Nitric Oxide 35 (2013) 186–192

Contents lists available at ScienceDirect

Nitric Oxide

journal homepage: www.elsevier .com/locate /yniox

Review

Phosphodiesterase 5 as target for adipose tissue disorders

1089-8603/$ - see front matter � 2013 Elsevier Inc. All rights reserved.http://dx.doi.org/10.1016/j.niox.2013.10.006

⇑ Corresponding author. Address: Federal University of Santa Catarina, Rua PastorWilliam Richard Schisler Filho, 900/711, Itacorubi, Florianópolis, Santa Catarina, CEP88034100, Brazil. Fax: +55 48 32229357.

E-mail addresses: gcolomboufsc@gmail.com (G. Colombo), mdanielahemosc@gmail.com (Maria Daniela H. Périco Colombo), leo-jf@uol.com.br (L.D.L. Schiavon),dacampora@gmail.com (A.J. d’Acampora).

Giovani Colombo a,⇑, Maria Daniela H. Périco Colombo b, Leonardo De Lucca Schiavon a,Armando José d’Acampora a

a Federal University of Santa Catarina, Brazilb Hematology and Transfusion Medicine Center of Santa Catarina (HEMOSC), Brazil

a r t i c l e i n f o

Article history:Received 5 May 2013Revised 26 September 2013Available online 28 October 2013

Keywords:AdipocyteAdipogenesisNitric oxidePhosphodiesterase 5Phosphodiesterase 5 inhibitorWhite adipose tissue

a b s t r a c t

Introduction: Adipose tissue as an endocrine organ is responsible for the release of multiple cytokines,which have the most diverse metabolic functions. Therefore, it is extremely important to preserve itsphysiological health in order to avoid local and systemic disorders. Experiments available in literatureshow the importance of the nitric oxide (NO)/guanosine 3050 cyclic monophosphate (cGMP)/proteinkinase G (PKG) pathway in adipocyte biology. Phosphodiesterase 5 (PDE5) is an enzyme responsiblefor cGMP inactivation, and the use of its inhibitors can be an alternative in the search of a more balancedadipose tissue.Objective: This review aims to describe the PDE5 role and the possibility of using PDE5 inhibitors in adi-pocyte physiology derangements and their consequences.Design and methods: Studies published in the last 10 years that related PDE5 and its inhibitors to adiposetissue were raised in major databases.Results: PDE5 is present in adipocyte, and PDE5 inhibitors can promote adipogenesis, interfere with adi-pokines secretion, decrease inflammatory markers expression, and increase the thermogenic potential ofwhite adipose tissue.Conclusions: PDE5 plays an important role in adipocyte physiology and the use of its inhibitors may provea useful tool to combat adipose tissue disorders and its highest expression, metabolic syndrome.

� 2013 Elsevier Inc. All rights reserved.

Contents

Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186Cyclic nucleotide phosphodiesterases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187Phosphodiesterase 5 (PDE5) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187Phosphodiesterase inhibitors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187Adipose tissue, adipokines, and inflammation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188NO/cGMP/PKG pathway and the adipocyte physiology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188NO pathway, PDE5 inhibitors and adipose tissue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191Conflicts of interest . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191

Introduction

While hunger was a constant presence during human evolution,nowadays 65% of the world’s population live in countries whereoverweight and obesity kill more people than underweight [1,2].

Over the last years adipose tissue rose from a simple energydeposit to an endocrine organ, producing hormones known asadipokines, and also cytokines related to the onset and perpetua-tion of inflammation [3–5]. On account of its role in metaboliccontrol, adipose tissue expansion leads to alteration of its normal

G. Colombo et al. / Nitric Oxide 35 (2013) 186–192 187

physiology, generating autocrine, paracrine, and finally systemicimbalances [6,7]. The dysfunction of these adipokines and cyto-kines is responsible for decreased systemic cellular sensitivity toinsulin, and consequently the onset of metabolic syndrome [6,7].

Nitric oxide (NO), a small gaseous molecule, is composed of onenitrogen atom and one oxygen atom, and has a half-life of severalseconds [8]. NO is produced by a family of NO synthases (NOS) andis released from many cell types in the body, where it acts either asa neurotransmitter or as a paracrine agent [8]. NO binds to solubleguanylyl cyclase (sGC), and causes a 100- to 200-fold activation ofthe enzyme [8]. Activation of sGC increases conversion of guano-sine-50-triphosphate (GTP) to guanosine 3050 cyclic monophosphate(cGMP), resulting in the elevation of cGMP, which initiates thecGMP-signaling pathway and subsequent physiological changes[8]. These changes are largely mediated through the activation ofcGMP dependent protein kinase (PKG) [8]. Some studies availablein the literature suggest the involvement of the NO/cGMP/PKGpathway in various aspects of adipose tissue physiology, and con-sequently a link with metabolic syndrome [9–15]. In fact, endothe-lial nitric oxide synthase (eNOS) deficient mice develop symptomsof human metabolic syndrome, with insulin resistance, hyperlipid-emia and hypertension [16,17]. Another way to increase the levelsof cGMP is through the atrial natriuretic peptide (ANP) action. Thisoccurs in a dose-dependent manner, and is mediated by the natri-uretic receptor-A (NPR-A). Thus, adipose tissue can be considered atarget for the natriuretic peptide, where adipocyte differentiationtakes place via NPR-A/cGMP/PKG [18–20].

Cyclic nucleotide phosphodiesterase 5 (PDE5) is the enzymeresponsible for cGMP hydrolysis to generate the product 50-GMP,which is inactive in the cyclic nucleotide (cN) pathway [21,22]. Sil-denafil, a PDE5 inhibitor, initially designated as UK-92,480, wasfirst synthesized by Pfizer in United Kingdom to treat hypertensionand angina pectoris, but interestingly, exhibited a different phar-macological effect, a marked penile erection, and became thefirst-line treatment option to erectile dysfunction (ED) [23–25]. Apopulation favored by the appearance of these drugs are diabetics,who frequently have dysfunction of the physiological mechanismof penile erection [24]. Initially cautious, physicians from variousspecialties gradually felt more comfortable prescribing this classof drugs to their diabetic patients, even with multiple comorbidi-ties [26,27]. Paradoxically, some researchers are evaluating the ef-fect of these drugs in disorders related to the genesis of metabolicsyndrome, such as adipose tissue inflammation, insulin resistance,atherosclerosis, and finally cardiovascular mortality [9–11].

This review aims to describe and analyze studies related to therole of PDE5 and the use of PDE5 inhibitors in adipose tissue disor-ders, and its consequences.

Cyclic nucleotide phosphodiesterases

� Phosphodiesterase (PDE) definition and functionThe PDEs are enzymes that selectively catalyze the hydrolysis of

30 cyclic phosphate bonds of adenosine (cAMP) and/or guanosine3050 cyclic monophosphate (cGMP) [21,22]. Cellular levels of cAMPand cGMP are regulated by the relative activities of adenylyl andguanylyl cyclases (AC and GC), which synthesize these cyclic nucle-otides, and by PDEs, which hydrolyze them [21]. Besides the PDEaction, which constitutes the main form for lowering active cyclicnucleotide (cN) levels in cells, other processes can do the same.Such examples are the extrusion of cN into the extracellular milieu,and/or transit of cN from one cell type to another [21].

� PDE family

Eleven families of mammalian PDEs are derived from 21 genesand classified based on amino acid sequences, regulatory properties,

and catalytic characteristics [21]. PDEs share a conserved catalyticdomain (C domain), but amino acid sequence outside this region dif-fers markedly [21]. Based on their sequence relatedness, kinetics,modes of regulation, and pharmacological properties is that theclass I PDEs (protozoa/metazoan) is divided into 11 families(PDE1-PDE11) [28]. Certain PDEs are highly specific for hydrolysisof cAMP (PDEs 4, 7, and 8) or cGMP (PDEs 5, 6, and 9), and othershydrolyze both (PDEs 1, 2, 3, 10, and 11) [21].

� PDE localization

One single cell type can express several different PDEs, and thenature and localization of these PDEs are likely to be a major reg-ulators of the local concentration of cAMP or cGMP in the cell [22].PDEs 1, 2, 3, and 4 are expressed in many tissues, whereas othersare more restricted [21]. In most cells, PDE 3 and 4 provide the ma-jor portion of cAMP-hydrolyzing activity [21]. PDE5 is abundant invascular and airway smooth muscle and platelets, but also is pres-ent in cerebellar Purkinje cells, gastrointestinal epithelial cells, andendothelial cells [21]. The PDE6 family is found primarily in retina,but also in pineal and certain melanoma cells, while PDEs 7, 8, 9,10, and 11 are not widely expressed [21].

Phosphodiesterase 5 (PDE5)

PDE 5 was originally identified, isolated, and characterized fromplatelets in 1978, and received little notoriety, until it was discov-ered to be a regulator of vascular smooth muscle contraction andtarget for the drug sildenafil [22,29]. Only one PDE 5 gene has beendiscovered, PDE5A, although several variants under the control ofdifferentially regulated promoters have been identified [22].PDE5 is composed of an amino-terminal regulatory domain (GAFA and B) and a carboxy-terminal metal-binding catalytic domain[30]. The structural basis for the high-affinity cGMP binding toPDE5 takes place because of the presence of these two highlyhomologous GAF domains [21,22,31]. In PDE5 high-affinity cGMPbinding occurs only to the GAF-A domain [22]. GAF-B is autoinhib-itory for cGMP binding to GAF-A, and consequently for activationof the catalytic site [21]. Under many physiological conditions, itis thought that GAF-A domain is likely to be occupied by cGMPand therefore fully active [22]. To compensate for that, there areshort-term regulatory mechanisms in PDE5 catalytic site function,like allosteric cGMP binding at GAF-A and phosphorylation at Ser-102, which increase Vmax and affinity for cGMP substrate or inhib-itors [21]. Phosphorylation that occurs in intact cells is mediatedby PKG, and facilitated by allosteric cGMP binding or ligand occu-pation of the catalytic site [21]. Therefore the product of GC andsubstrate for PDE5, cGMP, acts as a feed-forward activator of theenzyme [22]. Another form of PDE 5 regulation is the control ofits expression in tissues in response to a variety of stimuli, andinterestingly overexpression is a likely culprit in many vascularmaladies, including hypertension, angina, diabetic angiopathy,and ED [21].

Phosphodiesterase inhibitors

PDE inhibitors are drugs able to block phosphodiester hydroly-sis caused by PDEs, resulting in higher levels of cyclic nucleotides[32]. For inhibitor binding to PDEs there is a common conservedscheme that is important for selectivity toward individual PDEfamily members [32]. The first conserved feature is that the planarring portion of the inhibitor is held tightly by a P clamp, formed byhighly conserved hydrophobic residues, and the second is that theinhibitor always forms one or two hydrogen bonds with the pur-ine-selective glutamine [32].

188 G. Colombo et al. / Nitric Oxide 35 (2013) 186–192

Zaprinast was the first orally active PDE5 inhibitor available forhuman use [29]. It was originally administered to patients withexercise-induced asthma, and was shown to have moderate bron-chodilator effects [29]. As part of an investigation of NO pathwayregulating the relaxation of the corpus cavernosum in the penis,it was demonstrated that the PDE 5 inhibitor zaprinast could en-hance NO-induced relaxation of isolated corpus cavernosum [29].These findings suggested that PDE 5 inhibitors might be usefulfor treating impotence [29]. In 1998 the introduction of sildenafil,another PDE 5 inhibitor, revolutionized the treatment of ED, andtwo new drugs, vardenafil and tadalafil later expanded the numberof available options for the treatment of ED [24,29].

Besides the well-established use of PDE5 inhibitors in ED andpulmonary hypertension, the strong safety profile of the threePDE 5 inhibitors (sildenafil, vardenafil and tadalafil) has encour-aged investigators to test the efficacy of these drugs in treatmentof a number of ailments [21,33]. In some instances, effects maybe related to improved vascular health, but that does not appearto explain all of the benefits of these drugs [21,33] (Table 1).

Adipose tissue, adipokines, and inflammation

Over the last two decades research established the concept thatadipose tissue is an endocrine organ that releases factors known asadipokines [3–5]. The adipokines includes proteins involved in li-pid metabolism, insulin sensitivity, the complement system, bloodpressure and angiogenesis (leptin, adiponectin, resistin, visfatin,apelin. . .); a number of proteins involved in inflammation (tumornecrosis factor alfa (TNFa), interleukin (IL) IL-1b, IL-6, IL-8, IL-10,transforming growth factor-b, nerve growth factor); and acutephase response protein (plasminogen activator inhibitor-1 (PAI-1), haptoglobin, serum amyloid A) [7,34]. The list of molecules con-firmed as adipokines grows every year, as both novel and existingmolecules are seen to be secreted by adipose tissue or adipocytes[34,35]. However, there is still a major challenge to characterizethe function, mode of action and molecular targets for the growinglist of newly identified adipokines [34,35]. It is apparent that someadipokines have beneficial effects, whereas others have detrimen-tal properties [34].

Among the adipokines, adiponectin and leptin are well knownto have important roles in the lipid and glucose metabolisms aswell as inflammatory regulation [36]. Obesity is associated withincreased leptin serum concentration, but with no significant effecton appetite and body weight, which has been attributed to the

Table 1Possible applications for PDE5 inhibitors.

Neurological indications CognitionAntidepressive effectsCircadian adaptation

Muscular dystrophy StrokeNeonatal pulmonary hypertensionCardiovascular functions Atheroprotection

Improve endothelial healthAngiogenesis

Raynaud diseaseAutoimmune diseasesCystic fibrosisGenitourinary tract dysfunctions Benign prostate syndrome

Overactive bladderUrge incontinencePeyronie’s diseasePremature ejaculationUreteral relaxation for kidney stones

Anticancer agentsMetabolic syndrome, obesity?

central resistance to the hormone effects [35,37]. It is now gener-ally accepted, despite conflicting results, that leptin has a potentinhibitory effect on insulin secretion from pancreatic b-cellsin vitro and in vivo, and has the additional effect of reducing pre-proinsulin gene expression [34]. Adiponectin is the most abundantprotein secreted by adipose tissue, and directly sensitizes the bodyto insulin [38,39]. The promoter of the adiponectin gene containsmultiple binding sites for transcription factors that modulate itsactivity [40]. Among these factors we can highlight the activatedreceptor gamma activation of peroxisome proliferator-activatedreceptor gamma (PPAR-c) [40]. The low basal levels of adiponectinare strong predictors of future onset of diabetes mellitus and cor-onary disease in humans [41–43].

Adipose tissue contains bone marrow–derived macrophages,and the content of these macrophages tracks with the degree ofobesity [44]. In a link between adipokines and inflammation, themacrophage migration inhibitory factor (MMIF) is a cytokinewhich is high in obesity and insulin resistance, and is involved inmacrophage infiltration [45]. The inverse association of adiponec-tin levels, derived from visceral adipose tissue, and MMIF in obes-ity suggests that inflammation might be responsible for the lowlevels of this adipokine [45]. Activation of these tissue macro-phages leads to the release of a variety of chemokines, which inturn recruit additional macrophages, setting up a feed-forwardprocess that further increases adipose tissue macrophages (ATM)content and propagates the chronic inflammatory state [44]. TNFa,IL-6, IL-1b, and macrophage-secreted factors exert paracrine effectsto activate inflammatory pathways within insulin target cells[6,44]. Obesity also results in an increased flux of free fatty acidsinto the circulation, and subsequent uptake by the myocyte, hepa-tocyte or adipose tissue [7]. When the flux of fatty acid exceeds thecapacity of cellular metabolism it accumulates, as well as itsmetabolites (e.g., linoleic acid, diacyl glycerol (DAG), phosphatidicacid (PA), lysophosphatidic acid (LPA), ceramide) [7]. In turn, thefatty acid and its intermediates can activate a number of differentserine kinases that negatively affect insulin action [7]. Phosphory-lated protein kinase C theta (PKCH) starts a downstream activationof other two serine kinase, the c-JUN NH2-terminal kinase (JNK)and the inhibitor kB kinase (IKK). JNK and IKK associate with insu-lin receptor substrate (IRS-1), promoting its serine-phosphoryla-tion [7]. The serine phosphorylation is responsible for IRS-1blocking [7]. The insulin resistance occurs through interruptionof insulin receptor (IR)/IRS interaction and by promotion of IRS-1protein degradation [7,44].

NO/cGMP/PKG pathway and the adipocyte physiology

In target cells, NO interacts with the heme group of sGC, leadingto cGMP synthesis [8,46]. cAMP and cGMP are ubiquitous secondmessengers that mediate biological responses to a variety of extra-cellular cues, including hormones, neurotransmitters, chemokines,and cytokines [8,47]. The increased concentration of these cyclicnucleotides activates PKA and PKG, which phosphorylate a varietyof substrates, including transcription factors and ion channels[8,31,47]. Phosphorylation of these substrates regulates myriadphysiological processes, such as immune responses, cardiac andsmooth muscle contraction, visual response, glycogenolysis, plate-let aggregation, ion channel conductance, apoptosis, and growthcontrol [8,31,47]. To highlight the importance of PKG also in adi-pose tissue, PKGI-deficient (PKGI�/�) brown adipose tissue (BAT)cells show significant reductions in mitochondrial biogenesis.PKGI�/� cells show lower levels of PPARc coactivator-1a (PGC-1a) and uncoupling protein 1 (UCP-1), as well as severe defectsin BAT development and function reflected by inhibition of thethermogenic program [48]. Loss of PKGI stimulates RAS homolog

G. Colombo et al. / Nitric Oxide 35 (2013) 186–192 189

gene family member A (RhoA) and RhoA/Rho-associated proteinkinase (ROCK) activity leading to inhibition of insulin signalingby increasing serine phosphorylation of insulin receptor substrate1 (IRS-1) and consequently insulin resistance [48].

As mentioned earlier natriuretic peptides (NPs) can also stimu-late GC generating cGMP, to activate PKG [20]. It’s the heart thatacts as a regulator of the adipose tissue biology through the secre-tion of its peptide. In fact, in human and mouse adipocytes ANPand ventricular natriuretic peptide (BNP) lead to the same increasein PGC-1a and UCP1 expression, mitochondrial content, anduncoupled respiration as occurs in response to b-agonists, all in ap38 mitogen-activated protein kinase manner [20]. One of the sub-strates of PKG is the vasodilator-stimulated phosphoprotein (VASP)[46]. The VASP phosphorylation not only is a marker of PKG activa-tion, but also mediates relevant biological actions of the NO/cGMP/PKG pathway, such as the regulation of tone/vascular remodeling,influencing relevant steps of atherogenesis [46]. Although VASPsignaling is required for NO/cGMP antiinflammatory effects inadipose tissue [9], the loss of VASP in mice was associated withincreased levels of cGMP and development of brown-like adipocyte(‘‘browning’’) in WAT [49]. Increased ‘‘browning’’ of white adiposetissue could lead to increased energy expenditure and weight loss[50]. Another PKG substrate is RhoA [14]. The PKG-activated bycGMP enhances phosphorylation of RhoA at serine-188, whichresults in the inhibition of the ROCK pathway, leading to antihy-pertrophic and anti-inflammatory effects in white adipose tissue[14,51] (Fig. 1).

NO pathway, PDE5 inhibitors and adipose tissue

There are sparse studies available in the literature that reportthe role of PDE5 in the adipocyte physiology.

Fig. 1. Nitric oxide (NO)/guanosine 3050 cyclic monophosphate (cGMP)/protein kinase Garginine as the substrate (a). NO activates soluble guanylyl ciclase (sGC) (b). sGC activnucleotide phosphodiesterase 5 (PDE5) can hydrolyse cGMP to an inactive form 50-GMP(e). PKG phosphorylates a variety of substrates, including transcription factors and ionsvarious transcription factors: peroxisome proliferator-activated receptor gamma (PPAR-cmetabolism; CCAAT/enhancer-binding protein (CEBPa) terminal adipocyte differentiatrafficking of fatty acids, marker of terminal cell differentiation; fatty acid synthetase (Funcoupling protein 1 (UCP-1) mitochondrial proton leak, generates heat instead of ATP;and differentiation, sensitizes the body to insulin; Glucose transporter type 4 (GLUT4) ininto estrogens (E2); proinflammatory cytokines (PC), PKG decreases expression of chemoamotif) ligand 7 (CCL7), interleukin 6 (IL-6), and tumor necrosis factor a (TNFa) (g). Atriallevels and can also activate PKG (h). Inhibition of the RAS homolog gene family member Aanti-inflammatory effects in white fat (i). ⁄PDE5 site of action.

In 2002, an experimental study suggested that exogenous NOcould stimulate differentiation in rat white preadipocytes, as evi-denced by increased lipoprotein lipase (LPL), and glycerol-3-phos-phate dehydrogenase (GPDH) activities, as well as acceleratedtriacylglycerol (TG) accumulation [52]. Through the use of aninhibitor of nitric oxide synthase (NOS), it was clear that even un-der physiological conditions, endogenous NO was also positivelyinvolved in modulating adipocyte differentiation [52]. ThereforeNO, as a highly reactive and diffusible free radical gas that medi-ates autocrine/paracrine actions involved in cell proliferation anddifferentiation, was also linked to adipose tissue mass [52].

Another study, in turn, investigated whether stimulation of ni-tric oxide-cGMP signaling, by NOS substrate and PDE5 inhibitor,could improve the energetic balance and sensitivity to insulin[50]. Mice fed a high-fat diet (HF) for 12 weeks, were subjectedto three types of treatment: L-arginine plus sildenafil or sildenafilalone for 12 weeks; and thirdly L-arginine plus sildenafil foronly 5 and 18 h [50]. The chronic stimulation with sildenafil plusL-arginine or sildenafil alone resulted in increased energy expendi-ture/insulin sensitivity. Acute treatment with sildenafil plusL-arginine did not increase the insulin sensitivity [50]. This studyshows that the PDE5 is a potential target for dearrangementsdiet-induced, like energy imbalance and insulin resistance, andthat the effect of sildenafil is different depending on the time ofexposure to the drug, whether acute or chronic [50].

Study by Zhang et al. assessed the influence of increasing cGMPlevels, by different PDE5 inhibitors administration, in the adipocytemetabolism and glucose uptake mediated by insulin. Sildenafil,tadalafil and vardenafil were able to promote adipogenesis in3T3L1 cells, an adipocyte differentiation model, observed by theincrease of triglyceride content [10]. Treatment with 10 lM silde-nafil induced through a PKG way a significant up-regulation of the

(PKG) pathway in adipocyte. The nitric oxide synthase (NOS) produces NO usingated increases conversion of guanosine-50-triphosphate (GTP) to cGMP (c). Cyclicand inhibit the NO/cGMP/PKG pathway (d). The higher levels of cGMP activate PKG

channels (f). Adipocytes differentiation is regulated by coordinated expression of) ligand-activated transcription factor, regulates adipocyte differentiation and lipidtion; adipocyte fatty acid-binding protein (aP2) solubilization and intracellularAS) formation of long-chain fatty acids from acetyl-CoA, malonyl-CoA and NADPH;adiponectin abundant protein secreted by adipocyte, promotes fat cell proliferationsulin-mediated glucose transport into the cell; aromatase (ARO) converts androgensttractant protein-1 (MCP-1), chemokine (C-C motif) ligand 3 (CCL3), chemokine (C-Cnatriuretic peptide (ANP) binding to natriuretic receptor-A (NPR-A) increases cGMP(RhoA)/Rho-associated protein Kinase (ROCK) pathway exerts antihypertrophic and

190 G. Colombo et al. / Nitric Oxide 35 (2013) 186–192

expression of adipocyte-specific genes including PPARc, fatty acidsynthetase (FAS), adiponectin, CCAAT/enhancer-binding protein(CEBPa), adipocyte fatty acid-binding protein (aP2), and glucosetransporter type 4 (GLUT4) [10]. Another remarkable finding is thatsildenafil was able to augment basal and insulin-mediated glucoseuptake in adipocyte, associated with increased expression ofGLUT4 [10].

In order to verify the presence of PDE5 in primary humanvisceral adipocytes, and whether PDE5 inhibitors could directlymodulate aromatase (ARO), human visceral pre-adipocytes differ-entiated ex vivo were exposed to tadalafil or sildenafil [53]. Thisexperiment first evaluated PDE5 expression in human visceral adi-pocyte, and found relevant amounts of PDE5 mRNA, as well asdetectable levels of PDE5 protein [53]. A further finding of thisstudy was an increased expression of ARO following PDE5 inhibitoradministration, resulting in a net increase in estradiol (E2)/testos-terone (T) ratio [53]. This effect of the PDE5 inhibitor on ARO activ-ity might influence all the cells sensitive to serum E2 variation andactivity in vivo (i.e., adipose tissue, endothelial cells, bone cells,breast cells, and prostate cells), and be partly responsible for thebeneficial effects of PDE5 inhibitor on endothelial and metabolicfunctions [53]. In fact, lack of estrogen results in the developmentof a metabolic syndrome in humans and rodents, including excessadiposity, hepatic steatosis and insulin resistance [11]. It was alsodemonstrated that an increased androgen to estrogen ratio canpromote visceral fat accumulation in the rodent, by inhibiting 50

adenosine monophosphate-activated protein kinase (AMPK) acti-vation and stimulating lipogenesis [54].

A different study also explored the relevance of sexual differ-ence in the effect of PKGI overexpression [55]. Male and femalePKGI overexpressing mice were fed a low-fat (LF) or high-fat (HF)diet for 16 weeks [55]. Female transgenic PKGI mice exhibit in-creased oxygen consumption, cold-induced thermogenesis andlipolysis in WAT, but none of these effects were observed in malemice overexpressing PKGI [55]. The reasons for the marked genderdifferences are not clear, but may be caused by differences in sexhormones regarding the PKG signaling in adipocytes [55]. Estrogenplays an important role in adipose tissue regulation, is able to en-hance cGMP/PKG signaling, and also stimulates ANP release or ANPgene expression in heart [55].

In obesity there is mitochondrial dysfunction in adipose andmuscle tissues [56]. The proper operation of these mitochondriais required for glycemic control and to increase energy expenditure[56]. One study investigates the role of modulation of the NO path-way, by a PDE5 inhibitor, in mitochondrial biogenesis [56]. Sam-ples of omental adipose tissue ex vivo, removed from five malepatients undergoing elective surgery, were subjected to a rapid(24 h) and long-term (72 h) stimulation with either diethylenetri-amine NONOate (NO donor) or vardenafil (PDE5 inhibitor), andcompared to control samples [56]. After 24 h the NO donor andvardenafil increased the expression of PPARc and adiponectingenes, but after 72 h there were no differences in the expressionof all treated samples [56]. The NO donor resulted in a significantincrease in the expression of PGC-1a at 24 and 72 h of stimulation,while vardenafil stimulated the expression of the same gene onlyafter 72 h of exposure to the drug [56]. Mitochondrial DNA (mDNA)did not change after exposure to NO donor and vardenafil for 24 h,but an increased content of mitochondrial DNA could be detectedafter 72 h with both drugs [56].

To evaluate if obesity induced by HF diet could reduceendothelial nitric oxide signaling in adipose tissue, wild-type(WT), knockout for endothelial nitric oxide synthase (eNOS�/�),and VASP-deficient mice were submitted to a diet composed of10% lipids – LF diet, or 60% lipids – HF diet [9]. The study alsosought to examine whether the reduction of eNOS signaling wassufficient to generate inflammation independent of diet quality,

and if increased cGMP signaling could reduce adipose tissueinflammation in mice fed a HF diet [9]. In fact the HF diet for14 weeks in normal mice significantly decreased eNOS phosphory-lation relative to total eNOS, and also led to a reduction in the lev-els of both PKG and phosphor-VASP (pVASP) in epididymal whiteadipose tissue (eWAT), suggesting a decrease in signaling throughthe vascular NO-PKG–VASP pathway [9]. A HF diet was associatedwith reduced NO and increased TNFa mRNA contents in the sameeWAT samples, suggesting an association between nuclear factor-jB (NF-jB) activation and reduced vascular NO content and signal-ing [9]. The increase in levels of the macrophage marker mac-2 ineWAT of mice fed HF diet suggests an increased content of macro-phages in response to this diet [9]. Interestingly, the genetic defi-ciency of eNOS (eNOS�/�) mimicked the effect of the HF diet toincrease the expression of proinflammatory cytokines and therecruitment of macrophages to adipose tissue [9]. In turn, thetreatment with sildenafil was associated with a greater insulin-mediated decrease of blood glucose, implying increased insulinsensitivity [9]. Sildenafil also restored pVASP levels in the HF dietgroup eWAT, comparable to the LF diet group, demonstratingimprovement of NO-cGMP signaling in adipose tissue, and reducedadipose tissue macrophage infiltration and inflammation [9]. High-lighting the importance of VASP signaling for the anti-inflamma-tory effects of NO-cGMP pathway on adipose inflammation,levels of mRNA encoding proinflammatory cytokines monocytechemoattractant protein-1 (MCP-1), IL-6 and TNF-a were signifi-cantly elevated in VASP-deficient mice compared with WT con-trols, as was the phosphorylation of p65 subunit of NF-jB,indicative of increased NF-jB activation [9]. Inflammatory myeloidmarkers and levels of mac-2 protein also were upregulated ineWAT from VASP-null mice [9].

A recent study assessed the role of PKG in the differentiation ofWAT, as well as the effect of PDE5 inhibitor sildenafil in the estab-lishment of the brown phenotype (‘‘browning’’) of that tissue [14].In this experiment, mouse primary adipocytes and 3T3-L1 cellswere studied applying gain-and loss-of-function models [14]. PKGIwas detected in murine gonadal and inguinal WAT (gWAT/iWAT),BAT, as well as in 3T3-L1 cell lines, and was strongly reduced in fatdepots of adipocyte-restricted PKGI knockout mice [14]. PKGI stim-ulation with 8-pCPT-cGMP (cGMP analog) resulted in enhancedadipose differentiation in 3T3-L1 cell lines [14]. Conversely lentiv-iral knockout PKGI adipocytes (LV–Cre) were only poorly differen-tiated and displayed a significant decrease of lipid content, as wellas low expression of PPARc and aP2 [14]. In LV-Cre cells mitochon-drial biogenesis was impaired and UCP-1 mRNA levels were de-creased, indicating impaired potential for ‘‘browning’’ [14]. Toconfirm the role of PKGI in mitochondrial biogenesis and thermo-genic potential of WAT, cGMP was administered [14]. After 6 daysof treatment with cGMP, primary murine WAT isolated from iWATwas shifted to a brown-like phenotype, as shown by significantlyincreased up-regulation of UCP-1, PGC-1a, and PR domain contain-ing 16 (PRDM16) [14]. Sildenafil treatment also causes an increasein expression of UCP-1, PGC-1a, and PRDM16 comparable to cGMPin vitro [14]. The cGMP-treated LV-PKGI cells, as was to beexpected, presented an increased mitochondrial biogenesis con-firmed by quantitative real time polymerase chain reaction (qPCR)of genes encoded by mtDNA [14]. RhoA seems to be major down-stream molecule responsible for adipogenesis and for adipokinestimulation by PKGI [14]. Moreover, in vitro deletion of PKGI re-duced pRhoA and this could not be further increased by cGMPaddition [14]. Adiponectin expression was increased in 3T3-L1cells by cGMP treatment [14]. This effect was further increasedin PKGI-overexpressing cells [14]. With respect to inflammationthe amount of MCP-1, chemokine (C-C motif) ligant 7 (CCL7), che-mokine (C-C motif) ligant 3 (CCL3), IL-6, and TNF-a was decreasedin cGMP stimulated 3T3-L1 cells, and a further decrease was

Fig. 2. Flowchart describing research results of NO/cGMP/PKG pathway on WAT physiology. Nitric oxide (NO), guanosine 3050 cyclic monophosphate (cGMP), protein kinase G(PKG), white adipose tissue (WAT), lipoprotein lipase (LPL), triacylglycerol (TG), glycerol-3-phosphate dehydrogenase (GPDH), CCAAT/enhancer-binding protein (CEBPa),peroxisome proliferator-activated receptor gamma (PPARc), fatty acid synthetase (FAS), adipocyte fatty acid-binding protein (aP2), Glucose transporter type 4 (GLUT4), tumornecrosis factor alpha (TNFa), macrophage marker (mac-2), phosphorylation of p65 subunit of nuclear factor-jB (phospho NF-jB p65), monocyte Chemoattractant Protein-1(MCP-1), C-C motif ligant 3 (CCL3), C-C motif ligant 7 (CCL7), interleukin 6 (IL-6), mitochondrial DNA (mtDNA), uncoupling protein 1 (UCP-1), PPARc coactivator-1a (PGC-1a),PR domain containing 16 (PRDM16).

G. Colombo et al. / Nitric Oxide 35 (2013) 186–192 191

detected in LV-PKGI adipocytes treated with cGMP [14]. Pretreat-ment of cells with ROCK inhibitor Y-27632 also reduced activationof NF-jB signaling pathway, and IL-6/MCP-1 expression [14].Treatment with sildenafil, even for a short period of time (7 days)induced browning phenomenon in vivo [14]. The hematoxylinand eosin stain of iWAT from sildenafil-treated mice showed anappearance of multilocular adipocytes within white fat, suggestingBAT-like remodeling [14]. The iWAT of sildenafil-treated miceshowed an increased UCP-1/PGC-1a expression [14].

In summary, after reviewing the studies described above it isclear that NO/cGMP/PKG pathway can actually interfere with theadipocyte physiology. Such effect can also be achieved by theadministration of inhibitors of the enzyme that degrades cGMP(PDE5). The results indicate that PDE5 inhibitors, by increasingcGMP levels, can promote adipogenesis, interfere in adipokinessecretion, decrease the expression of proinflammatory proteins,and increase the thermogenic potential of white adipose tissue(Fig. 2).

Conclusion

Before concluding on the role of PDE5 in white adipocyte phys-iology, we must emphasize that this phosphodiesterase is part ofthe NP/NO/cGMP/PKGI pathway. Data available in the literatureshow results confirming that cascade activation initiated by eitherNO or NPs stimulate adipocyte differentiation. This differentiationin turn leads to beneficial modifications in adipocyte physiology,such as the release of adiponectin and increased expression ofGLUT4, changes related to decreased insulin resistance. Someexperiments show that another way to achieve these benefits isthrough PDE5 inhibitors and the resulting elevation of cGMP levels.Increased levels of cGMP in WAT lead to an increase in cellular up-take of glucose and triglycerides, activation of aromatase, decreaseof tissue infiltration by macrophages/inflammation, increase inthermogenesis, i.e., a counter point to the basic pathogenesis of

the metabolic syndrome. Thus, we can conclude that PDE5 adipo-cyte is a logical target for preserving adipose tissue health, and be-cause of the intimate relationship of this tissue to systemicmetabolism, a way to bring positive results in the control of manydisarrangements associated with metabolic syndrome. PDE5 inhib-itors currently in use are possible agents for the control of cGMPlevels and their beneficial effects, but perhaps new drugs willbecome available with greater potency and specificity for use inadipose tissue disorders.

Conflicts of interest

The authors have no conflict of interest.

References

[1] WHO. 10 Facts on obesity, 2012.[2] P. Zimmet, C.R. Thomas, Genotype, obesity and cardiovascular disease – has

technical and social advancement outstripped evolution?, J Intern. Med. 254(2) (2003) 114–125.

[3] G.R. Hajer, T.W. van Haeften, F.L. Visseren, Adipose tissue dysfunction inobesity, diabetes, and vascular diseases, Eur. Heart J. 29 (24) (2008) 2959–2971.

[4] A.E. Schutte, H.W. Huisman, R. Schutte, J.M. van Rooyen, L. Malan, C.M. Fourie,et al., Adipokines and cardiometabolic function: How are they interlinked?,Regul Pept. 164 (2–3) (2010) 133–138.

[5] M. Chandran, S.A. Phillips, T. Ciaraldi, R.R. Henry, Adiponectin: more than justanother fat cell hormone?, Diabetes Care 26 (8) (2003) 2442–2450.

[6] D.-E. Lee, S. Kehlenbrink, H. Lee, M. Hawkins, J.S. Yudkin, Getting the messageacross: mechanisms of physiological cross talk by adipose tissue, Am. J.Physiol. Endocrinol. Metab. 296 (6) (2009) E1210–E1229.

[7] C. Capurso, A. Capurso, From excess adiposity to insulin resistance the role offree fatty acids, Vascul. Pharmacol. 57 (2–4) (2012) 91–97.

[8] S.H. Francis, J.L. Busch, J.D. Corbin, D. Sibley, CGMP-dependent protein kinasesand cGMP phosphodiesterases in nitric oxide and cGMP action, Pharmacol.Rev. 62 (3) (2010) 525–563.

[9] P. Handa, S. Tateya, N.O. Rizzo, A.M. Cheng, V. Morgan-Stevenson, C.Y. Han,et al., Reduced vascular nitric oxide-cGMP signaling contributes to adiposetissue inflammation during high-fat feeding, Arterioscler. Thromb. Vasc. Biol.31 (12) (2011) 2827–2835.

192 G. Colombo et al. / Nitric Oxide 35 (2013) 186–192

[10] X. Zhang, J. Ji, G. Yan, J. Wu, X. Sun, J. Shen, et al., Sildenafil promotesadipogenesis through a PKG pathway, Biochem. Biophys. Res. Commun. 396(4) (2010) 1054–1059.

[11] M.E. Jones, K.J. McInnes, W.C. Boon, E.R. Simpson, Estrogen and adiposity –utilizing models of aromatase deficiency to explore the relationship, J. SteroidBiochem. Mol. Biol. 106 (1–5) (2007) 3–7.

[12] K. Hemmrich, C. Gummersbach, N.E. Paul, D. Goy, C.V. Suschek, K.D. Kroncke,et al., Nitric oxide and downstream second messenger cGMP and cAMPenhance adipogenesis in primary human preadipocytes, Cytotherapy 12 (4)(2010) 547–553.

[13] D. McGrowder, D. Ragoobirsingh, P. Brown, Modulation of glucose uptake inadipose tissue by nitric oxide-generating compounds, J. Biosci. 31 (3) (2006)347–354.

[14] M.M. Mitschke, L.S. Hoffmann, T. Gnad, D. Scholz, K. Kruithoff, P. Mayer, et al.,Increased cGMP promotes healthy expansion and browning of white adiposetissue, FASEB J. 27 (4) (2013) 1621–1630.

[15] A. Armani, V. Marzolla, G.M. Rosano, A. Fabbri, M. Caprio, PDE5 in theadipocyte: a novel player in fat metabolism?, Trends Endocrinol Metab. 22(10) (2011) 404–411.

[16] U. Razny, B. Kiec-Wilk, A. Polus, L. Wator, G. Dyduch, L. Partyka, et al., Theadipose tissue gene expression in mice with different nitric oxide availability,J. Physiol. Pharmacol. 61 (5) (2010) 607–618.

[17] U. Razny, B. Kiec-Wilk, L. Wator, A. Polus, G. Dyduch, B. Solnica, et al., Increasednitric oxide availability attenuates high fat diet metabolic alterations and geneexpression associated with insulin resistance, Cardiovasc. Diabetol. 10 (2011)68.

[18] T. Nishikimi, C. Iemura-Inaba, K. Akimoto, K. Ishikawa, S. Koshikawa, H.Matsuoka, Stimulatory and Inhibitory regulation of lipolysis by the NPR-A/cGMP/PKG and NPR-C/G(i) pathways in rat cultured adipocytes, Regul. Pept.153 (1–3) (2009) 56–63.

[19] K. Miyashita, H. Itoh, H. Tsujimoto, N. Tamura, Y. Fukunaga, M. Sone, et al.,Natriuretic peptides/cGMP/cGMP-dependent protein kinase cascades promotemuscle mitochondrial biogenesis and prevent obesity, Diabetes 58 (12) (2009)2880–2892.

[20] M. Bordicchia, D. Liu, E.Z. Amri, G. Ailhaud, P. Dessi-Fulgheri, C. Zhang, et al.,Cardiac natriuretic peptides act via p38 MAPK to induce the brown fatthermogenic program in mouse and human adipocytes, J. Clin. Invest. 122 (3)(2012) 1022–1036.

[21] S.H. Francis, M.A. Blount, J.D. Corbin, Mammalian cyclic nucleotidephosphodiesterases: molecular mechanisms and physiological functions,Physiol. Rev. 91 (2) (2011) 651–690.

[22] A.T. Bender, J.A. Beavo, Cyclic nucleotide phosphodiesterases: molecularregulation to clinical use, Pharmacol. Rev. 58 (3) (2006) 488–520.

[23] A. Kuthe, Phosphodiesterase 5 inhibitors in male sexual dysfunction, Curr.Opin. Urol. 13 (5) (2003) 405–410.

[24] R. Shamloul, H. Ghanem, Erectile dysfunction, Lancet (2012).[25] K.L. Saraiva, A.K. Silva, M.I. Wanderley, A.A. De Araujo, J.R. De Souza, C.A.

Peixoto, Chronic treatment with sildenafil stimulates Leydig cell andtestosterone secretion, Int. J. Exp. Pathol. 90 (4) (2009) 454–462.

[26] S.G. Chrysant, G.S. Chrysant, The pleiotropic effects of phosphodiesterase 5inhibitors on function and safety in patients with cardiovascular disease andhypertension, J. Clin. Hypertens. (Greenwich) 14 (9) (2012) 644–649.

[27] C. Mammi, D. Pastore, M.F. Lombardo, F. Ferrelli, M. Caprio, C. Consoli, et al.,Sildenafil reduces insulin-resistance in human endothelial cells, PLoS One 6 (1)(2011) e14542.

[28] M. Conti, J. Beavo, Biochemistry and physiology of cyclic nucleotidephosphodiesterases: essential components in cyclic nucleotide signaling,Annu. Rev. Biochem. 76 (2007) 481–511.

[29] V. Boswell-Smith, D. Spina, C.P. Page, Phosphodiesterase inhibitors, Br. J.Pharmacol. 147 (Suppl. 1) (2006) S252–S257.

[30] B.J. Sung, K.Y. Hwang, Y.H. Jeon, J.I. Lee, Y.S. Heo, J.H. Kim, et al., Structure of thecatalytic domain of human phosphodiesterase 5 with bound drug molecules,Nature 425 (6953) (2003) 98–102.

[31] K. Omori, J. Kotera, Overview of PDEs and their regulation, Circ. Res. 100 (3)(2007) 309–327.

[32] G.L. Card, B.P. England, Y. Suzuki, D. Fong, B. Powell, B. Lee, et al., Structuralbasis for the activity of drugs that inhibit phosphodiesterases, Structure 12(12) (2004) 2233–2247.

[33] M. Guazzi, M. Samaja, The role of PDE5-inhibitors in cardiopulmonarydisorders: from basic evidence to clinical development, Curr. Med. Chem. 14(20) (2007) 2181–2191.

[34] Dunmore S.J., J.E. Brown, The role of adipokines in beta-cell failure of type 2diabetes, J. Endocrinol. (2012).

[35] M. Bluher, Clinical relevance of adipokines, Diabetes Metab. J. 36 (5) (2012)317–327.

[36] K. Yamashita, H. Yatsuya, T. Muramatsu, H. Toyoshima, T. Murohara, K.Tamakoshi, Association of coffee consumption with serum adiponectin, leptin,inflammation and metabolic markers in Japanese workers: a cross-sectionalstudy, Nutr. Diabetes 2 (2012) e33.

[37] M.D. Klok, S. Jakobsdottir, M.L. Drent, The role of leptin and ghrelin in theregulation of food intake and body weight in humans: a review, Obes. Rev. 8(1) (2007) 21–34.

[38] K. Rabe, M. Lehrke, K.G. Parhofer, U.C. Broedl, Adipokines and insulinresistance, Mol. Med. 14 (11–12) (2008) 741–751.

[39] T. Kadowaki, T. Yamauchi, N. Kubota, K. Hara, K. Ueki, K. Tobe, Adiponectin andadiponectin receptors in insulin resistance, diabetes, and the metabolicsyndrome, J. Clin. Invest. 116 (7) (2006) 1784–1792.

[40] S.A. Phillips, J.T. Kung, Mechanisms of adiponectin regulation and use as apharmacological target, Curr. Opin. Pharmacol. 10 (6) (2010) 676–683.

[41] K.J. Mather, T. Funahashi, Y. Matsuzawa, S. Edelstein, G.A. Bray, S.E. Kahn, et al.,Adiponectin, change in adiponectin, and progression to diabetes in theDiabetes Prevention Program, Diabetes 57 (4) (2008) 980–986.

[42] S. Li, H.J. Shin, E.L. Ding, R.M. van Dam, Adiponectin levels and risk of type 2diabetes: a systematic review and meta-analysis, JAMA 302 (2) (2009) 179–188.

[43] D.M. Maahs, L.G. Ogden, G.L. Kinney, P. Wadwa, J.K. Snell-Bergeon, D. Dabelea,et al., Low plasma adiponectin levels predict progression of coronary arterycalcification, Circulation 111 (6) (2005) 747–753.

[44] J.M. Olefsky, C.K. Glass, Macrophages, inflammation, and insulin resistance,Annu. Rev. Physiol. 72 (2010) 219–246.

[45] K. Kos, S.P. Wong, M.S. Huda, M. Cakir, M. Jernas, L. Carlsson, et al., In humansthe adiponectin receptor R2 is expressed predominantly in adipose tissue andlinked to the adipose tissue expression of MMIF-1, Diabetes Obes. Metab. 12(4) (2010) 360–363.

[46] I. Russo, P. Del Mese, G. Doronzo, L. Mattiello, M. Viretto, A. Bosia, et al.,Resistance to the nitric oxide/cyclic guanosine 50-monophosphate/proteinkinase G pathway in vascular smooth muscle cells from the obese zucker rat, aclassical animal model of insulin resistance. Role of oxidative stress,Endocrinology 149 (4) (2008) 1480–1489.

[47] E. Bischoff, Potency, selectivity, and consequences of nonselectivity of PDEinhibition, Int. J. Impot. Res. 16 (Suppl. 1) (2004) S11–S14.

[48] B. Haas, P. Mayer, K. Jennissen, D. Scholz, M. Berriel Diaz, W. Bloch, et al.,Protein kinase G controls brown fat cell differentiation and mitochondrialbiogenesis, Sci. Signal. 2 (99) (2009) ra78.

[49] K. Jennissen, F. Siegel, M. Liebig-Gonglach, M.R. Hermann, S. Kipschull, S. vanDooren, et al., A VASP-Rac-soluble guanylyl cyclase pathway controls cGMPproduction in adipocytes, Sci. Signal. 5 (239) (2012) ra62.

[50] J.E. Ayala, D.P. Bracy, B.M. Julien, J.N. Rottman, P.T. Fueger, D.H. Wasserman,Chronic treatment with sildenafil improves energy balance and insulin actionin high fat-fed conscious mice, Diabetes 56 (4) (2007) 1025–1033.

[51] Y. Hara, S. Wakino, Y. Tanabe, M. Saito, H. Tokuyama, N. Washida, et al., Rhoand Rho-kinase activity in adipocytes contributes to a vicious cycle in obesitythat may involve mechanical stretch, Sci. Signal. 4 (157) (2011) ra3.

[52] H. Yan, E. Aziz, G. Shillabeer, A. Wong, D. Shanghavi, A. Kermouni, et al., Nitricoxide promotes differentiation of rat white preadipocytes in culture, J. LipidRes. 43 (12) (2002) 2123–2129.

[53] A. Aversa, M. Caprio, A. Antelmi, A. Armani, M. Brama, E.A. Greco, et al.,Exposure to phosphodiesterase type 5 inhibitors stimulates aromataseexpression in human adipocytes in vitro, J. Sex Med. 8 (3) (2011) 696–704.

[54] K.J. McInnes, A. Corbould, E.R. Simpson, M.E. Jones, Regulation of adenosine 50 ,monophosphate-activated protein kinase and lipogenesis by androgenscontributes to visceral obesity in an estrogen-deficient state, Endocrinology147 (12) (2006) 5907–5913.

[55] D.M. Nikolic, Y. Li, S. Liu, S. Wang, Overexpression of constitutively active PKG-I protects female, but not male mice from diet-induced obesity, Obesity (SilverSpring) 19 (4) (2011) 784–791.

[56] L. De Toni, G. Strapazzon, L. Gianesello, N. Caretta, C. Pilon, A. Bruttocao, et al.,Effects of type 5-phosphodiesterase inhibition on energy metabolism andmitochondrial biogenesis in human adipose tissue ex vivo, J. Endocrinol.Invest. 34 (10) (2011) 738–741.