Heme Oxygenase-1 DrivesMetaflammation and Insulin Resistance inMouse and ManAlexander Jais,1,12 Elisa Einwallner,1,12 Omar Sharif,1,2 Klaus Gossens,3 Tess Tsai-Hsiu Lu,3 Selma M. Soyal,4

David Medgyesi,3,5 Daniel Neureiter,4 Jamile Paier-Pourani,6 Kevin Dalgaard,3 J. Catharina Duvigneau,7

Josefine Lindroos-Christensen,1 Thea-Christin Zapf,1,11 Sabine Amann,1 Simona Saluzzo,1,2 Florian Jantscher,1

Patricia Stiedl,8 Jelena Todoric,1 Rui Martins,1,2 Hannes Oberkofler,4 Simone Muller,7 Cornelia Hauser-Kronberger,4

Lukas Kenner,1,7,8 Emilio Casanova,1,8 Hedwig Sutterluty-Fall,1 Martin Bilban,1 Karl Miller,9 Andrey V. Kozlov,6

Franz Krempler,9 Sylvia Knapp,1,2 Carey N. Lumeng,10 Wolfgang Patsch,4 Oswald Wagner,1 J. Andrew Pospisilik,3,*and Harald Esterbauer1,*1Medical University of Vienna, 1090 Vienna, Austria2CeMM, Research Center for Molecular Medicine of the Austrian Academy of Sciences, 1090 Vienna, Austria3Max Planck Institute of Immunobiology and Epigenetics, 79108 Freiburg, Germany4Paracelsus Medical University, 5020 Salzburg, Austria5BIOSS Centre of Biological Signalling Studies, University of Freiburg, 79104 Freiburg, Germany6Ludwig Boltzmann Institute for Experimental and Clinical Traumatology, 1200 Vienna, Austria7University of Veterinary Medicine Vienna, 1210 Vienna, Austria8Ludwig Boltzmann Institute for Cancer Research, 1090 Vienna, Austria9General Hospital Hallein, 5400 Hallein, Austria10University of Michigan, Ann Arbor, MI 48109, USA11Present address: Phillips University Marburg, 35043 Marburg, Germany12Co-first authors*Correspondence: pospisilik@ie-freiburg.mpg.de (J.A.P.), harald.esterbauer@meduniwien.ac.at (H.E.)

http://dx.doi.org/10.1016/j.cell.2014.04.043

SUMMARY

Obesity and diabetes affect more than half abillion individuals worldwide. Interestingly, thetwo conditions do not always coincide and themolecular determinants of ‘‘healthy’’ versus ‘‘un-healthy’’ obesity remain ill-defined. Chronic meta-bolic inflammation (metaflammation) is believed tobe pivotal. Here, we tested a hypothesized anti-inflammatory role for heme oxygenase-1 (HO-1)in the development of metabolic disease. Surpris-ingly, in matched biopsies from ‘‘healthy’’ versusinsulin-resistant obese subjects we find HO-1 tobe among the strongest positive predictors ofmetabolic disease in humans. We find that hepa-tocyte and macrophage conditional HO-1 deletionin mice evokes resistance to diet-induced insulinresistance and inflammation, dramatically reducingsecondary disease such as steatosis and livertoxicity. Intriguingly, cellular assays show thatHO-1 defines prestimulation thresholds for in-flammatory skewing and NF-kB amplification inmacrophages and for insulin signaling in hepa-tocytes. These findings identify HO-1 inhibition asa potential therapeutic strategy for metabolicdisease.

INTRODUCTION

The American Medical Association recently voted to recognize

obesity as a disease (http://www.ama-assn.org). Interestingly,

up to one in four individuals currently labeled as obese are actu-

ally metabolically healthy (Stefan et al., 2013), scoring normal on

indices of insulin resistance and systemic inflammation. Impor-

tantly, the factors determining ‘‘healthy’’ versus ‘‘unhealthy’’

obesity remain ill-defined.

Adipose tissuemass, and in particular ratios of ‘‘good’’ subcu-

taneous versus ‘‘bad’’ visceral depots are strongly correlated

with disease; whereas visceral fat expansion is associated with

metabolic disease (Tchkonia et al., 2013), expansion associated

with insulin sensitizers (Virtue and Vidal-Puig, 2008), transplanta-

tion of subcutaneous adipose (Gavrilova et al., 2000), and even

massive subcutaneous expansion (Kusminski et al., 2012) have

all been shown to be disease sparing. Indeed, pronounced

loss of white fat is—with rare exceptions (Pospisilik et al.,

2010)—linked to insulin resistance (Garg, 2011). Using insulin

sensitivity as the hallmark of healthy obesity several controlling

factors have emerged (Johnson and Olefsky, 2013). Visceral

adipocyte hypertrophy and associated hypoxia and fibrosis are

detrimental (Sun et al., 2013). Saturated free fatty acids and

ectopic lipid storage in liver and muscle impair insulin action

(Samuel and Shulman, 2012). Further, reactive oxygen species

(Houstis et al., 2006), mitochondrial function (Szendroedi et al.,

2012), endoplasmic reticulum stress (Hotamisligil, 2010), and

gut microbiota (Burcelin et al., 2012) all provide influence.

Cell 158, 25–40, July 3, 2014 ª2014 Elsevier Inc. 25

Importantly, many—if not all—insulin resistance drivers identi-

fied to date converge uniformly upon inflammation and proin-

flammatory signaling (Lumeng and Saltiel, 2011; Odegaard and

Chawla, 2013; Solinas and Karin, 2010). Such chronic inflamma-

tion, so-called ‘‘metaflammation,’’ currently provides a unifying

theory for unhealthy obesity (Gregor and Hotamisligil, 2011).

The concept of metaflammation was born out of the discovery

that tumor necrosis factor-alpha (TNF-a) potently inhibits adi-

pose insulin signaling (Hotamisligil et al., 1993). Subsequent

work implicated additional hallmark activities including inhibitor

of kB kinase (Yuan et al., 2001), c-Jun amino-terminal kinases

(Hirosumi et al., 2002), and protein kinase R (Nakamura et al.,

2010). Two landmark papers described macrophage infiltrates

in adipose as the source of insulin-resistance inducing TNF-a

(Weisberg et al., 2003; Xu et al., 2003). Such infiltrates have since

been characterized as classically activated, proinflammatory

M1-like macrophages (Lumeng et al., 2007; Solinas et al.,

2007), which have proven pivotal to the emergence of metabolic

disease (Patsouris et al., 2008). Importantly, studies now impli-

cate metaflammation as a central feature of human insulin resis-

tance (Capel et al., 2009; Qatanani et al., 2013).

Comparisons of ‘‘cold’’ metaflammation with classical ‘‘hot’’

inflammation reveal few differences (Calay and Hotamisligil,

2013). Here, we tested heme oxygenase-1 (HO-1), a historically

anti-inflammatory molecule that catalyzes heme breakdown to

CO, biliverdin, and free iron (Abraham and Kappas, 2008). In

addition to heme, HO-1 is induced by numerous stressors

including endotoxin, cytokines, and oxidants (Abraham and

Kappas, 2008). In humans, HO-1 loss-of-function leads to early

death (Yachie et al., 1999). Deletion in mice results in high peri-

natal mortality and increased susceptibility to inflammation

(Kapturczak et al., 2004). Interestingly, several studies have

shown that chemical HO-1 induction ameliorates obesity and

diabetes in rodents (Li et al., 2008; Ndisang et al., 2009). How-

ever, these results have failed validation in more specific genetic

models (Huang et al., 2013; Huang et al., 2012), and human

studies show similarly conflicting results (Bao et al., 2010; Shak-

eri-Manesch et al., 2009).

Here, using conditional mouse genetics we show that HO-1

is in fact proinflammatory, potently and independently driving

insulin resistance in the hepatic and macrophage compart-

ments. Against expectations, we find that HO-1 is among the

top predictors of metabolic disease in humans and mice and,

intriguingly, that hepatocyte and macrophage loss-of-function

leads to resistance to diet-induced metaflammation, insulin

resistance, and steatosis. The results indicate that HO-1 is in

fact necessary for the development of metaflammation and

metabolic disease, and call for a re-evaluation of numerous find-

ings in the field. They identify HO-1 as a candidate biomarker for

stratification of metabolically healthy and unhealthy obesity and

provide a framework for selective, personalized therapy.

RESULTS

HO-1 Expression Predicts Insulin Resistance in Mouseand ManOver the last years we collected a unique resource of matched

serum, liver, and adipose biopsies from well-characterized indi-

26 Cell 158, 25–40, July 3, 2014 ª2014 Elsevier Inc.

viduals for studying ‘‘healthy’’ versus ‘‘diseased’’ obesity. Re-

cruited subjects were required to be clinically obese (BMI >

30), to exhibit normal fasting glucose levels, to have no overt

inflammation, and to be free of any medications. Critically,

whereas obese insulin-resistant (obIR) individuals exhibited

clearly elevated insulin resistance (HOMA-IR R 5), obese insu-

lin-sensitive (obIS) individuals were required to show no signs

of systemic insulin resistance, i.e., HOMA-IR % 2. We thus bio-

psied 17 obIS and 27 obIR individuals well-matched for age and

BMI (Table 1) and complemented the samples with biopsies from

6 nonobese individuals. Consistent with their insulin resistance,

the obIR group displayed minimally elevated waist circumfer-

ence, visceral adiposity, triglyceride to HDL ratio, as well as

GLUT4 and HSD11B1 mRNA expression by qPCR (Table 1).

Thus, we assembled a clinical resource to probe early determi-

nants of insulin resistance in obesity.

Interestingly, and in contrast to our initial expectations, we

found elevated HO-1 expression in obIR individuals in both

the liver and visceral fat compartments (Figures 1A and 1B).

HO-1 levels correlated directly with metabolic dysregulation,

increasing progressively from lean to obIS and through to obIR

states (Figures 1A and 1B). Importantly, regression analyses

including seven ‘‘intermediate’’ insulin resistance subjects

(HOMA-IR > 2 and < 5) revealed that liver and adipose HO-1

expression explain 21.7% and 43.3% of HOMA-IR, relationships

maintained in both males and females, and independent of waist

circumference or visceral fat (Tables S1 and S2 available online).

These findings were confirmed on the protein level, and impor-

tantly, placed HO-1 within both the hepatocyte and Kupffer cell

compartments in liver and in macrophages in adipose (Figures

1C and 1D). Of note, HO-1 protein was elevated in both hepatic

compartments in obIR individuals (Figures 1C and 1E). In

adipose, HO-1 was particularly evident in inflammatory

macrophage crown-like structures (Figure 1D). Thus, liver and

adipose HO-1 are potent independent predictors of insulin resis-

tance in man.

Since a proinflammatory role for HO-1 contradicted bulk liter-

ature on the enzyme, we performed an unbiased search for data

sets to validate our findings. We identified five data sets, two in

liver and three in adipose, comparing obese and/or insulin-resis-

tant adult humans with controls, and mined them for any pattern

of HO-1 expression. Impressively, HO-1 was increased in the

metabolically compromised setting in five of five studies. HO-1

was increased in livers of obese versus lean individuals (Fig-

ure 1F) (Pihlajamaki et al., 2009) and of diabetic versus nondia-

betic subjects (Figure 1G) (Misu et al., 2010). In adipose, HO-1

was elevated in insulin-resistant obese subjects (Figure 1H)

(NCBI, GEO data set GDS3665). In a rare opportunity to control

for genetic background in the human setting, we mined adipose

expression profiles from monozygotic twins discordant for BMI

and observed HO-1 again to be consistently higher in the obese

sibling (Figure 1I) (Pietilainen et al., 2008). Finally, in a landmark

survey of adipose from 387 morbidly obese individuals, HO-1

scored 3rd of all insulin-resistance risk genes (Figure 1J) (Qata-

nani et al., 2013). Thus, liver and adipose tissue HO-1 expression

consistently predict metabolic dysfunction in humans.

To test conservation, we set out to generate a parallel data set

in mice. Similar to humans, mice exhibit substantial HO-1

Table 1. Characteristics of the Study Population

Trait Nonobese obIS obIR pa pb pc

Male/female (n) 3/3 2/15 10/17 0.088 0.089 0.090

Age (years) 40.5 (3.3) 39.2 (2.3) 38.8 (2.3) n.s. n.s. n.s.

BMI (kg/m2) 25.2 (0.9) 40.9 (1.2) 45.0 (0.9) n.d. n.d. n.s.

HOMA-IR 1.0 (0.2) 1.4 (0.1) 7.0 (0.3) n.d. 0.023 n.d.

Waist circumference (cm) 87.8 (2.8) 119.1 (3.4) 132.9 (2.9) n.d. n.d. 0.035

Hip circumference (cm) 101.2 (2.8) 130.5 (2.0) 135.1 (2.2) n.d. n.d. n.s.

Waist/hip ratio 0.87 (0.03) 0.91 (0.02) 0.99 (0.02) <0.001 0.042 0.078

SAT (cm) 2.3 (0.6) 3.1 (0.2) 3.5 (0.2) n.d. n.d. n.s.

VAT (cm) 1.8 (0.4) 5.3 (0.7) 6.9 (0.5) n.d. n.d. 0.025

Insulin (pmol/l) 31.8 (5.9) 44.9 (2.2) 201.1 (10.6) n.d. 0.027 n.d.

Glucose (mmol/l) 4.8 (0.2) 4.8 (0.1) 5.5 (0.1) n.d. n.s. n.d.

Cholesterol (mmol/l) 5.6 (0.4) 4.8 (0.2) 4.9 (0.2) n.s. n.s. n.s.

Triglycerides (mmol/l) 1.3 (0.4) 1.2 (0.1) 2.2 (0.3) 0.002 n.s. 0.006

HDL cholesterol (mmol/l) 1.8 (0.2) 1.5 (0.1) 1.2 (0.1) <0.001 n.s. 0.007

LDL cholesterol (mmol/l) 3.2 (0.2) 2.8 (0.2) 2.8 (0.2) n.s. n.s. n.s.

Triglyceride/HDL ratio 1.6 (0.4) 2.1 (0.3) 4.7 (0.7) <0.001 n.s. 0.001

C-reactive protein (mg/l) 4.5 (3.1) 5.6 (1.1) 7.2 (0.1) n.s. n.s. n.s.

GLUT4 mRNA (AU) 1.36 (0.27) 1.00 (0.10) 0.53 (0.05) <0.001 n.s. 0.001

HSD11B1 mRNA (AU) 0.63 (0.18) 1.00 (0.15) 2.28 (0.35) <0.001 n.s. <0.001

Data are numbers of observations or unadjusted and untransformed means (SEM). Two-sided p values obtained from Chi-square test or ANOVA

(Scheffe-Test for subgroup comparisons adjusted for age and sex) for categorical or continuous data, respectively.acomparison among all groups.bnonobese controls versus obese insulin-sensitive (obIS).cobIS versus obese insulin-resistant (obIR); participants (total) n = 50 (6/17/27; nonobese/obIS/obIR; n = 48 (6/16/26) for SAT and VAT; obIS, HOMA-IR

% 2.0; obIR, HOMA-IR > = 5.0; n.s., not significant; n.d., not determined; SAT, subcutaneous adipose tissue; VAT, visceral adipose tissue; AU, arbitrary

unit. See also Tables S1 and S2.

expression in the lung, kidney, and spleen and, interestingly, in

metabolic organs including the liver and white adipose tissue

(Figure S1A). To explore mouse HO-1 variation with metabolic

disease, we generated a large cohort of high-fat-diet (HFD)-

treated male C57BL/6J mice, stratified them according to

glucose (Figure S1B) and insulin tolerance (Figure S1C) and

selected subgroups exhibiting either insulin sensitivity (IS) or in-

sulin resistance (IR) despite comparable weight gain on the HFD

regimen (Figure S1D). Of note, and in complete agreement with

our human findings, HO-1 expression in liver (Figure 1K) and

adipose (Figure 1L) was directly proportional to severity of meta-

bolic dysregulation. Again, measures were confirmed on the pro-

tein level, in both liver (Figure 1M) and epididymal adipose tissue

(Figure 1N). Importantly, the obese-IR and -IS groups did not

differ in body weight (Figure S1D), and were matched for age

and developmental environment. Thus, hepatic and adipose tis-

sue HO-1 predict insulin resistance in mouse and man.

Hepatocyte HO-1 Knockout Mice Are InsulinHypersensitiveTo directly assess the role of HO-1 in the etiology of metabolic

disease, we generated mice bearing a conditional allele for

HO-1 (HO-1fl/fl) (Figures S2A and S2B). First, we crossed them

to Albumin-Cre transgenic mice (Alb-Cre) to obtain liver (hepato-

cyte)-specific HO-1 knockouts (Lhoko). Efficient deletion was

confirmed on the DNA, protein and activity levels (Figures

2A, 2B, and S2C). Functional deletion of HO-1 was confirmed

in vivo by examining HO-1 expression in whole livers upon hemin

injection, a heme analog known to induce HO-1 expression and

activity (Figure 2C). Thus, we generated hepatocyte-specific

HO-1 mutant mice.

Lhoko mice were born at Mendelian ratios, developed

normally, and exhibited normal liver morphology as well as

serum parameters (Figures S2D and S2E). Importantly, Lhoko

animals exhibited completely normal glucose and insulin toler-

ance relative to their littermate controls (Figures S2F and S2G),

showing no evidence of HO-1 imposed anti-inflammatory tone

in the naive state. Remembering that our observations of

disease predicting HO-1 were made in metabolically stressed

individuals, we challenged Lhoko mice with HFD. With respect

to adiposity, Lhoko mice showed no difference in HFD-induced

body weight gain relative to littermates (Figure S2H). Similarly,

they maintained unremarkable oral glucose tolerance (Fig-

ure 2D). Intriguingly though, insulin tolerance, fasting insulin,

and HOMA-IR all indicated significantly increased insulin sensi-

tivity in the HFD-treated Lhoko animals (Figures 2D and 2E), a

clear indication of improved metabolic health. In agreement

with these findings, serum measures of hepatotoxicity including

alanine transaminase (ALAT), aspartate transaminase (ASAT),

and ASAT/ALAT ratios were clearly improved (Figures 2F). To

understand the insulin sensitivity on a cellular level, an insulin

bolus was administered directly into the portal vein of

Cell 158, 25–40, July 3, 2014 ª2014 Elsevier Inc. 27

Figure 1. HO-1 Levels Predict Insulin Resistance in Mouse and Man

(A andB) qPCR analysis ofHO-1mRNA in (A) liver and (B) visceral adipose biopsies obtained from obese insulin-sensitive (obIS) and obese insulin-resistant (obIR)

humans.

(C and D) Immunohistochemical staining for HO-1 in (C) liver and (D) obIR visceral fat biopsies. Scale bar, 50 mm.

(E) Percentage of HO-1 positive hepatocytes in obIS and obIR liver biopsies.

(F–I) HO-1 mRNA levels obtained from public microarray data sets; (F and G) liver (H and I) adipose tissue. Numbers in bars represent study participants.

(legend continued on next page)

28 Cell 158, 25–40, July 3, 2014 ª2014 Elsevier Inc.

HFD-treated Lhoko and control animals, and liver biopsies

taken over 10 min. Intriguingly, loss of HO-1 induced stark

augmentation and acceleration of key upstream insulin

signaling activation events including insulin receptor (INSR)

and AKT phosphorylation (Figure 2G). The effects were highly

reproducible, and evident within 2 min of injection. Indicating

more widespread positive effects of hepatic HO-1 deletion,

examinations of muscle and adipose responses in the same

animals revealed moderately elevated pAKT (Figure 2G). In

keeping with these improvements in metabolic function, oil

red O staining and biochemical measures revealed attenuated

liver steatosis in HFD-treated Lhoko animals (Figures 2H and

2I). Gluconeogenic and lipogenic programs were largely unal-

tered (Figure S2I). Together, these findings comprise direct ge-

netic evidence that loss of hepatocyte HO-1 in vivo induces

proximal insulin signaling hypersensitivity and resistance to

HFD-induced metabolic sequelae.

To corroborate these data, we also performed the opposite

genetic experiment, acutely overexpressing HO-1 via tail vein

injection of adenoviral HO-1 (Figures S2J and S2K). Critically,

and in direct support of the human and Lhoko data, 1 week of

�5-fold HO-1 overexpression in unchallenged chow-fed

C57BL/6Jmice triggered glucose intolerance, insulin resistance,

and elevated HOMA-IR relative to Adeno-LacZ injected controls

(Figures 2J, 2K, and S2L). Thus, genetic gain- and loss-of-func-

tion studies in vivo indicate that hepatocyte HO-1 exacerbates

insulin resistance, steatosis, liver damage and metabolic

disease.

Macrophage HO-1 Knockout Mice Resist MetabolicDiseaseConsidering the observed increases in macrophage HO-1 in in-

sulin-resistant liver and adipose in both mouse and man, and

the central role of macrophages in metaflammation, we next

tested the contribution of macrophage HO-1 toward insulin

resistance. We crossed our conditional HO-1 mice to macro-

phage deleting LysM-Cre transgenic animals, creating macro-

phage HO-1 knockout (Macho) mice. Equal and efficient deletion

was confirmed in naive, as well as lipopolysaccharide (LPS)/

interferon (IFN)g-driven proinflammatory (M1) and interleukin

(IL)-4/IL-13-driven anti-inflammatory (M2) bone-marrow-derived

macrophages (BMDMs; Figures 3A–3C). Similar to Lhoko ani-

mals, Macho mice were born at Mendelian ratios, developed

normally, exhibited normal white blood cell counts, and ap-

peared robust and healthy (Figures S3A and S3B and data not

shown). Importantly, metabolic profiling was equally unremark-

able in the unchallenged state with no detectable alterations in

body weight, food intake, glucose tolerance, insulin sensitivity,

or plasma-free fatty acid levels (Figures S3A–S3F). Thus, macro-

phageHO-1 is dispensable for normal growth, development, and

metabolic homeostasis.

(J) Rank order of 968 visceral adipose tissue genes correlating with HOMA-IR and

genes) obesity. p values for Spearman correlation coefficients (HOMA-IR versus

significance cut-off at p < 0.01, corrected for multiple testing.

(K and L) qPCR analysis of HO-1 expression in (K) liver, and (L) epididymal adipos

(M and N) Immunoblot analysis for HO-1 in (M) liver, and (N) epididymal adipose

Results are mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001. See also Figure S1

As mentioned above, bulk literature would have predicted, if

anything, enhanced inflammation upon HO-1 deletion. In an

attempt to unmask such a phenotype, we again turned to HFD.

Intriguingly, and again in contrast to expectations, we observed

marked increases in glucose clearance and insulin sensitivity in

Macho animals after a 16 week HFD regimen (Figures 3D and

3E). Indeed, even body weight gain was slightly reduced

(Figure 3D)—despite no measurable change in food intake (Fig-

ure S3G). Importantly, all aspects of improved glucose homeo-

stasis were evident before body weight differences emerged

(data not shown) and, independently, when comparing body-

weight-matched animals at the experiment’s end (Figure 3F).

Interestingly, the metabolic improvements elicited by macro-

phage HO-1 loss were systemic. Both liver morphology and his-

tology revealed clear reductions in hepatic steatosis in Macho

animals (Figure 3G), improvements confirmed by reduced mea-

sures of lipid content and liver toxicity, as well as lipogenic and

gluconeogenic profiles (Figures 3H, 3I, and S3H). Specifically,

expression of the lipogenic factors Pparg, Srebf1, Cd36, and

Fasn were reduced in HFD-Macho livers as well as G6pc, which

mediates hepatic glucose output (Figure 3H). Expression of in-

flammatory Tnf, Il6, and Il1b were equally reduced (Figure 3H),

a contrast to elevated anti-inflammatory Il10 (Figure 3H). Impor-

tantly, HO-1 expression, whichwe show here to be proinflamma-

tory, was not only reduced in the myeloid compartment, but

importantly also in hepatocytes of the Macho mice (Figure 3G),

indicating that macrophage HO-1 is necessary for effective

transmission of inflammatory status to target tissues. Consistent

with our previous findings, liver, muscle, and fat responses to in-

traportal insulin were significantly enhanced in the HFD-treated

Macho mice (Figure S3I). These findings are in direct agreement

with our observations that HO-1 expression in both hepatocytes

and macrophages of human obese liver predicts disease

severity. An aside on the observed body weight difference,

Macho animals exhibited a minor increase in energy expenditure

(Figure 3J) independent of any altered activity (Figure 3K) or res-

piratory quotient (Figure S3J). These data are consistent with the

observed system-wide improvements in metabolic homeosta-

sis. In all, the data of both mouse models and the human cohort

implicate myeloid HO-1 as an effector of metabolic disease.

Thus, macrophage HO-1 is necessary for the manifestation of

HFD-induced metaflammation.

Reduced Metaflammation and Improved Adipose TissueArchitecture in Macho MiceSteatosis results in part from fatty acid spill-over resulting from

inflammation, insulin resistance, and improper lipolysis. In line

with the reduced steatosis, HFD-Macho mice exhibited reduced

fasted and fed serum-free fatty acids (Figure 4A). Further, exam-

ination of HFD-Macho epididymal fat revealed dramatic

improvements in metaflammation. Included were near-absence

predicting metabolically healthy (251 protective genes) and unhealthy (717 risk

expression values) obtained from public data set. Dotted line represents the

e obtained from obese IS and obese IR C57BL/6J mice after 16 weeks on HFD.

obtained from obese IR and obese IS C57BL/6J mice after 16 weeks of HFD.

and Tables S1 and S2.

Cell 158, 25–40, July 3, 2014 ª2014 Elsevier Inc. 29

(legend on next page)

30 Cell 158, 25–40, July 3, 2014 ª2014 Elsevier Inc.

of crown-like structures (Figure 4B), reduced adipocyte size

(Figures 4B and 4C), and a substantial reduction in macrophage

infiltration (Figure 4D). Importantly, these healthy metabolic attri-

butes were manifest in the face of clear obesity. Consistent with

this ‘‘healthy’’ obesity profile, HFD-Macho animals exhibited

clear reductions in the systemic indicators of metaflammation,

namely TNF-a, IL-1b, MCP-1, and RANTES (Figure 4E). Adipo-

nectin (Adipoq) and retinol binding protein-4 (Rbp4) mRNA levels

were increased in HFD-Macho mouse adipose, whereas leptin

(Lep) levels were not affected (Figure 4F). Notably, the Rbp4

mRNA changes were not manifest on the protein level (Fig-

ure 4G). Thus, Macho mice exhibit reduced systemic and adi-

pose tissue metaflammation.

The human andmouse data to this point suggested a predom-

inantly myeloid and hepatic role for HO-1 in metaflammation and

insulin resistance. We also largely ruled out the contribution of

HO-1 in muscle, adipocyte, and pancreatic b-cell compart-

ments, as targeted deletion in these depots did not affect

HFD-challenged body weight accumulation, glucose, or insulin

tolerance (Figures S4A–S4C).

Next, we asked how HO-1 elicits metaflammation. We purified

and FACS analyzed the nonadipocyte stromal vascular fraction

(SVF) from epididymal adipose of HFD-treated Macho and con-

trol mice. In line with the reduced evidence for metaflammation,

HFD-Macho adipose tissue contained a 2-fold reduction in

proinflammatory CD11c+ and CCR2+ macrophage infiltrates,

while protective M2-like MGL+ macrophage numbers were

increased (Figures 4H–4J and Figure S4D). These observations

were equally reflected in qPCR measures of Cd68, Itgax/

Cd11c and the chemokine receptors Ccr2 and Ccr5 in whole-

adipose RNA preparations (Figures S4E and S4F), as well as

confocal evidence of improved adipose tissue architecture

(Figure 4K). In support of a conserved system in humans, re-

examination of our human adipose biopsies revealed an increase

in total macrophage number and M1-like marker expression

(Table S3). Importantly, HO-1 exhibited a tight positive correla-

tion with M1-like marker expression (TNF, ITGAX/CD11c),

whereas M2-like markers showed no correlation (MRC1/

CD206) or an inverse relationship (CLEC10A/CD301, CCL18)

(Table S4). Importantly, these findingswere independent of waist

circumference and again maintained after correcting for visceral

fat. Thus, macrophage HO-1 is necessary to induce proinflam-

matory macrophage skewing and disruption of adipose tissue

architecture.

Figure 2. Hepatocyte HO-1 Knockout Mice are Insulin Hypersensitive

(A) HO-1 immunoblot of liver obtained from control and Lhoko mice.

(B) HO activity in liver homogenates, n R 6 per group.

(C) Immunohistochemical staining for HO-1 in livers of vehicle and hemin injecte

(D) Oral glucose and insulin tolerance tests on HFD.

(E) HOMA-IR of control and Lhoko mice after 16 weeks on HFD.

(F) Serum levels of liver enzymes after 16 weeks on 60% HFD, n R 4 per group.

(G) In vivo activation of liver insulin signaling. HFD-fed mice were injected with in

blotting. A representative result from three independent experiments is shown.

(H) Liver triglyceride content, n = 3 per group.

(I) H&E, oil red O and HO-1 staining of representative liver sections obtained from

(J) Insulin tolerance tests in AdLacZ and AdHO-1 injected C57BL/6J animals, n =

(K) HOMA-IR of AdLacZ and AdHO-1 injected C57BL/6J animals, n = 7 per grou

Results are mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001. See also Figure S2.

HO-1 Loss Induces an Anti-Inflammatory MetabolicSignature and Blunts NF-kB SignalingNext, to gain insight into the attenuated metaflammatory pheno-

type, we harvested and in vitro differentiated bone marrow

progenitors from unchallenged, chow-fed Macho animals and

controls intomacrophages (BMDMs). Using TNF-a, we activated

BMDMs and examined NF-kB signaling. Using phospho-IkBa as

a surrogate for activation, we found a clearly diminished NF-kB

response in Macho BMDMs upon induction (Figure 5A). Impor-

tantly, this finding was confirmed using more direct assays as

well: RelA/p65 nuclear translocation was reduced (Figure 5B),

its target-sequence binding by EMSA was reduced (Figure 5C),

and NF-kB target gene expression including Nfkbia/IkBa, Tnf,

and Nos2/iNos was reduced (Figure 5D). Further, chromatin

immunoprecipitation measures of endogenous NF-kB promoter

binding at known target genes Nos2 and Tnf also showed clear

reductions (Figure 5E). Thus HO-1-deficient BMDMs have

blunted NF-kB signaling.

We and others have previously shown that oxidative meta-

bolism is a signature of activated M2-like macrophages, that

activatedM1-like proinflammatory cells display amore glycolytic

metabolic profile, and that direct alterations in metabolism can

regulate NF-kB signaling potential (Haschemi et al., 2012; Rodrı-

guez-Prados et al., 2010; Vats et al., 2006). Intriguingly, when

assessing these parameters in naive BMDMs, we found clearly

elevated basal and spare respiratory capacity, as well as ATP-

turnover in the HO-1-deficient cells (Figures 5F and 5G). Finding

all three hallmarks of an M2 ‘‘aerobic signature’’ already in non-

stimulated Macho BMDMs indicated that HO-1 is required for

proper metabolic programming in the most basal cellular state,

and that such ‘‘pre-’’programming may potently influence

macrophage skewing and thus whole-body inflammation.

Indeed, we observed consistent reductions in TNF-a, IL-1b,

and IL-6 secretion in Macho BMDMs relative to controls (Fig-

ure 5H). Thus, HO-1 is required for metabolic programming of

naive macrophages, and thus, for proper TNF-a- and LPS-

induced skewing and activation.

HO-1 Suppresses OxPhos and ROS SignalingWe next asked whether mitochondrial changes might constitute

a common molecular link between the liver and myeloid pheno-

types (Cheng et al., 2009). Indeed, when repeating our measures

on both systems, parallels were evident for essentially all

measures. Like Macho BMDMs, primary Lhoko hepatocytes

d control and Lhoko mice. Scale bar, 100 mm.

sulin, tissue biopsies taken at the indicated times, and analyzed by immuno-

HFD-fed control and Lhoko mice. Scale bar, 200 mm.

7 per group, p value 1 sided.

p.

Cell 158, 25–40, July 3, 2014 ª2014 Elsevier Inc. 31

(legend on next page)

32 Cell 158, 25–40, July 3, 2014 ª2014 Elsevier Inc.

exhibited increased basal, spare as well as maximal respiratory

capacity (Figures 6A and S5A). Notably, these effects were inde-

pendent of any obvious underlying changes in coupling effi-

ciency (Figure S5B), mitochondrial DNA content (Figure S5C),

mitochondrial mass (Figure S5D), and respiratory chain stoichi-

ometry (Figure S5E). Interestingly, qPCR analysis revealed clear

andwidespread upregulation of the entire mitochondrial reactive

oxygen species (ROS) detoxification system, from glutathione

processing to the mitochondrial-specific superoxide dismutase

2 (Sod2), again in both systems (Figure 6B), indicating that the

increased mitochondrial flux was coupled to the induction of

ROS quenching machinery. Notably, these changes were

confirmed directly, with the findings of increased SOD activity

(Figure 6C) and indirectly, with an �50% increase in intracellular

H2O2, the product of SOD, and an increase of cytosolic ROS,

again in both knockout cell types (Figures 6D and S5F). Recent

studies have suggested that mild ROS elevations activate mito-

chondrial respiration (Mailloux et al., 2013). If ROS balance was

altering oxidative capacity (Figure 6A) then mitochondrial func-

tion in the knockout cells should be particularly sensitive to anti-

oxidants. In line with this idea, 1 hr preadministration of the

potent ROS quencher N-acetyl-cysteine (NAC) reverted HO-1-

deficient mitochondrial functional parameters to control levels

(Figure 6E). Thus, HO-1 regulates acute ROS thresholding of

mitochondrial respiration.

ROS have previously been linked to INSR activation via impair-

ment of protein tyrosine phosphatase nonreceptor type 1

(PTPN1/PTP1B) (Tiganis, 2011) and to nuclear factor-erythroid-

derived 2-like (NFE2L2/NRF2) function via impairment of the

inhibitory protein kelch-like ECH-associated protein 1 (KEAP1)

(Kobayashi et al., 2006). Because activation of INSR and NRF2

alone would explain the majority of phenotypes observed, we

tested their activation state in our most naive HO-1 knockout

models. Importantly, we found increased nuclear translocation

(Figure 6F) and target-binding activity (Figure 6G) of NRF2 in

Macho BMDMs and decreased PTP1B activity (Figure 6H) and

acute INSR activation (Figure 6I) in naive, chow-fed Lhoko livers.

Notably, in vitro reconstitution of HO-1 expression completely

reverted mitochondrial respiratory changes (Figure S5H), down-

stream pINSR increases (Figure S5G), and TNF-a and IL-1b

blunting (Figure S5I) on a timescale paralleling the 1–2 days ex-

pected for adenoviral HO-1 re-expression. This data set sup-

ports a mechanism where HO-1 continuously limits activation

of key signaling systems through an acute and ROS-dependent

Figure 3. Macrophage HO-1 Knockout Mice Resist Metabolic Disease

(A) qPCR analysis of HO-1mRNA levels in BMDMs generated from control and M

for 24 hr.

(B) Immunoblot for HO-1 in naive BMDMs derived from indicated genotypes.

(C) HO activity in naive BMDMs generated from LFD-fed control and Macho mic

(D) Body weight gain of control and Macho mice on 60% HFD.

(E) Oral glucose and insulin tolerance tests, corresponding blood glucose and in

(F) Oral glucose and insulin tolerance tests, corresponding blood glucose and in

(G) Liver aspects and stainings of representative liver sections obtained fromHFD

Scale bar, 50 mm.

(H) Relative mRNA levels of inflammatory, lipogenic, and gluconeogenic genes in

(I) Liver triglyceride and cholesterol content, n R 5 per group.

(J and K) Energy expenditure and activity of body-weight-matched control and M

Results are mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001. See also Figure S3.

mechanism. Again all key phenotypes were observed in naive

unchallenged cells, lending further support to the notion that

HO-1 contributes to the ‘‘pre’’-programming of cellular function.

DISCUSSION

Here, we show that HO-1 drives insulin resistance in mouse and

man. In the absence of macrophage HO-1 animals fail to develop

metaflammation and key hallmarks of metabolic disease. The

findings redefine the current view of HO-1, have substantial

implications for the stratification of healthy and unhealthy

obesity, and open the door for HO-1 inhibitors as a new thera-

peutic strategy for obesity and diabetes.

Intersecting strongpatient datawith genetic evidence fromfive

tissue-specific HO-1 deletion models, as well as overexpression

and genetic rescue scenarios, we find a clear proinflammatory

role for HO-1 in metabolic control and thus break dogma.

Induction of HO-1 by systemic cobalt-protoporphyrin (CoPP)

administration has previously been reported to ameliorate

obesity and diabetes in mice (Li et al., 2008). Indeed, much of

our knowledge of HO-1 biology has relied on CoPP. That said,

key studies raise concerns about CoPP specificity as an ‘‘HO-1

agonist.’’ Specifically, CoPP acts indirectly, by decreasing

BACH1 and increasing NRF2 protein levels (Shan et al., 2006).

CoPP inhibits ALAS1 function (Zheng et al., 2008), and recent

data also suggest effects on FOXO1 (Liu et al., 2013). Together,

NRF2, ALAS1 and FOXO1 impact tens if not hundreds of down-

stream targets including heme synthesis, mitochondrial respira-

tion, cytochrome function, and much more. Also, CoPP blocks

HO-1’s interaction with the key proinflammatory mediator IFN

regulatory factor 3 (IRF3) (Koliaraki and Kollias, 2011).

In the liver, we found that HO-1 deficiency promotes insulin

sensitivity and reduces diet-induced fatty liver disease. This

overall finding is in line with a study showing that in the context

of IRS1/IRS2 double-knockout, HO-1 links insulin signaling dys-

regulation to impaired mitochondrial function (Cheng et al.,

2009). Critically, we observed a brake-like, negative regulator

effect on insulin activation even in healthy, chow-fed animals.

Also in macrophages, HO-1 loss blunted NF-kB signaling poten-

tial in the unstimulated state, even in naive, not-yet polarized

cells. These findings indicate a novel role for HO-1 in the meta-

bolic regulation of cellular signaling thresholds. It will be inter-

esting to see whether additional signaling pathways exhibit

similar threshold control via HO-1.

acho mice, naive and polarized with LPS/IFNg (M1-like) or IL-4/IL-13 (M2-like)

e, n = 3 per group.

sulin levels, n = 10–15 per genotype.

sulin levels in body-weight-matched animals, n = 7 per genotype.

-fed control andMachomice for H&E, oil red O, HO-1 (red), andMAC-2 (brown).

liver samples, n = 4 per group.

acho mice, n = 7 per genotype.

Cell 158, 25–40, July 3, 2014 ª2014 Elsevier Inc. 33

(legend on next page)

34 Cell 158, 25–40, July 3, 2014 ª2014 Elsevier Inc.

Mechanistically, the signal thresholding effects of HO-1

appear to have common ground. HO-1 loss increases mitochon-

drial respiration, antioxidant capacity, and consequently, H2O2

production. Our findings are consistent with detailed examina-

tions of redox regulation of NRF2 and PTP1B in the literature

and their downstream effects on INSR and NF-kB responses

(Kobayashi et al., 2006; Thimmulappa et al., 2006; Tiganis,

2011). Here, we add HO-1 as a new upstream effector of such

redox-dependent signal ‘‘preconditioning.’’

Corroborating these notions, macrophage HO-1 knockout

mice have been independently reported elsewhere to exhibit

an anti-inflammatory phenotype (Tzima et al., 2009). Using

polyI:C and LPS as stimuli, the authors found that HO-1 was

necessary for full activation of thioglycollate elicited peritoneal

macrophages. Our study and theirs, together, provide definitive

genetic evidence that HO-1 is necessary for both ‘‘cold’’ and

‘‘hot’’ inflammation. Intriguingly, Tzima et al. reported no change

of NF-kB activation, a contrast perhaps attributable to differ-

ences in M1-like activation between the BMDMs used here,

and thioglycollate elicited peritoneal macrophages (Ghosn

et al., 2010). Our data are also in line with improvements in

glucose handling observed in HFD-fed recipient mice trans-

planted with HO-1 haploinsufficient bone marrow (Huang et al.,

2012). Notably, our observation of an M2-like aerobic signature

in unstimulated Macho BMDMs indicates that HO-1 regulates

cell fate skewing even prior to stimulation, and therefore that

HO-1 may fine tune pro- versus anti-inflammatory cell ratios. It

will be interesting to see if HO-1 and associatedmetabolic skew-

ing in precursor cell populations will predict outcome in classical

inflammatory disorders.

Along similar lines, our data suggest a trigger-like role for HO-1

in metaflammation and insulin resistance and thus indicate high

prognostic value for detecting disease onset. The observations

are certainly generalizable, corroborated by multiple indepen-

dent patient cohorts including human diabetes (Misu et al.,

2010), generalized obesity (Qatanani et al., 2013), and obese

monozygotic twins (Pietilainen et al., 2008). These data and our

own, show that liver and adipose HO-1 predict disease severity

independently of age, sex, and degree of adiposity, and that they

correlate specifically with pro- but not anti-inflammatory

markers. Careful examination reveals that even our obIS group

showed earliest signs of insulin resistance relative to nonobese

Figure 4. Macrophage HO-1 Knockout Mice Resist Metaflammation

(A) Serum-free fatty acid levels measured in fed and fasted states on HFD, n R

(B) H&E staining of epididymal adipose of HFD fed control and Macho mice. Sca

(C) Adipocyte size distribution in epididymal fat of HFD fed control and Macho mi

n = 6 per group. n = number of adipocytes.

(D) Immunofluorescence staining for MAC-2 and confocal imaging of whole-m

experiments is shown. Scale bar, 200 mm.

(E) Serum levels of proinflammatory cytokines, n R 3 per group.

(F) qPCR analysis of adipokines in epididymal fat isolated from control and Mach

(G) Serum adipokine levels in control and Macho mice, n = 4 per group.

(H–J) Flow cytometric analysis of the SVF isolated from epididymal white adipose t

cells and examined for CD11b coexpression to obtain total macrophage numbe

subpopulations is presented; (J) distribution of CD11c+ MGL� and CD11c� MGL

(K) Immunofluorescence staining for perilipin (red) andMAC-2 (green) of whole-mo

is shown. Scale bar, 100 mm.

Results are mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001. See also Figure S4,

controls with elevations in HOMA-IR, insulin and nonsignifcant

increases in GLUT4 and HSD11B1 mRNA (Table 1). HO-1 had

significant predictive power in distinguishing these groups.

Thus, it will be most interesting to evaluate the predictive power

of blood HO-1 levels in detecting pre- and early disease states.

Ending actually with a theorectical and technical unknown, our

study leaves unanswered the potential role of Kupffer cells, and

Kupffer cell HO-1, in the etiology of steatosis, particularly in the

Macho animals. While reported to delete strictly in the circulating

macrophage compartment of the liver (Maeda et al., 2005), we

findwhat appears to be LysM-Cremediated deletion in the entire

liver myeloid compartment, including Kupffer cells, specifically

after long-term HFD. Increased promiscuity of Cre-drivers

upon HFD would certainly not be a new concept. The findings

indicate that the stark improvements observed in HFD-Macho

livers could well be supported by HO-1 deletion in Kupffer cells.

EXPERIMENTAL PROCEDURES

Human Study Population

Study subjects included 51 obese patients and 6 nonobese controls that

underwent weight-reducing surgery or elective surgical procedures such as

cholecystectomy. Participants were included if they had fasting plasma

glucose levels <7.0 mmol/l, no history of diabetes, no medications, no weight

changes >3% during the previous 2 months, C-reactive protein (CRP) levels

<20 mg/l, and normal blood leukocyte counts. For further information, see

the Extended Experimental Procedures.

Metabolically Healthy and Unhealthy Obese Mice

Age- and body-weight-matched C57BL/6J mice were selected based on their

response in glucose and insulin tolerance tests after 16 weeks on HFD.

Glucose and Insulin Tolerance Tests

Glucose and insulin tolerance tests were performed as described in the

Extended Experimental Procedures.

Generation of Tissue-Specific HO-1 Knockout Mice

All mouse models are described in the Extended Experimental Procedures.

Animals were kept on a 12 hr light/dark cycle with free access to food and

water and housed in accordance with international guidelines. Dietary inter-

ventions started at 6 weeks of age using diets which contained 10% and

60% calories of fat (Research Diets, D12450B and D12492).

Heme Oxygenase Activity

HO activity was determined as outlined in the Extended Experimental

Procedures.

7 per group.

le bar, 200 mm.

ce as determined by quantitative morphometry of H&E stained tissue sections,

ount epididymal fat pads. A representative result from three independent

o mice, n = 4 per group.

issue (WAT) of HFD fed control andMachomice; (H) Cells were gated for F4/80+

rs; (I) The number of CD11c+, CCR2+ and MGL+ adipose tissue macrophage+ adipose macrophages, n = 3 mice per group.

unt epididymal fat. A representative result from three independent experiments

Tables S4 and S5.

Cell 158, 25–40, July 3, 2014 ª2014 Elsevier Inc. 35

(legend on next page)

36 Cell 158, 25–40, July 3, 2014 ª2014 Elsevier Inc.

In Vivo Liver Insulin Signaling

For examination of in vivo insulin signaling, mice were fasted for 2 hr, anesthe-

tized and injected with insulin. Liver, muscle and adipose were removed at the

indicated times.

Indirect Calorimetry

Indirect calorimetry and activity measurements were performed as reported

elsewhere (Pospisilik et al., 2010) and as described in the Extended Experi-

mental Procedures.

Flow Cytometry

SVF cells were obtained by collagenase digestion from epididymal adipose

tissue as described (Todoric et al., 2011). FACS analysis was performed as

described (Lumeng et al., 2007) and outlined in the Extended Experimental

Procedures.

Nuclear Translocation Assay

Naive BMDMswere stimulated with 10 ng/ml TNF-a or medium only for 30 and

60 min, fixed in ice-cold methanol, washed, and stained with anti-RelA/p65

antibody (Santa Cruz Biotechnology). Nuclei were counterstained with DAPI.

Electrophoretic Mobility Shift Assay

Nuclear extracts were prepared from BMDMs as described in the Extended

Experimental Procedures. EMSA was performed using the NF-kB EMSA

kit (LI-COR Biosciences) according to the manufacturer’s instructions. For

supershift assays anti-RelA/p65 (Santa Cruz Biotechnology), anti-p50

(Millipore), and anti-rabbit IgG (Vector Laboratories) was used. Unlabeled

double-stranded NF-kB probe was used to test specific binding. Samples

were separated on a nondenaturing polyacrylamide gel. NF-kB DNA binding

was visualized using the LI-COR Odyssey Imaging System.

Chromatin Immunoprecipitation Assay

ChIP assayswere performed using theMAGnify ChIP Kit (Life Technologies) as

detailed by the manufacturer’s instruction and as described in the Extended

Experimental Procedures.

Oxygen Consumption Assay and Bioenergetic Profile

OCR was measured using the XF24 Flux Analyzer (Seahorse Bioscience) as

described (Teperino et al., 2012). For details see Extended Experimental

Procedures.

PTP1B Activity Assay

PTP1B activity was determined as outlined in the Extended Experimental

Procedures.

Hydrogen Peroxide Measurements

Hydrogen peroxide (H2O2) production was measured using the Amplex Red

Hydrogen Peroxide/Peroxidase Assay Kit (Invitrogen).

Figure 5. Attenuated NF-kB Signaling and Anti-Inflammatory Metaboli

(A) pIkBa and IkBaweremeasured by immunoblotting in BMDMs from control and

the indicated times. b-Actin was used as internal loading control.

(B) RelA/p65 nuclear translocation assay. Naive BMDMs were stimulated with TN

immunofluorescent staining. Nuclei were stained with DAPI, and the percentage o

Scale bar, 50 mm.

(C) EMSA. Nuclear extracts were prepared from control and Macho BMDMs stimu

binding was verified by adding a 100-fold excess of unlabeled double-stranded

assays using anti-RelA/p65 and anti-p50 antibodies.

(D) RelativemRNA expression of NF-kB target genes in BMDMs generated from LF

for 60 min, n = 3 per group.

(E) qPCR analysis of NF-kB targets Tnf andNos2 in chromatin immunoprecipitatio

(F and G) Oxygen consumption rates (OCR) and mitochondrial function of naive

(H) Cytokine secretion of LPS stimulated control and Macho BMDMs, n = 3 per

Results are mean ± SEM n = 2–3 independent experiments. *p < 0.05, **p < 0.01

Superoxide Dismutase Assay

Superoxide dismutase (SOD) activity was measured using a commercially

available test kit (Cayman) according to manufacturer’s instructions.

Adenovirus Experiments

BMDM rescue experiments were performed using RGD-fiber-modified adeno-

virus particles (Vector Biolabs). In vivo grade viruses were produced to target

livers and isolated primary Lhoko hepatocytes. For details, refer to the

Extended Experimental Procedures.

NRF2 DNA-Binding Activity

The NRF2 TransAM ELISA-kit (Active Motif) was used to evaluate NRF2 DNA-

binding activity according to the manufacturer’s protocol.

Statistical Analysis

Data are expressed as mean ± standard error of the mean (SEM) unless other-

wise specified. Statistical analyses were performed as described in the

Extended Experimental Procedures. All reported p values are two-tailed

unless stated otherwise. p < 0.05 was considered to indicate statistical

significance.

Other Methods

See Extended Experimental Procedures.

SUPPLEMENTAL INFORMATION

Supplemental Information includes Extended Experimental Procedures, six

tables, and five figures and can be found with this article online at http://dx.

doi.org/10.1016/j.cell.2014.04.043.

AUTHOR CONTRIBUTIONS

H.E. conceptualized the study. W.P., O.W., and J.A.P. contributed to the study

design. H.E. and J.A.P. supervised the study and wrote the manuscript. A.J.

analyzed the macrophage model. E.E. analyzed liver and muscle phenotypes.

O.S. and S.K. performed and analyzed NF-kB immunoblots and bandshift as-

says. K.G. performed in vivo adenovirus rescue experiments. S.M.S., D.N.,

J.T., H.O., C.H.-K, L.K., and W.P. contributed human data. D.M. performed

PTP1B activity assays. K.M., F.K., and W.P. contributed human samples.

T.T.-H.L., K.D., S.S., P.S., R.M., S.M., E.C., and M.B. contributed to some of

the experiments. J.P-P. and A.V.K. measured mitochondrial ROS production.

J.C.D. measured HO activity. J.L.-C. and T.-C.Z. analyzed adipose HO-1

knockout mice. S.A. contributed b-cell data. F.J. and H.S.-F. generated

adenoviral particles. C.N.L. helped with FACS andwhole-mount immunohisto-

chemistry. A.J., E.E., O.S., W.P., and O.W. were involved in manuscript

preparation.

c Signature in Macrophage HO-1 KO Mice

Machomice fedwith LFD. Naive cells were stimulatedwith TNF-a (10 ng/ml) for

F-a (10 ng/ml) for the indicated times and analyzed for RelA/p65 localization by

f nuclear RelA/p65 positive cells was counted. Arrows show nuclear RelA/p65.

lated with TNF-a (10 ng/ml) for the indicated times. Specific RelA/p65 and p50

probe (cold oligo) harboring the NF-kB consensus site as well as supershift

D fed control andMachomice after stimulation with vehicle or TNF-a (10 ng/ml)

ns prepared from control andMacho BMDMs stimulatedwith TNF-a (10 ng/ml).

BMDMs derived from control and Macho mice, n = 9 per group.

group.

, ***p < 0.001.

Cell 158, 25–40, July 3, 2014 ª2014 Elsevier Inc. 37

(legend on next page)

38 Cell 158, 25–40, July 3, 2014 ª2014 Elsevier Inc.

ACKNOWLEDGMENTS

This research was supported by the Vienna Science and Technology Fund

(J.A.P. and H.E.), the Max-Planck Society (J.A.P.), the Austrian Diabetes

Association (H.E.), the Medical Scientific Fund of the Mayor of the City of

Vienna (E.E.), the Osterreichische Gesellschaft fur Laboratoriumsmedizin

und Klinische Chemie (E.E.), and a grant from the National Institutes of Health

(DK090262, C.N.L.). The authors are grateful to J. Husa, J. Kacerovsky, I.

Miller, G. Mitterer, M. Ozsvar Kozma, andM. Ruf for critical technical and theo-

retical help.

Received: October 7, 2013

Revised: March 3, 2014

Accepted: April 18, 2014

Published: July 3, 2014

REFERENCES

Abraham, N.G., and Kappas, A. (2008). Pharmacological and clinical aspects

of heme oxygenase. Pharmacol. Rev. 60, 79–127.

Bao,W., Song, F., Li, X., Rong, S., Yang,W., Zhang, M., Yao, P., Hao, L., Yang,

N., Hu, F.B., and Liu, L. (2010). Plasma heme oxygenase-1 concentration is

elevated in individuals with type 2 diabetes mellitus. PLoS ONE 5, e12371.

Burcelin, R., Garidou, L., and Pomie, C. (2012). Immuno-microbiota cross and

talk: the new paradigm of metabolic diseases. Semin. Immunol. 24, 67–74.

Calay, E.S., and Hotamisligil, G.S. (2013). Turning off the inflammatory, but not

the metabolic, flames. Nat. Med. 19, 265–267.

Capel, F., Klimcakova, E., Viguerie, N., Roussel, B., Vıtkova, M., Kovacikova,

M., Polak, J., Kovacova, Z., Galitzky, J., Maoret, J.J., et al. (2009). Macro-

phages and adipocytes in human obesity: adipose tissue gene expression

and insulin sensitivity during calorie restriction and weight stabilization.

Diabetes 58, 1558–1567.

Cheng, Z., Guo, S., Copps, K., Dong, X., Kollipara, R., Rodgers, J.T., Depinho,

R.A., Puigserver, P., and White, M.F. (2009). Foxo1 integrates insulin signaling

with mitochondrial function in the liver. Nat. Med. 15, 1307–1311.

Dmitriev, I., Krasnykh, V., Miller, C.R., Wang, M., Kashentseva, E., Mikheeva,

G., Belousova, N., and Curiel, D.T. (1998). An adenovirus vector with geneti-

cally modified fibers demonstrates expanded tropism via utilization of a

coxsackievirus and adenovirus receptor-independent cell entry mechanism.

J. Virol. 72, 9706–9713.

Garg, A. (2011). Clinical review#: Lipodystrophies: genetic and acquired body

fat disorders. J. Clin. Endocrinol. Metab. 96, 3313–3325.

Gavrilova, O., Marcus-Samuels, B., Graham, D., Kim, J.K., Shulman, G.I.,

Castle, A.L., Vinson, C., Eckhaus, M., and Reitman, M.L. (2000). Surgical

Figure 6. Lhoko and Macho Mice Display Skewed OxPhos, ROS Signa(A) Oxygen consumption rates (OCR) and mitochondrial function of primary hep

littermate controls. Note that for macrophages the same experiment as in Figure

(B) Relative mRNA levels of mitochondrial ROS detoxification genes in HO-1 kno

n = 3–4 per group. *, NRF2 target genes.

(C) SOD activity measured in lysates obtained from primary hepatocytes and na

(D) Intracellular H2O2 levels in primary hepatocytes and naive BMDMs, n = 4 per

(E) Oxygen consumption rates (OCR) and mitochondrial function of primary hepa

their littermate controls. N-acetyl-cysteine (NAC, 10 mM)was added 1 hr beforeme

is shown, n = 4–6 per group.

(F) Nuclear and cytosolic extracts were prepared from naive Macho and contro

(cytosolic marker) by immunoblotting. A representative result is shown.

(G) NRF2 binding activity in nuclear extracts derived from naive Macho and cont

(H) Normalized PTP1B activity in protein extracts obtained from primary hepato

animals, n = 3–4 per group.

(I) In vivo activation of liver insulin signaling. LFD fedmice were injected with insulin

representative result from three independent experiments is shown.

Results are mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001. See also Figure S5.

implantation of adipose tissue reverses diabetes in lipoatrophic mice.

J. Clin. Invest. 105, 271–278.

Ghosn, E.E., Cassado, A.A., Govoni, G.R., Fukuhara, T., Yang, Y., Monack,

D.M., Bortoluci, K.R., Almeida, S.R., Herzenberg, L.A., and Herzenberg, L.A.

(2010). Two physically, functionally, and developmentally distinct peritoneal

macrophage subsets. Proc. Natl. Acad. Sci. USA 107, 2568–2573.

Gregor, M.F., and Hotamisligil, G.S. (2011). Inflammatory mechanisms in

obesity. Annu. Rev. Immunol. 29, 415–445.

Haschemi, A., Kosma, P., Gille, L., Evans, C.R., Burant, C.F., Starkl, P., Knapp,

B., Haas, R., Schmid, J.A., Jandl, C., et al. (2012). The sedoheptulose kinase

CARKL directs macrophage polarization through control of glucose meta-

bolism. Cell Metab. 15, 813–826.

Hirosumi, J., Tuncman, G., Chang, L., Gorgun, C.Z., Uysal, K.T., Maeda, K.,

Karin, M., and Hotamisligil, G.S. (2002). A central role for JNK in obesity and

insulin resistance. Nature 420, 333–336.

Hotamisligil, G.S. (2010). Endoplasmic reticulum stress and the inflammatory

basis of metabolic disease. Cell 140, 900–917.

Hotamisligil, G.S., Shargill, N.S., and Spiegelman, B.M. (1993). Adipose

expression of tumor necrosis factor-alpha: direct role in obesity-linked insulin

resistance. Science 259, 87–91.

Houstis, N., Rosen, E.D., and Lander, E.S. (2006). Reactive oxygen species

have a causal role in multiple forms of insulin resistance. Nature 440, 944–948.

Huang, J.Y., Chiang, M.T., Yet, S.F., and Chau, L.Y. (2012). Myeloid heme oxy-

genase-1 haploinsufficiency reduces high fat diet-induced insulin resistance

by affecting adipose macrophage infiltration in mice. PLoS ONE 7, e38626.

Huang, J.Y., Chiang, M.T., and Chau, L.Y. (2013). Adipose overexpression of

heme oxygenase-1 does not protect against high fat diet-induced insulin resis-

tance in mice. PLoS ONE 8, e55369.

Johnson, A.M., and Olefsky, J.M. (2013). The origins and drivers of insulin

resistance. Cell 152, 673–684.

Kapturczak, M.H., Wasserfall, C., Brusko, T., Campbell-Thompson, M., Ellis,

T.M., Atkinson, M.A., and Agarwal, A. (2004). Heme oxygenase-1 modulates

early inflammatory responses: evidence from the heme oxygenase-1-deficient

mouse. Am. J. Pathol. 165, 1045–1053.

Kobayashi, A., Kang, M.I., Watai, Y., Tong, K.I., Shibata, T., Uchida, K., and

Yamamoto, M. (2006). Oxidative and electrophilic stresses activate Nrf2

through inhibition of ubiquitination activity of Keap1. Mol. Cell. Biol. 26,

221–229.

Koliaraki, V., and Kollias, G. (2011). A new role for myeloid HO-1 in the innate to

adaptive crosstalk and immune homeostasis. Adv. Exp. Med. Biol. 780,

101–111.

Kusminski, C.M., Holland, W.L., Sun, K., Park, J., Spurgin, S.B., Lin, Y., Askew,

G.R., Simcox, J.A., McClain, D.A., Li, C., and Scherer, P.E. (2012).

ling, and NRF2/INSR Signaling Toneatocytes and naive BMDMs derived from Lhoko and Macho animals and their

5F is shown, n = 5 for hepatocytes, n = 9 for macrophages.

ckout hepatocytes and macrophages as well as their corresponding controls,

ive BMDMs, n = 4 per group.

group.

tocytes and naive BMDMs isolated from Lhoko and Macho animals as well as

asurements. Note that for untreated hepatocytes the same experiment as in (A)

l BMDMs and analyzed for NRF2, Histone H3 (nuclear marker) and GAPDH

rol BMDMs, n = 4 per genotype.

cytes (Hep) as well as liver biopsies of LFD and HFD fed Lhoko and control

, liver biopsies taken at the indicated times, and analyzed by immunoblotting. A

Cell 158, 25–40, July 3, 2014 ª2014 Elsevier Inc. 39

MitoNEET-driven alterations in adipocyte mitochondrial activity reveal a

crucial adaptive process that preserves insulin sensitivity in obesity. Nat.

Med. 18, 1539–1549.

Li, M., Kim, D.H., Tsenovoy, P.L., Peterson, S.J., Rezzani, R., Rodella, L.F.,

Aronow, W.S., Ikehara, S., and Abraham, N.G. (2008). Treatment of obese

diabetic mice with a heme oxygenase inducer reduces visceral and subcu-

taneous adiposity, increases adiponectin levels, and improves insulin sensi-

tivity and glucose tolerance. Diabetes 57, 1526–1535.

Liu, X., Cui, Y., Li, M., Xu, H., Zuo, J., Fang, F., and Chang, Y. (2013). Cobalt

protoporphyrin induces HO-1 expression mediated partially by FOXO1 and

reduces mitochondria-derived reactive oxygen species production. PLoS

ONE 8, e80521.

Lumeng, C.N., and Saltiel, A.R. (2011). Inflammatory links between obesity and

metabolic disease. J. Clin. Invest. 121, 2111–2117.

Lumeng, C.N., Bodzin, J.L., and Saltiel, A.R. (2007). Obesity induces a pheno-

typic switch in adipose tissue macrophage polarization. J. Clin. Invest. 117,

175–184.

Maeda, S., Kamata, H., Luo, J.L., Leffert, H., and Karin, M. (2005). IKKbeta

couples hepatocyte death to cytokine-driven compensatory proliferation

that promotes chemical hepatocarcinogenesis. Cell 121, 977–990.

Mailloux, R.J., McBride, S.L., and Harper, M.E. (2013). Unearthing the secrets

of mitochondrial ROS and glutathione in bioenergetics. Trends Biochem. Sci.

38, 592–602.

Misu, H., Takamura, T., Takayama, H., Hayashi, H., Matsuzawa-Nagata, N.,

Kurita, S., Ishikura, K., Ando, H., Takeshita, Y., Ota, T., et al. (2010). A liver-

derived secretory protein, selenoprotein P, causes insulin resistance. Cell

Metab. 12, 483–495.

Nakamura, T., Furuhashi, M., Li, P., Cao, H., Tuncman, G., Sonenberg, N.,

Gorgun, C.Z., and Hotamisligil, G.S. (2010). Double-stranded RNA-dependent

protein kinase links pathogen sensing with stress and metabolic homeostasis.

Cell 140, 338–348.

Ndisang, J.F., Lane, N., and Jadhav, A. (2009). Upregulation of the heme

oxygenase system ameliorates postprandial and fasting hyperglycemia in

type 2 diabetes. Am. J. Physiol. Endocrinol. Metab. 296, E1029–E1041.

Odegaard, J.I., and Chawla, A. (2013). Pleiotropic actions of insulin resistance

and inflammation in metabolic homeostasis. Science 339, 172–177.

Patsouris, D., Li, P.P., Thapar, D., Chapman, J., Olefsky, J.M., and Neels, J.G.

(2008). Ablation of CD11c-positive cells normalizes insulin sensitivity in obese

insulin resistant animals. Cell Metab. 8, 301–309.

Pietilainen, K.H., Naukkarinen, J., Rissanen, A., Saharinen, J., Ellonen, P.,

Keranen, H., Suomalainen, A., Gotz, A., Suortti, T., Yki-Jarvinen, H., et al.

(2008). Global transcript profiles of fat in monozygotic twins discordant for

BMI: pathways behind acquired obesity. PLoS Med. 5, e51.

Pihlajamaki, J., Boes, T., Kim, E.Y., Dearie, F., Kim, B.W., Schroeder, J., Mun,

E., Nasser, I., Park, P.J., Bianco, A.C., et al. (2009). Thyroid hormone-related

regulation of gene expression in human fatty liver. J. Clin. Endocrinol. Metab.

94, 3521–3529.

Pospisilik, J.A., Knauf, C., Joza, N., Benit, P., Orthofer, M., Cani, P.D., Ebers-

berger, I., Nakashima, T., Sarao, R., Neely, G., et al. (2007). Targeted deletion

of AIF decreases mitochondrial oxidative phosphorylation and protects from

obesity and diabetes. Cell 131, 476–491.

Pospisilik, J.A., Schramek, D., Schnidar, H., Cronin, S.J., Nehme, N.T., Zhang,

X., Knauf, C., Cani, P.D., Aumayr, K., Todoric, J., et al. (2010). Drosophila

genome-wide obesity screen reveals hedgehog as a determinant of brown

versus white adipose cell fate. Cell 140, 148–160.

Qatanani, M., Tan, Y., Dobrin, R., Greenawalt, D.M., Hu, G., Zhao, W., Olefsky,

J.M., Sears, D.D., Kaplan, L.M., and Kemp, D.M. (2013). Inverse regulation of

inflammation and mitochondrial function in adipose tissue defines extreme

insulin sensitivity in morbidly obese patients. Diabetes 62, 855–863.

Rodrıguez-Prados, J.C., Traves, P.G., Cuenca, J., Rico, D., Aragones, J.,

Martın-Sanz, P., Cascante, M., and Bosca, L. (2010). Substrate fate in acti-

vated macrophages: a comparison between innate, classic, and alternative

activation. J. Immunol. 185, 605–614.

40 Cell 158, 25–40, July 3, 2014 ª2014 Elsevier Inc.

Samuel, V.T., and Shulman, G.I. (2012). Mechanisms for insulin resistance:

common threads and missing links. Cell 148, 852–871.

Shakeri-Manesch, S., Zeyda, M., Huber, J., Ludvik, B., Prager, G., and Stulnig,

T.M. (2009). Diminished upregulation of visceral adipose heme oxygenase-1

correlates with waist-to-hip ratio and insulin resistance. Int J Obes (Lond)

33, 1257–1264.

Shan, Y., Lambrecht, R.W., Donohue, S.E., and Bonkovsky, H.L. (2006). Role

of Bach1 and Nrf2 in up-regulation of the heme oxygenase-1 gene by cobalt

protoporphyrin. FASEB J. 20, 2651–2653.

Solinas, G., and Karin, M. (2010). JNK1 and IKKbeta: molecular links between

obesity and metabolic dysfunction. FASEB J. 24, 2596–2611.

Solinas, G., Vilcu, C., Neels, J.G., Bandyopadhyay, G.K., Luo, J.L., Naugler,

W., Grivennikov, S., Wynshaw-Boris, A., Scadeng, M., Olefsky, J.M., and

Karin, M. (2007). JNK1 in hematopoietically derived cells contributes to diet-

induced inflammation and insulin resistance without affecting obesity. Cell

Metab. 6, 386–397.

Stefan, N., Haring, H.U., Hu, F.B., and Schulze, M.B. (2013). Metabolically

healthy obesity: epidemiology, mechanisms, and clinical implications. Lancet

Diabetes Endocrinol 1, 152–162.

Sun, K., Tordjman, J., Clement, K., and Scherer, P.E. (2013). Fibrosis and

adipose tissue dysfunction. Cell Metab. 18, 470–477.

Szendroedi, J., Phielix, E., and Roden, M. (2012). The role of mitochondria in

insulin resistance and type 2 diabetes mellitus. Nat. Rev. Endocrinol. 8,

92–103.

Tchkonia, T., Thomou, T., Zhu, Y., Karagiannides, I., Pothoulakis, C., Jensen,

M.D., and Kirkland, J.L. (2013). Mechanisms and metabolic implications of

regional differences among fat depots. Cell Metab. 17, 644–656.

Teperino, R., Amann, S., Bayer, M., McGee, S.L., Loipetzberger, A., Connor,

T., Jaeger, C., Kammerer, B., Winter, L., Wiche, G., et al. (2012). Hedgehog

partial agonism drives Warburg-like metabolism in muscle and brown fat.

Cell 151, 414–426.

Thimmulappa, R.K., Lee, H., Rangasamy, T., Reddy, S.P., Yamamoto, M.,

Kensler, T.W., and Biswal, S. (2006). Nrf2 is a critical regulator of the innate

immune response and survival during experimental sepsis. J. Clin. Invest.

116, 984–995.

Tiganis, T. (2011). Reactive oxygen species and insulin resistance: the good,

the bad and the ugly. Trends Pharmacol. Sci. 32, 82–89.

Tzima, S., Victoratos, P., Kranidioti, K., Alexiou, M., and Kollias, G. (2009).

Myeloid heme oxygenase-1 regulates innate immunity and autoimmunity by

modulating IFN-beta production. J. Exp. Med. 206, 1167–1179.

Vats, D., Mukundan, L., Odegaard, J.I., Zhang, L., Smith, K.L., Morel, C.R.,

Wagner, R.A., Greaves, D.R., Murray, P.J., and Chawla, A. (2006). Oxidative

metabolism and PGC-1beta attenuate macrophage-mediated inflammation.

Cell Metab. 4, 13–24.

Virtue, S., and Vidal-Puig, A. (2008). It’s not how fat you are, it’s what you do

with it that counts. PLoS Biol. 6, e237.

Weisberg, S.P., McCann, D., Desai, M., Rosenbaum, M., Leibel, R.L., and

Ferrante, A.W., Jr. (2003). Obesity is associated with macrophage accumula-

tion in adipose tissue. J. Clin. Invest. 112, 1796–1808.

Xu, H., Barnes, G.T., Yang, Q., Tan, G., Yang, D., Chou, C.J., Sole, J., Nichols,

A., Ross, J.S., Tartaglia, L.A., and Chen, H. (2003). Chronic inflammation in fat

plays a crucial role in the development of obesity-related insulin resistance.

J. Clin. Invest. 112, 1821–1830.

Yachie, A., Niida, Y., Wada, T., Igarashi, N., Kaneda, H., Toma, T., Ohta, K.,

Kasahara, Y., and Koizumi, S. (1999). Oxidative stress causes enhanced endo-

thelial cell injury in human heme oxygenase-1 deficiency. J. Clin. Invest. 103,

129–135.

Yuan, M., Konstantopoulos, N., Lee, J., Hansen, L., Li, Z.W., Karin, M., and

Shoelson, S.E. (2001). Reversal of obesity- and diet-induced insulin resistance

with salicylates or targeted disruption of Ikkbeta. Science 293, 1673–1677.

Zheng, J., Shan, Y., Lambrecht, R.W., Donohue, S.E., and Bonkovsky, H.L.

(2008). Differential regulation of human ALAS1 mRNA and protein levels by

heme and cobalt protoporphyrin. Mol. Cell. Biochem. 319, 153–161.

Supplemental Information

EXTENDED EXPERIMENTAL PROCEDURES

ReagentsUnless otherwise stated all chemicals and reagents were obtained from SIGMA.

Human Study PopulationStudy subjects included 51 obese patients and 6 nonobese controls that underwent weight reducing surgery or elective surgical

procedures such as cholecystectomy. Participants were included if they had fasting plasma glucose levels < 7.0 mmol/l, no history

of diabetes, no medications, no weight changes > 3% during the previous 2 months, C-reactive protein (CRP) levels < 20 mg/l and

normal blood leukocyte counts. All study subjects provided informed consent and study protocols were approved by the local Ethics

Committee. Tissue biopsies from visceral adipose tissue, obtained during surgery, were collected in RNA-later (Life Technologies)

and stored at�80�C until further processing. In a subset of 43 obese study participants liver biopsies were obtained at the beginning

of the surgical intervention. Glucose, insulin, lipid parameters, high-sensitive CRP andHOMA-IR indexwere determined as described

(Oberkofler et al., 2004; Todoric et al., 2011). Abdominal subcutaneous and visceral fat thickness were determined by ultrasonogra-

phy using the HDI 3000 CV System (ATL) (Oberkofler et al., 2010; Pontiroli et al., 2002).

Quantitative PCRqPCRwas performed as described (Todoric et al., 2011). In brief, total RNA andDNAwere extracted from respective tissues and cells

using RNA and DNA isolation kits (RNeasy, QIAGEN; TRIzol, Invitrogen). Isolated total RNAwas reverse-transcribed into cDNA using

commercially available kits (Applied Biosystems). qPCR reactions were performed using the iQ SYBR Green Supermix (Bio-Rad

Laboratories). Postamplification melting curve analysis was performed to check for unspecific products and primer-only controls

were included to ensure the absence of primer dimers. For normalization threshold cycles (Ct-values) were normalized to acidic

ribosomal phosphoprotein P0 (Rplp0) within each sample to obtain sample-specific DCt values ( = Ct gene of interest - Ct Rplp0).

2�DDCt values were calculated to obtain fold expression levels, where DDCt = (DCt treatment - DCt control). Human GLUT4 and

CD68 transcripts were quantified using TaqMan gene expression assays Hs00168966_1 and Hs00154355_m1 (Applied Biosystems,

Foster City, CA), respectively. Other human and mouse primers used are listed in Tables S5 and S6.

ImmunohistochemistryHuman and mouse tissues were fixed in 4% phosphate-buffered formalin and embedded into paraffin. Tissue sections were depar-

affinized and rehydrated prior to antigen unmasking using Target Retrieval Solution, pH 6.0 (Dako). Endogenous peroxidase activity

was quenched with incubation of 3% hydrogen peroxide for 10 min. Sections were blocked with Biotin-Blocking System (Dako) and

normal serum obtained from the secondary antibody host animal. Blocked sections were incubated with anti-HO-1 (1:500; ab13243,

Abcam), anti-CD68 (1:500; M0718, Dako) and anti-MAC-2 (1:500; CL8942AP, Cedarlane). Secondary antibody staining was per-

formed using the EnVision Detection System (Dako) and 3,30-diaminobenzidine as chromogenic substrate (Roche Molecular

Biochemicals) according to the manufacturer’s instructions. HO-1 positive cells were evaluated using the particle analysis module

incorporated in ImageAccess 9 (IMAGIC). HO-1 positive cells per all hepatocytes were counted on three different high-power fields

(4003 magnification).

Mouse and Human Multiple Tissue RNA LibrariesTotal RNA was isolated from tissues of three 6 week old male C57BL/6J mice to analyze HO-1 tissue distribution. A human multiple

tissue RNA library (Life Technologies) was used to analyze the tissue distribution of human HO-1 expression in tissues corresponding

to the mouse library.

Metabolically Healthy and Unhealthy Obese MiceObese age- and body-weight-matched C57BL/6J mice were selected based on their response in oral glucose and insulin tolerance

tests after 16 weeks on HFD. Epididymal fat pads and livers were removed and RNA and protein isolated to measure HO-1

expression.

Glucose and Insulin Tolerance TestsFollowing an overnight fast, mice were administered glucose (1 g/kg) by oral gavage, and blood samples for glucose and insulin

determination were collected from the tail vein at the indicated times. Insulin tolerance was assessed after a 2 hr fast by intraperito-

neal administration of human regular insulin (0.75 U/kg) and blood glucose monitoring. Low-dose insulin tolerance tests (0.1 U/kg)

were performed in adenovirus injectedmice to assess hepatic insulin sensitivity. Glycemia was assessed using a Accu-Chek (Roche)

glucometer. Plasma insulin levels were determined using the Ultrasensitive Mouse Insulin ELISA kit (Mercodia).

Western Blot AnalysisProteins were extracted from tissues by homogenizing in RIPA buffer (0.5% NP-40, 0.1% sodium deoxycholate, 150 mM NaCl,

50 mM Tris-HCl, pH 7.5) containing protease inhibitors (Complete Mini, Roche). The homogenate was cleared by centrifugation at

Cell 158, 25–40, July 3, 2014 ª2014 Elsevier Inc. S1

4�C for 30 min at 15,000 g and the supernatant containing the protein fraction recovered. Protein concentration in the supernatant

was determined using the BCA Protein Assay Kit (Pierce). 20 mg of proteins were resolved by SDS-PAGE and transferred to PVDF

membranes (GE Healthcare). Membranes were blocked with 5% BSA in Tris-buffered saline containing 0.2% Tween-20 (TBS-T),

and incubated with primary antibodies at 4�C over night. The following antibodies were used (all from Cell Signaling unless indicated

otherwise): anti-HO-1 (1:5,000; ab13243, Abcam), anti-phospho-Insulin receptor (pTyr1162/1163) (1:1,500; sc-25103, Santa Cruz

Biotechnology), anti-insulin receptor (1:1,500; Cat.-No. 3025), anti-phospho-AKT (pSer473) (1:1,500; Cat.-No. 9271), anti-phos-

pho-AKT (pThr308) (1:1,500; Cat.-No. 9275), anti-AKT (1:1,500; Cat.-No. 9272), anti-phospho-IkBa (pSer32/36) (1:1,000; Cat.-No.

9246), anti-ikBa (1:1,000; sc-371, Santa Cruz Biotechnology), anti-NRF2 (1:1,000; Cat.-No. 12721), anti-Histone H3 (1:1,000; Cat-

No. 4499), anti-b-Actin (1:500; A5441, Sigma), anti-GAPDH (1:2,500; ab9485, Abcam), anti-HSC-70 (1:5,000; sc-7298, Santa Cruz

Biotechnology). Expression levels of 5 different OXPHOScomplexes (CI subunit NDUFB8, CII-30kDa, CIII-Core protein 2, CIV subunit

I and CV alpha subunit) were analyzed using the MitoProfile Total OXPHOS Rodent WB Antibody Cocktail (Abcam). Incubated

membranes were washed and probed with the appropriate anti-igG-horseradish peroxidase-linked (HRP) secondary antibody

(NA 934, anti-rabbit IgG, 1:20,000; NA 931, anti-mouse IgG, 1:20,000; GE Healthcare). Antigen-specific binding of antibodies was

detected with SuperSignal West Femto and Pico Kits (Pierce) using a ChemiDoc XRS Imager (Bio-Rad). Image analysis was per-

formed using Image Lab Software Version 3.0.1 (Bio-Rad).

Generation of Tissue-Specific HO-1 Knockout MiceMicewith aconditionalHO-1alleleweregeneratedbyhomologous recombination. The targeting vectorwasdesigned to introduce two

LoxPsitesflankingexon2of theHO-1gene locus.Eliminationof exon2 leads toaconsecutive frameshift in exon3, anearly stopcodon,

thus resulting in a truncatedpeptideconsistingof 9aminoacids. The targetingconstructwaselectroporated intoC57BL/6Nembryonic

stemcells. TheprimaryEScellswerescreenedandpotentially targetedcloneswereexpanded.Southernblot confirmationwasused to

select correctly targeted cells that were subjected to removal of the Frt-flanked neomycin cassette by Flp-recombinase, and sub-

sequently injected into blastocysts. Offsprings were tested for germline transmission and two correctly targeted mouse lines were

established. C57BL/6N founder mice were backcrossed onto the more insulin resistance prone C57BL/6J background using speed

congenics (purity > 99.9%). Hepatocyte-specific HO-1 knockout (Lhoko) mice were generated by crossing the conditional HO-1

line with Alb-Cre transgenic mice expressing Cre recombinase under the control of the albumin promoter (Postic et al., 1999). Mice

with a targeted deletion in macrophages (Macho) were generated by crossing the conditional HO-1 line with LysM-Cre knockin

mice having the Cre recombinase inserted into the first coding ATG of the lysozyme 2 gene (Clausen et al., 1999). Muscle-, adipose-

and beta-cell targeted HO-1 knockout mice were generated using Mck-cre (Bruning et al., 1998), Fabp4/aP2-cre (He et al., 2003)

and Rip-cre (Postic et al., 1999) knockin mice, respectively. Heterozygous breeding schemes (male HO-1fl/+;Cre+/� X female HO-

1fl/+;Cre�/�) were used to generate Lhoko andMacho animals, muscle (Muhoko), adipose (Fahoko) and b-cell (Bhoko) HO-1 knockout

mice as well as their respective wild-type (HO-1+/+;Cre�/�), conditional (HO-1fl/fl;Cre�/�) and Cre control littermates (HO-1+/+;Cre+/�).Offspringswere found tobebornat expectedMendelian ratiosandnodifference in survival rateswasobserved.Animalswerekeptona

12 hr light/dark cycle with free access to food andwater and housed in accordancewith international guidelines. Dietary interventions

started at 6 weeks of age using standardized low- and high-fat diets which contained 10% and 60% calories of fat, respectively

(Research Diets Inc., D12450B and D12492). Animal studies were approved by Austrian and German governments.

Heme Oxygenase ActivityHO activity was determined as reported elsewhere (Kozlov et al., 2010). In brief, aliquots of frozen liver (Lhoko) or cell pellets (Macho)

werehomogenizedusingaElvehjempotterwithPTFEpestle in abuffer 1:10 (wt/vol) containing300mMsucrose, 20mMTris, and2mM

EDTAat pH7.4. Protein concentrationswere determinedusingBradford reagent andBSAas standard. For the activity assay, 100 ml of

homogenate (containing about 1 mg of protein) was added to a reaction mixture containing 500 nmol NADPH in a 100mMpotassium

phosphate buffer (pH 7.4) supplementedwith 1mMEDTAand 20 nmol of hemin. 30 ml rat kidney cytosolic (RKC) fractionwas added to

the reaction to provide sufficient biliverdin reductase activity. RKCwas prepared as described elsewhere (McCoubrey, 2001). The re-

action mixture was incubated under constant agitation for 30 min at 37�C in darkness. After adding 0.2 assay volumes of saturated

potassiumchloride, the formedbilirubinwas extracted into chloroform (4 assay volumes). Sampleswere then centrifuged, the organic

phase harvested for subsequent spectrophotometric determination of bilirubin concentrations. Sampleswere scanned three times to

calculate the absorption difference between 450 and 520 nm. All samples were run in duplicates and corrected for the absorption

measured in corresponding samples incubatedon ice.Bilirubin standardcurvesweregeneratedbyspiking knownamountsof bilirubin

into sample homogenates. Heme oxygenase activity was calculated as nanomol bilirubin formed per milligram protein per 30 min.

Hemin InjectionHemin (2.5 mg/ml) was dissolved in 10% ammonium hydrochloride in 0.15M sodium chloride and injected intraperitoneally into mice

(10 ml/g body weight). Livers were harvested 12 hr postinjection to test HO-1 induction.

Mouse Laboratory Parameters and CytokinesAlanine transaminase (ALT), Aspartate transaminase (AST), cholinesterase (ChE), alkaline phosphatase (ALKP), triglycerides and

cholesterol were quantified with tests certified for in vitro diagnostics at the Department of Laboratory Medicine of the Medical

S2 Cell 158, 25–40, July 3, 2014 ª2014 Elsevier Inc.

University of Vienna. Free fatty acids were measured using the nonesterified fatty acid (NEFA) kit (Wako Chemicals) according to the

provided protocol. Blood samples were analyzed on an automated hematology analyzer (CELL-DYN 3500, Abbott Laboratories) and

validated manually with Giemsa stained blood smears. IL-6, TNF-a, IL-1b and IL-10 ELISA kits (all from Biolegend) were used to

determine the respective cytokine levels in mouse serum and supernatants derived from cultured macrophages according to the

protocol provided by the manufacturer. For hepatic triglyceride and cholesterol measurements, total lipids were extracted from

50 mg liver as previously described (Folch et al., 1957). Serum leptin, adiponectin and RBP4 levels were quantified using commer-

cially available ELISA kits (Millipore, R&D Systems).

Oil Red O StainingFor staining of neutral lipids, liver cryosections were stained with oil red O (1% w/v isopropanol, diluted 3:2 in PBS) for 1 hr at room

temperature according to standard procedures.

In Vivo Insulin SignalingFor examination of in vivo insulin signaling, mice were fasted for 2 hr, anesthetized and injected with human regular insulin (1 U/kg,

Novo Nordisk). Liver, muscle and adipose biopsies were removed at the indicated times, flash-frozen in liquid nitrogen and stored at

�80�C until further processing.

Indirect CalorimetryIndirect calorimetry and activity measurements were performed as reported elsewhere (Pospisilik et al., 2010). In brief, body-weight-

matched mice were placed in metabolic cages connected to an open-circuit, indirect calorimetry system combined with the deter-

mination of spontaneous activity by beam breaking (Oxylet, Panlab-Bioseb). The animals were accustomed to the apparatus during

the first 24 hr, followed by measurements. Oxygen consumption and carbon dioxide production were recorded using a computer-

assisted data acquisition program (Metabolism 2.2.01, Panlab Harvard Apparatus).

Generation of Bone-Marrow-Derived MacrophagesBone marrow cells were obtained from femurs and tibias of 6-10 week old control and Macho mice. Bone-marrow-derived macro-

phages (BMDMs) were generated in RPMI-1640 supplemented with 10% fetal bovine serum (FBS) and 10% supernatant derived

fromM-CSF transduced L929 confluent cells as described (Weischenfeldt and Porse, 2008; Zanoni et al., 2009). At day 7 postinduc-

tion macrophages were collected and cultured in the presence or absence of stimuli in RPMI-1640 medium supplemented with 10%

FBS. BMDMs were polarized for 24 hr with 10 ng/ml LPS and 20 ng/ml Interferon-g (R&D Systems) to generate classically activated

M1 macrophages. Alternatively activated M2 macrophages were generated using IL-4/IL-13 (20 ng/ml each; R&D Systems). For

NF-kB signaling experiments macrophages were treated with 10 ng/ml TNF-a (Peprotech) for the indicated times.

Histology, Adipocyte Size, and NumberFor tissue sections, hematoxylin and eosin (H&E) staining was performed on 5 mm paraffin sections of tissues fixed for 16 hr in 4%

phosphate-buffered formalin at 4�C. Adipocyte size distribution was determined by semi-automated morphometry. In brief, three

fields of view of 3 different sections (200 mm intervals) per animal were quantified. Epididymal fat pads from 4 animals per group

were analyzed. Semi-automated morphometry (Axiovision morphometry software, Zeiss) was used to define and quantify fat cells

based on shape, size and presence of a lipid droplet.

Whole-Mount ImmunofluorescenceWhole-mount epididymal fat samples were processed and stained as described (Lumeng et al., 2007; Todoric et al., 2011). In brief,

mice (three per group) were perfused with 1% paraformaldehyde for fixation before dissecting fat pads. Samples were incubated for

1 hr at room temperature with 5% normal goat serum (Dako) in 0.3% PBS-T to block unspecific binding sites. After blocking, tissues

were incubated overnight at 4�C with primary and secondary antibodies diluted in PBS containing 5% BSA. Anti-perilipin (Cat.-Nr.

9349, Cell Signaling Technology) and anti-MAC-2 antibodies (Cat.-Nr. CL8942AP, Cedarlane Laboratories) were used to stain adipo-

cyte lipid droplets and macrophages, respectively. The following secondary antibodies were used for immunofluorescence staining:

fluorescein anti-rat IgG (Cat.-Nr. FI-4001, VECTOR Laboratories), Alexa Fluor-594 goat anti-rabbit IgG (Cat.-Nr. A11037, Life Tech-

nologies). For control experiments, the primary antibody was substituted with preimmune serum. Stained tissues were mounted in

chamber slides and subjected to confocal imaging analysis on a LSM 700 Laser Scanning Microscope (Zeiss).

Flow CytometryStromal vascular cells (SVCs) were obtained by collagenase digestion from epididymal adipose tissue as described previously

(Lumeng et al., 2007; Pospisilik et al., 2010). In brief, minced fat depots were digested at 37�C for 30 min in a shaking water bath

with a cocktail consisting of 1 mg/ml collagenase II (Worthington) in DMEM containing 3% fatty-acid-free BSA and DNase I

(100 Units/ml). After digestion, the slurry was passed through a 100 mm cell strainer (Becton Dickinson) and centrifuged at 200 g

for 5 min to separate SVCs and adipocyte fractions. Adipocyte fractions were examined by microscopy to detect adherent cells.

Digestion was continued for another 15 min until adipocyte fractions were free of adherent cells to guarantee isolation of most

Cell 158, 25–40, July 3, 2014 ª2014 Elsevier Inc. S3

adipose macrophages. Cells were then pelleted by centrifugation at 200 g for 5 min, resuspended in 0.5 ml red blood cell lysis buffer

and incubated for 5 min prior to resuspension in sorting buffer (PBS with 2% FBS) at a concentration of 106 cells/ml. Isolated cells

were incubated with Fc Block (BD Biosciences) prior to staining with conjugated antibodies for 30 min at 4�C followed by 2 washes in

sorting buffer. Cells were resuspended in sorting buffer supplemented with Fixable Viability Dye eFluor (eBioscience) and subjected

to FACS analysis (LSR Fortessa, BDBiosciences). The following antibodies were used for staining adipose tissuemacrophages: anti-

F4/80 PE/Cy7 (Cat.-Nr. 123113, BioLegend), anti-CD11b Percp/Cy5.5 (Cat.-Nr. 101228, BioLegend), anti-CD11c Alexa Fluor700

(Cat.-Nr. 56-0114-82, eBioscience), anti-CCR2 APC (Cat.-Nr. FAB5538A, R&D Systems) and anti-MGL Alexa Fluor 647 (Cat.-Nr.

MCA2392A647T, AbD Serotec). Viable cells were gated and analyzed using FlowJo (Tree Star).

Nuclear Translocation AssayTo assess nuclear RelA/p65 translocation naive BMDMs were seeded in 8-well chamber slides (BD Biosciences). Cells were either

stimulated with 10 ng/ml TNF-a or medium only as control for 30 and 60 min, fixed in ice-cold methanol for 10 min, washed and

stained overnight at 4�Cwith anti-RelA/p65 antibody (sc-372, Santa Cruz Biotechnology). After several washes cells were incubated

with a fluorescent secondary antibody (Alexa Fluor 488, Life Technologies) for 1 hr at room temperature. Nuclei were counterstained

with DAPI. Images were obtained using a Zeiss Axio Imager A1 microscope and evaluated in a double-blinded manner by two

different researchers.

Electrophoretic Mobility Shift AssayNuclear extracts were prepared from 1x107 BMDMs. Adherent cells were washed with ice-cold PBS, transferred and washed three

times in 1 ml of ice-cold hypotonic buffer (10 mMHEPES pH 7.9, 1.5 mMMgCl2, 10mMKCl) to induce swelling. Thereafter, the pellet

was resuspended in hypotonic buffer supplemented with 0.1% Nonidet P-40 and incubated on ice for 5 min to release nuclei. Sub-

sequently, samples were centrifuged at 4�C to pellet nuclei and the cytoplasmic fraction was discarded. Nuclei were resuspended in

50 ml of ice-cold, high-salt buffer (20mMHEPES pH7.9, 1.5mMMgCl2, 420mMNaCl, 25%glycerol) and incubated for 15min at 4�C.Disrupted nuclei were centrifuged at full speed in a table top centrifuge for 15 min at 4�C to isolate the nuclear protein fraction. EMSA

was performed using the NF-kB EMSA kit and an IRDye 700 labeled double-stranded probe harboring the NF-kB consensus site (LI-

COR Biosciences) according to the manufacturer’s instructions. For the binding assay, 5 mg of nuclear extracts were added to bind-

ing buffer (10 mM Tris, 50 mM KCl, 1 mM DTT, pH 7.5) together with 50 ng poly (dI:dC), 2.5 mM DTT/0.25% Tween-20 and 0.5 pmol

IRDye labeled probe in a total volume of 20 ml. For supershift assays 0.2 mg of the following antibodies were used: anti-RelA/p65 (sc-

372, Santa Cruz Biotechnology), anti-p50 (06-886, Millipore) and anti-rabbit IgG (S-5000, Vector Laboratories). Unlabeled double-

stranded NF-kB probe was used at a 100-fold molar excess to test specific binding. Samples were incubated at room temperature

for 15min and separated on a prerun nondenaturing polyacrylamide gel in TGE buffer (25mMTris-HCl pH 8.0, 190mMglycine, 1mM

EDTA) for 3 hr at 100 V. NF-kB DNA binding was visualized using the Odyssey Imaging System (LI-COR Biosciences).

Chromatin Immunoprecipitation AssayChromatin immunoprecipitation (ChIP) assayswere performed using theMAGnifyChIPKit (Life Technologies) as detailed by theman-

ufacturer’s instruction. In brief, 2 3 106 cells were crosslinked for 10 min at 37�C by adding formaldehyde (methanol-free, Thermo

Scientific) into the culturing media to a final concentration of 1%. Reactions were stopped by adding glycine to a final concentration

of 0.125 M. Cells were harvested, washed twice with ice-cold PBS and lysed in 100 ml lysis buffer containing protease inhibitors. Cell

lysates were transferred to microTUBEs (Covaris) and subjected to shearing on the Covaris S2 sonicator (4 cycles; cycle time: 60 s,

peak power: 140, duty factor: 5, cycles/burst: 200). Dynabeads (Life Technologies) were precoupled with anti-RelA/p65 (sc-372,

Santa Cruz Biotechnology) or rabbit IgG according to the protocol provided. 3 mg of antibody was used for each ChIP experiment.

Sheared chromatin DNA was diluted to 200,000 cells per ChIP reaction and incubated with antibody-coupled beads at 4�Cwith con-

stant rotation for 4 hr. Input controls were set aside for each IP sample. Bound chromatin waswashedwith a series of washing buffers

included in the kit and formaldehyde crosslinkingwas reversed using Proteinase K. DNAwas purified usingDNApurificationmagnetic

beads provided in the kit and eluted in 150 ml of DNA Elution Buffer. Precipitated DNA was subjected to quantitative PCR using

primers 50-CCCCAGATTGCCACAGAATC-30 and 50-CCAGTGAGTGAAAGGGACAG-30 for the Tnf promoter (NF-kB site 1) (Kuwata

et al., 2006) and 50-CCTAGTGAGTCCCAGTTTTGAAGT-30 and 50-CATCAGGTATTTATACCCCTCCAG-30 for the proximal NF-kB

site in the Nos2/iNos promoter (Guo et al., 2008). Data were analyzed according to the MAGnify ChIP Kit (Life Technologies) manual.

Oxygen Consumption Assay and Bioenergetic ProfileAnalysis of oxygen consumption rates (OCR) was performed using the XF24 Flux Analyzer (Seahorse Bioscience) as reported

previously (Haschemi et al., 2012; Teperino et al., 2012). In brief, naive bone-marrow-derivedmacrophages and primary hepatocytes

were seeded into XF 24-well cell culture microplates and allowed to recover for 24 hr. A final volume of 600 ml of buffer-free Assay

Medium (Seahorse Bioscience) was added to each well prior to the experimental protocol. Cells were then transferred to a CO2-free

incubator and maintained at 37�C for 1 hr before starting the assay. Following instrument calibration, cells were transferred to the

XF24 Flux Analyzer to record cellular oxygen consumption rates. The measurement protocol consisted of 2 min mixture, 2 min

wait and 4 min OCR measurement times (macrophages), and 4 min mixture, 0 min wait and 3 min OCR measurement times (hepa-

tocytes). For the mitochondrial stress test ATP synthase was inhibited by injection of 1 mM oligomycin, followed by 3 mM

S4 Cell 158, 25–40, July 3, 2014 ª2014 Elsevier Inc.

FCCP-induced mitochondrial uncoupling to determine the spare/maximal respiratory capacity. Nonmitochondrial respiration was

determined after rotenone/antimycin A injection (1 mMeach). At the end of the assay, the mediumwas carefully aspirated and cellular

DNA measured using CyQuant (Life Technologies) to adjust for potential differences in cell densities. For acute ROS-quenching

experiments, 10 mMbuffered N-acetyl-cysteine (NAC) was added to naive BMDMs or primary hepatocytes 1 hr prior to mitochondrial

stress tests and OCR measurements. For adenoviral rescue experiments, primary hepatocytes isolated from Lhoko animals were

transduced with Adeno-LacZ (AdLacZ) or Adeno-HO-1 (AdHO-1) virus particles (10 pfu/cell). Mitochondrial stress tests and OCR

measurements were initiated 48 hr after transduction.

Primary Hepatocyte IsolationHepatocytes were isolated and cultured according to the protocols provided by the manufacturer of Liver Perfusion Medium and

Liver Digest Medium, respectively (Invitrogen). In brief, mice were anaesthetized by intraperitoneal injection of ketamine-xylazine

(10% ketamine, 5% xylazine). The liver was perfused in situ with Liver Perfusion Medium (Invitrogen), and digested using Liver Digest

Medium (Invitrogen). Digested livers were removed, minced, and filtered through a 70 mm cell strainer (BD Biosciences). Filtered

hepatocytes were washed in William’s E medium supplemented with 5% FBS, purified from the nonparenchymal cells by centrifu-

gation and plated in the respective experimental dishes. After 4 hr of incubation adherent hepatocytes were switched to serum-free

RPMI-1640 and incubated over night. Experiments were performed on the following day.

PTP1B Activity AssaySamples were homogenized in immunoprecipitation buffer (20 mM Tris pH 7.6, 150mMNaCl, 0.5 mMEDTA, 10% glycerol, protease

inhibitors, and 1% Triton X-100. Lysates were cleared by centrifugation for 15 min at 20.000 g. Liver (2 mg) and hepatocytes (1 mg)

samples were precleared for 30 min at 4�C by incubation with 20 ml protein G coupled Dynabeads (Life Technologies). Precleared

samples were incubated for 1 hr at 4�C with goat anti-PTP1B (Santa Cruz Biotechnology). Next, 20 ml protein G coupled Dynabeads

were added to capture PTP1B and incubated for 2 hr at 4�C. Samples were washed two times with immunoprecipitation buffer and

two times with phosphatase assay buffer (25 mM HEPES pH 7.2, 50 mM NaCl, 2 mM EDTA). Total PTP1B activity was measured

using 50 mM of para-Nitrophenylphosphate (New England BioLabs) as a substrate according to the manufacturer’s instructions.

Subsequently, immunoprecipitated PTP1B samples were boiled and subjected to immunoblotting using rabbit anti-PTP1B (Abcam).

PTP1B signals were quantified by densitometry. Total PTP1B activity results were normalized to the amount of PTP1B in the precip-

itates to obtain specific PTP1B activity values (1 Unit = hydrolysis of 1 nmol of para-Nitrophenylphosphate in 1 min).

Hydrogen Peroxide MeasurementsHydrogen peroxide (H2O2) production was measured using the Amplex Red Hydrogen Peroxide/Peroxidase Assay Kit (Invitrogen)

according to the protocol provided by the manufacturer.

Superoxide Dismutase AssaySamples were sonicated and centrifuged at 1,500 g for 5 min at 4�C in ice-cold HEPES buffer (20 mM) containing EGTA (1 mM),

mannitol (210 mM) and sucrose (70 mM). Superoxide dismutase (SOD) activity was measured using a commercially available test

kit (Cayman) according to manufacturer’s instructions.

Cytosolic ROS and Mitochondrial ContentBMDMs and primary hepatocytes were cultured in Lab-Tek II Chambered Coverglass polystyrene media chambers mounted to ex-

trathin borosilicate cover glass for optimum high-power inverted microscopy (Nalge Nunc). Cytosolic ROS were quantified as

communicated recently (Kuznetsov et al., 2011). In brief, cells were loaded for 20 min with 5 mM 20,70-Dichlorofluorescin diacetate

(DCF-DA; Invitrogen). Stained cells were imaged with an inverted confocal microscope (LSM 510, Zeiss), the AxioVision software

package (Zeiss) was used to analyze recorded images as described (Kuznetsov et al., 2011). Results were expressed as arbitrary

fluorescence intensity units. MitoTracker Deep Red 633 (500 nM, staining time 40min at 37�C) was used to analyze themitochondrial

content in isolated primary hepatocytes. In BMDMs mitochondrial content was determined using MitoTracker Green FM (Life Tech-

nologies) according to the manufacturer’s instructions. Stained cells were washed once with serum-free RPMI-1640 and incubated

at 37�C for 30min. Thereafter, cells were washed extensively with PBS, resuspended in FACS buffer (PBS, 2mMEDTA, 2%FBS) and

analyzed with the LSR II Flow Cytometer (Becton Dickinson). Result were analyzed using the FlowJo software package (Tree Star).

Adenovirus ExperimentsThe macrophage tropic RGD-fiber-modified (Dmitriev et al., 1998) Ad(RGD)-HO-1 and Ad(RGD)-GFP obtained commercially from

Vector Biolabs were used for Macho-BMDM rescue experiments. BMDMs were infected with 100 pfu/cell on day 4 of differentiation.

Experiments were performed on day 7 of differentiation, i.e., 3 days after adenoviral transduction. For adenovirus experiments target-

ing livers and isolatedprimary Lhokohepatocytes, in vivogradeHO-1 (AdHO-1) andLacZ (AdLacZ) expressing viruseswereproduced

according to published protocols (Hardy et al., 1997). In brief, Sfi I-digested Adlox plasmid DNAwas cotransfected with psi5 DNA into

Cre8 cells using FuGENE 6 Transfection Reagent (Roche). Three days after transfection cells were collected by centrifugation and

recombined viruses extracted from cell pellets by four freeze and thaw cycles. Cell debris was removed by centrifugation. HEK293

Cell 158, 25–40, July 3, 2014 ª2014 Elsevier Inc. S5

cells (ATCC) were used to amplify adenoviral particles. Amplified AdHO-1 and AdLacZ were purified by cesium chloride density-

gradient ultracentrifugation, collected from the gradient, diluted 1:1 in 2x storage buffer (10 mM Tris pH 8.0, 100 mM NaCl, 0.1%

BSA, 50% glycerol) and stored in small aliquots at �20�C. Primary Lhoko hepatocytes were infected with 10 plaque forming units

(pfu)/cell. Transduced cells were incubated for 48 hr before experiments were started. In vivo tail vein injections were performed

as described (Pospisilik et al., 2007). In brief, a total volumeof 200ml (0.23 109 pfu/g bodyweight) of PBS-dilutedAdHO-1 andAdLacZ

particles was injected into the tail vein of recipient mice. OGTT and ITT were performed on day 5 and 7 postinjection.

NRF2 DNA-Binding ActivityThe NRF2 TransAM ELISA-kit (Active Motif) was used to evaluate NRF2 DNA-binding activity according to the manufacturer’s pro-

tocol. In brief, 8 mg of nuclear extracts isolated from naive BMDMswere incubated for 1 hr in 96 well plates on which oligonucleotides

containing the antioxidant response element (ARE) consensus-binding site were immobilized. After binding and washing, NRF2 was

detected using an antibody specific for active, i.e., DNA-bound NRF2. Addition of a horseradish peroxidase (HRP) coupled second-

ary antibody was used to quantitate-binding activity by spectrophotometry.

Statistical AnalysisData are expressed as mean ± standard error of the mean (SEM) unless otherwise specified. Statistical significance was tested by

Student’s t test or ANOVAwith Scheffe’s post hoc testing where appropriate. Correlations were tested by linear regression. Categor-

ical datawere summarizedby frequencies andanalyzedby theChi-square test. TheWilcoxon ranksum testwasused toanalyzepublic

human data sets and energy expenditure data. Heteroskedasticity of human data was tested by the Breusch-Pagan/Cook-Weisberg

test, normal distribution of quantitative traits were tested by the Shapiro-Francia normality test. Logarithmic transformations were

made if the equal variance and normality assumptions were rejected. Adjustments for confounding variables were performed as indi-

cated. All figures andmouse statistical analyses were generated using Prism 6 (GraphPad). Human data were analyzed with Stata 12

(StataCorp). All reported p values are two-tailed unless stated otherwise. p < 0.05 was considered to indicate statistical significance.

SUPPLEMENTAL REFERENCES

Bruning, J.C., Michael, M.D., Winnay, J.N., Hayashi, T., Horsch, D., Accili, D., Goodyear, L.J., and Kahn, C.R. (1998). Amuscle-specific insulin receptor knockout

exhibits features of the metabolic syndrome of NIDDM without altering glucose tolerance. Mol. Cell 2, 559–569.

Clausen, B.E., Burkhardt, C., Reith, W., Renkawitz, R., and Forster, I. (1999). Conditional gene targeting in macrophages and granulocytes using LysMcre mice.

Transgenic Res. 8, 265–277.

Folch, J., Lees, M., and Sloane Stanley, G.H. (1957). A simple method for the isolation and purification of total lipides from animal tissues. J. Biol. Chem. 226,

497–509.

Guo, H., Mi, Z., and Kuo, P.C. (2008). Characterization of short range DNA looping in endotoxin-mediated transcription of the murine inducible nitric-oxide

synthase (iNOS) gene. J. Biol. Chem. 283, 25209–25217.

Hardy, S., Kitamura, M., Harris-Stansil, T., Dai, Y., and Phipps, M.L. (1997). Construction of adenovirus vectors through Cre-lox recombination. J. Virol. 71, 1842–

1849.

He, W., Barak, Y., Hevener, A., Olson, P., Liao, D., Le, J., Nelson, M., Ong, E., Olefsky, J.M., and Evans, R.M. (2003). Adipose-specific peroxisome proliferator-

activated receptor gamma knockout causes insulin resistance in fat and liver but not in muscle. Proc. Natl. Acad. Sci. USA 100, 15712–15717.

Kozlov, A.V., van Griensven, M., Haindl, S., Kehrer, I., Duvigneau, J.C., Hartl, R.T., Ebel, T., Jafarmadar, M., Calzia, E., Gnaiger, E., et al. (2010). Peritoneal inflam-

mation in pigs is associated with early mitochondrial dysfunction in liver and kidney. Inflammation 33, 295–305.

Kuwata, H., Matsumoto, M., Atarashi, K., Morishita, H., Hirotani, T., Koga, R., and Takeda, K. (2006). IkappaBNS inhibits induction of a subset of Toll-like recep-

tor-dependent genes and limits inflammation. Immunity 24, 41–51.

Kuznetsov, A.V., Kehrer, I., Kozlov, A.V., Haller, M., Redl, H., Hermann, M., Grimm, M., and Troppmair, J. (2011). Mitochondrial ROS production under cellular

stress: comparison of different detection methods. Anal. Bioanal. Chem. 400, 2383–2390.

McCoubrey, W.K., Jr. (2001). Detection of heme oxygenase 1 and 2 proteins and bilirubin formation. Curr Protoc Toxicol Chapter 9, Unit 9 3.

Oberkofler, H., Linnemayr, V., Weitgasser, R., Klein, K., Xie, M., Iglseder, B., Krempler, F., Paulweber, B., and Patsch,W. (2004). Complex haplotypes of the PGC-

1alpha gene are associated with carbohydrate metabolism and type 2 diabetes. Diabetes 53, 1385–1393.

Oberkofler, H., Pfeifenberger, A., Soyal, S., Felder, T., Hahne, P., Miller, K., Krempler, F., and Patsch, W. (2010). Aberrant hepatic TRIB3 gene expression in

insulin-resistant obese humans. Diabetologia 53, 1971–1975.

Pontiroli, A.E., Pizzocri, P., Giacomelli, M., Marchi, M., Vedani, P., Cucchi, E., Orena, C., Folli, F., Paganelli, M., and Ferla, G. (2002). Ultrasound measurement of

visceral and subcutaneous fat in morbidly obese patients before and after laparoscopic adjustable gastric banding: comparison with computerized tomography

and with anthropometric measurements. Obes. Surg. 12, 648–651.

Postic, C., Shiota, M., Niswender, K.D., Jetton, T.L., Chen, Y., Moates, J.M., Shelton, K.D., Lindner, J., Cherrington, A.D., and Magnuson, M.A. (1999). Dual roles

for glucokinase in glucose homeostasis as determined by liver and pancreatic beta cell-specific gene knock-outs using Cre recombinase. J. Biol. Chem. 274,

305–315.

Todoric, J., Strobl, B., Jais, A., Boucheron, N., Bayer, M., Amann, S., Lindroos, J., Teperino, R., Prager, G., Bilban,M., et al. (2011). Cross-talk between interferon-

g and hedgehog signaling regulates adipogenesis. Diabetes 60, 1668–1676.

Weischenfeldt, J., and Porse, B. (2008). Bone Marrow-Derived Macrophages (BMM): Isolation and Applications. CSH Protoc 2008, pdb prot5080.

Zanoni, I., Ostuni, R., Capuano, G., Collini, M., Caccia, M., Ronchi, A.E., Rocchetti, M., Mingozzi, F., Foti, M., Chirico, G., et al. (2009). CD14 regulates the dendritic

cell life cycle after LPS exposure through NFAT activation. Nature 460, 264–268.

S6 Cell 158, 25–40, July 3, 2014 ª2014 Elsevier Inc.

Figure S1. Conserved HO-1 Tissue Expression Profile and Metabolically Healthy and Unhealthy Obese C57BL/6J mice, Related to Figure 1

(A) Quantitative PCR analysis of HO-1 mRNA in human and mouse tissue libraries.

(B and C) Stratification of obese insulin-sensitive (IS) and obese insulin-resistant (IR) C57BL/6J mice fed a 60% fat diet for 16 weeks; (B) oral glucose tolerance

tests, and (C) intraperitoneal insulin tolerance tests, n = 6 per group.

(D) Body weight of obese IS and obese IR mice after 16 weeks on high fat diet. Results presented as mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001.

Cell 158, 25–40, July 3, 2014 ª2014 Elsevier Inc. S7

Figure S2. Generation of Conditional HO-1 Mice, Lhoko Low-Fat Diet Studies, Lipogenic and Gluconeogenic Program, and Adenoviral

Injections, Related to Figure 2

(A) Shown (top to bottom) are wild-type, targeted, conditional (floxed) and knockout HO-1 gene loci.

(B) Southern blot of correctly targeted ES cell clone.

(C) PCR for tissue-specific HO-1 deletion in Lhoko mice.

(D) Hematoxylin-eosin (H&E) staining of liver sections obtained from control and Lhoko mice fed a LFD for 16 weeks.

(E) Serum levels of liver enzymes, n = 4 per genotype.

(F) Oral glucose tolerance tests on LFD.

(G) Insulin tolerance tests on LFD.

(H) Body weight gain of control and Lhoko mice on HFD.

(I) Gluconeogenic and lipogenic program in HFD fed Lhoko livers, n = 4 per genotype.

(J) Relative HO-1 mRNA levels in AdLacZ and AdHO-1 transduced liver biopsies, n = 4 per group.

(K) HO-1 immunoblot of AdLacZ and AdHO-1 transduced liver biopsies, n = 3 per group.

(L) Oral glucose tolerance tests and corresponding blood glucose and insulin levels in AdLacZ and AdHO-1 transduced C57BL/6J animals, n = 7 per group. Two-

sided p values obtained from unpaired t test. Results are mean ± SEM. *p < 0.05.

S8 Cell 158, 25–40, July 3, 2014 ª2014 Elsevier Inc.

Figure S3. Characterization of Macrophage HO-1 Knockout Mice, Related to Figure 3

(A) Body weight gain of control and Macho mice on LFD.

(B) White blood cell counts (WBC) in LFD fed control and Macho mice, n = 4 per group.

(C) Daily food intake (% of body weight, BW) of lean control and Macho mice, n = 8 per group.

(D) Oral glucose tolerance tests and corresponding blood glucose and insulin levels on LFD.

(E) Insulin tolerance tests on LFD.

(F) Serum-free fatty acid levels measured in fed and overnight fasted states of control and Macho mice on LFD, n = 6 per group. (G) Daily food intake (% BW) of

control andMachomice on HFD, n = 8 per group. (H) Serum levels of liver enzymes determined in control andMachomice after 16weeks onHFD, n = 3 per group.

(I) In vivo activation of liver insulin signaling. HFD fed mice were injected with insulin, tissue biopsies taken at the indicated times, and analyzed for pINSR, INSR,

pAKT and AKT by immunoblotting. (J) Respiratory quotient (RQ) of body-weight-matched control and Macho animals, n = 7 per genotype. Results are mean ±

SEM. *p < 0.05.

Cell 158, 25–40, July 3, 2014 ª2014 Elsevier Inc. S9

Figure S4. Muscle-, Adipose-, and Beta-Cell HO-1 Knockout Animals and Macrophage Markers in Macrophage HO-1 Knockout Mice,

Related to Figure 4

(A) Muscle-HO1 knockout mice (Muhoko) on HFD: body weight gain, glucose and insulin tolerance tests.

(B) Adipose-HO1 knockout mice (Fahoko) on HFD: body weight gain, glucose and insulin tolerance tests.

(C) Beta-cell-HO1 knockout mice (Bhoko) on HFD: body weight gain, glucose and insulin tolerance tests.

(D) Gating strategy used in flow cytometry quantification of total macrophage numbers and their subpopulations. Stromal vascular cells obtained from epididymal

fat pads were gated for F4/80+/CD11b+ cells to obtain total macrophage numbers. Double-positive cells were examined for CD11c, MGL and CCR2 expression.

The percentage of CD11c+ and MGL+ cells within the F4/80+/CD11b+ macrophage population is indicated. A representative result from three independent

experiments is shown.

(E) Relative mRNA expression of HO-1/Hmox1 and macrophage markers Cd68, Cd11c/Itgax and Mgl2 in epididymal adipose tissue from HFD fed control and

Macho mice, n = 4 per group. (F) Relative mRNA expression of cytokine receptors in epididymal adipose tissue from HFD fed control and Macho mice, n = 4 per

group. Results are mean ± SEM. *p < 0.05, **p < 0.01.

S10 Cell 158, 25–40, July 3, 2014 ª2014 Elsevier Inc.

Figure S5. Mitochondrial Profiling of HO-1 Knockout Animals and HO-1 Rescue Studies, Related to Figure 6

(A and B) Basal respiration, maximal respiration, spare respiratory capacity (spare) and coupling efficiency of primary hepatocytes as well as naive BMDMs

derived from Lhoko and Macho animals and their littermate controls, n = 5 for hepatocytes, n = 9 for macrophages.

(C) Relative mitochondrial DNA (mtDNA) levels of HO-1 knockout hepatocytes and macrophages as well as their corresponding controls, n = 3 per group.

(D)MitoTracker green andMitoTracker deep red fluorescencewas recorded tomeasuremitochondrial mass of HO-1 knockoutmacrophages and hepatocytes as

well as their corresponding controls. AU; arbitrary units.

(E) Protein extracts from hepatocytes andmacrophages were analyzed using an antibody cocktail for protein components of the electron transport chain. b-Actin

was used as internal loading control, n = 4 per genotype.

(F) Cytosolic ROS measurements in HO-1 knockout macrophages and hepatocytes as well as their corresponding controls; FI, fluorescence intensity.

(G) In vitro activation of hepatocyte insulin signaling in AdLacZ and AdHO-1 transduced primary hepatocytes isolated from hepatocyte HO-1 knockout animals.

Primary hepatocytes were stimulated with insulin and analyzed for pINSR, INSR, pAKT and AKT by immunoblotting. HO-1 was used to control for adenoviral

HO-1 expression, n = 3 per group.

(H) Oxygen consumption rates (OCR) and mitochondrial function of AdLacZ and AdHO-1 transduced primary hepatocytes. Cells were derived from hepatocyte

HO-1 knockout animals (Lhoko) and transduced in vitro, n = 6 per group.

(I) Cytokine secretion of LPS (10 ng/ml) stimulated Macho BMDMs transduced with fiber type AdGFP and AdHO-1 virus particles, n = 3 per group. Results are

mean ± SEM. *p < 0.05, **p < 0.001.

Cell 158, 25–40, July 3, 2014 ª2014 Elsevier Inc. S11