RESEARCH ARTICLE

The lysyl oxidase inhibitor β-aminopropionitrile reduces bodyweight gain and improves the metabolic profile in diet-inducedobesity in ratsMarıa Miana1,2,*, Marıa Galan3,*, Ernesto Martınez-Martınez1,2,4, Saray Varona3, Raquel Jurado-Lopez1,2,Belen Bausa-Miranda1,2, Alfonso Antequera5, Marıa Luaces6, Jose Martınez-Gonzalez3, Cristina Rodrıguez3,*,‡

and Victoria Cachofeiro1,2,*,‡

ABSTRACTExtracellular matrix (ECM) remodelling of the adipose tissueplays a pivotal role in the pathophysiology of obesity. The lysyloxidase (LOX) family of amine oxidases, including LOX andLOX-like (LOXL) isoenzymes, controls ECM maturation, andupregulation of LOX activity is essential in fibrosis; however, itsinvolvement in adipose tissue dysfunction in obesity is unclear. Inthis study, we observed that LOX is the main isoenzyme expressedin human adipose tissue and that its expression is stronglyupregulated in samples from obese individuals that had beenreferred to bariatric surgery. LOX expression was also induced in theadipose tissue from male Wistar rats fed a high-fat diet (HFD).Interestingly, treatment with β-aminopropionitrile (BAPN), a specificand irreversible inhibitor of LOX activity, attenuated the increase inbody weight and fat mass that was observed in obese animals andshifted adipocyte size toward smaller adipocytes. BAPN alsoameliorated the increase in collagen content that was observed inadipose tissue from obese animals and improved several metabolicparameters – it ameliorated glucose and insulin levels, decreasedhomeostasis model assessment (HOMA) index and reducedplasma triglyceride levels. Furthermore, in white adipose tissuefrom obese animals, BAPN prevented the downregulation ofadiponectin and glucose transporter 4 (GLUT4), as well as theincrease in suppressor of cytokine signaling 3 (SOCS3) anddipeptidyl peptidase 4 (DPP4) levels, triggered by the HFD.Likewise, in the TNFα-induced insulin-resistant 3T3-L1 adipocytemodel, BAPN prevented the downregulation of adiponectin andGLUT4 and the increase in SOCS3 levels, and consequentlynormalised insulin-stimulated glucose uptake. Therefore, our dataprovide evidence that LOX plays a pathologically relevant role in themetabolic dysfunction induced by obesity and emphasise theinterest of novel pharmacological interventions that target adipose

tissue fibrosis and LOX activity for the clinical management ofthis disease.

KEY WORDS: Lysyl oxidase, Extracellular matrix, Adipose tissue,Fibrosis, Obesity, Insulin resistance

INTRODUCTIONIn recent years, there has been increasing evidence that theextracellular matrix (ECM) plays a key role in adipose tissuedevelopment and function, and fibrosis is now recognised as acrucial player in adipose tissue dysfunction in obesity. In fact, earlydeposition of connective tissue occurs in subcutaneous whiteadipose tissue after moderate weight gain and further increases withobesity (Mutch et al., 2009; Divoux et al., 2010; Henegar et al.,2008; Alligier et al., 2012). Interestingly, data from animal modelsof obesity suggest that this abnormal ECM remodelling of whiteadipose tissue could contribute to local and systemic metabolicalterations (Halberg et al., 2009; Khan et al., 2009).

Fibrosis arises from the imbalance between ECM synthesis anddegradation that culminates in an excessive accumulation of ECMcomponents. Lysyl oxidase (LOX) activity is crucially involved inECM synthesis and processing. To date, five isoenzymes belongingto the LOX family have been identified, including LOX, LOX-like1(LOXL1), LOXL2, LOXL3 and LOXL4. They are copper-dependent amine oxidases that exhibit a high degree of homologyin the C-terminal domain, which contains the catalytic site. LOX,the archetypal member of this family and the best characterised,catalyses the covalent cross-linking of collagens and elastin fibres(Rodríguez et al., 2008a,b). This enzyme oxidises lysine andhydroxylysine residues in these substrates, leading to the synthesisof highly reactive peptidylsemialdehydes that can spontaneouslycondense to form the intra- and intermolecular covalent cross-linkages that are responsible for ECM stability. The disturbance ofLOX activity induces connective tissue abnormalities related topathological processes, including cardiovascular diseases, as wehave previously described (Miana et al., 2011; Orriols et al., 2014;Raposo et al., 2004; Rodriguez et al., 2002, 2009; Stefanon et al.,2013).

Changes in LOX activity could be involved in the disturbanceof adipocyte function in obesity. In fact, LOX expression isupregulated in the white adipose tissue from ob/ob mice (Halberget al., 2009), and it is inhibited during adipocyte differentiation(Dimaculangan et al., 1994). Furthermore, LOX participates in thecommitment of pluripotent stem cells to the adipocyte lineage(Huang et al., 2009). However, the potential role of LOX activity inhuman obesity has been poorly characterised. A transcriptomicstudy revealed an increase in LOX expression in subcutaneous whiteReceived 20 January 2015; Accepted 28 March 2015

1Departamento de Fisiologıa, Facultad de Medicina, UniversidadComplutense, Madrid 28040, Spain. 2Instituto de Investigacion SanitariaGregorio Maranon (IiSGM), Madrid 28007, Spain. 3Centro de InvestigacionCardiovascular (CSIC-ICCC), IIB-Sant Pau, Barcelona 08025, Spain.4Cardiovascular Translational Research, NavarraBiomed (Fundacion MiguelServet), Pamplona 31008, Spain. 5Upper Gastroenterology & Bariatric SurgeryDepartment, Fuenlabrada University Hospital, Madrid 28942, Spain.6Cardiology Department, Cardiovascular Institute, Hospital Clınico San Carlos,Madrid 28040, Spain.*These authors contributed equally to this work

‡Authors for correspondence (crodriguezs@csic-iccc.org; vcara@ucm.es)

This is an Open Access article distributed under the terms of the Creative Commons AttributionLicense (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use,distribution and reproduction in any medium provided that the original work is properly attributed.

543

© 2015. Published by The Company of Biologists Ltd | Disease Models & Mechanisms (2015) 8, 543-551 doi:10.1242/dmm.020107

Disea

seModels&Mechan

isms

adipose tissue from obese subjects, but its pathophysiologicalrelevance remains unclear (Henegar et al., 2008). Therefore, the aimof this study was to explore the role of LOX activity in adiposetissue remodelling and in the metabolic disturbances associatedwith obesity. For this purpose, we have evaluated the impact ofβ-aminopropionitrile (BAPN), a specific and irreversible inhibitorof LOX activity, in a model of diet-induced obesity.

RESULTSLOX expression is induced in adipose tissue from morbidlyobese patientsThe expression of LOX and LOXL isoenzymes in visceral adiposetissue was investigated in samples from control individuals andmorbidly obese patients. The mean age of the obese group was

40.2±9.5 (range 30-55) vs 36.0±9.5 (range 26-51) years inthe normoweight group. The average BMI was 44.6±3.7 (range41.5-50.9) in the obese group vs 22.7±3.6 (range 16.4-25.0) in thecontrol group (P=0.008). No significant differences were observed inthe fasting plasma glucose levels between both groups (114.2±26.0 vs95.8±10.8 mg/dl; P=0.344). However, obese patients showedhigher insulin plasma levels (486.6±173.1 vs 268.4±138.4 pg/dl;P=0.0929), homeostasis model assessment (HOMA) index (3.21±1.49 vs 1.44±0.75; P=0.075) and plasma triglyceride levels (169.4±44.8 vs 106.6±24.9 mg/dl; P=0.0554) than controls, althoughneither difference reached statistical significance. As expected,histological analysis of visceral adipose tissue showed an enhancedadipocyte area (cross-sectional area of single adipocytes) (Fig. 1A;P=0.05) and an increase in pericellular collagen (Fig. 1B;P=0.007) inobesepatients as comparedwith control subjects.AmongLOXfamilymembers, LOX exhibited the highest expression level in the adiposetissue from control individuals, whereas that of LOXLs was almostnegligible. Interestingly, LOX mRNA levels were significantlyupregulated (about threefold) in the visceral adipose tissue fromobese patients compared with that of controls. A consequent increasein LOX protein levels, assessed by western blotting andimmunohistochemistry (Fig. 1D,E), was detected in samples fromobese subjects. LOX protein levels were correlated with collagencontent (r=0.70; P=0.0433). As previously described (Kern et al.,2003; Sell et al., 2013), visceral adipose tissue from obese patientsshowed an increase in dipeptidyl peptidase 4 (DPP4) protein levels(220.9±62.6 vs 100.0±10.0 AU;P=0.015; Fig. 1D) and a reduction inthose of adiponectin (22.7±2.8 vs 100.0±52.4 AU; P=0.05; Fig. 1D)compared with normoweight subjects.

LOX expression is induced in adipose tissue from HFD-fedratsThe expression of LOX in white adipose tissuewas investigated in ananimal model of diet-induced obesity that we have previouslydescribed (Martínez-Martínez et al., 2014a,b). As shown in Fig. 2A,LOX mRNA levels were increased approximately 15-fold in theepididymal adipose tissue of rats fed a high-fat diet (HFD) comparedwith control animals fed normal chow. In agreement, LOX proteinlevels, assessed by western blotting and immunohistochemistry(Fig. 2B,C), were higher in the white adipose tissue from obeseanimals. A correlation was observed between protein levels of LOXand collagen content (r=0.8389; P=0.003). The HFD did not affectthe expression of bonemorphogenetic protein 1 (BMP1), the enzymethat processes pro-LOX to its active form (Fig. 2D). Furthermore, thisdietary intervention did not modify the expression of LOX in brownadipose tissue (data not shown).

BAPN ameliorates the increase in bodyweight, adiposity andadipose tissue fibrosis in HFD-fed ratsTo investigate whether LOX contributes to the adipose tissuedysfunction associated with obesity, rats subjected to a HFD weretreated with BAPN, an irreversible and specific inhibitor of LOXactivity. As shown in Fig. 3A, the HFD induced a significant increasein bodyweight that reached a significant difference from the thirdweekonwards. After three weeks of treatment, BAPN significantlyprevented the rise in body weight in HFD rats, but not in animalsthat were fed a standard diet (Fig. 3A). These differences weresustaineduntil the endof the study (Table 1, Fig. 3A). Similarly,BAPNreduced the increase in the weight of white adipose tissue (bothepididymal and lumbar) in obese animals (Table 1) and attenuatedtheir enhanced adiposity (Table 1). It should be noted that the changes

TRANSLATIONAL IMPACT

Clinical issueObesity is one of the most serious public health challenges of the 21stcentury. The risk of a number of major diseases, including type2 diabetes mellitus, ischemic heart disease, ischemic stroke andseveral common forms of cancers, is dramatically increased in obeseindividuals. The epidemic proportions achieved by obesity makesit mandatory to reach a deeper understanding of its underlyingpathophysiological mechanisms, which could provide novel therapeutictargets. In recent years, fibrosis has been recognised as a crucial playerin adipose tissue dysfunction in obesity. Lysyl oxidase (LOX) activity,which governs extracellular matrix maturation, is essential for tissuefibrosis. However, its contribution to adipose tissue dysfunction inobesity has not been clearly established.

ResultsThis study analyses the role of LOX in adipose tissue remodelling byusing three experimental systems: adipose tissue samples from obeseindividuals that had been referred for bariatric surgery (weight losssurgery), an animal model of diet-induced obesity and cell-basedstudies. The authors demonstrate that LOX is the main lysyl oxidaseisoenzyme expressed in human adipose tissue and that it is upregulatedin samples from both obese individuals and rats fed a high-fat diet. Inobese rats, the inhibition of LOX activity through β-aminopropionitrile(BAPN, a specific inhibitor of LOX activity) reduces adipose tissuefibrosis, partially corrects the adipocyte-size distribution pattern (shiftingit toward smaller sizes) and attenuates the increase in body weight andfat mass. Furthermore, LOX inhibition improves multiple metabolicparameters: normalizing glucose, insulin and triglyceride levels andreducing the homeostatic model assessment (HOMA) index – ameasure of insulin resistance. Likewise, in these animals, BAPNameliorates the disturbances in the adipose tissue expression ofadiponectin, glucose transporter 4 (GLUT4), suppressor of cytokinesignaling 3 (SOCS3) and dipeptidyl peptidase 4, all proteins involved inthe control of insulin sensitivity. Finally, BAPN also normalises theinsulin-stimulated glucose uptake and protein levels of GLUT4,adiponectin and SOCS3 in the TNFα-induced insulin-resistant 3T3-L1adipocyte model.

Implications and future directionsThe results reported in this study demonstrate the upregulation of theadipose tissue expression of LOX in human obesity and provideevidence that LOX inhibition prevents adipose tissue dysfunction,reduces bodyweight gain and improves metabolic disturbances in diet-induced obesity. These findings uncover the pathologically relevantcontribution of LOX to the metabolic dysfunction induced by obesity andemphasise the potential of novel pharmacological interventions targetingadipose tissue fibrosis and LOX activity for the clinical management ofobesity. Future studies will help to further clarify the mechanismsunderlying the beneficial consequences of LOX inhibition in obesity.

544

RESEARCH ARTICLE Disease Models & Mechanisms (2015) 8, 543-551 doi:10.1242/dmm.020107

Disea

seModels&Mechan

isms

in body weight triggered by BAPN in HFD-fed rats were not relatedto differences in food intake (Table 1).The changes in adipose tissue mass elicited by BAPN in HFD-fed

rats prompted us to determine whether LOX inhibition couldmodulate the adipocyte area. Histological analysis of epididymaladipose tissue revealed an increased adipocyte area in the HFDgroup. A trend towards an attenuation of this parameter wasobserved in obese animals that had been treated with BAPN(Fig. 3B,C). Furthermore, a shift toward smaller adipocytes wasdetected in BAPN-treated obese animals compared with HFD-fedrats (Fig. 3D). Interestingly, LOX inhibition prevented the increasein pericellular collagen content observed in obese rats, as analysedby using Picrosirius red staining (Fig. 3E,F).

BAPN improves the metabolic alterations observed in obeseratsNext, we examined whether the inhibition of LOX activitycould modify metabolic parameters in obese animals. Treatmentwith BAPN improved fasted glucose and insulin levels andconsequently reduced HOMA index in the HFD group (Table 1).LOX inhibition also reduced plasma triglycerides in obese animals,but no significant differences were observed in the total cholesterollevels (Table 1).

BAPN improves insulin signalling in adipose tissue fromobese animalsIn order to understand how BAPN improves insulin sensitivity inobese animals, we analysed the levels of proteins involved in thecontrol of insulin sensitivity in epididymal adipose tissue. Thereduction in both glucose transporter 4 (GLUT4) and adiponectinexpression observed in the HFD group was normalised through theinhibition of LOX activity (Fig. 4A,B). Furthermore, the increase inthe protein levels of DPP4 and suppressor of cytokine signaling 3(SOCS3) triggered by the HFDwas completely prevented by BAPN(Fig. 4C,D).

BAPN normalises the expression of GLUT4 and adiponectin,and improves glucose uptake in an in vitro model of insulinresistanceTo determinewhether BAPN directly affects adipocyte function, weperformed in vitro experiments in the TNFα-induced insulinresistance model in differentiated 3T3-L1 adipocytes. As shownin Fig. 5 and as previously described (Stephens et al., 1997; Sethiand Hotamisligil, 1999), TNFα reduced expression of GLUT4 andadiponectin and increased SOCS3 protein levels in these cells.Interestingly, these effects were prevented by BAPN (Fig. 5A-C).Accordingly, BAPN normalised the TNFα-induced decrease in

Fig. 1. Adipocyte size, total collagen content and lysyl oxidase expression in visceral adipose tissue from control subjects and morbidly obeseindividuals. (A) Quantification of adipocyte area (left) and representative microphotographs of visceral adipose tissue sections from control subjects (CT) andmorbidly obese individuals (OB) stained with haematoxylin and eosin and then examined by using light microscopy (magnification 40×; right). (B) Quantification ofcollagen volume fraction (CVF) normalised for adipocyte number. Right panel shows representative microphotographs of visceral adipose tissue sections fromthese individuals stained with Picrosirius red that were then examined by using light microscopy (magnification 40×). (C) mRNA levels of lysyl oxidase (LOX) andLOX-like (LOXL) isoenzymes 1, 2, 3 and 4 evaluated using real-time PCR of these samples. Results were normalised to the expression of 18S RNA. (D) LOX,DPP4 and adiponectin (Adipo) protein levels determined by western blotting. β-actin was analysed as loading control. (E) Representative images of theimmunohistochemical analysis for LOX are presented (magnification 40×). Histograms represent the mean±s.d. of five subjects. Scale bars: 50 µm. *P<0.05 and**P<0.01 vs control group.

545

RESEARCH ARTICLE Disease Models & Mechanisms (2015) 8, 543-551 doi:10.1242/dmm.020107

Disea

seModels&Mechan

isms

insulin-stimulated glucose uptake observed in differentiated 3T3-L1adipocytes (Fig. 5D).

DISCUSSIONIt is becoming apparent that fibrosis is an essential factor in adiposetissue dysfunction in obesity (Divoux et al., 2010; Halberg et al.,2009; Khan et al., 2009). ECM remodelling impacts on themetabolic function of adipose tissue, and the control of LOXactivity, a major player in ECM maturation, could be crucial in thiscontext. The research presented here reveals that LOX is the mainlysyl oxidase isoenzyme expressed in adipose tissue and that itsexpression is upregulated in the adipose tissue of morbidly obesepatients and obese rats. More interestingly, we observed thatinhibition of LOX activity protects against diet-induced obesity,thereby limiting weight gain, attenuating the disturbances in adiposetissue and improving the insulin resistance observed in obeseanimals.Increased expression of ECM components in adipose tissue has

been repeatedly reported in murine models of obesity and in obesesubjects associated with either body mass index (BMI), insulinsensitivity and/or resistance, and/or inflammation (Abdennour et al.,2014; Divoux et al., 2010; Halberg et al., 2009; Khan et al., 2009;Tam et al., 2012; Walker et al., 2014). A transcriptomic analysis ofadipose tissue from obese individuals revealed the pathologicalrelevance of ECM. However, the increase in the expression of LOXand other matrix components observed in these microarray studieswas not subsequently validated using other techniques (Henegaret al., 2008). Accordingly, we found that LOX mRNA levels areincreased in adipose tissue samples from obese individuals, whichwas confirmed by western blotting and immunohistochemistry.Furthermore, obese individuals exhibited a marked fibrosis ofadipose tissue, a pathological process that reduces tissue plasticityand alters adipocyte function (Khan et al., 2009; Virtue and Vidal-Puig, 2010). In agreement with previous reports in the ob/ob model(Halberg et al., 2009); we also observed an upregulation of LOX in

the adipose tissue of rats fed a HFD that was associated withenhanced pericellular fibrosis. The strong correlation of LOXexpression with pericellular fibrosis in the adipose tissue from bothhumans and rats, as well as the ability of BAPN to limit pericellularfibrosis in HFD-fed animals support the notion that LOX makes arelevant contribution to the fibrotic response in obesity.

LOX and LOXL isoenzymes differ from other copper aminooxidases in their quinone cofactor, the lysine tyrosylquinone(LTQ), which confers particular catalytic properties to theseenzymes (Rodriguez et al., 2008b). BAPN is a well-knownspecific and irreversible inhibitor of LOX activity (Tang et al.,1983; Rodriguez et al., 2008b) that targets the active site of LOX orLOXL isoenzymes. BAPN is a useful tool to study cellularprocesses involving LOX and LOXL isoenzymes; however, itspharmacological use as a systemic drug is not indicated. It has beenobserved that prolonged administration of BAPN evokes a weakneurotoxicity and reduces the mechanical strength of bones (Ahsanet al., 1999; Spencer and Schaumburg, 1983; Turecek et al., 2008).Further, lathyrism, a disease characterised by extensive disruptionof connective tissue, is caused by the excessive and long-termconsumption of certain legumes of the genus Lathyrus, whichcontain high amounts of BAPN and related compounds (Dasler,1954). Nevertheless, owing to the lack of viable LOX-deficientmice (Maki et al., 2002; Hornstra et al., 2003), this inhibitor hasbeen extensively used in short-term in vivo studies to support theinvolvement of lysyl oxidases in multiple biological processes,including tumour progression and metastasis (Levental et al., 2009;Erler et al., 2006), pulmonary arterial hypertension (Nave et al.,2014) and arterial stiffening (Kothapalli et al., 2012). Similarly, wehave observed that BAPN prevents the increase in both body weightand fat-pad weight found in obese animals without affecting foodintake. These changes were accompanied by a shift toward smalleradipocytes. Likewise, in mice of an ob/ob background, theknockdown of ECM components, such as collagen VI, limited thegain of body weight and reduced the increase in fat mass (Khan

Fig. 2. Expression of LOX and BMP1 inepididymal adipose tissue in control andobese rats. (A) Lysyl oxidase (LOX) mRNAlevels analysed by using real-time PCR ofepididymal adipose tissue from control ratsfed a normal chow (CT) and rats fed a high-fatdiet (HFD). Results were normalised tocyclophilin expression. (B) LOX protein levelsdetermined by western blotting of the samesamples described in A. β-actin was analysedas loading control. (C) Representative imagesof immunostaining of LOX are presented(magnification 40×). (D) Bone morphogeneticprotein-1 (BMP1) mRNA levels werequantified by using real-time PCR analyses ofepididymal adipose tissue from control ratsfed a normal chow (CT) and rats fed a HFD.Expression was normalised to that ofcyclophilin. Graphs represent the mean±s.d.of eight animals. Scale bars: 50 µm. *P<0.05vs control group.

546

RESEARCH ARTICLE Disease Models & Mechanisms (2015) 8, 543-551 doi:10.1242/dmm.020107

Disea

seModels&Mechan

isms

et al., 2009). Therefore, the composition and structure of ECMseems to modulate the metabolism of fat tissue in different animalmodels.Dysfunction of adipose tissue metabolism has a negative impact

on lipid and glucose homeostasis and plays a crucial role in thedevelopment of insulin resistance in obesity (Goossens, 2008;Guilherme et al., 2008; Trayhurn, 2013). As in obese individuals,HFD-fed animals also exhibited a decline in insulin sensitivity, asindicated by an increase in HOMA index. This was associated withaltered expression of several mediators involved in the control ofmetabolic functions in the adipose tissue, including adiponectin,GLUT4, SOCS3 and DPP4, as previously described in both humansand animal models (Emanuelli et al., 2001; Kadowaki et al., 2006;Kanatani et al., 2007; Kern et al., 2003; Lamers et al., 2011;Palanivel et al., 2012; Sell et al., 2013; Shepherd and Kahn, 1999;Wang et al., 2000), and with a significant hypertriglyceridemia, a

sign frequently observed in the setting of obesity linked to insulinresistance (Eckel, 2011; Taskinen et al., 2011).

More interestingly, we have observed that inhibition of LOXactivity attenuates insulin resistance in obese animals. In fact,BAPN reduced insulin and glucose levels and consequently HOMAindex, prevented the upregulation of plasma triglycerides andameliorated the changes in protein expression observed in theadipose tissue from HFD rats. Inhibition of LOX activity reversedthe decrease in GLUT4 and adiponectin levels, and the increase inboth SOCS3 and DPP4 expression. In view of the low expressionof LOXL isoenzymes in human adipose tissue, these data suggestthat the upregulation of LOX observed in obesity has multipleconsequences for the physiology of adipose tissue. This functionalrepercussion of LOX upregulation in systemic and adipose tissuemetabolism has been suggested previously in a transgenic mousemodel that specifically overexpresses HIF-1α in the adipose tissue

Fig. 3. Effects of lysyl oxidase inhibition on body weight evolution, adipocyte size and distribution and total collagen content in epididymal adiposetissue from control and obese rats. (A) Body weight evolution along the study. (B) Adipocyte area of epididymal adipose tissue analysed in haematoxylin-eosin-stained sections from control rats fed a normal chow (CT) and rats fed a high-fat diet (HFD) treated with vehicle or with the inhibitor of LOX activity(β-aminopropionitrile, BAPN, 100 mg/kg/day). (C) Representativemicrophotographs of sections stained with haematoxylin and eosin and then examined by usinglight microscopy (magnification 40×). (D) Size distribution of adipocytes from epididymal adipose tissue of the animals described in B. (E) Quantification ofcollagen volume fraction (CVF) normalised for the number of adipocytes analysed in epididymal adipose tissue sections stained with Picrosirius Red.(F) Representative microphotographs of sections stained with Picrosirius Red that were examined by using light microscopy (magnification 40×). Scale bars:50 µm. Values are mean±s.d. of six to eight animals. *P<0.05, **P<0.01, ***P<0.001 vs control group. †P<0.05, ††P<0.01 vs HFD group.

547

RESEARCH ARTICLE Disease Models & Mechanisms (2015) 8, 543-551 doi:10.1242/dmm.020107

Disea

seModels&Mechan

isms

(Halberg et al., 2009). In agreement with our data, the authors of thatstudy demonstrated that LOX contributes, at least in part, to themetabolic disturbances observed in the model, which wereimproved by the administration of BAPN. Therefore, our dataconfirm and extend previous findings on the relevance of LOX inadipocyte function.The role of LOX in the control of adipocyte function has been

poorly characterised. LOX is downregulated during adipocytedifferentiation (Dimaculangan et al., 1994), and our data shows thatinhibition of LOX impacts adipocyte homeostasis and improvesinsulin resistance in the well-established model of TNFα-inducedinsulin resistance. In fact, BAPN normalised the expression ofGLUT4, adiponectin and SOCS3, and improved glucose uptake,supporting a direct and more complex role of LOX in adipose tissuephysiology. Beyond its role in EMC maturation, LOX exhibits other

cellular functions, including the control of epithelial-to-mesenchymaltransition, cell proliferation, migration, adhesion and transformation(Huang et al., 2009). The molecular mechanisms underlying theconsequences of LOX inhibition in local and systemicmetabolism areunknown and, in view of the complex biology of this enzyme, furtherresearch is warranted.

In summary, this study demonstrates that LOX could play arelevant role in the metabolic dysfunction induced by obesity andsuggests that the inhibition of LOX activity could be a valuablestrategy to ameliorate obesity-related metabolic disturbances. Theseaspects highlight the benefits as well as the complexity of targetingfibrosis as a potential therapeutic intervention in obesity.

MATERIALS AND METHODSSubjectsVisceral adipose tissue was obtained from five individuals that had beenreferred for bariatric surgery (all women). Inclusion criteria were:age≥18 years and universally accepted indications: long-term obesity(more than 4 years); BMI≥40 despite other weight loss strategies orBMI≥35 in the presence of obesity-related comorbidities (diabetes mellitus,obesity hypoventilation syndrome, obstructive sleep apnoea syndrome andhypertension). Exclusion criteria were: age>60 years and unacceptablesurgical risk. Visceral adipose tissue was also obtained from fivenormoweight BMI≤25 individuals that had been referred to non-bariatricsurgery who were included as controls (40% women). The BMI wascalculated according to the formula: weight (kg)/square of the height(metres). In accordance with institutional guidelines, all individuals gaveinformed consent. The study was performed in accordance with theDeclaration of Helsinki, and the Ethical Committee of the FuenlabradaUniversity Hospital approved the protocol.

AnimalsMale Wistar rats of 150 g (Harlan Ibérica, Barcelona, Spain) were fed eithera HFD (33.5% fat; Harlan Teklad no. TD.03307, Haslett, MI, USA; n=16)or a standard diet (3.5% fat; Harlan Teklad no. TD.2014, Haslett, MI, USA;n=16) for 6 weeks. Half of the animals of each group received theirreversible inhibitor of LOX activity BAPN (100 mg/kg/day; Sigma-

Table 1. Effect of the inhibition of lysyl oxidase activity with β-aminopropionitrile (BAPN; 100 mg/kg/day) on general characteristicsand metabolic parameters in rats fed a normal chow (CT) and rats feda high-fat diet (HFD)

CT BAPN HFD HFD+BAPN

BW (g) 320±24 304±15 374±11*** 332±24††

FI (g/day) 20±5 23±2 13±3*** 14±2**EAT (mg/cm/tibia) 0.09±0.03 0.07±0.03 0.25±0.05** 0.19±0.04**,†

LAT (mg/cm/tibia) 0.11±0.03 0.09±0.03 0.3±0.05*** 0.23±0.05**,†

Adiposity (%) 2.3±0.8 2.4±0.8 6.1±1.3** 4.4±1.0*,†

Glucose (mg/dl) 97±10 91±9 115±6* 97±8†

Insulin (pg/ml) 64±20 50±13 184±91 79±39HOMA index 2.1±0.6 1.6±0.5 7.6±3.8* 2.4±1.1†

TG (mg/dl) 50±11 52±7 71±7* 51±10†

TC (mg/dl) 55±7 55±11 61±7 53±6

BW, body weight; EAT, epididymal adipose tissue; FI, food intake; HOMAindex, the homeostasis model assessment; LAT, lumbar adipose tissue;TG, triglycerides; TC, total cholesterol. Data values represent mean±s.d. ofeight animals. *P<0.05, **P<0.01, ***P<0.001 compared with control group.†P<0.05, ††P<0.01 compared with HFD group.

Fig. 4. Effects of lysyl oxidaseinhibition on the protein levels offactors involved in the control ofinsulin sensitivity in epididymaladipose tissue from control andobese rats. Protein levels of (A)glucose transporter type 4 (GLUT4),(B) adiponectin (Adipo), (C)dipeptidylpeptidase 4 (DPP4) and (D)suppressor of cytokine signaling 3(SOCS3) in epididymal adipose tissuefrom control rats fed a normal chow (CT)and rats fed a high-fat diet (HFD) thathad been treated with vehicle or with theinhibitor of lysyl oxidase (LOX) activity(β-aminopropionitrile, BAPN, 100 mg/kg/day). Bars on graphs represent themean±s.d. of six to eight animals, valueswere normalised to the expression of β-actin. *P<0.05, **P<0.01 vs controlgroup. †P<0.05, ††P<0.01 vs HFDgroup. ‘C’, control; ‘B’, BAPN; ‘H’, HFD;‘B/H’, HFD+BAPN.

548

RESEARCH ARTICLE Disease Models & Mechanisms (2015) 8, 543-551 doi:10.1242/dmm.020107

Disea

seModels&Mechan

isms

Aldrich, St Louis, MO, USA) in the drinking water for the same period, aspreviously described (Brasselet et al., 2005). The amount of BAPNeffectively taken daily per animal was calculated from the amount of waterconsumed on a daily basis. Animal weight was periodically controlled toadjust the target dose of BAPN. Food and water intake were determinedthroughout the experimental period. Animals were fasted the day beforeeuthanasia by anaesthesia with a cocktail of ketamine (Imalgene 1000,70 mg/kg; intraperitoneal, Merial Laboratorios, Barcelona, Spain) andxilacine (Rompun 2%, 6 mg/kg; Bayer Hispania, Barcelona, Spain). Serumand plasma were collected and abdominal adipose tissue was dissected forfurther analysis. Adiposity index was calculated as: sum of fat pads/[(bodyweight-fat pad weight)×100]. The Animal Care and Use Committee ofUniversidad Complutense de Madrid approved all experimental proceduresaccording to the Spanish Policy for Animal Protection RD53/2013, whichmeets the European Union Directive 2010/63/UE.

Real-time PCRTotal RNAwas isolated using Ultraspec™ (Biotecx, Houston, TX, USA) andwas reverse transcribed into cDNA using the High Capacity cDNA ReverseTranscription kit (Applied Biosystems, Thermo Fisher Scientific Inc,Waltham, MA, USA). Quantification of mRNA levels was performed byusing real-time PCR and an ABI PRISM 7900HT sequence detection system(Applied Biosystems, Thermo Fisher Scientific Inc,Waltham,MA, USA) andTaqMan™ gene expression assays-on-demand (Applied Biosystems, ThermoFisher Scientific Inc,Waltham,MA, USA) for rat (Rn00566984_m1; RefSeq:NM_017061.2) and human LOX (Hs00184700_m1; RefSeq:NM_002317.5),human LOXL1 (Hs00173746_m1; RefSeq: NM_005576.2), human LOXL2(Hs00158757_m1; RefSeq: NM_002318.2), human LOXL3 (Hs00261671_m1; RefSeq: NM_032603.2) and human LOXL4 (Hs00260059_m1;RefSeq:NM_032211.6).RatBMP1mRNAwasdeterminedbyusing real-timePCR with SYBR Green PCR Master mix (Applied Biosystems, ThermoFisher Scientific Inc, Waltham, MA, USA) and the following primers

(www.Biomers.net, Donau,Germany): 5′-GTCCTTCACGACAACAAAC-3′and 5′-GGGTACTTGTCAGGCCAGTT-3′ (positions 2331 to 2434corresponding to RefSeq: NM_031323.1; amplicon length: 103). 18S rRNA(4319413E;RefSeq:X03205.1) was used as an endogenous control for humansamples. Results from rat samples were normalised according to expression ofadenine phosphoribosyltransferase (Aprt) or cyclophilin using SYBR Greenand specific oligonucleotides (www.Biomers.net, Donau, Germany) asfollows: rat Aprt 5′-CGGGCGTGCTGTTCAGGGAT-3′ and 5′-TCAGGT-GACCGGCCAGGAGG-3′ (positions 92 to 309 corresponding to RefSeq:L04970.1; amplicon length: 217); rat cyclophilin 5′-TCACCAGGGGAGAT-GGCACAGG-3′ and 5′-CCATGCTCACCCATCCGGGC-3′ (positions 319to 419 corresponding to RefSeq: NM_022536.2; amplicon length: 101).Similar results were obtained after normalisation to both housekeeping genes.The ΔΔCt (or ΔΔCq following MIQE nomenclature) method was used fornormalisation. In order to appropriately apply this method, we checked thatboth the target and the reference genes were amplified with comparableefficiencies (>95%, calculated on the basis of the slope of calibration curves).Each sample was amplified in duplicate.

Morphological and histological evaluationVisceral and epididymal adipose tissue sampleswere dehydrated, embedded inparaffin and cut into 5-μm-thick sections. Sections were stained withPicrosirius Red in order to detect collagen fibres. The area of pericellularfibrosis was identified as the ratio of collagen deposition to the total tissue areaafter excluding the vessel area from the region of interest. This value wasnormalised by the number of adipocytes. Adipocytes (80-100 per subject oranimal) with intact cellular membranes were chosen for determination of thecross-sectional area in hematoxylin-eosin stained sections. For each sample,10-15 fields were analysed using a 40× objective (Leica DM 2000; LeicaCamera AG, Wetzlar, Germany) and quantified (Leica Q550 IWB; LeicaCamera AG, Wetzlar, Germany). A single researcher that was unaware of theexperimental groups performed the analysis.

Fig. 5. LOX inhibition influencesglucose uptake and the expressionof factors involved in the control ofinsulin sensitivity in an in vitromodel of insulin resistance. Proteinlevels of (A) glucose transporter type4 (GLUT4), (B) adiponectin (Adipo)and (C) suppressor of cytokinesignaling 3 (SOCS3). 3T3-L1adipocytes were stimulated with orwithout TNFα (1.15 and 2.87 nM,equivalent to 20 and 50 ng/ml,respectively) for 72 h in the presenceof either vehicle orβ-aminopropionitrile, an inhibitor ofLOX activity (BAPN, 200 µM). Datathat were normalised to theexpression of β-actin are expressedas mean±s.d. of four assays inarbitrary units. (D) Insulin-stimulatedglucose uptake in 3T3-L1 adipocytes.Bars on graphs represent themean±s.d. of the counts of [3H]-2-deoxyglucose normalised to theamount of total protein. Values aremean±s.d. of four assays. *P<0.05,**P<0.01 vs unstimulated cells(TNFα=0); †P<0.05, ††P<0.01 vs cellsstimulated with the sameconcentration of TNFα in the absenceof BAPN.

549

RESEARCH ARTICLE Disease Models & Mechanisms (2015) 8, 543-551 doi:10.1242/dmm.020107

Disea

seModels&Mechan

isms

Immunohistological evaluationSections were deparaffinised in xylene and rehydrated in graded ethanolsolutions. After blocking, slides were incubated with antibodies againstLOX (Abcam, Cambridge, UK). Afterwards, samples were incubated witha biotinylated secondary antibody (Vector Laboratories Inc, Burlingame,CA, USA), washed and treated with the Vectastain (ABC) avidin-biotinperoxidase complex (Vector Laboratories Inc, Burlingame, CA, USA).Colour was developed using 3,3′-diaminobenzidine (DAB) and sectionswere counterstained with haematoxylin. Negative controls, in which theprimary antibodywas omitted, were included to test for non-specific binding(Orriols et al., 2014).

Cell culture and differentiationMurine 3T3-L1 preadipocytes were cultured to confluence in Dulbecco’smodified Eagle’s medium (DMEM; Life Technologies, Thermo FisherScientific Inc, Waltham, MA, USA) that was supplemented with 10% (v/v)calf serum (Biological Industries, Kibbutz Beit-Haemek, Israel). At 2 dayspost-confluence (designated day 0), cells were induced to differentiate withDMEM containing a standard induction cocktail of 10% (v/v) fetal bovineserum (FBS; Biological Industries, Kibbutz Beit-Haemek, Israel), 1 μMdexamethasone, 0.5 mM 3-isobutyl-1-methylxanthine (IBMX) and 100 nMinsulin (all from Sigma-Aldrich, St Louis, MO, USA). After 48 h, thismedium was replaced with DMEM supplemented with 10% FBS and100 nM of insulin. This medium was changed every 48 h. As a model ofinsulin resistance, murine TNFα (20 or 50 ng/ml, equivalent to 1.15 and2.87 nM, respectively, Sigma-Aldrich, St Louis, MO, USA) was added tothe cell culture medium 7 days after the induction of differentiation, whenmore than 95% of the cells had the morphological and biochemicalproperties of adipocytes. For treatment with TNFα (72 h), fullydifferentiated 3T3-L1 adipocytes were treated every 24 h with thecytokine as previously described (Stephens et al., 1997). BAPN (200 µM)was added to TNFα-treated cells during the last 24 h of incubation. Thistreatment did not induce cytotoxicity, as analysed by using the XTT-basedassay for cell viability (Roche Diagnostics, Indianapolis, IN, USA).

Preparation of whole cell and tissue extracts3T3-L1 preadipocytes were washed with PBS and cell monolayers wereharvested in a non-denaturing buffer containing 150 mM NaCl, 10 mM Tris,pH 7.4, 1 mM EGTA, 1 mM EDTA, 1% Triton X-100, 0.5% Nonidet P-40,1 mM Na3VO4, 1 µg/ml leupeptin and 1 mM DTT. Samples were extractedfor 30 min on ice and centrifuged at 16,000 g at 4°C for 15 min. Supernatantswere analysed for protein content using the BCA kit (Thermo FisherScientific Inc, Waltham, MA, USA). Total protein from epididymal adiposetissue was obtained by homogenisation in lysis buffer and centrifugation for5 min at 12,700 g (4°C). The tissue extract was then separated from thefat and cellular debris, and analysed for protein content.

Western blottingCell and tissue lysates were separated using SDS-PAGE and transferredto 0.45-μm polyvinylidene difluoride membranes (Immobilon, MerckMillipore, Darmstadt, Germany). Blots were incubated with antibodiesagainst GLUT4 (Santa Cruz Biotechnology Inc, Heidelberg, Germany),adiponectin (Chemicon, Merck Millipore, Darmstadt, Germany), DDP4(Abcam, Cambridge, UK), SOCS3 (Cell Signaling Technology Inc,Danvers, MA, USA) and LOX (Abcam, Cambridge, UK). Boundantibodies were detected after incubation with a horseradish peroxidase(HRP)-conjugated IgG and then using the Super SignalWest Dura ExtendedDuration Substrate (Thermo Fisher Scientific Inc, Waltham, MA, USA).Equal loading of protein in each lane was verified by Ponceau staining andby western blotting for β-actin (Sigma).

Glucose uptake measurementsFully differentiated 3T3-L1 adipocytes were insulin-deprived and pre-treatedwith BAPN in the presence or absence of TNFα for 24 h. Adipocytes wereserum-deprived for 3 h in DMEM supplemented with 2% of fatty-acid-freebovine serum albumin (BSA) with or without TNFα and BAPN. Serum-freemedium was then removed and the cells were washed with 1 ml of Krebs-Ringer-HEPES (KRH) buffer pH 7.4 plus 0.2% BSA. Glucose uptake was

initiated with 0.9 ml of KRH buffer containing 100 mM insulin for 30 minfollowed by the addition of 100 µM 2-deoxy-D-glucose and 1 µCi/ml of [3H]-2-deoxy-D-glucose/ml. After 10 min, cells were washed with an ice-coldsolution of 50 mMD-glucose in PBS three times. Cells were lysedwith a buffercontaining 0.5 N NaOH and 0.1% sodium dodecyl sulfate (SDS), and theradioactivity retained by the cell lysatewasmeasuredusing a liquid scintillationcounter (LS 6500, Beckman Coulter, Fullerton, CA, USA). Measurementsweremade in triplicate and corrected for nonspecific diffusion. Counts of [3H]-2-deoxyglucose were normalised to protein levels.

Statistical analysisData are expressed as means and standard deviations. Normality ofdistributions and homogeneity of the variance were verified by means ofeither the Kolmogorov–Smirnov or Levene’s test. Either Pearson or Spearmancorrelation analysis was used to examine the association among differentvariables according to whether they were normally distributed or not,respectively. Differences between two groups were analysed by eitherunpaired Student’s t-test test or Mann–Whitney as parametric and non-parametric tests, respectively. Specific differences between more groups wereanalysed using Kruskall–Wallis followed by Dunn’s test for non-normaldistribution variables. For normal distribution variables, one-way analysis ofvariance followed by Newman–Keuls or Tamahane tests were used whetherthe variance was homogenous or not, respectively. Data analysis wasperformed using the statistical program SPSS version 22.0 (SPSS Inc.,Chicago, IL, USA). The predetermined significance level was α=0.05.

AcknowledgementsWe thank Encarna Mun oz-Ferrero, Silvia Aguilo and Blanca Martınez for theirtechnical help, Anthony DeMarco for his help in editing the manuscript and CristinaFernandez for her help in the statistical analysis of the data.

Competing interestsThe authors declare no competing or financial interests.

Author contributionsV.C. and C.R. designed the study, performed experiments and data analysis andwrote the manuscript. M.M. and M.G. performed experiments and data analysis andhelped to write the manuscript. E.M.-M., S.V., R.J.-L. and B.B.-M. performedexperiments and contributed to the discussion of the manuscript. J.M.-G.contributed to the discussion and the writing of the manuscript. A.A. and M.L. wereresponsible for the clinical study and contributed to discussion of the manuscript.V.C. and C.R. had full access to all the data in the study and take responsibility for theintegrity of the data and the accuracy of the data analysis.

FundingThis work was supported by Plan Estatal I+D+I 2013-2016 [PI12/01729, PI12/01952] and Red de Investigacion Cardiovascular [RD12/0042/0033 and RD12/0042/0053] of the Instituto de Salud Carlos III (ISCIII), as well as the Ministerio deEconomıa y Competitividad (MINECO) [SAF2012-40127]. The study was cofundedby Fondo Europeo de Desarrollo Regional (FEDER), a way to make Europe. M.G.was supported by the Sara Borrell Program from ISCIII (CD12/00589).

ReferencesAbdennour, M., Reggio, S., Le Naour, G., Liu, Y., Poitou, C., Aron-Wisnewsky, J.,

Charlotte, F., Bouillot, J.-L., Torcivia, A., Sasso, M. et al. (2014). Association ofadipose tissue and liver fibrosis with tissue stiffness in morbid obesity: links withdiabetes and BMI loss after gastric bypass. J. Clin. Endocrinol. Metab. 99, 898-907.

Ahsan, T., Lottman, L. M., Harwood, F., Amiel, D. and Sah, R. L. (1999). Integrativecartilage repair: inhibition by beta-aminopropionitrile. J. Orthop. Res. 17, 850-857.

Alligier, M., Meugnier, E., Debard, C., Lambert-Porcheron, S., Chanseaume, E.,Sothier, M., Loizon, E., Hssain, A. A., Brozek, J., Scoazec, J.-Y. et al. (2012).Subcutaneous adipose tissue remodelling during the initial phase of weight gaininduced by overfeeding in humans. J. Clin. Endocrinol. Metab. 97, E183-E192.

Brasselet, C., Durand, E., Addad, F., Al Haj Zen, A., Smeets, M. B., Laurent-Maquin, D., Bouthors, S., Bellon, G., de Kleijn, D., Godeau, G. et al. (2005).Collagen and elastin cross-linking: a mechanism of constrictive remodelling afterarterial injury. Am. J. Physiol. Heart Circ. Physiol. 289, H2228-H2233.

Dasler, W. (1954). Isolation of toxic crystals from sweet peas (Lathyrus odoratus).Science 120, 307-308.

Dimaculangan, D. D., Chawla, A., Boak, A., Kagan, H. M. and Lazar, M. A.(1994). Retinoic acid prevents downregulation of ras recision gene/lysyl oxidaseearly in adipocyte differentiation. Differentiation 58, 47-52.

Divoux, A., Tordjman, J., Lacasa, D., Veyrie, N., Hugol, D., Aissat, A.,Basdevant, A., Guerre-Millo, M., Poitou, C., Zucker, J.-D. et al. (2010).

550

RESEARCH ARTICLE Disease Models & Mechanisms (2015) 8, 543-551 doi:10.1242/dmm.020107

Disea

seModels&Mechan

isms

Fibrosis in human adipose tissue: composition, distribution, and link with lipidmetabolism and fat mass loss. Diabetes 59, 2817-2825.

Eckel, R. H. (2011). The complex metabolic mechanisms relating obesity tohypertriglyceridemia. Arterioscler. Thromb. Vasc. Biol. 31, 1946-1948.

Emanuelli, B., Peraldi, P., Filloux, C., Chavey, C., Freidinger, K., Hilton, D. J.,Hotamisligil, G. S. and Van, O. E. (2001). SOCS-3 inhibits insulin signaling andis up-regulated in response to tumor necrosis factor-alpha in the adipose tissue ofobese mice. J. Biol. Chem. 276, 47944-47949.

Erler, J. T., Bennewith, K. L., Nicolau, M., Dornhofer, N., Kong, C., Le, Q.-T., Chi,J.-T. A., Jeffrey, S. S. and Giaccia, A. J. (2006). Lysyl oxidase is essential forhypoxia-induced metastasis. Nature 440, 1222-1226.

Goossens, G. H. (2008). The role of adipose tissue dysfunction in the pathogenesisof obesity-related insulin resistance. Physiol. Behav. 94, 206-218.

Guilherme, A., Virbasius, J. V., Puri, V. and Czech, M. P. (2008). Adipocytedysfunctions linking obesity to insulin resistance and type 2 diabetes. Nat. Rev.Mol. Cell Biol. 9, 367-377.

Halberg, N., Khan, T., Trujillo, M. E., Wernstedt-Asterholm, I., Attie, A. D.,Sherwani, S., Wang, Z. V., Landskroner-Eiger, S., Dineen, S., Magalang, U. J.et al. (2009). Hypoxia-inducible factor 1alpha induces fibrosis and insulinresistance in white adipose tissue. Mol. Cell. Biol. 29, 4467-4483.

Henegar, C., Tordjman, J., Achard, V., Lacasa, D., Cremer, I., Guerre-Millo, M.,Poitou, C., Basdevant, A., Stich, V., Viguerie, N. et al. (2008). Adipose tissuetranscriptomic signature highlights the pathological relevance of extracellularmatrix in human obesity. Genome Biol. 9, R14.

Hornstra, I. K., Birge, S., Starcher, B., Bailey, A. J., Mecham, R. P. and Shapiro,S. D. (2003). Lysyl oxidase is required for vascular and diaphragmaticdevelopment in mice. J Biol Chem. 278, 14387-14393.

Huang, H., Song, T.-J., Li, X., Hu, L., He, Q., Liu, M., Lane, M. D. and Tang, Q.-Q.(2009).BMPsignalingpathway is required forcommitment ofC3H10T1/2pluripotentstem cells to the adipocyte lineage. Proc. Natl. Acad. Sci. USA 106, 12670-12675.

Kadowaki, T., Yamauchi, T., Kubota, N., Hara, K., Ueki, K. and Tobe, K. (2006).Adiponectin and adiponectin receptors in insulin resistance, diabetes, and themetabolic syndrome. J. Clin. Invest. 116, 1784-1792.

Kanatani, Y., Usui, I., Ishizuka, K., Bukhari, A., Fujisaka, S., Urakaze, M., Haruta,T., Kishimoto, T., Naka, T. and Kobayashi, M. (2007). Effects of pioglitazone onsuppressor of cytokine signaling 3 expression: potential mechanisms for itseffects on insulin sensitivity and adiponectin expression. Diabetes 56, 795-803.

Kern, P. A., Di Gregorio, G. B., Lu, T., Rassouli, N. and Ranganathan, G. (2003).Adiponectin expression from human adipose tissue: relation to obesity, insulinresistance, and tumor necrosis factor-alpha expression. Diabetes 52, 1779-1785.

Khan, T., Muise, E. S., Iyengar, P., Wang, Z. V., Chandalia, M., Abate, N., Zhang,B. B., Bonaldo, P., Chua, S. and Scherer, P. E. (2009). Metabolic dysregulationand adipose tissue fibrosis: role of collagen VI. Mol. Cell. Biol. 29, 1575-1591.

Kothapalli, D., Liu, S.-L., Bae, Y. H., Monslow, J., Xu, T., Hawthorne, E. A.,Byfield, F. J., Castagnino, P., Rao, S., Rader, D. J. et al. (2012). Cardiovascularprotection by ApoE andApoE-HDL linked to suppression of ECMgene expressionand arterial stiffening. Cell Rep. 2, 1259-1271.

Lamers, D., Famulla, S., Wronkowitz, N., Hartwig, S., Lehr, S., Ouwens, D. M.,Eckardt, K., Kaufman, J. M., Ryden, M., Muller, S. et al. (2011). Dipeptidylpeptidase 4 is a novel adipokine potentially linking obesity to the metabolicsyndrome. Diabetes 60, 1917-1925.

Levental, K. R., Yu, H., Kass, L., Lakins, J. N., Egeblad, M., Erler, J. T., Fong,S. F. T., Csiszar, K., Giaccia, A., Weninger, W. et al. (2009). Matrix crosslinkingforces tumor progression by enhancing integrin signaling. Cell 139, 891-906.

Maki, J. M., Rasanen, J., Tikkanen, H., Sormunen, R., Makikallio, K., Kivirikko,K. I. and Soininen, R. (2002). Inactivation of the lysyl oxidase gene Lox leads toaortic aneurysms, cardiovascular dysfunction, and perinatal death in mice.Circulation 106, 2503-2509.

Martınez-Martınez, E., Jurado-Lopez, R., Valero-Mun oz, M., Bartolome, M. V.,Ballesteros, S., Luaces, M., Briones, A. M., Lopez-Andres, N., Miana, M. andCachofeiro, V. (2014a). Leptin induces cardiac fibrosis through galectin-3, mTORand oxidative stress: potential role in obesity. J. Hypertens. 32, 1104-1114.

Martınez-Martınez, E., Miana, M., Jurado-Lopez, R., Bartolome, M. V., SouzaNeto, F. V., Salaices, M., Lopez-Andres, N. and Cachofeiro, V. (2014b). Thepotential role of leptin in the vascular remodelling associated with obesity.Int. J. Obes. (Lond.) 38, 1565-1572.

Miana, M., de Las Heras, N., Rodriguez, C., Sanz-Rosa, D., Martin-Fernandez,B., Mezzano, S., Lahera, V., Martinez-Gonzalez, J. and Cachofeiro, V. (2011).Effect of eplerenone on hypertension-associated renal damage in rats: potentialrole of peroxisome proliferator activated receptor gamma (PPAR-gamma).J. Physiol. Pharmacol. 62, 87-94.

Mutch, D. M., Rouault, C., Keophiphath, M., Lacasa, D. and Clement, K. (2009).Using gene expression to predict the secretome of differentiating humanpreadipocytes. Int. J. Obes. (Lond.) 33, 354-363.

Nave, A. H., Mizıkova, I., Niess, G., Steenbock, H., Reichenberger, F., Talavera,M. L., Veit, F., Herold, S., Mayer, K., Vadasz, I. et al. (2014). Lysyl oxidases playa causal role in vascular remodelling in clinical and experimental pulmonaryarterial hypertension. Arterioscler. Thromb. Vasc. Biol. 34, 1446-1458.

Orriols, M., Guadall, A., Galan, M., Martı-Pamies, I., Varona, S., Rodrıguez-Calvo, R., Briones, A. M., Navarro, M. A., de Diego, A., Osada, J. et al. (2014).Lysyl oxidase (LOX) in vascular remodelling. Insight from a new animal model.Thromb. Haemost. 112, 812-824.

Palanivel, R., Fullerton, M. D., Galic, S., Honeyman, J., Hewitt, K. A., Jorgensen,S. B. and Steinberg, G. R. (2012). Reduced Socs3 expression in adipose tissueprotects female mice against obesity-induced insulin resistance. Diabetologia 55,3083-3093.

Raposo, B., Rodrıguez, C., Martınez-Gonzalez, J. and Badimon, L. (2004). Highlevels of homocysteine inhibit lysyl oxidase (LOX) and downregulate LOXexpression in vascular endothelial cells. Atherosclerosis 177, 1-8.

Rodrıguez, C., Raposo, B., Martınez-Gonzalez, J., Casanı, L. and Badimon, L.(2002). Low density lipoproteins downregulate lysyl oxidase in vascularendothelial cells and the arterial wall. Arterioscler. Thromb. Vasc. Biol. 22,1409-1414.

Rodrıguez, C., Martinez-Gonzalez, J., Raposo, B., Alcudia, J. F., Guadall, A.and Badimon, L. (2008a). Regulation of lysyl oxidase in vascular cells: lysyloxidase as a new player in cardiovascular diseases. Cardiovasc. Res. 79, 7-13.

Rodrıguez, C., Rodrıguez-Sinovas, A. and Martınez-Gonzalez, J. (2008b). Lysyloxidase as a potential therapeutic target. Drug News Perspect. 21, 218-224.

Rodriguez, C., Alcudia, J. F., Martinez-Gonzalez, J., Guadall, A., Raposo, B.,Sanchez-Gomez, S. and Badimon, L. (2009). Statins normalize vascular lysyloxidase down-regulation induced by proatherogenic risk factors.Cardiovasc. Res.83, 595-603.

Sell, H., Bluher, M., Kloting, N., Schlich, R., Willems, M., Ruppe, F., Knoefel,W. T., Dietrich, A., Fielding, B. A., Arner, P. et al. (2013). Adipose dipeptidylpeptidase-4 and obesity: correlation with insulin resistance and depot-specificrelease from adipose tissue in vivo and in vitro. Diabetes Care. 36, 4083-4090.

Sethi, J. K. and Hotamisligil, G. S. (1999). The role of TNF alpha in adipocytemetabolism. Semin. Cell Dev. Biol. 10, 19-29.

Shepherd, P. R. and Kahn, B. B. (1999). Glucose transporters and insulin action–implications for insulin resistance and diabetes mellitus. N. Engl. J. Med. 341,248-257.

Spencer, P. S. and Schaumburg, H. H. (1983). Lathyrism: a neurotoxic disease.Neurobehav. Toxicol. Teratol. 5, 625-629.

Stefanon, I., Valero-Mun oz, M., Fernandes, A. A., Ribeiro, R. F., Jr, Rodrıguez,C., Miana, M., Martınez-Gonzalez, J., Spalenza, J. S., Lahera, V., Vassallo,P. F. et al. (2013). Left and right ventricle late remodelling following myocardialinfarction in rats. PLoS ONE 8, e64986.

Stephens, J. M., Lee, J. and Pilch, P. F. (1997). Tumor necrosis factor-alpha-induced insulin resistance in 3T3-L1 adipocytes is accompanied by a loss ofinsulin receptor substrate-1 and GLUT4 expression without a loss of insulinreceptor-mediated signal transduction. J. Biol. Chem. 272, 971-976.

Tam, C. S., Tordjman, J., Divoux, A., Baur, L. A. and Clement, K. (2012). Adiposetissue remodelling in children: the link between collagen deposition and age-related adipocyte growth. J. Clin. Endocrinol. Metab. 97, 1320-1327.

Tang, S. S., Trackman, P. C. and Kagan, H. M. (1983). Reaction of aortic lysyloxidase with beta-aminopropionitrile. J. Biol. Chem. 258, 4331-4338.

Taskinen, M.-R., Adiels, M., Westerbacka, J., Soderlund, S., Kahri, J.,Lundbom, N., Lundbom, J., Hakkarainen, A., Olofsson, S.-O., Orho-Melander, M. et al. (2011). Dual metabolic defects are required to producehypertriglyceridemia in obese subjects. Arterioscler. Thromb. Vasc. Biol. 31,2144-2150.

Trayhurn, P. (2013). Hypoxia and adipose tissue function and dysfunction inobesity. Physiol. Rev. 93, 1-21.

Turecek, C., Fratzl-Zelman, N., Rumpler, M., Buchinger, B., Spitzer, S., Zoehrer,R., Durchschlag, E., Klaushofer, K., Paschalis, E. P. and Varga, F. (2008).Collagen cross-linking influences osteoblastic differentiation. Calcif. Tissue Int.82, 392-400.

Virtue, S. and Vidal-Puig, A. (2010). Adipose tissue expandability, lipotoxicity andthe Metabolic Syndrome–an allostatic perspective. Biochim. Biophys. Acta 1801,338-349.

Walker, R. W., Allayee, H., Inserra, A., Fruhwirth, R., Alisi, A., Devito, R., Carey,M. E., Sinatra, F., Goran, M. I. and Nobili, V. (2014). Macrophages and fibrosis inadipose tissue are linked to liver damage and metabolic risk in obese children.Obesity(Silver Spring) 22, 1512-1519.

Wang, Z., Zhou, Y.-T., Kakuma, T., Lee, Y., Kalra, S. P. and Kalra, P. S. (2000).Leptin resistance of adipocytes in obesity: role of suppressors of cytokinesignaling. Biochem. Biophys. Res. Commun. 277, 20-26.

551

RESEARCH ARTICLE Disease Models & Mechanisms (2015) 8, 543-551 doi:10.1242/dmm.020107

Disea

seModels&Mechan

isms