This is the accepted version of the article: Casaña Lorente, Estefania; Jimenez, Veronica; Sacristan, Victor; [et al.]. «BMP7 overexpression in adipose tissue induces white adipogenesis and improves insulin sensitivity in ob». International Journal of Obesity, Vol. 45, Issue 2 (February 2021), p. 449-460. 12 pàg. DOI 10.1038/s41366-020-00700-6 This version is available at https://ddd.uab.cat/record/239890 under the terms of the license 1 BMP7 overexpression in adipose tissue induces white 2 adipogenesis and improves insulin sensitivity in ob/ob mice 3 4 Estefania Casana1,2,3*, Veronica Jimenez1,2,3*, Victor Sacristan1,2, Sergio 5 Muñoz1,2,3, Claudia Jambrina1,2,3, Jordi Rodó1,2, Miquel Garcia1,2,3, Cristina 6 Mallol1,2, Xavier León1,2,3, Sylvie Franckhauser1,2,3 and Fatima Bosch1,2,3# 7 1 8 Center of Animal Biotechnology and Gene Therapy (CBATEG), 2Department of 9 Biochemistry and Molecular Biology, Universitat Autònoma de Barcelona, 10 08193 Bellaterra, and 3CIBER de Diabetes y Enfermedades Metabólicas 11 Asociadas (CIBERDEM), 28029 Madrid, Spain. 12 13 14 * These authors contributed equally to this work. 15 16 # To whom correspondence should be addressed: 17 Center of Animal Biotechnology and Gene Therapy 18 Edifici H, Universitat Autònoma de Barcelona 19 08193 Bellaterra (Barcelona) 20 SPAIN 21 Tel: +34 93 581 41 82 22 E-mail: fatima.bosch@uab.cat 23 24 Running title: BMP7 induces white adipogenesis 1 25 ABSTRACT 26 Background/Objectives: During obesity, hypertrophic enlargement of white 27 adipose tissue (WAT) promotes ectopic lipid deposition and development of 28 insulin resistance. In contrast, WAT hyperplasia is associated with preservation 29 of insulin sensitivity. The complex network of factors that regulate white 30 adipogenesis is not fully understood. Bone morphogenic protein 7 (BMP7) can 31 induce brown adipogenesis, but its role on white adipogenesis remains to be 32 elucidated. Here, we assessed BMP7-mediated effects on white adipogenesis 33 in ob/ob mice. 34 Methods: BMP7 was overexpressed in either WAT or liver of ob/ob mice using 35 adeno-associated viral (AAV) vectors. Analysis of gene expression, histological 36 and morphometric alterations, and metabolites and hormones concentrations 37 were carried out. 38 Results: Overexpression of BMP7 in adipocytes of subcutaneous and visceral 39 WAT increased fat mass, the proportion of small-size adipocytes and the 40 expression of adipogenic and mature adipocyte genes, suggesting induction of 41 adipogenesis irrespective of fat depot. These changes were associated with 42 reduced hepatic steatosis and improved insulin sensitivity. In contrast, liver- 43 specific overproduction of BMP7 did not promote WAT hyperplasia despite 44 BMP7 circulating levels were similar to those achieved after genetic engineering 45 of WAT. 46 Conclusions: This study unravels a new autocrine/paracrine role of BMP7 on 47 white adipogenesis and highlights that BMP7 may modulate WAT plasticity and 48 increase insulin sensitivity. 49 2 50 INTRODUCTION 51 52 Obesity and type 2 diabetes (T2D) are strongly associated and have become an 53 alarming growing health problem worldwide. Obesity has been causally linked 54 to the development of insulin resistance, type 2 diabetes (T2D), arthritis, cancer, 55 cardiovascular diseases and Alzheimer’s disease. These obesity-linked 56 complications lead to reduced life expectancy and poor quality of life, thus 57 representing a massive burden for the health-care systems. 58 Obesity is a condition where adipose tissue mass is increased due to an 59 imbalance between energy intake and expenditure. Expandability of white 60 adipose tissue (WAT) may result from an increase in the size (hypertrophy) 61 and/or in the number of adipocytes (hyperplasia), by the differentiation of new 62 adipocytes from undifferentiated preadipocytes (adipogenesis) [1–3]. Adipocyte 63 hypertrophy is closely linked to adipose dysfunction and inflammation, abnormal 64 secretion patterns of adipokines, ectopic lipid deposition in non-adipose tissues 65 such as liver, skeletal muscle and heart, and whole-body insulin resistance and 66 T2D, not only in obese but also in lean individuals [4,5]. Conversely, WAT 67 expansion through hyperplasia has been associated with improved insulin 68 sensitivity [6–8]. 69 Under physiological conditions, in a situation of positive energy balance, 70 in both humans and rodents excess lipids is stored primarily via hyperplasia in 71 the subcutaneous adipose tissue (SAT), since this depot has greater adipogenic 72 differentiation capacity than visceral adipose tissue (VAT) [4,9,10]. In obesity, 73 recruitment and adipogenic differentiation of the stromal vascular precursor 74 cells in SAT are impaired. Therefore, subcutaneous adipocytes become 3 75 hypertrophic and adipogenesis is restricted to VAT, which ultimately also 76 expands through hypertrophy when its hyperplasic capacity is exceeded [4,11– 77 17]. Enhancement of WAT hyperplasia may limit the deleterious metabolic 78 effects mediated by dysfunctional white adipocytes and help to preclude hepatic 79 steatosis and insulin resistance. Nevertheless, the complex network of factors 80 that regulate white adipogenesis has not been fully elucidated. 81 The bone morphogenetic protein (BMP) family belongs to the 82 transforming growth factor β (TGFβ) superfamily of cytokines that regulate an 83 array of fundamental cellular processes, including proliferation, differentiation, 84 apoptosis and morphogenesis [18]. BMPs play major roles in adipogenesis, not 85 only regulating progenitor cell determination, but also promoting terminal 86 adipogenic differentiation [19,20]. Among them, BMP7 was considered to be 87 essential for brown adipogenesis in both committed brown preadipocytes and 88 uncommitted multipotent mesenchymal precursors [19,21]. In contrast, BMP2 89 and BMP4 had been described as master regulatory factors that drive the 90 commitment and differentiation of adipocyte precursors into white adipocytes 91 [19]. Recent studies have evidenced that BMP2 and BMP4, as well as BMP9 92 and BMP14, can trigger both white and brown adipogenesis [20,22–30]. 93 Nevertheless, whether the brown adipogenic factor BMP7 may also induce 94 white adipogenesis is unknown. 95 In this study, specific overexpression of BMP7 in WAT of adult obese 96 mice resulted in redistribution of the size of white adipocytes, with a greater 97 proportion of small-size adipocytes both in SAT and VAT, and amelioration of 98 insulin resistance. In contrast, overexpression of BMP7 in the liver led to similar 99 levels of circulating BMP7 but did not promote WAT hyperplasia. These results 4 100 are consistent with a paracrine/autocrine role of BMP7 inducing white 101 adipogenesis. 102 5 103 1. MATERIAL AND METHODS 104 105 2.1. Animals. Eleven-week-old B6.V-Lepob/OlaHsd (ob/ob) male mice were 106 used. Mice were kept in a specific pathogen-free facility (SER-CBATEG, UAB) 107 and maintained under a light-dark cycle of 12 h at 22 ºC. Mice were fed ad 108 libitum with a standard diet (2018S Teklad Global Diets®, Envigo). For tissue 109 sampling, mice were anesthetized with inhalational anesthetic isoflurane 110 (IsoFlo®, Abbott Laboratories, Abbott Park, IL, US) and sacrificed. Tissues of 111 interest were excised and kept at -80°C or in formalin until analysis. Animal care 112 and experimental procedures were approved by the Ethics Committee in Animal 113 and Human Experimentation of the Universitat Autònoma de Barcelona. 114 115 2.2. Recombinant AAV vectors. AAV expression cassettes were obtained by 116 cloning, between the ITRs of AAV2, a murine optimized BMP7 coding-sequence 117 (moBMP7) under the control of either: i) the liver-specific human α1-antitrypsin 118 promoter (hAAT) (AAV-hAAT-BMP7); or ii) the ubiquitous early CMV 119 enhancer/chicken beta actin promoter (CAG) with the addition of 4 tandem 120 repeats of miRT122a sequence (5’CAAACACCATTGTCACACTCCA3’) and 121 miRT1 sequence (5’TTACATACTTCTTTACATTCCA3’) in the 3’ untranslated 122 region of the expression cassette. The moBMP7 sequence comprised in the 123 expression cassettes was a murine BMP7 coding-sequence that was codon- 124 optimized to enhance production of wild-type BMP7 protein using the 125 GeneOptimizer algorithm (GeneArt; Life Technologies), which relies on a 126 multifactorial approach. Non-coding cassettes, carrying either the hAAT or the 127 CAG promoter but no transgene, were used to produce null vectors. Single- 6 128 stranded AAV vectors of serotype 8 were produced by triple transfection in 129 HEK293 cells. HEK293 cells were kindly provided by K.A. High, Children’s 130 Hospital of Philadelphia. AAV vectors were purified using an optimized CsCl 131 gradient–based purification protocol that renders vector preps of high purity and 132 devoid of empty capsids [31]. Viral vectors were determined by fluorescence 133 using the Quant-iT™ PicoGreen™ dsDNA Assay Kit (Invitrogen). A phage 134 lambda DNA was used as standard curve to calculate the titer of viral vectors. 135 136 2.3. Administration of AAV vectors. Systemic intravenous and intra- 137 epididimal WAT (eWAT) administration of AAV vectors were performed as 138 previously described [32,33]. 139 140 2.4. Immunohistochemistry. Tissues were fixed for 12-24 h in 10% formalin, 141 embedded in paraffin and sectioned. Sections were incubated overnight at 4ºC 142 with rat anti-Mac2 antibody (CL8942AP; Cedarlane). Biotinylated rabbit anti-rat 143 (E0467; Dako) was used as secondary antibody. The ABC peroxidase kit 144 (Pierce) was used for immunodetection, and sections were counterstained in 145 Mayer’s hematoxylin. Morphometric analysis of adipocyte size was performed in 146 WAT sections stained with hematoxylin–eosin as previously described [34]. A 147 minimum of four animals per group was used and at least 250 adipocytes per 148 animal were analyzed. 149 150 2.5. RNA analysis. Total RNA was obtained from different tissues using 151 isolation reagent (Tripure, Roche, for liver samples and QIAzol, Qiagen, for 152 adipose depots) and a RNeasy Minikit (Qiagen) and treated with DNAseI 7 153 (Qiagen). One µg of RNA was reverse-transcribed using the Transcriptor First 154 Strand cDNA Synthesis kit (Roche). Real-time quantitative PCR (qRT-PCR) 155 was performed in a LightCycler (Roche) using 1) the LightCycler 480 SYBR 156 Green I Master Mix (Roche) and murine primers (Table S1) or 2) the 157 LightCycler 480 Probes Master Mix (Roche) and murine primers and probes 158 (IDT, Leuven, Germany) (Table S2). Data were normalized to Rplp0 expression. 159 160 2.6. Hormone and metabolite assays. Hepatic triglyceride content was 161 determined after chloroform:methanol (2:1 vol/vol) extraction of total lipids, as 162 previously described [35]. Triglycerides were quantified spectrophotometrically 163 using an enzymatic assay (Horiba-ABX) in a Pentra 400 Analyzer (Horiba-ABX). 164 Glycemia was determined using a Glucometer EliteTM (Bayer) and insulin levels 165 were measured using the Rat Insulin ELISA kit (90010, Crystal Chem). Serum 166 BMP7 and adiponectin levels were determined using the Human BMP7 ELISA 167 kit (DBP700, R&D Systems) and the Mouse Adiponectin ELISA kit (80569, 168 Crystal Chem). 169 170 2.7. Insulin tolerance test. Insulin (Humulin Regular; Eli Lilly) was injected 171 intraperitoneally at a dose of 0.75 IU/kg body weight to fed mice. Glycemia was 172 measured in blood samples from tail vein at the indicated time points. 173 174 2.8. Indirect calorimetry and activity. An indirect open circuit calorimeter 175 (Oxylet, Panlab) was used to monitor O2 consumption, CO2 production and 176 activity. Mice were individualized and acclimated to the metabolic chambers for 177 24h. O2 consumption and CO2 production data were collected in each cage for 3 8 178 min, every 15 min, for 24 h during the light and dark cycles and adjusted by 179 body weight. Activity was recorded continuously for 24 h during the light and 180 dark cycles. 181 182 2.9. Statistical analysis. Sample size determination was based on previous 183 experience with similar studies. Mice were randomly divided into groups (n=8- 184 10 per group). In addition, we tested that the mean body weight and the mean 185 glycemia were statistically not different for each experimental group prior to 186 assignment to treatment groups. Furthermore, each experimental group was 187 caged separately to avoid any caging effects. All tests and analyses were 188 performed by investigators blinded to the treatment. All results are expressed as 189 mean ± SEM. Values higher than 1.5IQR were considered atypical and were 190 excluded from analyses. Differences between groups were compared by two- 191 sided Student’s t-test. Statistical significance was considered if P< 0.05. 192 9 193 3. RESULTS 194 195 3.1. BMP7 promotes white adipose tissue expansion 196 Intra-adipose depot delivery of AAV8 vectors in both lean and obese diabetic 197 mice leads to long-term efficient transduction of WAT and is a useful tool to 198 study adipose pathophysiology and adipocyte function [32,33,36]. To examine 199 the role of BMP7 on white adipogenesis, we chose ob/ob mice as a well- 200 established model of obesity with WAT hypertrophy, insulin resistance and a 201 significant accumulation of hepatic triglycerides relatively early in life [37,38]. A 202 cohort of ob/ob mice received an intra-epidydimal WAT (eWAT) injection of 203 1x1012 vg/mouse of AAV8 vectors encoding murine BMP7 under the 204 transcriptional control of the ubiquitous CAG promoter. To avoid expression of 205 the transgene in other main organs for which AAV8 shows strong tropism, such 206 as liver and heart [39–41], we took advantage of microRNAs (miRs). Target 207 sequences for miR-122a and miR-1, which selectively de-target transgene 208 expression from liver and heart when included into AAV vectors [32,42], were 209 added in tandem repeats of four copies to the 3’-UTR of the murine BMP7 210 expression cassette (AAV8-CAG-BMP7-miRT122-miRT1; AAV-BMP7). Another 211 cohort of ob/ob mice administered with 1x1012 vg of non-coding null vectors 212 (AAV8-null) served as controls. 213 Animals treated intra-eWAT with AAV-BMP7 vectors showed high BMP7 214 overexpression mainly in this depot but also in retroperitoneal (rWAT), 215 mesenteric (mWAT) and inguinal (iWAT) depots (Figure 1A), as previously 216 reported [33]. Marginal BMP7 expression was detected in iBAT (Figure 1A). As 217 expected, microRNA target sequences efficiently prevented BMP7 expression 10 218 in the liver (Figure 1A) and heart (data not shown). In agreement with previous 219 reports demonstrating the high secretory capacity of AAV-modified WAT 220 [32,33], increased BMP7 circulating levels were observed in mice receiving 221 AAV-BMP7 vectors (Figure 1B). 222 As animals aged, ob/ob mice overexpressing BMP7 in white adipocytes 223 showed increased body weight compared with AAV-null-treated counterparts 224 (Figure 1C) although no differences in food intake were observed (Figure S1A). 225 A tendency towards a decrease in spontaneous activity during the dark phase 226 may have contributed to body weight gain (Figure S1B). Treatment with AAV- 227 BMP7 vectors also led to a specific increase of the main WAT depots weight 228 (Figure 1D). This was parallel to a redistribution of the size of white adipocytes, 229 with a greater proportion of small-size adipocytes in both eWAT and iWAT 230 (Figure 1E,F and Figure S1C). Altogether, these observations suggested that 231 BMP7 induced white adipogenesis, irrespective of fad pad. 232 233 3.2. BMP7 induces expression of adipogenic markers and decreases 234 WAT inflammation 235 The expression of the preadipocyte marker Preadipocyte factor 1 (Pref1) as well 236 as that of the final adipogenic inducers Peroxisome Proliferator Activated 237 Receptor Gamma (Pparγ) and CCAAT/enhancer binding protein alpha (Cebpα) 238 was induced in eWAT and iWAT of ob/ob mice overexpressing BMP7 in white 239 adipocytes (Figure 2A,B). These results suggested induction of white 240 adipogenesis in SAT and VAT of these animals. In addition, the expression of 241 proteins involved in lipid accumulation, which are markers of mature adipocytes, 242 such as sterol regulatory element binding transcription factor 1 (Srebf1), fatty 11 243 acid synthase (Fasn), acetyl-CoA carboxylase 1 (Acc1), acetyl-CoA carboxylase 244 2 (Acc2), fatty acid-binding protein 4 (Fabp4), perilipin 1 (Plin1) and glucose 245 transporter type 4 (Slc2a4), was also increased in eWAT and iWAT of AAV- 246 BMP7-treated animals (Figure 2C,D). 247 In agreement with expansion of WAT, ob/ob mice overexpressing BMP7 248 showed increased adiponectin levels (Figure 2E). WAT inflammation was also 249 decreased in these mice, evidenced by lower presence of macrophages, 250 revealed as “crown-like” structures, and reduced expression of the macrophage 251 markers F480 and Cd68 and of the pro-inflammatory cytokines Mcp1 and Tnfα 252 compared with AAV-null treated mice (Figure 2F-I). 253 254 3.3. BMP7 overexpression in WAT does not induce brown adipogenesis 255 In contrast to the observations made in WAT, ob/ob mice treated with AAV- 256 BMP7 or AAV-null showed similar iBAT weight and lipid deposition in this depot 257 (Figure 1D and Figure 3A) likely due to the marginal expression of BMP7 in 258 BAT (Figure 1A). BMP7 can induce brown adipocyte differentiation in vitro and 259 non-shivering thermogenesis [21,43]. However, no differences in the expression 260 of pro-adipogenic and mature adipocyte markers were observed in iBAT of 261 AAV-BMP7-treated mice (Figure 3B,C). Moreover, multilocular beige adipocytes 262 were not detected in iWAT (Figure 1E) and the expression of the thermogenic 263 markers uncoupling protein 1 (Ucp1) and peroxisome proliferator-activated 264 receptor gamma coactivator 1-alpha (Ppargc1a) remained unchanged in iBAT 265 and iWAT (Figure 3D,E). Consistent with these findings, WAT-derived BMP7 266 failed to induce energy expenditure (Figure 3F). 267 12 268 3.4. BMP7 overproduction in WAT ameliorates hepatic steatosis and 269 improves insulin resistance 270 Histological analysis of the liver revealed that null-treated ob/ob mice developed 271 marked hepatic steatosis (Figure 4A). In contrast, hepatic lipid deposition was 272 decreased in ob/ob mice overexpressing BMP7 in WAT (Figure 4A,B). This was 273 parallel to reduced hepatic inflammation, evidenced by decreased number of 274 Mac2+ cells (Figure 4C) and lower liver expression of Cd68 (Figure 4D). 275 Ob/ob mice treated with AAV-null vectors showed normal fed glycaemia 276 but were hyperinsulinemic (Figure 5A,B). In contrast, ob/ob mice treated with 277 BMP7 were normoglycemic and presented a marked reduction of serum insulin 278 levels (Figure 5A,B), suggesting improved insulin sensitivity. The intraperitoneal 279 insulin tolerance test (ITT) confirmed amelioration of insulin resistance in AAV- 280 BMP7-treated ob/ob mice (Figure 5C). 281 282 3.5. BMP7 overexpression in the liver does not induce white 283 adipogenesis 284 To elucidate whether the BMP7-mediated induction of WAT hyperplasia was 285 due to the paracrine/autocrine action of BMP7 in WAT or to the increased 286 circulating levels of the factor, the liver of ob/ob mice was genetically 287 engineered to overproduce BMP7. To this end, ob/ob mice were administered 288 intravenously (IV) with 5x1011 vg of AAV8 vectors encoding murine BMP7 under 289 the control of the liver-specific hAAT promoter (AAV-hAAT-BMP7). As controls, 290 another cohort of ob/ob mice received the same dose of non-coding null vectors 291 (AAV-hAAT-null). 13 292 Ob/ob mice treated with AAV-hAAT-BMP7 vectors showed specific 293 hepatic overexpression of BMP7 (Figure 6A). This led to increased BMP7 294 circulating levels (Figure 6B), which were similar to those observed in ob/ob 295 mice overexpressing BMP7 in white adipocytes (Figure 1B). However, body 296 weight of ob/ob mice overproducing BMP7 in the liver was indistinguishable 297 from that of AAV-hAAT-null-treated mice (Figure 6C). In addition, the weight of 298 liver and adipose depots was similar in both groups of animals (Figure 6D). 299 Moreover, white adipocyte size of ob/ob treated with AAV-hAAT-BMP7 vectors 300 remained unchanged (Figure 6E and Figure S2A,B). In agreement, the 301 expression of genes involved in adipogenesis and of markers of mature 302 adipocytes (Pref1, Wnt10b, Cebpβ, Pparγ, Cebpα, Fabp4, Plin1, Slc2a4) was 303 similar between BMP7 and null-treated ob/ob mice in both eWAT and iWAT 304 (Figure 6F and Figure S2C). These results suggested that BMP7-mediated 305 effects in white adipogenesis likely resulted from autocrine/paracrine effects of 306 BMP7 in WAT. 307 According to the lack of hyperplasic expansion of WAT, both adiponectin 308 levels and the degree of WAT inflammation remained unchanged in ob/ob mice 309 overexpressing BMP7 in the liver compared with control mice (Figure 6G-I). 310 In addition, no differences in lipid deposition or in the expression levels of 311 the adipogenic and mature adipocyte markers Pref1, Pparγ, Cebpα, Prdm16, 312 Fabp4 and Plin (Figure S2D,E) were observed in iBAT of AAV-hAAT-BMP7- 313 treated mice. Multilocular adipocytes were neither detected in iWAT of these 314 mice (Supplementary Figure 2B) and the expression of the thermogenic 315 markers Ucp1 and Ppargc1a remained unchanged in iWAT and iBAT (Figure 316 S2F,G). 14 317 Moreover, similar hepatic lipid deposition, TG content or inflammation 318 were detected in AAV-hAAT-BMP7-treated mice compared with null-treated 319 ob/ob mice (Figures 6H and 6J,K and Figure S2H), indicating lack of 320 amelioration of hepatosteatosis. In agreement, animals treated with AAV-hAAT- 321 BMP7 or AAV-null vectors showed similar glycemia and hyperinsulinemia 322 (Figure 6L,M). 323 15 324 4. DISCUSSION 325 326 The results reported in this study suggest that BMP7 is able to induce white 327 adipogenesis in vivo. Enhanced expression of adipogenic genes increased both 328 adiposity and the proportion of small-size adipocytes in subcutaneous and 329 visceral WAT depots of ob/ob mice in which BMP7 was overexpressed. In 330 contrast, when BMP7 was overproduced in the liver, no changes in fat mass, 331 adipocyte size distribution or expression levels of adipogenic markers were 332 observed despite similar BMP7 circulating levels. This was consistent with 333 BMP7-mediated induction of hyperplasic expansion of WAT in an 334 autocrine/paracrine manner. Compared with the low BMP7 circulating levels 335 reached in this study after engineering WAT or liver with AAV-BMP7 vectors 336 (approximately 350 or 550 pg/ml, respectively), short-term treatment with 337 adenoviral vectors overexpressing BMP7 resulted in very high serum 338 concentration of the factor [21,44]. In these studies, serum levels of BMP7 in 339 the first publication ranged 3000-4000 pg/ml [21] and several hundred-fold 340 higher in the second one [44]. Such very high levels of BMP7 decreased fat 341 mass, increased energy expenditure and attenuated hyperglycemia and obesity 342 [21,44], suggesting that BMP7 may elicit different effects depending on the 343 circulating levels. 344 Despite an increase in body weight and fat mass, obese ob/ob mice 345 overexpressing BMP7 in WAT presented reduced hepatic steatosis, WAT and 346 liver inflammation, and increased insulin sensitivity, together with increased 347 proportion of small-size adipocytes. Although the absolute number of new 348 adipocytes was not quantified in SAT and VAT, our results showed an 16 349 increased proportion of smaller adipocytes in BMP7-treated mice together with 350 increased expression of adipogenic markers, suggesting hyperplasia in both 351 WAT depots. These results suggested that BMP7 was able to shift the 352 unhealthy obese phenotype of ob/ob mice towards an improved metabolic 353 phenotype. It has been observed that a subset of obese humans can also 354 develop metabolically healthy obesity [45], with reduced hepatic fat deposition 355 and increased insulin sensitivity despite high BMI [46,47]. 356 Similar to the observations made in ob/ob mice overexpressing BMP7 in 357 WAT, hyperplasic expansion of SAT and improved insulin sensitivity is also 358 observed in several animal models, such as ob/ob mice overexpressing 359 mitoNEET (a mitochondrial membrane protein) in adipose tissue [48]; HFD-fed 360 mice treated with adipogenic cocktails [49], or FGF21 knock-out mice treated 361 with recombinant FGF21 [8]. Moreover, treatment of obese insulin-resistance 362 patients with pioglitazone increases WAT adipogenesis, particularly in SAT [7]. 363 However, our results suggest that not only subcutaneous but also visceral 364 adipose hyperplasia may be responsible for the metabolic benefit induced by 365 BMP7. In agreement, obese mouse models displaying VAT hyperplasia, such 366 as transgenic mice overexpressing GLUT4 in adipose tissue [50], ob/ob mice 367 lacking the liver X receptors α and β (LXRαβ) [51], or transgenic mice 368 overexpressing CIDEA in adipose tissue [52], also show improved glucose 369 homeostasis and reduced hepatic steatosis. 370 Increased circulating adiponectin levels in ob/ob mice overexpressing 371 BMP7 in white adipocytes may also contribute to improve insulin sensitivity. 372 Adiponectin production is closely linked to adipose tissue hyperplasia, as 373 indicated by previous reports [8,53,54]. Transgenic mice overexpressing 17 374 adiponectin in adipose tissue presented increased fat mass, WAT hyperplasia 375 and improved insulin sensitivity [37]. Furthermore, the induction of SAT and/or 376 VAT hyperplasia observed in the previously mentioned animal models led to 377 increased plasma adiponectin levels [8,48,51,52]. Likewise, T2D patients 378 treated with TZD show increased plasma adiponectin concentration [55]. 379 Adiponectin also has insulin-sensitizing and anti-inflammatory properties and is 380 additionally associated with decreased hepatic steatosis [37,56,57]. In humans, 381 adiponectin levels are inversely correlated with development of non-alcoholic 382 fatty liver disease and with the degree of insulin resistance and T2D [58,59]. 383 Treatment of mice or rats fed a high fat diet (HFD) with recombinant adiponectin 384 decreased liver steatosis and increased insulin sensitivity [60,61]. Similarly, 385 muscular or hepatic gene transfer of adiponectin using AAV vectors enhanced 386 insulin sensitivity and reduced inflammation and hepatic lipid deposition in HFD- 387 fed diabetic rats [56,57]. Moreover, decreased inflammation and hepatic lipid 388 deposition were observed in animal models that develop WAT hyperplasia and 389 showed increased adiponectin levels [37,48,51]. All these results suggest that 390 the increased circulating adiponectin levels observed in ob/ob mice 391 overexpressing BMP7 in white adipocytes may play an important role in the 392 amelioration of liver and WAT inflammation, thus contributing to reduce pro- 393 inflammatory immune cells and cytokine production, as well as, liver steatosis. 394 These phenotypic benefits together with increased WAT hyperplasia would in 395 turn improve the insulin sensitivity observed in these ob/ob mice. However, 396 given that the ob/ob model is deficient in leptin, it would be of particular interest 397 to study whether a similar BMP7-mediated metabolic benefit would be also 398 obtained in dietary mouse models of obesity. 18 399 Altogether, our study unravels for the first time a new autocrine/paracrine 400 role of BMP7 on white adipogenesis and highlights that BMP7, when locally 401 expressed in WAT may be a good candidate to modulate adipose tissue 402 plasticity in order to reduce obesity-associated fatty liver and insulin resistance. 403 19 404 AUTHOR CONTRIBUTIONS 405 EC, VJ and FB designed and supervised experiments and analyzed data. EC, 406 VJ, VS, SM, CJ, JR, MG and CM generated reagents and performed 407 experiments. XL produced AAV vectors. EC, VJ, SF, and FB, contributed to 408 discussion, and wrote the manuscript. 409 410 ACKNOWLEDGMENTS 411 This work was supported by grants from Ministerio de Economía y 412 Competitividad (MINECO) and FEDER, Plan Nacional I+D+I (SAF2014-54866R 413 and SAF2017-86266R), and Generalitat de Catalunya (2014 SGR 1669, 2017 414 SGR 1508, ICREA Academia Award to F.B.), Spain, and the European 415 Foundation for the Study of Diabetes (EFSD/MSD European Research 416 Programme on Novel Therapies for Type 2 Diabetes, 2013). V.J. was recipient 417 of a post-doctoral research fellowship from EFSD/Lilly. E.C., V.S., and C.M. 418 received a predoctoral fellowship from Ministerio de Educación, Cultura y 419 Deporte, J.R. from Ministerio de Economía y Competitividad, Spain. The 420 authors thank Marta Moya, Sara Darriba, Tura Ferré, Maria Molas, Jennifer 421 Barrero and Lídia Hernández for technical assistance. 422 423 424 COMPETING INTERESTS 425 The authors declare no competing interests. 426 427 428 Supplementary information is available at International Journal of Obesity’s 429 website. 20 430 REFERENCES 431 1. Wang QA, Tao C, Gupta RK, Scherer PE (2013) Tracking adipogenesis during 432 white adipose tissue development, expansion and regeneration. Nat Med 19: 433 1338–1344. 434 2. Spalding KL, Arner E, Westermark PO, Bernard S, Buchholz BA, Bergmann O, 435 Blomqvist L, Hoffstedt J, Näslund E, Britton T, et al. (2008) Dynamics of fat cell 436 turnover in humans. Nature 453: 783–787. 437 3. Arner P, Andersson DP, Thörne A, Wirén M, Hoffstedt J, Näslund E, Thorell A, 438 Rydén M (2013) Variations in the Size of the Major Omentum Are Primarily 439 Determined by Fat Cell Number. J Clin Endocrinol Metab 98: E897–E901. 440 4. Hammarstedt A, Gogg S, Hedjazifar S, Nerstedt A, Smith U (2018) Impaired 441 adipogenesis and dysfunctional adipose tissue in human hypertrophic obesity. 442 Physiol Rev 98: 1911–1941. 443 5. Acosta JR, Douagi I, Andersson DP, Bäckdahl J, Rydén M, Arner P, 444 Laurencikiene J (2016) Increased fat cell size: a major phenotype of 445 subcutaneous white adipose tissue in non-obese individuals with type 2 446 diabetes. Diabetologia 59: 560–570. 447 6. Hoffstedt J, Arner E, Wahrenberg H, Andersson DP, Qvisth V, Löfgren P, Rydén 448 M, Thörne A, Wirén M, Palmér M, et al. (2010) Regional impact of adipose tissue 449 morphology on the metabolic profile in morbid obesity. Diabetologia 53: 2496– 450 2503. 451 7. McLaughlin TM, Liu T, Yee G, Abbasi F, Lamendola C, Reaven GM, Tsao P, 452 Cushman SW, Sherman A (2010) Pioglitazone increases the proportion of small 453 cells in human abdominal subcutaneous adipose tissue. Obesity (Silver Spring) 454 18: 926–931. 455 8. Li H, Wu G, Fang Q, Zhang M, Hui X, Sheng B, Wu L, Bao Y, Li P, Xu A, et al. 456 (2018) Fibroblast growth factor 21 increases insulin sensitivity through specific 21 457 expansion of subcutaneous fat. Nat Commun 9: 272. 458 9. Fried SK, Lee M-J, Karastergiou K (2015) Shaping fat distribution: New insights 459 into the molecular determinants of depot- and sex-dependent adipose biology. 460 Obesity (Silver Spring) 23: 1345–1352. 461 10. Carobbio S, Pellegrinelli V, Vidal-Puig A (2017) Adipose Tissue Function and 462 Expandability as Determinants of Lipotoxicity and the Metabolic Syndrome. Adv 463 Exp Med Biol 960: 161–196. 464 11. Wang QA, Tao C, Gupta RK, Scherer PE (2013) Tracking adipogenesis during 465 white adipose tissue development, expansion and regeneration. Nat Med 19: 466 1338–1344. 467 12. Kim SM, Lun M, Wang M, Senyo SE, Guillermier C, Patwari P, Steinhauser ML 468 (2014) Loss of white adipose hyperplastic potential is associated with enhanced 469 susceptibility to insulin resistance. Cell Metab 20: 1049–1058. 470 13. Jeffery E, Church CD, Holtrup B, Colman L, Rodeheffer MS (2015) Rapid depot- 471 specific activation of adipocyte precursor cells at the onset of obesity. Nat Cell 472 Biol 17: 376–385. 473 14. Permana PA, Nair S, Lee Y-H, Luczy-Bachman G, Vozarova De Courten B, 474 Tataranni PA (2004) Subcutaneous abdominal preadipocyte differentiation in 475 vitro inversely correlates with central obesity. Am J Physiol Endocrinol Metab 476 286: E958-62. 477 15. Tchoukalova Y, Koutsari C, Jensen M (2007) Committed subcutaneous 478 preadipocytes are reduced in human obesity. Diabetologia 50: 151–157. 479 16. Almuraikhy S, Kafienah W, Bashah M, Diboun I, Jaganjac M, Al-Khelaifi F, 480 Abdesselem H, Mazloum NA, Alsayrafi M, Mohamed-Ali V, et al. (2016) 481 Interleukin-6 induces impairment in human subcutaneous adipogenesis in 482 obesity-associated insulin resistance. Diabetologia 59: 2406–2416. 483 17. Arner P, Arner E, Hammarstedt A, Smith U (2011) Genetic predisposition for 484 type 2 diabetes, but not for overweight/obesity, is associated with a restricted 22 485 adipogenesis. PLoS One 6: e18284. 486 18. Carreira ACO, Zambuzzi WF, Rossi MC, Filho RA, Sogayar MC, Granjeiro JM 487 (2015) Bone Morphogenetic Proteins: Promising Molecules for Bone Healing, 488 Bioengineering, and Regenerative Medicine. In, Vitamins and Hormones pp 489 293–322. 490 19. Schulz TJ, Tseng YH (2009) Emerging role of bone morphogenetic proteins in 491 adipogenesis and energy metabolism. Cytokine Growth Factor Rev 20: 523– 492 531. 493 20. Blázquez-Medela AM, Jumabay M, Rajbhandari P, Sallam T, Guo Y, Yao J, 494 Vergnes L, Reue K, Zhang L, Yao Y, et al. (2019) Noggin depletion in adipocytes 495 promotes obesity in mice. Mol Metab 25: 50–63. 496 21. Tseng Y-H, Kokkotou E, Schulz TJ, Huang TL, Winnay JN, Taniguchi CM, Tran 497 TT, Suzuki R, Espinoza DO, Yamamoto Y, et al. (2008) New role of bone 498 morphogenetic protein 7 in brown adipogenesis and energy expenditure. Nature 499 454: 1000–1004. 500 22. Salisbury EA, Lazard ZW, Ubogu EE, Davis AR, Olmsted-Davis EA (2012) 501 Transient brown adipocyte-like cells derive from peripheral nerve progenitors in 502 response to bone morphogenetic protein 2. Stem Cells Transl Med 1: 874–885. 503 23. Qian S-W, Tang Y, Li X, Liu Y, Zhang Y-Y, Huang H-Y, Xue R-D, Yu H-Y, Guo 504 L, Gao H-D, et al. (2013) BMP4-mediated brown fat-like changes in white 505 adipose tissue alter glucose and energy homeostasis. Proc Natl Acad Sci U S A 506 110: E798-807. 507 24. Kim M, Kim JI, Kim JB, Choe S (2017) The activin-βA/BMP-2 chimera AB204 is 508 a strong stimulator of adipogenesis. J Tissue Eng Regen Med 11: 1524–1531. 509 25. Gustafson B, Hedjazifar S, Gogg S, Hammarstedt A, Smith U (2015) Insulin 510 resistance and impaired adipogenesis. Trends Endocrinol Metab 26: 193–200. 511 26. Elsen M, Raschke S, Tennagels N, Schwahn U, Jelenik T, Roden M, Romacho 512 T, Eckel J (2014) BMP4 and BMP7 induce the white-to-brown transition of 23 513 primary human adipose stem cells. Am J Physiol Cell Physiol 306: C431-40. 514 27. Hinoi E, Nakamura Y, Takada S, Fujita H, Iezaki T, Hashizume S, Takahashi S, 515 Odaka Y, Watanabe T, Yoneda Y (2014) Growth differentiation factor-5 516 promotes brown adipogenesis in systemic energy expenditure. Diabetes 63: 517 162–175. 518 28. Lord E, Bergeron E, Senta H, Park H, Faucheux N (2010) Effect of BMP-9 and 519 its derived peptide on the differentiation of human white preadipocytes. Growth 520 Factors 28: 149–156. 521 29. Pei Z, Yang Y, Kiess W, Sun C, Luo F (2014) Dynamic profile and adipogenic 522 role of growth differentiation factor 5 (GDF5) in the differentiation of 3T3-L1 523 preadipocytes. Arch Biochem Biophys 560: 27–35. 524 30. Kuo MM-C, Kim S, Tseng C-Y, Jeon Y-H, Choe S, Lee DK (2014) BMP-9 as a 525 potent brown adipogenic inducer with anti-obesity capacity. Biomaterials 35: 526 3172–3179. 527 31. Ayuso E, Mingozzi F, Montane J, Leon X, Anguela XM, Haurigot V, Edmonson S 528 a, Africa L, Zhou S, High K a, et al. (2010) High AAV vector purity results in 529 serotype- and tissue-independent enhancement of transduction efficiency. Gene 530 Ther 17: 503–510. 531 32. Jimenez V, Muñoz S, Casana E, Mallol C, Elias I, Jambrina C, Ribera A, Ferre 532 T, Franckhauser S, Bosch F (2013) In vivo AAV-mediated Genetic Engineering 533 of White and Brown Adipose Tissue in Adult Mice. Diabetes 62: 1–12. 534 33. Jimenez V, Jambrina C, Casana E, Sacristan V, Muñoz S, Darriba S, Rodó J, 535 Mallol C, Garcia M, León X, et al. (2018) FGF21 gene therapy as treatment for 536 obesity and insulin resistance. EMBO Mol Med 10:. 537 34. Muñoz S, Franckhauser S, Elias I, Ferré T, Hidalgo A, Monteys AM, Molas M, 538 Cerdán S, Pujol A, Ruberte J, et al. (2010) Chronically increased glucose uptake 539 by adipose tissue leads to lactate production and improved insulin sensitivity 540 rather than obesity in the mouse. Diabetologia 53: 2417–2430. 24 541 35. Carr TP, Andresen CJ, Rudel LL (1993) Enzymatic determination of triglyceride, 542 free cholesterol, and total cholesterol in tissue lipid extracts. Clin Biochem 26: 543 39–42. 544 36. Lagarrigue S, Lopez-Mejia IC, Denechaud P-D, Escoté X, Castillo-Armengol J, 545 Jimenez V, Chavey C, Giralt A, Lai Q, Zhang L, et al. (2016) CDK4 is an 546 essential insulin effector in adipocytes. J Clin Invest 126: 335–348. 547 37. Kim JY, Van De Wall E, Laplante M, Azzara A, Trujillo ME, Hofmann SM, 548 Schraw T, Durand JL, Li H, Li G, et al. (2007) Obesity-associated improvements 549 in metabolic profile through expansion of adipose tissue. J Clin Invest 117: 550 2621–2637. 551 38. Lindström P (2007) The physiology of obese-hyperglycemic mice [ob/ob mice]. 552 ScientificWorldJournal 7: 666–685. 553 39. Gao G-P, Alvira MR, Wang L, Calcedo R, Johnston J, Wilson JM (2002) Novel 554 adeno-associated viruses from rhesus monkeys as vectors for human gene 555 therapy. Proc Natl Acad Sci U S A 99: 11854–11859. 556 40. Zincarelli C, Soltys S, Rengo G, Rabinowitz JE (2008) Analysis of AAV 557 serotypes 1-9 mediated gene expression and tropism in mice after systemic 558 injection. Mol Ther 16: 1073–1080. 559 41. Wang L, Wang H, Bell P, McCarter RJ, He J, Calcedo R, Vandenberghe LH, 560 Morizono H, Batshaw ML, Wilson JM (2010) Systematic Evaluation of AAV 561 Vectors for Liver directed Gene Transfer in Murine Models. Mol Ther 18: 118– 562 125. 563 42. Mallol C, Casana E, Jimenez V, Casellas A, Haurigot V, Jambrina C, Sacristan 564 V, Morró M, Agudo J, Vilà L, et al. (2017) AAV-mediated pancreatic 565 overexpression of Igf1 counteracts progression to autoimmune diabetes in mice. 566 Mol Metab 6: 664–680. 567 43. Schulz TJ, Huang TL, Tran TT, Zhang H, Townsend KL, Shadrach JL, Cerletti 568 M, McDougall LE, Giorgadze N, Tchkonia T, et al. (2011) Identification of 25 569 inducible brown adipocyte progenitors residing in skeletal muscle and white fat. 570 Proc Natl Acad Sci U S A 108: 143–148. 571 44. Townsend KL, Suzuki R, Huang TL, Jing E, Schulz TJ, Lee K, Taniguchi CM, 572 Espinoza DO, McDougall LE, Zhang H, et al. (2012) Bone morphogenetic 573 protein 7 (BMP7) reverses obesity and regulates appetite through a central 574 mTOR pathway. FASEB J 7: 1–10. 575 45. Häring H-U (2016) Novel phenotypes of prediabetes? Diabetologia 59: 1806– 576 1818. 577 46. Stefan N, Kantartzis K, Machann J, Schick F, Thamer C, Rittig K, Balletshofer B, 578 Machicao F, Fritsche A, Häring H-U (2008) Identification and characterization of 579 metabolically benign obesity in humans. Arch Intern Med 168: 1609–1616. 580 47. Stefan N, Häring H-U, Schulze MB (2018) Metabolically healthy obesity: the low- 581 hanging fruit in obesity treatment? lancet Diabetes Endocrinol 6: 249–258. 582 48. Kusminski CM, Holland WL, Sun K, Park J, Spurgin SB, Lin Y, Askew GR, 583 Simcox JA, McClain DA, Li C, et al. (2012) MitoNEET-driven alterations in 584 adipocyte mitochondrial activity reveal a crucial adaptive process that preserves 585 insulin sensitivity in obesity. Nat Med 18: 1539–1549. 586 49. Lu Q, Li M, Zou Y, Cao T (2014) Induction of adipocyte hyperplasia in 587 subcutaneous fat depot alleviated type 2 diabetes symptoms in obese mice. 588 Obesity (Silver Spring) 22: 1623–1631. 589 50. Shepherd PR, Gnudi L, Tozzo E, Yang H, Leach F, Kahn BB (1993) Adipose cell 590 hyperplasia and enhanced glucose disposal in transgenic mice overexpressing 591 GLUT4 selectively in adipose tissue. J Biol Chem 268: 22243–22246. 592 51. Beaven SW, Matveyenko A, Wroblewski K, Chao L, Wilpitz D, Hsu TW, Lentz J, 593 Drew B, Hevener AL, Tontonoz P (2013) Reciprocal Regulation of Hepatic and 594 Adipose Lipogenesis by Liver X Receptors in Obesity and Insulin Resistance. 595 52. Abreu-Vieira G, Fischer AW, Mattsson C, de Jong JMA, Shabalina IG, Rydén M, 596 Laurencikiene J, Arner P, Cannon B, Nedergaard J, et al. (2015) Cidea improves 26 597 the metabolic profile through expansion of adipose tissue. Nat Commun 6: 7433. 598 53. Li P, Song Y, Zan W, Qin L, Han S, Jiang B, Dou H, Shao C, Gong Y (2017) 599 Lack of CUL4B in Adipocytes Promotes PPARγ-Mediated Adipose Tissue 600 Expansion and Insulin Sensitivity. Diabetes 66: 300–313. 601 54. Grünberg JR, Hoffmann JM, Hedjazifar S, Nerstedt A, Jenndahl L, Elvin J, 602 Castellot J, Wei L, Movérare-Skrtic S, Ohlsson C, et al. (2017) Overexpressing 603 the novel autocrine/endocrine adipokine WISP2 induces hyperplasia of the 604 heart, white and brown adipose tissues and prevents insulin resistance. Sci Rep 605 7: 43515. 606 55. Hammarstedt A, Sopasakis VR, Gogg S, Jansson P-A, Smith U (2005) Improved 607 insulin sensitivity and adipose tissue dysregulation after short-term treatment 608 with pioglitazone in non-diabetic, insulin-resistant subjects. Diabetologia 48: 96– 609 104. 610 56. Long W, Hui Ju Z, Fan Z, Jing W, Qiong L (2014) The effect of recombinant 611 adeno-associated virus-adiponectin (rAAV2/1-Acrp30) on glycolipid 612 dysmetabolism and liver morphology in diabetic rats. Gen Comp Endocrinol 206: 613 1–7. 614 57. Shklyaev S, Aslanidi G, Tennant M, Prima V, Kohlbrenner E, Kroutov V, 615 Campbell-Thompson M, Crawford J, Shek EW, Scarpace PJ, et al. (2003) 616 Sustained peripheral expression of transgene adiponectin offsets the 617 development of diet-induced obesity in rats. Proc Natl Acad Sci U S A 100: 618 14217–14222. 619 58. Cnop M, Havel PJ, Utzschneider KM, Carr DB, Sinha MK, Boyko EJ, Retzlaff 620 BM, Knopp RH, Brunzell JD, Kahn SE (2003) Relationship of adiponectin to 621 body fat distribution, insulin sensitivity and plasma lipoproteins: evidence for 622 independent roles of age and sex. Diabetologia 46: 459–469. 623 59. Ma H, Gomez V, Lu L, Yang X, Wu X, Xiao SY (2009) Expression of adiponectin 624 and its receptors in livers of morbidly obese patients with non-alcoholic fatty liver 27 625 disease. J Gastroenterol Hepatol 24: 233–237. 626 60. Xu A, Wang Y, Keshaw H, Xu LY, Lam KSL, Cooper GJS (2003) The fat-derived 627 hormone adiponectin alleviates alcoholic and nonalcoholic fatty liver diseases in 628 mice. J Clin Invest 112: 91–100. 629 61. Ma H, Cui F, Dong J-J, You G-P, Yang X-J, Lu H-D, Huang Y-L (2014) 630 Therapeutic effects of globular adiponectin in diabetic rats with nonalcoholic fatty 631 liver disease. World J Gastroenterol 20: 14950–14957. 632 28 633 FIGURE LEGENDS 634 635 Figure 1. BMP7 increases fat mass and reduces white adipocyte size. 636 Ob/ob mice were administered intra-eWAT with 1x1012 vg/mouse of AAV-BMP7 637 vectors at 11 weeks of age. Control ob/ob mice received 1x1012 vg of AAV-null 638 vectors. (A) AAV-derived BMP7 expression in epididymal (eWAT), 639 retroperitoneal (rWAT), mesenteric (mWAT) and inguinal (iWAT) white adipose 640 depots, interscapular BAT (iBAT) and the liver 3 months after AAV-treatment. 641 The qPCR was performed with primers that specifically detected the murine 642 optimized-BMP7 (moBMP7) coding sequence. n=7-10. (B) Circulating levels of 643 BMP7 3 months after vector administration. n=8-10. (C) Evolution of body 644 weight in animals treated with AAV-BMP7 or AAV-null vectors. n=8-10. (D) 645 Weight of several WAT and BAT depots and the liver in the same cohorts of 646 mice as in (C). n=8-10. (E) Representative images of the hematoxylin-eosin 647 staining of eWAT and iWAT sections. Scale bars: 100 μm. (F) Mean area of 648 white adipocytes in eWAT and iWAT. n=8-10 (eWAT) and n=4 (iWAT). All 649 values are expressed as mean±SEM. ND, non-detected. AU, arbitrary units. 650 **P<0.01 and ***P<0.001 versus the AAV-Null-treated group. 651 652 Figure 2. BMP7 induces white adipogenesis and decreases WAT 653 inflammation. (A,B) Expression of adipogenic markers in eWAT (A) and iWAT 654 (B). (C,D) Expression of markers of mature adipocytes in eWAT (C) and iWAT 655 (D). (E) Serum concentration of adiponectin. (F) Immunohistochemistry for the 656 macrophage-specific marker Mac2 in eWAT sections. Scale bars: 100 μm. (G) 657 Quantification by qRT-PCR of the expression levels of the macrophage markers 29 658 Cd68 and F4/80 in eWAT. (H,I) Expression of the pro-inflammatory markers 659 Mcp1(H) and Tnfα (I) in eWAT. All values are expressed as mean±SEM. n=8- 660 10. **P<0.01 and ***P<0.001 versus the AAV-null-injected group. 661 662 Figure 3. WAT-derived BMP7 does neither induce brown adipogenesis nor 663 enhance non-shivering thermogenesis. (A) Representative images of the 664 hematoxylin-eosin staining of iBAT sections of ob/ob mice treated intra-eWAT 665 with AAV-BMP7 or AAV-null vectors. Scale bars: 100 μm. (B,C) Quantification 666 by qRT-PCR of the expression levels of markers of adipogenesis (B) and of 667 mature adipocytes (C) in iBAT. n=7-10. (D,E) Quantification by qRT-PCR of the 668 expression of the thermogenic markers Ucp1 and Ppargc1a in iBAT (D) and 669 iWAT (E). n=7-9 (F) Energy expenditure was measured with an indirect open 670 circuit calorimeter 6 weeks after AAV vector delivery. n=8-9. Data were taken 671 during the light and dark cycles. All values are expressed as mean±SEM. 672 673 Figure 4. BMP7 ameliorates hepatic steatosis. (A) Representative images of 674 the hematoxylin-eosin staining of liver sections of ob/ob mice treated intra- 675 eWAT with AAV-BMP7 or AAV-null vectors. Scale bars: 200 μm and 50 μm 676 (inset). (B) Fed hepatic triglyceride content. n=8-10. (C) Immunostaining for the 677 macrophage-specific marker Mac-2 in liver sections. Red arrowheads indicate 678 Mac2+ cells. Scale bars: 100 μm and 50 μm (inset). (D) Quantification by qRT- 679 PCR of the expression of the macrophage marker Cd68 in the liver. n=7-10. All 680 values are expressed as mean±SEM. **P<0.01 and ***P<0.001 versus the 681 AAV-null injected group. 30 682 Figure 5. BMP7 improves insulin sensitivity. (A) Fed blood glucose levels 3 683 months after vector administration. n=8-10. (B) Fed serum insulin levels 3 684 months after vector delivery. n=7-10. (C) Insulin tolerance test was performed 685 after an intraperitoneal injection of insulin (0.75 units/kg body weight) 2 months 686 post-AAV delivery. n=7-9. Results were calculated as the percentage of initial 687 blood glucose levels. All values are expressed as mean±SEM. *P<0.05 and 688 **P<0.01 versus the AAV-null injected group. 689 690 Figure 6. Liver-derived BMP7 does not induce white adipogenesis. Ob/ob 691 mice were administered intravenously with 5x1011 vg/mouse of AAV-hAAT- 692 BMP7 vectors at 11 weeks of age. Control ob/ob mice received 5x1011 vg of 693 AAV-hAAT-null vectors. (A) AAV-derived BMP7 expression in the liver, eWAT 694 and iWAT depots 5 months after AAV administration. Analysis by qPCR was 695 performed with primers that specifically detected the murine optimized-BMP7 696 (moBMP7) coding sequence. n=8-9. (B) BMP7 serum circulating levels 5 697 months after vector administration. n=7-9. (C) Body weight of animals treated 698 with either AAV-hAAT-BMP7 or AAV-hAAT-null vectors. n=8-9. (D) Weight of 699 the liver, epididymal (eWAT), inguinal (iWAT), retroperitoneal (rWAT) and 700 mesenteric (mWAT) white adipose tissue depots and interscapular brown 701 adipose tissue (iBAT). n=8-9. (E) Mean area of white adipocytes in eWAT and 702 iWAT. n=5. (F) Quantification by qRT-PCR of the expression levels of markers 703 of adipogenesis and of mature adipocytes in eWAT. n=7-9. (G) Serum levels of 704 adiponectin. n=8-9. (H) Immunohistochemistry for the macrophage-specific 705 marker Mac2 in sections of eWAT and liver. Scale bars: 100 μm and 50 μm 706 (inset). (I,J) Quantification by qRT-PCR of the expression of the macrophage 31 707 markers Cd68 and F4/80 in eWAT (I) and in the liver (J). n=7-9. (K) Fed hepatic 708 triglyceride content. n=8-9. (L,M) Fed blood glucose (L) and insulin (M) levels 5 709 months after vector administration. n=8-9. ND, non-detected. AU, arbitrary units. 710 All values are expressed as mean±SEM. *P<0.05 versus the AAV-hAAT-null- 711 injected group. 712 32 713 SUPPLEMENTAL FIGURE LEGENDS 714 715 Figure S1. Redistribution of the size of white adipocytes by WAT-derived 716 BMP7. (A) Histogram depicting the food intake of ob/ob mice treated with 717 1x1012 vg/mouse of AAV-BMP7 or AAV-null vectors. n=8-10. (B) Activity was 718 measured 6 weeks after AAV vector delivery. n=8-9. Data were taken during the 719 light and dark cycles. (C) Frequency distribution of white adipocyte area in 720 eWAT of the same groups as in (A). n=8-10. All values are expressed as 721 mean±SEM. 722 723 Figure S2. Liver-derived BMP7 failed to induce adipogenesis and did not 724 decrease hepatic steatosis. (A,B) Representative images of the hematoxylin- 725 eosin staining of eWAT (A) and iWAT (B) sections of ob/ob mice treated 726 intravenously with 5x1011 vg/mouse of AAV-hAAT-BMP7 or AAV-hAAT-null 727 vectors. Scale bars: 100 μm. (C) Expression levels of adipogenic markers in 728 iWAT. n=7-9. (D) Representative images of hematoxylin-eosin staining of iBAT 729 sections. Scale bars: 100 μm. (E) Expression of markers of adipogenesis and of 730 mature adipocyte markers in iBAT. n=7-9. (F,G) Expression of markers of 731 thermogenesis in iWAT (F) and iBAT (G). n=7-9. (H) Representative images of 732 hematoxylin-eosin staining of liver sections. Scale bars: 200 μm and 50 μm 733 (inset). All values are expressed as mean±SEM. 33 FIGURE 1 A B C moBMP7 expression (AU) 8 AAV-null 600 80 *** 7 ****** *** Body Weight (g) 70 6 AAV-BMP7 500 ****** 60 ***** 5 400 * 4 300 50 3 200 40 2 AAV-null 100 30 AAV-BMP7 1 ND ND ND ND ND ND ND 20 0 0 eWAT rWAT mWAT iWAT iBAT Liver 11 13 15 17 19 21 23 25 Age (weeks) AAV D *** *** E AAV-null AAV-BMP7 Tissue weight (g) *** *** eWAT eWAT iWAT rWAT mWAT iBAT Liver F Mean adipocyte area iWAT (µm2) *** ** eWAT iWAT FIGURE 2 eWAT eWAT A Adipogenesis markers C Mature adipocyte markers 25 *** AAV-null 20 7 ** Relative expression Relative expression AAV-BMP7 15 6 *** 7 *** 5 6 *** ** 5 4 4 3 3 *** ** 2 *** 2 *** 1 1 0 0 Pref1 Wnt10b Cebpβ Pparγ Cebpα Srebf1 Fasn Acc1 Acc2 Fabp4 Plin1 Slc2a4 iWAT iWAT B Adipogenesis markers D Mature adipocyte markers 5 * 5 *** Relative expression Relative expression 4 4 3 3 *** ** *** 2 2 ** ** ** ** 1 1 0 0 Pref1 Wnt10b Cebpβ Pparγ Cebpα Srebf1 Fasn Acc1 Acc2 Fabp4 Plin1 Slc2a4 E F AAV-null AAV-BMP7 *** Adiponectin (µg/mL) eWAT G 1.2 H 0.40 I p=0.056 0.14 Relative expression Relative expression Relative expression 1.0 0.12 0.30 0.10 0.8 ** 0.08 0.6 0.20 ** *** 0.06 0.4 0.10 0.04 0.2 0.02 0 0 0 Cd68 F4/80 Mcp1 Tnfα FIGURE 3 A AAV-null AAV-BMP7 B Adipogenesis markers C Mature adipocytes markers 2.0 AAV-null 1.5 Relative expression AAV-BMP7 Relative expression 1.5 1.0 1.0 0.5 0.5 0 0 Pref1 Wnt10b Cebpβ Pparγ Cebpα Prdm16 Fabp4 Plin1 Slc2a4 D iBAT E iWAT F 1.5 1.5 Energy expenditure Relative expression Relative expression (kcal/day/kg0.75) 1.0 1.0 0.5 0.5 0 0 Ucp1 Ppargc1a Ucp1 Ppargc1a Day Night FIGURE 4 AAV-null A AAV-null B AAV-BMP7 AAV-BMP7 160 140 120 *** 100 80 60 40 20 0 C AAV-null AAV-BMP7 D 1.50 Relative expression 1.25 1.00 0.75 ** 0.50 0.25 0 Cd68 FIGURE 5 AAV-null AAV-BMP7 A B C 200 80 150 Blood glucose (% of initial) Blood glucose (mg/dL) 175 70 ** 150 60 125 125 50 100 40 100 * * 75 30 50 20 75 AAV-null 25 10 AAV-BMP7 0 0 50 0 15 30 45 60 75 90 Time (minutes) FIGURE 6 A B C moBMP7 expression (AU) 70 0.14 AAV-hAAT-null Body weight (g) 800 65 0.12 AAV-hAAT-BMP7 700 0.10 60 600 0.08 500 55 0.06 400 50 300 AAV-hAAT-null 0.04 45 200 AAV-hAAT-BMP7 0.02 40 100 0 ND ND ND ND 11 15 19 23 27 31 0 Liver eWAT iWAT Age (weeks) AAV D E G 8000 Tissue weight (g) Adiponectin (µg/mL) 7000 6000 5000 4000 3000 2000 1000 Liver eWAT iWAT rWAT mWAT iBAT 0 eWAT iWAT F Adipogenesis markers Mature adipocytes 1.50 markers H AAV-hAAT-null AAV-hAAT-BMP7 Relative expression 1.25 1.00 0.75 eWAT 0.50 0.25 0 Pref1 Cebpβ Pparγ Cebpα Fabp4 Plin1 Slc2a4 I 1.2 eWAT J 1.50 Liver Relative expression Relative expression 1.0 1.25 * 0.8 1.00 Liver 0.6 0.75 0.4 0.50 0.2 0.25 0 0 Cd68 F4/80 Cd68 K L M 180 300 160 250 140 120 200 100 150 80 60 100 40 50 20 0 0