr61/CCN1, CTGF/CCN2, an HFD displayed improved glucose tolerance and insulin NOV/CCN3) family. The CCN genes encode matricial mul- sensitivity. Interestingly, the absence of NOV was associ- tifunctional proteins that are involved in organogenesis fl ated with a change in macrophages profile (M1-like to M2- (9). In addition, these proteins play key roles in in amma- like), in a marked decrease in adipose tissue expression of tion (10), wound healing, fibrosis, and cancers (9). Several several proinflammatory cytokines and chemokines, and in studies have also shown that NOV is involved in the ad- enhanced insulin signaling. Conversely, NOV treatment of hesion, migration, proliferation, differentiation, and sur- adipocytes increased chemokine expression. Altogether, vival of different cell types (11–16). NOV also modulates these results show that NOV is a new adipocytokine that the expression of inflammatory molecules and their effects could be involved in obesity-associated insulin-resistance. (17–20). These different functions are mediated through

1Sorbonne Universities, Pierre and Marie Curie University Paris 06, INSERM, 9Sorbonne Paris City University, Paris Diderot University, BFA, CNRS, Paris, France Saint-Antoine Research Center, Saint-Antoine Hospital, Paris, France 10Department of Endocrinology, Paris, Assistance Publique-Hôpitaux de Paris, 2Hospitalo-Universitary Institute, ICAN, Paris, France Saint-Antoine Hospital, Paris, France 3 Department of Pharmacy, Assistance Publique-Hôpitaux de Paris, Saint-Antoine Corresponding author: Bruno Fève, [email protected]. Hospital, Paris, France Received 9 May 2015 and accepted 23 May 2016. 4Paris-Sud University, EA 4123, Châtenay-Malabry, France 5Sorbonne Universities, Pierre and Marie Curie University Paris 06, INSERM, This article contains Supplementary Data online at http://diabetes Cordeliers Research Center, Paris, France .diabetesjournals.org/lookup/suppl/doi:10.2337/db15-0617/-/DC1. 6RIKEN Research Center for Allergy and Immunology (RCAI), RIKEN Yokohama © 2016 by the American Diabetes Association. Readers may use this article as Institute, Yokohama, Japan long as the work is properly cited, the use is educational and not for profit, and 7Department of Biochemistry, Kobe Pharmaceutical University, Kobe, Japan the work is not altered. More information is available at http://diabetesjournals 8Sorbonne Universities, Pierre and Marie Curie University Paris 06, INSERM, .org/site/license. Tenon Hospital, Paris, France diabetes.diabetesjournals.org Martinerie and Associates 2503 the interaction of NOV with specific integrins and, in some Mice were kept on a 12-h light/12-h dark cycle with ad cases, by the Notch pathway (9). libitum access to food and tap water. Male mice (6 weeks Measurements of plasma NOV concentrations in a large old) were fed a SD (4% fat; Genestil 1326 Royaucourt, cohort of patients screened for cardiometabolic diseases, in France) or an HFD (42% fat; TD.88137; Harlan Teklad, which 60% of the patients had a BMI above 30 kg/m2, Gannat, France) for 16 weeks. Mice were bred according revealed, for the first time, a strong relationship between to the GuidefortheCareandUseofLaboratoryAnimals plasma NOV concentrations and obesity (21). Plasma NOV (U.S. National Institutes of Health Publication No. 85-23, re- levels were also correlated with weight loss in patients who vised 1996). The animal facility was granted approval (B-75- had undergone bariatric surgery. Accordingly, we found 12-01) given by the French Administration. All procedures that the NOV protein was synthesized and secreted by were approved by a local ethic committee (No. Ce5/2012/091). adipocytes and macrophages in adipose tissue from obese Mice genotypes were performed as previously described patients and in human primary cultures (21). Moreover, we (22). The genotypes were further checked by NOV mRNA demonstrated that the induction of NOV during adipose levels in epididymal white adipose tissue (eWAT) and sub- tissue expansion was not restricted to humans. Indeed, in cutaneous WAT (scWAT), the plasma level of the protein mice fed a high-fat diet (HFD), plasma NOV levels and its and its expression in heart where NOV is normally highly expression in adipose tissue were also increased compared expressed (Supplementary Fig. 1). with mice fed a standard diet (SD) (21). Body Composition Even though a high BMI is often associated with im- Body weight (BW) was measured using a laboratory scale paired glucose tolerance and type 2 diabetes, we did not (Ohaus CS 200 Series; Sigma-Aldrich, l’Isle d’Abeau, France). find any correlation between NOV and fasting glycemia. Body mass composition was analyzed using an EchoMRI However, there was a positive correlation between plasma 100 Whole Body Composition Analyzer (EchoMRI, Houston, NOV and HbA (21). Thus, it is conceivable that a rela- 1c TX) according to the manufacturer’s instructions (26). tionship could exist between NOV and carbohydrate metabolism. In this context, a possible link between NOV Indirect Calorimetry Measurements and insulin was suggested by Shimoyama et al. (22), who For measurement of energy expenditure, respiratory ex- reported that NOV expression was reduced in rat aortas change ratio (RER), and spontaneous locomotor activity, after streptozotocin injection and was increased by insulin mice were placed in a LabMaster indirect calorimetry treatment. Moreover, the nov gene is localized on chromo- system (TSE Systems GmbH, Bad Homburg, Germany). some 8q24 (23), which is a susceptibility locus controlling Metabolic Parameter Exploration b-cell function in linkage studies of patients with diabetes Oral glucose tolerance tests (OGTTs) and insulin tolerance (24). Taken together, these results and a recent report (25) tests (ITTs) were performed after 14 and 15 weeks of the showing that NOV is a direct target of transcription factor diet, respectively, on food-deprived (6 h and 4 h, respec- FOXO1 and that NOV impairs insulin secretion in pancre- tively) nonanesthetized mice. For the OGTT, animals atic b-cells reinforce the possibility that this protein could were weighed and orally fed using gavage needles (Ref be involved in energy homeostasis. However, no data have 18060-020; Phymep, Paris, France) with a 1 g/kg BW of a been reported to date on the function of NOV in adipose 20% glucose solution. Whole-tail vein blood (30 mL) was tissue or in whole-body metabolism. Thus, the main focus sampled at baseline and 15 min to measure plasma in- of this study was to identify whether targeted disruption of sulin. For the ITT, animals were weighed and intraperito- the nov gene modulates energy balance in mice fed a nor- neally injected with human insulin (0.75 mU/kg). For mal or obesogenic diet. both tests, glucose levels were obtained from whole- tail vein blood using an automatic Accu-Chek Performa RESEARCH DESIGN AND METHODS glucometer (Roche, Meylan, France) at indicated times. Mice For insulin-signaling experiments, mice were intraperitone- 2/2 NOV mice were generated by Shimoyama et al. (22) ally injected with 0.5 units of regular human insulin using a targeting vector to delete the genomic region (Actrapid Penfill; Novo Nordisk, Paris, France). Tissues encompassing exons 1, 2, and a part of 3 for complete were snap frozen in liquid nitrogen 10 min later. elimination of the NOV transcription. Mice heterozygously transmitting the NOV-knockout genome were generated, Plasma Measurements and their descendants were backcrossed to C57Bl/6J for Total cholesterol, HDL and LDL cholesterol, triglycerides, fi at least six generations and maintained as heterozy- and nonesteri ed fatty acids (NEFA) levels were de- gotes. C57Bl/6J heterozygous males were then mated with termined by enzymatic colorimetric assays (SOBIODA, – 129sv/PAS females. Embryos derived from these crossings Montbonnot Saint-Martin, France). Leptin, adiponectin, fi were reimplanted in pseudopregnant females to get an F1 and NOV levels were measured using ELISA-speci c kits on a mixed background (C57Bl/6J-SV129) with the “free of (R&D Systems, Lille, France). specific pathogenic microorganisms” status. Pancreatic Measurements 2 2 C57Bl/6-SV129 NOV / mice and the wild-type (WT) The pancreatic and islet insulin content was extracted controls were derived from the same heterozygous crossings. at 220°C in acidic ethanol (1.5% [vol/vol] HCl in 75% 2504 NOV/CCN3 Is Involved in Insulin Resistance Diabetes Volume 65, September 2016

2 2 [vol/vol] ethanol). Pancreatic islets from NOV / and WT cells were analyzed using a Gallios flow cytometer mice were isolated after collagenase digestion and cultured (Beckman Coulter, Inc., Villepinte, France), and data were in RPMI-1640 containing 10% FBS at 37°C for 24 h. Batches processed using Kaluza software (Beckman Coulter). of 30 islets were incubated in Krebs buffer containing Adipose Tissue and Liver Biochemical Measurements 2.8 mmol/L glucose, 16.7 mmol/L glucose, or 50 mmol/L 2 2 eWAT from WT and NOV / mice was removed and KCl for 1 h at 37°C. Plasma insulin, total pancreatic and islet digested as described above. Homogenates were centri- insulin content, and insulin secreted in response to glu- fuged, and isolated adipocytes were obtained after cen- cose were measured using the Ultra Sensitive Insulin trifugation. Glycerol secreted from isolated adipocytes ELISA kit (Crystal Chem, Inc., Downers Grove, IL). was used as an index of lipolysis and measured with the Histomorphological Procedures Free-Glycerol reagent (Sigma-Aldrich). Triacylglycerols were Paraffin-embedded adipose sections were analyzed after extracted from adipocytes by the Dole method (28) and hematoxylinandeosinstaining.Fivefields were randomly from liver by acetone and then tested with the triglycer- observed at original magnification 310 by an optical micro- ides reagent (Sigma-Aldrich). scope (BX61 Olympus). Adipocyte quantification was per- Primary Preadipocytes and 3T3-L1 Culture formed using ImageJ software (http://rsbweb.nih.gov/ij/) Primary preadipocytes derived from SVF of scWAT or on ;500–600 cells per mouse (4 animals per group and per genotype were analyzed). eWAT were plated on six-well culture dishes and cultured in DMEM (4.5 g/L D-glucose) with 10% FCS (Eurobio, Les Considering that adipocytes constituted 90% of the Ulis, France). Cell proliferation was assessed by BrdU in- adipose tissue volume, the number of cells was estimated corporation according to the manufacturer’s instructions by dividing the tissue mass by the mean volume of cells. (Cell Signaling, Danvers, MA). Adipocyte differentiation was The results are expressed in cells per gram of tissue. initiated in confluent cultures of primary and 3T3-L1 pre- Sections (3-mm thick) were stained with Sirius Red, and adipocytes with DMEM supplemented with 10% FCS and interstitial fibrosis was assessed semiquantitatively by 0.1 mmol/L 3-isobutyl-1-methylxanthine, 200 nmol/L dexa- randomly selecting 10 fields per section that covered the entire surface at original magnification 3200. Fibrosis methasone, and 100 nmol/L insulin (Sigma-Aldrich) for 3 days. Adipocytes were further maintained for 6 days was then quantified using computer-based morphometric in DMEM supplemented with 10% FCS and 100 nmol/L Analysis software (Olympus) that allowed the formation of insulin. a binary image in which the stained area could be automat- ically calculated as a percentage of the image area (27). Primary Macrophages Culture Bone marrow mononuclear (BMM) phagocytic precursor Isolation of Stromal Vascular Fraction and Flow cells were isolated from femurs and tibiae of HFD-fed WT Cytometry Analysis 2/2 2 2 and NOV mice, as previously described (29). These eWAT from WT and NOV / mice was removed, minced, precursors were differentiated into adherent mature mac- and digested with collagenase A (1 mg/mL in PBS; rophages (BMM) for 6 days in noncoated Petri dishes in Sigma-Aldrich) at 37°C for 45 min. Homogenates were DMEM (4.5 g/L D-glucose) with 20% decomplemented passed through a 70-mm mesh, and the effluent was FCS (Eurobio) containing 20 ng/mL macrophage colony- centrifuged at 500g for 3 min. Cells from pellets defined stimulating factor (PeproTech, Neuilly-sur-Seine, France). as the stromal vascular fraction (SVF) were resuspended BMM were seeded at 2 3 106 cells/well in six-well plates in FACS buffer (PBS with 5% FBS). Extracellular staining and polarized toward a M1 phenotype with 10 ng/mL was preceded by incubation with FcR-blocking antibody lipopolysaccharide (LPS; Sigma-Aldrich) or toward a M2 (2.4G2;BDBiosciences,FranklinLakes,NJ)toavoid phenotype with 10 ng/mL interleukin 4 (IL-4; PeproTech) nonspecific staining. For lymphocytes cytometry analy- for 24 h. Cytokine secretion into the cell culture medium sis, cells were first stained with monoclonal antibodies to of BMM was subsequently analyzed using tumor necrosis cell-surface markers purchased from BD Biosciences; fluo- factor a (TNF-a) for M1 or IL-10 for M2 phenotype ELISA rescein isothiocyanate–conjugated anti-CD3 (145-2C11), kit (PeproTech). allophycocyanin-conjugated anti-CD4 (RM4-5), and peridinin- Peritoneal macrophages (PEM) were derived from HFD- chlorophyll protein–conjugated anti-CD8 (53-6.7). Cells 2 2 fed WT and NOV / mice after an intraperitoneal ad- were then fixed, permeabilized, and incubated with phyco- ministration of PBS 13 (5 mL). The peritoneal wash was erythrin (PE)-conjugated anti-Foxp3 (FJK-16s, eBioscience) recovered from injected mice, and PEM were collected by for intracellular staining according to the manufacturer’s centrifugation at 600g at4°Cfor5min.PEM,atadensityof specifications. For staining analysis of macrophages, cells 106 cells/well in a six-well plate, were polarized toward an were surface-stained with BV421-conjugated anti-F4/80 M1 or an M2 phenotype as described for BMM. BMM or (T45-2342), Alexa Fluor 647–conjugated anti-CD206 (19.2), PEM were then processed as mentioned for FACS analysis. and biotin-conjugated anti-CD11b (M1/70) revealed with PE-conjugated streptavidin all from BD Bioscience. Fluo- Gene Expression Analysis rescein isothiocyanate-conjugated anti-CD11c (N418) and Total RNA was extracted from mouse tissues, primary anti–NOS2-PE cyanine 7 were from eBioscience. Stained preadipocytes, adipocyte cultures, and from 3T3-L1 cells diabetes.diabetesjournals.org Martinerie and Associates 2505 using RNeasy Lipid Mini or RNeasy Mini Kits (Qiagen, Les was observed in mice fed the SD between the two 2 2 Ulis, France). Reverse transcription and real-time semi- groups of mice (Fig. 1A and B). However, NOV / mice quantitative PCR (qPCR) amplification on an ABI 7300 ap- fedtheHFDgainedlessweightthanWTcontrols(Fig. 2 2 paratus (Applied Biosystems) were performed as previously 1C). The BW gain of NOV / mice was markedly lower 2 2 described (13). Specific primers (Proligo; Sigma-Aldrich) compared with WT mice (Fig. 1D). NOV / mice fed used for the amplification of target genes are presented in the HFD also displayed a decrease in fat and relative Supplementary Table 1. The comparative Ct method (30) fat mass (Fig. 1E and F) and a slight increase in relative was used to calculate gene expression values. The ribosomal lean body mass compared with their respective con- S26 was used for normalization as a housekeeping gene. trols; however, the absolute lean mass was unchanged (Fig. 1G and H), and there was no difference in mouse- Small Interfering RNA and Transfection tail length (96 6 0.83 mm for WT and 96.5 6 1.32 mm Small interfering RNAs (siRNA) sequences predesigned by 2 2 for NOV / ). Qiagen are listed in Supplementary Table 2. 3T3-L1 cells Indirect calorimetry was used to intensively analyze plated at 9 3 103 cells/cm2 in DMEM containing 10% FCS energy intake and energy expenditure of the mice. No were transfected, 1 day later, with siRNA (120 pmol in 5 mL) significant change in food intake (Fig. 1I)orinsponta- using the RNAi Max reagent (Qiagen). Differentiation was neous locomotor activity (Fig. 1K and Supplementary initiated 72 h after transfection (day 0). On day 3 after the Fig. 2C) was observed. A slight increase in energy ex- induction of differentiation, total RNA extracted from the 2 2 penditure was detected in NOV / mice but was limited cells was subjected to real-time qPCR. For the study of insulin to the daylight period (Fig. 1J and Supplementary Fig. signaling, 3T3-L1 cells were transfected on day 3 with siRNA 2B). The RER (VCO2-to-VO2 ratio) did not change be- (240 pmol in 10 mL); then, on day 5, the cultures were de- 2 2 tween NOV / and WT mice (Fig. 1L and Supplemen- prived of serum and insulin and were further treated on day 2 tary Fig. 2D). 6 with insulin (10 8 mol/L) for 8 min. RNA was obtained To examine whether the marked differences in BW from parallel cultures and underwent real-time qPCR. gain and adiposity between the two genotypes were asso- Protein Analysis ciated with changes in carbohydrate metabolism, we 3T3-L1 adipocytes or mouse tissues were lysed, and equal evaluated glycemia and insulinemia under basal and protein amounts (20 mg) were separated by SDS-PAGE. dynamic conditions (Fig. 2). During the OGTT, HFD-fed 2 2 The conditioned medium corresponding to 106 3T3-L1 ad- (Fig. 2B) but not SD-fed (Fig. 2A)NOV / animals ipocytes was used for ELISA tests or for NOV expression. showed improved glucose tolerance compared with WT mice. This correlated with a significant decrease in the Antibodies and Chemicals 2 2 area under the curve for NOV / mice compared with Recombinant human NOV protein was obtained as pre- controls (Fig. 2B, upper panels). Of note, fasting glucose viously described (15,17). For NOV protein detection, we 2 2 levels were slightly lower in HFD-fed NOV / mice com- used a homemade affinity-purified rabbit antibody (re- pared with their WT controls. We then conducted an in- ferred as CT-m, anti-mouse NOV antibody) (31). traperitoneal ITT (Fig. 2C and D), which demonstrated an Rabbit polyclonal antibodies against total Akt, phos- 2 2 improved insulin sensitivity in HFD-fed NOV / mice phorylated (p)-Akt (P-Ser 473), total Erk, and p-Erk, were compared with their WT littermates (Fig. 2D). The im- from Cell Signaling; uncoupling protein 1 (UCP-1), per- 2 2 proved glucose tolerance observed in HFD-fed NOV / oxisome proliferator-activated receptor g (PPAR-g)coactiva- mice was not accompanied by an increased insulinogenic tor 1-a (PGC1-a), poly(ADP-ribose) polymerase 1 (Parp-1) index after the oral glucose challenge compared with the were from Abcam (Paris, France), and mouse mono- WT controls (Fig. 2E). Because a previous study reported clonal antibodies against b-actin, b-tubulin, and GAPDH decreased insulin release subsequent to NOV overexpres- were from Sigma-Aldrich. sion in a rat b-cell line (25), insulin secretion was tested 2 2 Statistics in islets isolated from WT and NOV / mice fed the The results are expressed as the mean 6 SEM. Variance HFD. Insulin secretion similarly increased in control and 2 2 equality was analyzed by an F test (Excel; Microsoft Corp., NOV / islets when exposed to 16.7 mmol/L glucose or Issy-Les-Moulineaux, France). Comparisons between groups 50 mmol/L KCl compared with 5.5 mmol/L glucose (Fig. in the animal studies were carried out using a para- 2F). The pancreatic and islet insulin contents were also metric Student t test or by a nonparametric Mann- similar between the two genotypes. Further pancreatic Whitney-Wilcoxon test (Minitab, Paris, France). A P value morphometry analysis showed that b-cell fraction, b-cell of ,0.05 was considered statistically significant. mass, mean islet density, and mean islet size were similar 2 2 in NOV / and WT mice fed the HFD (Supplementary RESULTS Fig. 3). NOV2/2 Mice Are Protected Against HFD-Induced Collectively, these observations show that on an HFD, Obesity, Glucose Intolerance, and Insulin Resistance NOV deficiency decreased weight gain and improved glu- We addressed the effect of NOV deficiency on weight gain cose tolerance and insulin sensitivity, while insulin pro- in mice fed the SD or HFD for 16 weeks. No difference duction remained unchanged. 2506 NOV/CCN3 Is Involved in Insulin Resistance Diabetes Volume 65, September 2016

Figure 1—BW of WT and NOV2/2 (KO) mice (n =5–9 for each group) during a 16-week time course with the SD (A) or HFD (C). BW gain in WT and NOV2/2 mice fed the SD (B)orHFD(D). Mean fat content (E), lean body mass content (G), relative fat content (g/BW) (F), and relative lean body mass (g/BW) (H) of WT and NOV 2/2 mice fed the HFD for 14 weeks (n =5foreachgroup).I: Food intake. Whole energy expenditure was expressed as kcal/kg/h (J), spontaneous locomotor activity (beam break/h) (K), and RER (VCO2-to-VO2)(L)ofWT and NOV2/2 mice fed the HFD for 14 weeks (n = 5 for each group) by the mean of 4 days and 4 nights. Values are expressed as the means 6 SEM. *P < 0.05 and **P < 0.01 for NOV2/2 vs. WT mice.

2 2 HFD-Fed NOV / Mice Display Less Adipose Tissue, a with the assumption that adipocytes represent ;90% of adi- Higher Proportion of Small Adipocytes, and Improved pose tissue volume (32). The results revealed no difference in 2 2 Lipid and Hepatic Parameters the number of adipocytes in scWAT between WT and NOV / – – 6 On the SD, the organ weights to BW ratios were not mice (3.10 6 0.41 and 3.1 6 0.46 3 10 cells/g of tissue, fi signi cantly different between the two genotypes (Table respectively). In contrast, the adipocyte number in eWAT 2 2 1). However, on the HFD, a striking and similar decrease was significantly increased in NOV / mice (3.3 6 0.47 vs. 2/2 of these ratios was observed in NOV mice compared 2.1 6 0.32 3 106 cells/g of tissue, respectively; P , 0.05). with the WT controls for all the WAT depots. We also We also analyzed several plasma metabolic parameters, assessed the adipocyte size in hematoxylin and eosin– which are summarized in Table 1. On the SD, no differ- 2/2 stained eWAT and scWAT tissue from WT and NOV ences could be detected between the two genotypes. In WT mice fed the HFD. Of interest, we observed that NOV mice, the HFD induced a significant increase in the con- deficiency influenced the adipocyte size distribution in centrations of triglycerides, cholesterol (C), HDL-C, NEFA, 2 2 both fat tissues (Fig. 3). The mean adipocyte surface area and leptin. In NOV / mice, the HFD also provoked an 2 2 in eWAT and scWAT was markedly reduced in NOV / increase in cholesterol, triglycerides, HDL-C, and leptin, compared with WT mice (Fig. 3A and B). Moreover, x2 but not NEFA. However, compared with WT, the increases analysis on more than 1,500 independent determina- in cholesterol, HDL-C, and leptin were significantly reduced 2 2 2 2 tions showed that in both eWAT and scWAT, NOV / in NOV / mice. A trend (P , 0.1) toward a lower con- displayed a higher proportion of small adipocytes compared centration of LDL-C was also observed in these mice. with those of the WT mice (P , 0.0001) (Fig. 3C and D). Basal or induced lipolysis was also investigated in adipo- Adipocyte cellularity was also evaluated in eWAT and scWAT cytes isolated from eWAT fat depots from HFD-fed mice. diabetes.diabetesjournals.org Martinerie and Associates 2507

Figure 2—Glucose tolerance, insulin sensitivity, and insulin secretion in WT and NOV2/2 (KO) mice. OGTT (A and B) and ITT (C and D) were performed after 14 and 15 weeks, respectively, for animals fed the SD (A and C) or HFD (B and D). The area under the curve (AUC) for OGTT mice (n =10–14 per group) fasted for 6 h is presented in the upper panels of A and B. The results are expressed as the means 6 SEM. C and D: ITT mice (n =8–11 per group) were fasted for 4 h. The results are expressed as the percentage of basal glycemia, with the AUC presented in the upper panels. E: The insulinogenic index was determined at 0 and 15 min after oral feeding with glucose and expressed as the ratio Ins15min-Ins0 (mU/L)–to–G15min-G0 (mmol/L). F: Insulin secretion assayed after in vitro incubation of islets in 5.5 mmol/L, 16.7 mmol/L glucose, or 50 mmol/L KCl (n =4–5pergroup).*P < 0.05 and ***P < 0.001 for NOV2/2 vs. WT mice.

As shown in Supplementary Fig. 4, basal and isoproterenol- content was reduced by approximately threefold (Supple- stimulated lipolysis was stimulated in adipocytes from mentary Fig. 5B). This observation was consistent with 2 2 NOV / compared with those of WT animals. decreased mRNA levels of the fatty acid transporter 2 2 The liver phenotype of WT and NOV / mice fed the CD36 and its main transactivator PPAR-g (Supplementary HFD was also characterized. Analysis of hepatic histology, Fig. 5C). By contrast, no signiﬁcant difference was de- illustrated in Supplementary Fig. 5A, revealed less steato- tected for genes involved in fatty acid oxidation (PPAR-a, 2 2 sis in NOV / mice. Indeed, the hepatic triacylglycerol ACOX1), lipogenesis (SREBP-1c, ACC) or neoglucogenesis 2508 NOV/CCN3 Is Involved in Insulin Resistance Diabetes Volume 65, September 2016

Table 1—Organ weight and biological parameters in WT and NOV2/2 mice under SD and HFD SD HFD WT (n = 7) NOV2/2 (n = 10) WT (n = 9) NOV2/2 (n =9) Organ weight (mg)/BW (g) Liver 43.53 6 1.6 43.88 6 1.0 40.92 6 1.4 42.95 6 1.4 Spleen 4.16 6 0.7 4.14 6 0.4 2.82 6 0.4 3.22 6 0.3 Kidney 7.91 6 0.6 7.82 6 0.5 6.23 6 0.3 7.53 6 0.4* Pancreas 7.23 6 0.6 7.44 6 0.5 8.15 6 0.4 7.59 6 0.6 prWAT 2.88 6 0.6 3.44 6 0.6 9.44 6 1.1‡‡‡ 4.84 6 0.9** eWAT 6.45 6 0.9 5.90 6 0.9 22.81 6 1.71‡‡‡ 13.19 6 1.4***‡‡‡ scWAT 6.71 6 0.8 7.72 6 0.7 16.74 6 2.2‡‡ 9.13 6 1.7* BAT 4.1 6 0.7 3.83 6 0.6 6.56 6 0.7‡ 4.8 6 0.5 Biological parameters Glucose (mmol/L) 6.88 6 0.1 6.71 6 0.1 9.23 6 0.25‡‡‡ 8.35 6 0.3*‡‡‡ Insulin (mUI/L) ND ND 16.04 6 2.82 15.21 6 1.29 Cholesterol (g/L) 0.748 6 0.05 0.683 6 0.02 1.79 6 0.1‡‡‡ 1.51 6 0.1*‡‡‡ Triglycerides (g/L) 0.481 6 0.03 0.472 6 0.03 2.65 6 0.1‡‡‡ 2.35 6 0.1‡‡‡ HDL-C (g/L) 0.726 6 0.06 0.715 6 0.054 1.55 6 0.1‡‡‡ 1.29 6 0.1**‡‡‡ LDL-C (g/L) ND ND 0.118 6 0.05 0.063 6 0.04 NEFA (mmol/L) 168 6 16 165 6 31 259 6 30‡ 158 6 20** Leptin (ng/mL) 0.08 6 0.04 0.04 6 0.03 1.5 6 0.3‡‡‡ 0.66 6 0.2*‡‡ Adiponectin (mg/mL) 2.11 6 0.2 1.93 6 0.3 2.42 6 0.2 2.25 6 0.2 All of these parameters were obtained at sacriﬁce after 16 weeks of the SD or HFD diet. Results are expressed as means 6 SEM. ND, not determined. *P , 0.05, **P , 0.01, and ***P , 0.001 for NOV2/2 vs. WT mice (within the same diet); ‡P , 0.05, ‡‡P , 0.01, and ‡‡‡P , 0.001 for HFD vs. SD mice (within the same genotype).

(PEPCK1, G6Pase). Interestingly, we did not detect any in SREBP-1c, PPAR-g, and C/EBP-a expression. In line change in several liver inflammation markers, including with this result, when NOV expression was suppressed chemokine (C-C motif) ligand 2 (CCL2), TNF-a,CD68, in 3T3-L1 preadipocytes by transfection with an siRNA and F4/80 (Supplementary Fig. 5D). specific for NOV, the expression of SREBP-1c, PPAR-g, a In Vitro Preadipocyte Differentiation Is Increased and C/EBP- was also increased compared with the by NOV Deficiency same cells transfected with a nonsilencing siRNA (Fig. Because a difference in adipogenesis could at least partly 4F). We did not observe any difference in the proliferation capacity of preadipocytes derived from WT or explain the reduced adipocyte size and adipose mass 2 2 2 2 / found in HFD-fed NOV / mice, we first investigated NOV mice (Supplementary Fig. 6C). These results show that under HFD, the relative re- NOV expression in differentiating 3T3-L1 cells. As shown 2/2 in Supplementary Fig. 6A, NOV mRNA and protein ex- sistance to obesity and insulin resistance of NOV pression was strongly induced during differentiation. In mice might be related to a higher proportion of small agreement with this observation, there was an ;10-fold adipocytes that likely exhibit better intrinsic adipogenic increase in the NOV transcript level after adipocyte dif- ability in the absence of NOV. ferentiation in primary cultures (Supplementary Fig. 6B). Inflammation Is Reduced in Adipose Tissue From These findings suggested that NOV could be involved in NOV2/2 Mice adipogenesis. Thus, we then analyzed the expression of NOV has been involved in the regulation of various in- the three major transcription factors induced during adi- flammatory molecules (10,20) that are known to impair pogenesis in eWAT, perirenal WAT (prWAT), and scWAT. insulin signaling and promote insulin resistance in meta- As shown in Fig. 4A–C, we did not find any significant bolic tissues. Consistent with enhanced insulin action and differences in the expression of SREBP-1c, PPAR-g, and sensitivity (Fig. 2), we observed a specific pattern of al- 2 2 C/EBP-a between WT and NOV / mice in the three fat tered inflammatory gene expression in different adipose 2 2 depots, but the pattern of these mRNAs did not exclude tissues from HFD-fed NOV / mice (Fig. 5). In eWAT and the possibility that NOV could modulate adipogenesis. To prWAT, transcript levels of TNF-a, CCL2, and F4/80 were examine whether an alteration in adipocyte differentia- decreased (Fig. 5A and B), whereas the expression of these tion was an intrinsic property of NOV-deficient mice, we genes did not differ from WT in scWAT (Fig. 5C). We also isolated SVFs from eWAT and scWAT and tested their detected a decrease in leptin levels in eWAT and scWAT 2 2 in vitro adipogenic capacity. Interestingly, we showed in NOV / animals, consistent with their reduced leptin 2 2 (Fig. 4D and E) that preadipocytes originating from plasma levels (Table 1). Thus, in NOV / mice fed the 2 2 NOV / mice had a greater adipocyte differentiation abil- HFD, the reduced proinflammatory pattern of gene ex- ity than those from WT mice, as assessed by the increase pression essentially affects the visceral fat depots, which diabetes.diabetesjournals.org Martinerie and Associates 2509

Figure 3—Adipocyte size and cellularity in WT and NOV2/2 (KO) mice fed the HFD for 16 weeks. A and B: Representative hematoxylin and eosin–stained sections of eWAT and scWAT. Adipocyte size distribution determined from eWAT (C) and scWAT (D) histomorphometry and image analysis; insets show the mean adipocyte surface 6 SEM. ***P < 0.001. A x2 test was performed for each histogram; ***P < 0.0001 for NOV2/2 vs. WT animals. are generally considered to be crucial in the control of no difference in percentages of CD4+Foxp3+ regulatory metabolic status. Consistent with these data, we found T cells (Fig. 5G, right panel). Interestingly, a large pro- that treatment of 3T3-L1 cells with NOV recombinant portion of infiltrating CD3+ lymphocytes were identified – – protein led to a striking increase in CCL2 expression at CD4 CD8 (double-negative) T cells, the frequency of which 2 2 both the mRNA and protein level (Fig. 5D and E). was significantly increased in NOV / compared with WT It is now largely recognized that rodent or human mice (Fig. 5G,leftpanel). obesity is associated with a low-grade inflammation of We further analyzed whether NOV deletion can affect adipose tissue and its infiltration by immune cells, includ- M1 versus M2 polarization in myeloid/macrophage cells. ing lymphocytes and macrophages, that shift from a Macrophages were derived from BMM (or PEM) of WT 2 2 M2-like anti-inflammatory to a M1-like proinflammatory and NOV / HFD-fed mice. NOV expression was report- phenotype (33). In addition, these changes in macrophage ed in human macrophages (21); however, its expression in phenotypes are likely involved in obesity-linked insulin macrophages from HFD-fed WT mice has not been inves- resistance (34). Given the metabolic phenotype and the tigated so far. In our experimental conditions, the levels 2 2 profile of gene expression in deep fat depots of NOV / of NOV mRNA in BMM or PEM derived from HFD-fed animals, we further characterized the immune cell popu- WT mice were below the threshold of detection (data not 2 2 lations present in eWAT of NOV / and WT mice. The shown). Nevertheless, because a transient expression of absence of NOV was associated with an increase in a NOV during macrophage maturation or differentiation population of macrophages expressing high levels of the could affect their susceptibility to proinflammatory mole- 2 2 M2-like marker CD206 (Fig. 5F) and with a trend to fewer cules, we subjected WT and NOV / BMM or PEM cultures macrophages displaying an apparent M1-like profile to polarization to a M1-like or M2-like phenotype with LPS (CD11c+). No difference in the overall frequency of CD3+ and IL-4, respectively (32). Compared with untreated con- lymphocytes population was detected between the two trol cultures, the mRNA levels of M1-like proinflammatory genotypes (data not shown). Although no difference markers, such as TNF-a and CCL2, were significantly in- was observed in the frequency of CD8+Tcells,theper- duced in BMM cultures treated with LPS, whereas IL-4 centages of CD4+ lymphocytes among infiltrating T cells markedly enhanced the M2-like marker CD206 (Supple- 2 2 were strongly decreased in NOV / compared with WT mentary Fig 7A). We also observed in our experimental (Fig. 5G, left panel). This was related to significantly re- conditions, as previously described (35), that the mRNA – duced frequency of conventional CD4+Foxp3 T cells, with levels of the M2-like marker IL-10 was considerably 2510 NOV/CCN3 Is Involved in Insulin Resistance Diabetes Volume 65, September 2016

Figure 4—Expression of genes involved in adipocyte differentiation in eWAT (A), prWAT (B), and scWAT (C) in WT and NOV2/2 (KO) mice fed the HFD for 16 weeks (n =6–8 per group). SVF derived from eWAT (D) and scWAT (E) from mice fed the HFD for 16 weeks were cultured and underwent adipocyte differentiation for 7 days. F: Adipocyte differentiation was induced for 3 days in 3T3-L1 cells transfected with control nonsilencing siRNA (SC) or a specific siRNA against NOV (Si NOV). In the inset, the inhibition of NOV expression was evaluated on day 3 after adipocyte differentiation. Data represent the mean values of four measurements and correspond to a representative experiment performed at least three times with similar results. **P < 0.01 and ***P < 0.001 for NOV2/2 or siNOV conditions vs. control. induced by LPS. Analysis of TNF-a and IL-10 proteins con- adipose cells in these mice could be energetically more firmed the mRNA results (Supplementary Fig. 7B). expensive, we first analyzed markers of apoptosis in adi- Consistent with these results, we also found that among pose tissue. Western blots did not reveal any significant the CD11b+F4/80+ population of BMM treated with LPS differences in the Parp-1 cleavage in scWAT of both ge- there was a significant increase of macrophages expressing notypes on the HFD, whereas Parp-1 cleavage in eWAT 2 2 the M1-like marker inducible nitric oxide synthase 2, was even decreased in NOV / animals (Supplementary whereas IL-4 markedly enhanced the percentage of cells Fig. 9). Thus, these results could not account for the lower 2 2 expressing the M2-like marker CD206 (Supplementary BW gain of NOV / mice. Fig. 7C). Overall, mRNA, secretion, or FACS analyses de- We next analyzed the expression of genes in the brown 2 2 tected no significant difference between WT and NOV / adipose tissue (BAT) and WAT known to be involved in 2 2 BMM under these different treatments (Supplementary energy expenditure. In HFD-fed NOV / mice, the ex- Fig. 7A–C). Similar results were also obtained with PEM pression of UCP-1 was increased in BAT and eWAT at 2 2 (data not shown). Thus, whereas NOV / were protected both mRNA and protein levels (Fig. 6). The expression against inflammation in adipose tissue, it does not seem to of PGC1-a, a major coregulator that cooperates with be due to the absence of NOV in macrophages. PPAR-g to transactivate the UCP-1 gene promoter (38), Because limited inflammation and reduced fibrosis were was also significantly increased in the BAT of these mice 2 2 previously described in NOV / mice in the unilateral ure- (Fig. 6). teral obstruction experimental model of nephropathy (36) fi fi In Vitro and In Vivo NOV De ciency Increases Insulin and because brosis in adipose tissue is also linked to Sensitivity insulin-resistance (37), we also studied the effect of NOV To further delineate the molecular mechanisms involved fi 2 2 deletion on brosis of the adipose tissue. Analysis of the in the improved insulin sensitivity in NOV / mice, we interstitial collagen accumulation, assessed by Sirius Red col- analyzed the effect of NOV deficiency on insulin signaling. oration, revealed a trend for decreased fibrosis in eWAT of 2/2 We found that in 3T3-L1 adipocytes transfected with two NOV animals and no significant difference between WT fi 2/2 different siRNAs speci c for NOV, the decrease in NOV and NOV mice in the scWAT (Supplementary Fig. 8). expression was accompanied by a marked increase in Expression of Genes Involved in Energy Expenditure insulin-stimulated p-Akt and p-Erk compared with the Is Increased by NOV Deficiency in Adipose Tissue same cells transfected with a nonsilencing siRNA as a con- A slight increase in energy expenditure was observed in trol (sc) (Fig. 7A–C). In vivo, we also observed that an acute 2 2 NOV / mice (Fig. 1J). Because an increased turnover of exposure to insulin resulted in a significant increase in diabetes.diabetesjournals.org Martinerie and Associates 2511

Figure 5—Expression of inﬂammatory genes and analysis of inﬁltrating immune cell populations in eWAT (A), prWAT(B), and scWAT (C)of WT and NOV2/2 (KO) mice fed the HFD for 16 weeks. After 24 h of serum starvation, 3T3-L1 adipocytes were incubated with vehicle (Ctr) or NOV (10 mg/mL) for 24 h, and the level of CCL2 expression was evaluated in cells by real-time qPCR (D) or in the conditioned media by ELISA (E). Data represent the mean values 6 SEM (n =4–9 experiments). *P < 0.05, **P < 0.01, and ***P < 0.001 for KO mice or NOV- treated cells vs. control. F and G: Flow cytometry analysis of macrophages and lymphocytes isolated from eWAT of WT and NOV2/2 mice. F: Percentages of CD11c+ (left panel), or CD206+high (right panel) cells among CD11b+F4/80+ macrophages. G: Percentages of total CD8+CD4+ or CD4–CD8– double negative (DN) T cells (left and middle panels) and CD4+Foxp3+ regulatory T cells or CD4+Foxp3– conventional T cells (right panel) among stromal CD3+ lymphocytes. Data represent mean values 6 SEM (n = 4 per group). *P < 0.05.

2 2 2 2 p-Akt in eWAT of NOV / compared with WT mice (Fig. that in NOV / mice on an HFD, the inflammation of the 7D and E) but not in scWAT (data not shown). eWAT and prWAT was reduced compared with WT mice. 2 2 Taken together, these differences between WT and NOV / DISCUSSION mice likely account for the improved glucose tolerance, 2 2 In this study, we report for the first time that NOV insulin sensitivity, and metabolic profile of NOV / mice. deficiency protects mice fed an HFD against obesity, likely The increased population of small adipocytes in eWAT through an increase in energy expenditure. We also report could be due to the better proliferation or the better 2512 NOV/CCN3 Is Involved in Insulin Resistance Diabetes Volume 65, September 2016

Figure 6—Expression of genes and proteins involved in energy expenditure in BAT (A) and eWAT (B) of WT and NOV2/2 (KO) mice fed the HFD for 16 weeks (n =6–8 per group). C and D: Representative Western blots of BAT and eWAT. E and F: Bar graphs show fold change in UCP1 and PGC1-a in KO vs. WT mice. The results are expressed as the means 6 SEM. *P < 0.05 and ***P < 0.001 for NOV2/2 vs. WT mice.

2 2 differentiation ability of the resident preadipocytes. Our of fat mass in NOV / mice or whether this is overcome data are in favor of the second hypothesis because we did over time by increased cellularity. not observe any difference in the proliferation rate of Taking into account the whole body, this fat mass loss 2 2 primary preadipocytes derived from WT or NOV / mice. occurs while there is no decrease in food intake, and In contrast, the absence of NOV had a positive effect on preliminary analysis of lipids in feces did not reveal any the differentiation ability of these cells. Moreover, differ- signiﬁcant difference in gut absorption of lipids between 2 2 ent data have reported that NOV could play an effective NOV / and control mice (Supplementary Fig. 10). How- role in the control of cell differentiation because it can ever, there is a slight but signiﬁcant increased energy inhibit myoblast and osteoblast differentiation (13,39,40). expenditure during the daylight period. Remarkably, this It is also crucial for primitive hematopoietic progenitor cell enhanced energy expenditure is associated with an induction activity (14). However, although the improved ability to in BAT and eWAT in the expression of UCP1 and PGC1-a, 2 2 differentiate in NOV / preadipocytes was observed in two key effectors of the thermogenic pathway. This could 2 2 both eWAT and scWAT, the cellularity was only increased represent a major phenomenon to protect NOV / mice in eWAT. This observation is in accordance with recent from HFD-induced obesity and insulin resistance. However, data (41) showing that in mice fed an HFD for over the molecular mechanisms by which the absence of NOV 1 month, adipogenesis is preferentially initiated in eWAT leads to UCP1 induction remain unknown. For instance, the 2 2 compared with scWAT. This potentially improved ability increased lipolytic activity of NOV / adipocytes could also for preadipocyte differentiation, in which NOV is disrupted promote substrate availability for thermogenesis. or inhibited, is apparently contradictory with the decreased In this study, we report that the effect of the absence 2 2 fat mass in NOV / mice and with the increase in NOV of NOV is only detectable when mice are challenged with expression in vitro during the differentiation process. Our an HFD. Indeed, two reports describing the phenotype of 2 2 results suggest that the absence of NOV favors the recruit- NOV / mice revealed that they are viable and appar- ment of smaller adipocytes that are more sensitive to in- ently normal and fertile (22,40). However, after a lesion 2 2 sulin. However, a longer follow-up of WT and NOV / of the femoral artery, the null mice presented an enhance- mice would be required to examine whether the recruitment of neointimal thickening compared with WT mice ment of small adipocytes maintains the limited expansion during vascular repair (22). Another example showed that diabetes.diabetesjournals.org Martinerie and Associates 2513

Figure 7—Insulin signaling in siNOV-transfected 3T3-L1 adipocytes and in eWAT from WT and NOV2/2 (KO) mice fed the HFD. A: Representative Western blots of 3T3-L1 adipocytes transfected with a control nonsilencing siRNA (sc) or with two different speciﬁc siRNAs against NOV (si1 and si4) from day 3 to 6 after adipocyte differentiation. B: Bar graphs show fold change in p-Akt/Akt and p-Erk/Erk in si1- and si4- vs. sc-transfected cells. Values are the means 6 SEM of three experiments. C: NOV gene expression as determined in parallel cultures. *P < 0.05 and ***P < 0.001 for si1- or si4-treated vs. sc-treated cells. D: Representative Western blots of eWAT from WT and KO mice after or not a 10-min exposure to insulin (0.5 units). E: Bar graphs show fold change in p-Akt/Akt. The results are expressed as the means 6 SEM. ###P < 0.001 for NOV2/2 vs. WT mice; ***P < 0.001 for insulin-treated vs. untreated animals.

2 2 – – after an injury to the femur, bone regeneration in NOV / CD4 CD8 . These particular T cells that lack the expres- mice was accelerated compared with WT mice (40). sion of both CD4 and CD8 T-cell coreceptors have been More recently, we also found that during the progression studied in mice and humans for their contribution to peri- 2 2 of obstructive nephropathy, NOV / mice displayed lim- pheral tolerance and disease prevention (44). Our cur- ited inflammation and renal interstitial fibrosis compared rent findings are consistent with these observations. The with WT mice (36). reduced TNF-a and F4/80 expression in two visceral WAT 2 2 An HFD is known to induce low-grade inflammation in (eWAT and prWAT) depots from HFD-fed NOV / adipose tissue (42) and is associated with the increased ex- mice reflects the decreased inflammation in these fat pression of NOV mRNA in adipose tissue and NOV plasma depots and is in agreement with the involvement of levels (21). Under these specific nutritional conditions, the these tissues in insulin sensitivity (45–47). Interestingly, absence of NOV not only limits adipose tissue expansion but adecreaseinTNF-a and F4/80 expression was not de- is also associated with improved glucose tolerance. This is tected in scWAT, which is generally considered to have primarily related to better insulin sensitivity, as supported an accessory role in the insulin resistance phenotype. by ITTs and by the normalization of plasma NEFA levels in FACS analysis of eWAT suggests that under HFD, the 2 2 NOV / mice fed the HFD, suggesting the restoration of absence of NOV was associated with a switched overbal- the antilipolytic effect of insulin. These results are strength- ance from M1-like to M2-like macrophage population. ened by in vitro and in vivo experiments showing that NOV However, our in vitro analysis revealed that the absence deficiency in adipocytes promotes insulin signaling. of NOV did not influence the polarization of macro- Our results also strongly suggest that NOV could be a phages toward a M1 or M2 phenotype. This strongly key player in adipose tissue inflammation and that this in suggests that the switch from M1-like to M2-like mac- turn could promote insulin resistance. Recent studies rophage population does not result from an autocrine have shown that CD4+ and CD8+ T cells play a major role action of NOV on macrophages but rather from an in- in promoting adipose tissue inflammation by secreting direct role of NOV on the expression of proinflammatory cytokines and recruiting macrophages during HFD (43). molecules by adipocytes and possibly also by other cells NOV deficiency led to a strong decrease in the frequency present in the adipose tissue. Given the documented of infiltrating conventional CD4+ T cells, whereas no dif- protective role of such M2-like populations against ference was observed for CD8+ or CD4+ regulatory T cells. obesity-linked insulin resistance, this suggests that these 2 2 Of note, NOV / mice displayed a significantly in- changes may also contribute to the improved insulin 2 2 creased subpopulation of CD3+ T lymphocytes that were sensitivity and glucose tolerance of NOV / mice. 2514 NOV/CCN3 Is Involved in Insulin Resistance Diabetes Volume 65, September 2016

Our in vitro and in vivo experiments revealed that Duality of Interest. No potential conflicts of interest relevant to this article CCL2 expression is modulated by NOV. This regulation were reported. may be considered to be pivotal in the development of Author Contributions. C.M. and B.F. designed the research. M.G., T.T.H.D., insulin resistance because targeted deletions in mice for B.A., M.M., G.D., C.K., M.A., M.B., T.L., P.-O.M., M.F., A.B., B.B., and R.G.D. performed the research. H.K., S.H., and C.E.C. contributed new reagents/analytic the CCL2 and/or CCR2 genes result in reduced macro- fl tools. C.M., B.A., M.M., G.D., B.B., R.G.D., S.L., and B.F. analyzed data. C.M. and phage content, decreased in ammation in fat, and pro- B.F. wrote the paper. B.F. is the guarantor of this work and, as such, had full tection from HFD-induced insulin resistance; conversely, access to all the data in the study and takes responsibility for the integrity of the mice overexpressing CCL2 in adipose tissue have in- data and the accuracy of the data analysis. creased local macrophage infiltration along with insulin resistance (47–49). The specific macrophage marker References 2/2 F4/80 expression is reduced in NOV mice WAT. This 1. DeMaria EJ. Bariatric surgery for morbid obesity. N Engl J Med 2007;356: suggests that the absence of NOV could decrease CCL2 2176–2183 with a subsequent reduction in macrophage number within 2. Yoon KH, Lee JH, Kim JW, et al. Epidemic obesity and type 2 diabetes in the adipose tissue. However, the regulation of CCL2 by Asia. Lancet 2006;368:1681–1688 NOV could also have an effect on insulin sensitivity 3. He J, Klag MJ, Whelton PK, Zhoa Y, Weng X. Short- and long-term prog- through additional pathways because it has also been nosis after acute myocardial infarction in Chinese men and women. Am J Epi- – shownthatanincreaseintheplasmalevelsofCCL2is demiol 1994;139:693 703 – sufficient to induce insulin resistance independent of 4. Haslam DW, James WP. Obesity. Lancet 2005;366:1197 1209 fi 5. Renehan AG, Tyson M, Egger M, Heller RF, Zwahlen M. Body-mass index macrophage in ltration into adipose tissue (50). Ac- and incidence of cancer: a systematic review and meta-analysis of prospective cordingly, CCL2, through its angiogenic effect on endo- observational studies. Lancet 2008;371:569–578 thelial cells, could also contribute to the expansion and 6. Renehan A, Smith U, Kirkman MS. Linking diabetes and cancer: a con- remodeling of adipose tissue (51). sensus on complexity. Lancet 2010;375:2201–2202 During the OGTT, we did not observe a higher plasma 7. Romero-Corral A, Montori VM, Somers VK, et al. Association of bodyweight 2 2 insulin concentration in NOV / mice compared with WT with total mortality and with cardiovascular events in coronary artery disease: controls. Insulin release in isolated islets was also compa- a systematic review of cohort studies. Lancet 2006;368:666–678 rable between the two genotypes, and the total content of 8. Joliot V, Martinerie C, Dambrine G, et al. Proviral rearrangements and insulin in the whole pancreas or in isolated islets was overexpression of a new cellular gene (nov) in myeloblastosis-associated virus 2 2 – similar in NOV / and WT mice. These results are appar- type 1-induced nephroblastomas. Mol Cell Biol 1992;12:10 21 9. Chen CC, Lau LF. Functions and mechanisms of action of CCN matricellular ently discordant with those reported by Paradis et al. (25). proteins. Int J Biochem Cell Biol 2009;41:771–783 Nevertheless, the experimental conditions used by that 10. Kular L, Pakradouni J, Kitabgi P, Laurent M, Martinerie C. The CCN family: a study and ours substantially differ because they used a new class of inflammation modulators? Biochimie 2011;93:377–388 rat cell line overexpressing NOV. Taken together, our re- 11. Laurent M, Martinerie C, Thibout H, et al. NOVH increases MMP3 expression sults suggest that the increase in insulin sensitivity is the and cell migration in glioblastoma cells via a PDGFR-alpha-dependent mecha- major mechanism by which glucose tolerance is improved nism. FASEB J 2003;17:1919–1921 2 2 in NOV / mice. 12. Lin CG, Chen CC, Leu SJ, Grzeszkiewicz TM, Lau LF. Integrin-dependent In summary, the current study demonstrates that NOV functions of the angiogenic inducer NOV (CCN3): implication in wound healing. participates in the development of obesity-induced insu- J Biol Chem 2005;280:8229–8237 lin resistance. These results indicate that the use of coun- 13. Calhabeu F, Lafont J, Le Dreau G, et al. NOV/CCN3 impairs muscle cell – teracting agents against NOV represents a potential strategy commitment and differentiation. Exp Cell Res 2006;312:1876 1889 14. Gupta R, Hong D, Iborra F, Sarno S, Enver T. NOV (CCN3) functions as a reg- for new therapies against obesity and its metabolic ulator of human hematopoietic stem or progenitor cells. Science 2007;316:590–593 comorbidities. 15. Doghman M, Arhatte M, Thibout H, et al. Nephroblastoma overexpressed/ cysteine-rich protein 61/connective tissue growth factor/nephroblastoma overexpressed gene-3 (NOV/CCN3), a selective adrenocortical cell proapoptotic factor, Acknowledgments. The authors acknowledge L. Dinard from the animal is down-regulated in childhood adrenocortical tumors. J Clin Endocrinol Metab facilities platform (PHEA Saint-Antoine) for her technical support, A. Munier from 2007;92:3253–3260 the cytometry platform (UMS, Lumic 030) for her help in FACS analysis, 16. McCallum L, Lu W, Price S, Lazar N, Perbal B, Irvine AE. CCN3: a key growth R. Morichon from the image platform (UMS, Lumic 030) for his help in micro- regulator in chronic myeloid leukaemia. J Cell Commun Signal 2009;3:115–124 scopy analysis, and S. Dumont and F. Merabetene for their contribution in 17. Lafont J, Jacques C, Le Dreau G, et al. New target genes for NOV/CCN3 in adipose tissue histomorphological analysis (UMS, Lumic 030). The authors also chondrocytes: TGF-beta2 and type X collagen. J Bone Miner Res 2005;20:2213–2223 acknowledge the technical platform Functional and Physiological Exploration 18. Riser BL, Najmabadi F, Perbal B, et al. CCN3 (NOV) is a negative regulator of (FPE) of the Unit “Biologie Fonctionnelle et Adaptative” (Sorbonne Paris City CCN2 (CTGF) and a novel endogenous inhibitor of the fibrotic pathway in an University, Paris Diderot University, BFA, CNRS, Paris, France) for metabolic in vitro model of renal disease. Am J Pathol 2009;174:1725–1734 and behavioral analysis. 19. Chen CC, Young JL, Monzon RI, Chen N, Todorovic V, Lau LF. Cytotoxicity of Funding. This work was financed by INSERM, Pierre et Marie Curie University, TNFalpha is regulated by integrin-mediated matrix signaling. EMBO J 2007;26: and by a grant from the French Society of Diabetes. T.T.H.D. received financial 1257–1267 support from the French Society of Endocrinology and the French Embassy in 20. Le Dréau G, Kular L, Nicot AB, et al. NOV/CCN3 upregulates CCL2 and Vietnam, and P.-O.M. received financial support from ANRT (National Association CXCL1 expression in astrocytes through beta1 and beta5 integrins. Glia 2010;58: of Research and Technology). 1510–1521 diabetes.diabetesjournals.org Martinerie and Associates 2515

21. Pakradouni J, Le Goff W, Calmel C, et al. Plasma NOV/CCN3 levels are 37. Sun K, Tordjman J, Clément K, Scherer PE. Fibrosis and adipose tissue closely associated with obesity in patients with metabolic disorders. PLoS One dysfunction. Cell Metab 2013;18:470–477 2013;8:e66788 38. Puigserver P, Wu Z, Park CW, Graves R, Wright M, Spiegelman BM. A cold- 22. Shimoyama T, Hiraoka S, Takemoto M, et al. CCN3 inhibits neointimal inducible coactivator of nuclear receptors linked to adaptive thermogenesis. Cell hyperplasia through modulation of smooth muscle cell growth and migration. 1998;92:829–839 Arterioscler Thromb Vasc Biol 2010;30:675–682 39. Sakamoto K, Yamaguchi S, Ando R, et al. The nephroblastoma overex- 23. Martinerie C, Viegas-Pequignot E, Guénard I, et al. Physical mapping of pressed gene (NOV/ccn3) protein associates with Notch1 extracellular domain human loci homologous to the chicken nov proto-oncogene. Oncogene 1992;7: and inhibits myoblast differentiation via Notch signaling pathway. J Biol Chem 2529–2534 2002;277:29399–29405 24. An P, Freedman BI, Rich SS, et al. Quantitative trait loci on chromosome 40. Matsushita Y, Sakamoto K, Tamamura Y, et al. CCN3 protein participates 8q24 for pancreatic beta-cell function and 7q11 for insulin sensitivity in obese in bone regeneration as an inhibitory factor. J Biol Chem 2013;288:19973– nondiabetic white and black families: evidence from genome-wide linkage scans 19985 in the NHLBI Hypertension Genetic Epidemiology Network (HyperGEN) study. 41. Wang QA, Tao C, Gupta RK, Scherer PE. Tracking adipogenesis during white Diabetes 2006;55:551–558 adipose tissue development, expansion and regeneration. Nat Med 2013;19: 25. Paradis R, Lazar N, Antinozzi P, Perbal B, Buteau J. Nov/Ccn3, a novel 1338–1344 transcriptional target of FoxO1, impairs pancreatic b-cell function. PLoS One 42. Chandalia M, Abate N. Metabolic complications of obesity: inflated or in- 2013;8:e64957 flamed? J Diabetes Complications 2007;21:128–136 26. Taicher GZ, Tinsley FC, Reiderman A, Heiman ML. Quantitative magnetic 43. Huh JY, Park YJ, Ham M, Kim JB. Crosstalk between adipocytes and im- resonance (QMR) method for bone and whole-body-composition analysis. Anal mune cells in adipose tissue inflammation and metabolic dysregulation in obe- Bioanal Chem 2003;377:990–1002 sity. Mol Cells 2014;37:365–371 27. Abed A, Toubas J, Kavvadas P, et al. Targeting connexin 43 protects against 44. Hillhouse EE, Lesage S. A comprehensive review of the phenotype and the progression of experimental chronic kidney disease in mice. Kidney Int 2014; function of antigen-specific immunoregulatory double negative T cells. J Auto- 86:768–779 immun 2013;40:58–65 28. Dole VP, Meinertz H. Microdetermination of long-chain fatty acids in plasma 45. Weisberg SP, McCann D, Desai M, Rosenbaum M, Leibel RL, Ferrante AW and tissues. J Biol Chem 1960;235:2595–2599 Jr. Obesity is associated with macrophage accumulation in adipose tissue. J Clin 29. Stanley ER. Murine bone marrow-derived macrophages. Methods Mol Biol Invest 2003;112:1796–1808 1997;75:301–304 46. Xu H, Barnes GT, Yang Q, et al. Chronic inflammation in fat plays a crucial 30. Pfaffl MW. A new mathematical model for relative quantification in real-time role in the development of obesity-related insulin resistance. J Clin Invest 2003; RT-PCR. Nucleic Acids Res 2001;29:e45 112:1821–1830 31. Kular L, Rivat C, Lelongt B, et al. NOV/CCN3 attenuates inflammatory pain 47. Weisberg SP, Hunter D, Huber R, et al. CCR2 modulates inflammatory and through regulation of matrix metalloproteinases-2 and -9. J Neuroinflammation metabolic effects of high-fat feeding. J Clin Invest 2006;116:115–124 2012;9:36 48. Kanda H, Tateya S, Tamori Y, et al. MCP-1 contributes to macrophage 32. Eto H, Suga H, Matsumoto D, et al. Characterization of structure and cellular infiltration into adipose tissue, insulin resistance, and hepatic steatosis in obesity. components of aspirated and excised adipose tissue. Plast Reconstr Surg 2009; J Clin Invest 2006;116:1494–1505 124:1087–1097 49. Kamei N, Tobe K, Suzuki R, et al. Overexpression of monocyte chemo- 33. Lumeng CN, Bodzin JL, Saltiel AR. Obesity induces a phenotypic switch in attractant protein-1 in adipose tissues causes macrophage recruitment and in- adipose tissue macrophage polarization. J Clin Invest 2007;117:175–184 sulin resistance. J Biol Chem 2006;281:26602–26614 34. Gregor MF, Hotamisligil GS. Inflammatory mechanisms in obesity. Annu Rev 50. Tateya S, Tamori Y, Kawaguchi T, Kanda H, Kasuga M. An increase in the Immunol 2011;29:415–445 circulating concentration of monocyte chemoattractant protein-1 elicits systemic 35. Saraiva M, O’Garra A. The regulation of IL-10 production by immune cells. insulin resistance irrespective of adipose tissue inflammation in mice. Endocri- Nat Rev Immunol 2010;10:170–181 nology 2010;151:971–979 36. Marchal PO, Kavvadas P, Abed A, et al. Reduced NOV/CCN3 expression 51. Salcedo R, Ponce ML, Young HA, et al. Human endothelial cells express limits inflammation and interstitial renal fibrosis after obstructive nephropathy in CCR2 and respond to MCP-1: direct role of MCP-1 in angiogenesis and tumor mice. PLoS One 2015;10:e0137876 progression. Blood 2000;96:34–40