Glucoraphanin Ameliorates Obesity and InsulinResistance Through Adipose Tissue Browning andReduction of Metabolic Endotoxemia in MiceNaoto Nagata,1 Liang Xu,1 Susumu Kohno,2 Yusuke Ushida,3 Yudai Aoki,3 Ryohei Umeda,3 Nobuo Fuke,3

Fen Zhuge,1 Yinhua Ni,1 Mayumi Nagashimada,1 Chiaki Takahashi,2 Hiroyuki Suganuma,3 Shuichi Kaneko,4

and Tsuguhito Ota1,4

Diabetes 2017;66:1222–1236 | DOI: 10.2337/db16-0662

Low-grade sustained inflammation links obesity to insulinresistance and nonalcoholic fatty liver disease (NAFLD).However, therapeutic approaches to improve systemicenergy balance and chronic inflammation in obesity arelimited. Pharmacological activation of nuclear factor (ery-throid-derived 2)–like 2 (Nrf2) alleviates obesity and insulinresistance in mice; however, Nrf2 inducers are not clinicallyavailable owing to safety concerns. Thus, we examinedwhether dietary glucoraphanin, a stable precursor of theNrf2 inducer sulforaphane, ameliorates systemic energy bal-ance, chronic inflammation, insulin resistance, and NAFLDin high-fat diet (HFD)–fed mice. Glucoraphanin supplemen-tation attenuated weight gain, decreased hepatic steatosis,and improved glucose tolerance and insulin sensitivity inHFD-fed wild-type mice but not in HFD-fed Nrf2 knockoutmice. Compared with vehicle-treated controls, glucoraphanin-treated HFD-fed mice had lower plasma lipopolysaccha-ride levels and decreased relative abundance of thegram-negative bacteria family Desulfovibrionaceae in theirgut microbiomes. In HFD-fed mice, glucoraphanin increasedenergy expenditure and the protein expression of uncouplingprotein 1 (Ucp1) in inguinal and epididymal adipose depots.Additionally, in this group, glucoraphanin attenuated hepaticlipogenic gene expression, lipid peroxidation, classically ac-tivatedM1-likemacrophage accumulation, and inflammatorysignaling pathways. By promoting fat browning, limiting met-abolic endotoxemia-related chronic inflammation, and mod-ulating redox stress, glucoraphanin may mitigate obesity,insulin resistance, and NAFLD.

Low-grade sustained inflammation, triggered by chroni-cally high levels of proinflammatory cytokines and gutmicrobiota–derived circulatory lipopolysaccharide (LPS),links obesity with comorbidities such as insulin resistanceand nonalcoholic fatty liver disease (NAFLD) (1,2). Al-though a number of pharmacological treatments for obe-sity and NAFLD have been tested, few drugs are clinicallyavailable owing to the lack of long-term efficacy and safetyconcerns (3,4). Thus, a novel therapeutic approach thatwould improve energy metabolism and reduce chronic in-flammation in obesity is sorely needed.

Nuclear factor (erythroid-derived 2)–like 2 (Nrf2), a ba-sic leucine zipper transcription factor, is widely expressedin human and mouse tissues and serves as a defense re-sponse against extrinsic and intrinsic stressors (5). Uponexposure to electrophilic and oxidative stress, Nrf2 de-taches from its repressor Kelch-like ECH-associated protein1 nuclear factor (Keap1), and is translocated from the cy-toplasm into the nucleus. This translocation leads to thetranscriptional activation of genes encoding phase 2 detox-ifying and antioxidant enzymes (6). In addition to theubiquitous induction of cytoprotective genes, Nrf2 regu-lates a large number of genes involved in glucose and lipidmetabolism. In the liver, the constitutive activation of Nrf2through Keap1 knockdown represses the expression ofgenes involved in gluconeogenesis (7) and lipogenesis (8),thereby alleviating obesity, diabetes, and hepatic steatosis.Accordingly, synthetic Nrf2 inducers, such as synthetic

1Department of Cell Metabolism and Nutrition, Brain/Liver Interface MedicineResearch Center, Kanazawa University, Kanazawa, Ishikawa, Japan2Division of Oncology and Molecular Biology, Cancer Research Institute, Kana-zawa University, Kanazawa, Ishikawa, Japan3Research and Development Division, Kagome Co., Ltd., Nasushiobara, Tochigi,Japan4Department of Disease Control and Homeostasis, Kanazawa University GraduateSchool of Medical Science, Kanazawa, Ishikawa, Japan

Corresponding author: Tsuguhito Ota, tota@staff.kanazawa-u.ac.jp.

Received 25 May 2016 and accepted 8 February 2017.

This article contains Supplementary Data online at http://diabetes.diabetesjournals.org/lookup/suppl/doi:10.2337/db16-0662/-/DC1.

© 2017 by the American Diabetes Association. Readers may use this article aslong as the work is properly cited, the use is educational and not for profit, and thework is not altered. More information is available at http://www.diabetesjournals.org/content/license.

1222 Diabetes Volume 66, May 2017

OBESITY

STUDIES

triterpenoid 2-cyano-3,12-dioxoolean-1,9-dien-28-oic acid(CDDO)-imidazolide (9), CDDO-methyl ester (known asbardoxolone methyl) (10), and dithiolethione analogoltipraz (11), have been shown to ameliorate high-fatdiet (HFD)–induced obesity and diabetes. These syntheticNrf2 inducers also decrease liver and adipose tissue lipo-genesis and enhance glucose uptake in skeletal muscles.However, the mechanisms by which Nrf2 enhances energymetabolism in response to an HFD remain largely un-known. Although enhanced Nrf2 signaling has shownpromising results in several animal studies, the syntheticNrf2 inducers have caused adverse cardiac events and gas-trointestinal toxicities in clinical trials (12,13). These ob-servations prompted us to explore a safer Nrf2 inducer forthe treatment of obesity, insulin resistance, and NAFLD.

Sulforaphane, an isothiocyanate derived from crucifer-ous vegetables, is one of the most potent naturallyoccurring Nrf2 inducers; this compound exhibits antican-cer activity in cancer cell lines and in carcinogen-inducedrodent models (14). Among the cruciferous vegetables,broccoli sprouts are the best source of glucoraphanin, astable glucosinolate precursor of sulforaphane (15). Inboth rodents and humans, glucoraphanin is hydrolyzedby gut microbiota-derived myrosinase into bioactive sul-foraphane before intestinal absorption (16). A recent clin-ical study demonstrated the safety of orally administeredglucoraphanin (17). In the current study, we examined thedietary glucoraphanin-mediated modulation of systemicenergy balance and the mitigation of chronic inflamma-tion, insulin resistance, and NAFLD in diet-induced obesemice.

RESEARCH DESIGN AND METHODS

Glucoraphanin PreparationThe sulforaphane precursor glucoraphanin was preparedas previously described (17) with minor modifications.Briefly, 1 day after germination from broccoli seeds (Cau-dill Seed Company, Louisville, KY), sprouts were boiled inwater for 30 min. The water extract was mixed with dex-trinized cornstarch and subsequently spray dried to yieldan extract powder containing 135 mg of glucoraphaninper gram (0.31 mmol/g) (18). The total glucoraphanintiter in the resulting powder was determined by high-performance liquid chromatography as previously reported(19).

Mice and DietsMale C57BL/6JSlc mice were purchased from Japan SLC(Hamamatsu, Japan) at 7 weeks of age. The Nrf2 knockout(Nrf22/2) mouse strain (RBRC01390; C57BL/6J back-ground) was provided by RIKEN BioResource Center (Tsu-kuba, Japan) (6). After 1 week of acclimation, mice werefed normal chow (NC) (containing 2.2% dextrinized corn-starch, 10% kcal from fat, #D12450B; Research Diets, NewBrunswick, NJ), NC containing 0.3% glucoraphanin (NC-GR) (containing 2.2% extract powder), an HFD (containing2.2% dextrinized cornstarch, 60% kcal from fat, #D12492;

Research Diets), or an HFD containing 0.3% glucoraphanin(HFD-GR) (containing 2.2% extract powder) for 14 weeks.Both the NC and the HFD containing cornstarch or glucor-aphanin were prepared by Research Diets. All mice studiedwere maintained on a 12-h light/dark cycle at 24–26°Cwith free access to water and food. All animal procedureswere performed in accordance with the Guidelines for theCare and Use of Laboratory Animals at Kanazawa Univer-sity, Japan.

Indirect CalorimetryAfter 3 weeks of feeding, mice were individually housed inan indirect calorimeter chamber at 24–26°C (Oxymax;Columbus Instruments, Columbus, OH). Calorimetry,daily body weight, and daily food intake data were ac-quired during a 3-day acclimation period followed by a2-day experimental period. VO2 and VCO2 were measuredin each chamber every 20 min. The respiratory exchangeratio (RER = VCO2/VO2) was calculated with use of Oxy-max software. Energy expenditure was calculated asVO2 3 (3.815 + [1.232 3 RER]) and normalized to thebody mass of each mouse.

Metabolic Measurements and Biochemical AnalysesMetabolic parameters, body fat composition, insulinsensitivity, and glucose tolerance were assessed as pre-viously described (20). Plasma LPS levels were analyzedwith a Limulus Amebocyte Lysate assay kit (QCL-1000;Lonza, Allendale, NJ). Plasma LPS-binding protein (LBP)levels were determined with an ELISA kit (Enzo Life Sci-ences, Farmingdale, NY). Immunoblotting was performedwith primary antibodies (Supplementary Table 1) as pre-viously described (20). mRNA expression levels were de-termined by quantitative real-time PCR that used SYBRGreen with the primers (Supplementary Table 2) as pre-viously described (20).

Isolation and Differentiation of Inguinal White AdiposeTissue–Derived Primary Beige AdipocytesStromal vascular fractions from inguinal white adiposetissue (WAT) of 7-week-old wild-type and Nrf22/2 micewere prepared as previously reported (21). At confluence,stromal vascular fraction cells were induced for 2 dayswith differentiation medium containing DMEM/F-12 sup-plemented with 10% FBS, 20 nmol/L insulin, 1 nmol/LT3, 5 mmol/L dexamethasone, 500 mmol/L isobutylme-thylxanthine, 125 mmol/L indomethacin, and 0.5 mmol/Lrosiglitazone (all from Sigma-Aldrich, St. Louis, MO). In-duced cells were subsequently cultured in maintenancemedium (DMEM/F-12 containing 10% FBS, 20 nmol/Linsulin, and 1 nmol/L T3) for 5 days and treated withDMSO or sulforaphane (Toronto Research Chemicals,Toronto, Ontario, Canada) at the indicated concentra-tions for 48 h.

FACSCells from the liver and epididymal WAT were preparedas previously described (22). Isolated cells were incu-bated with Fc-Block (BD Bioscience, San Jose, CA) and

diabetes.diabetesjournals.org Nagata and Associates 1223

subsequently incubated with fluorochrome-conjugated an-tibodies (Supplementary Table 1). Flow cytometry wasperformed on a FACSAria II (BD Bioscience), and thedata were analyzed with FlowJo software (Tree Star, Ash-land, OR).

Analysis of Gut Microbiota Through Pyrosequencingof the 16S rRNA GeneMetagenomic DNA was extracted from mouse cecalcontent with a QIAamp DNA Stool Mini Kit (QIAGEN,Hilden, Germany). The V1–V2 region of the 16S rRNAgene was amplified by using primer sets as previouslyreported (23). Mixed samples were prepared by poolingapproximately equal amounts of PCR amplicons fromeach sample and subjected to a GS Junior System (RocheDiagnostics, Basel, Switzerland) for subsequent 454 se-quencing. Preprocessing and taxonomic assignment of se-quencing reads were conducted as described previously(23) and separated by unique bar codes. The 16S rRNAsequence database was constructed by retrieving 16S se-quences of bacterial isolates (1,200–2,384 bases in length)from the Ribosomal Database Project Release 10.27. Weused 4,000 filter-passed reads of 16S sequences for theoperational taxonomic unit (OTU) analysis of each sam-ple. Clustering of 16S sequence reads with identityscores .96% into OTUs was performed by using theUCLUST algorithm (www.drive5.com). Representative se-quences with identity scores .96% for each OTU wereassigned to bacterial species by using the BLAST algo-rithm. Principal component analysis with EZR software(www.jichi.ac.jp/saitama-sct/SaitamaHP.files/statmedEN.html) was applied for assessment of alterations of cecalbacterial phylum associated with the diets.

Statistical AnalysesData were expressed as mean 6 SEM. P , 0.05 wasconsidered statistically significant. Statistical differencesbetween groups were determined by a two-tailed Studentt test. An overall difference among more than two groupswas determined by one-way ANOVA. If one-way ANOVAswere significant, differences between individual groupswere estimated by Bonferroni post hoc test. All calcula-tions were performed with SPSS version 19.0 statisticalsoftware (IBM Corporation, Armonk, NY).

RESULTS

Glucoraphanin Decreases Weight Gain and Adiposityand Increases Energy Expenditure in HFD-Fed MiceTo investigate the effects of glucoraphanin on systemicenergy balance, we examined the body weight of wild-typemice fed NC or an HFD supplemented with glucoraphanin orvehicle (i.e., cornstarch only). Glucoraphanin reduced weightgain only in HFD-fed mice without affecting food intake(Fig. 1A and Supplementary Fig. 1A). This reduction was notaccompanied by evidence of gross toxicity. We determinedthe plasma concentration of sulforaphane in NC-GR andHFD-GR mice, but not in NC or HFD mice, indicating thatglucoraphanin was absorbed as a sulforaphane after food

consumption (Supplementary Fig. 1B). The reduction ofweight gain in HFD-GR mice was largely attributed to de-creased fat mass, not to lean mass (Fig. 1B). To assess energyexpenditure, we placed the mice in indirect calorimetry cagesafter 3 weeks of feeding before an evident change in thebody mass of HFD-fed mice was observed (29.9 6 0.5 vs.28.5 6 0.5 g in HFD vs. HFD-GR, respectively). HFD-GRmice exhibited consistently higher VO2 and VCO2 than HFDcontrols (Fig. 1C and D), leading to increased energy expen-diture (Fig. 1E); however, they displayed a similar RER (Fig.1F), suggesting that glucoraphanin supplementation en-hanced sugar and fat use under HFD conditions. In NC-GRmice, these parameters of energy balance were not affected(Fig. 1B–F). Consistent with increased energy expenditure,glucoraphanin increased the core body temperature of HFD-fed mice by ;0.5°C (Fig. 1G).

Glucoraphanin Improves Diet-Induced InsulinResistance and Glucose ToleranceAfter 14 weeks of feeding, glucoraphanin supplementa-tion did not affect plasma triglyceride, total cholesterol,and free fatty acid (FFA) levels in either NC- or HFD-fedmice (Table 1). In NC-GR mice, blood glucose levels werenot altered by glucoraphanin, but HFD-GR mice exhibitedsignificantly lower fasted blood glucose compared withvehicle-treated controls (Table 1). Additionally, glucora-phanin significantly decreased plasma insulin concentra-tions in HFD-fed mice under both fasted and fedconditions, resulting in lower HOMA of insulin resistance(HOMA-IR) (Table 1). During the insulin tolerance test(ITT), glucoraphanin significantly enhanced the reductionin blood glucose levels in HFD-fed mice but not in NC-fedmice compared with the vehicle-treated controls (Fig. 2A).Glucoraphanin improved glucose tolerance in HFD-fedmice during the glucose tolerance test (GTT), but hadno effect in NC-fed mice (Fig. 2B). Insulin secretion dur-ing the GTT was not affected by glucoraphanin (data notshown). In line with increased insulin sensitivity, insulin-stimulated Akt phosphorylation on Ser473 was enhancedby glucoraphanin in the liver, muscle, and epididymalWAT of HFD-fed mice (Fig. 2C).

Glucoraphanin Does Not Exert Antiobesity andInsulin-Sensitizing Effects in Nrf22/2 MiceAlthough the Keap1-Nrf2 pathway is a well-known target ofsulforaphane, this isothiocyanate has also been reportedto modulate different biological pathways independent ofthe Keap1-Nrf2 pathway (24,25). To determine whetherthe antiobesity and insulin-sensitizing effects of glucora-phanin are mediated through Nrf2, the effects of glucor-aphanin on energy balance and glucose metabolism wereassessed in NC- and HFD-fed Nrf22/2 mice. Althoughfood intake and plasma concentration of sulforaphanein NC-GR or HFD-GR Nrf22/2 mice were comparablewith that detected in wild-type NC-GR or HFD-GR mice(Supplementary Fig. 1C and D), the effects of glucorapha-nin on weight gain after HFD feeding (Fig. 3A), VO2 (Fig.3B), VCO2 (Fig. 3C), energy expenditure (Fig. 3D), RER

1224 Antiobesity Effect of a Natural Nrf2 Inducer Diabetes Volume 66, May 2017

Figure 1—Glucoraphanin reduces weight gain and increases energy expenditure in HFD-fed mice. A: Body weight of mice fed NC, NC-GR,HFD, or HFD-GR (n = 9/group). B: Representative computed tomographic images of abdominal regions of mice fed the indicated diet for13 weeks. Pink and yellow areas represent visceral and subcutaneous fat, respectively. Bar graphs represent body fat mass and lean masscalculated from the computed tomography scan data. C–F: VO2 (C), VCO2 (D), energy expenditure (E ), and RER (F ) during light and darkcycles in mice fed the indicated diet for 3 weeks. G: Rectal temperature in mice fed the indicated diet for 6 weeks. Data are mean 6 SEM.*P < 0.05, **P < 0.01 vs. NC; #P < 0.05, ##P < 0.01 vs. HFD.

diabetes.diabetesjournals.org Nagata and Associates 1225

(Fig. 3E), rectal temperature (Fig. 3F), insulin sensitivity(Fig. 3G), and glucose tolerance (Fig. 3H) were abolishedby the Nrf2 deficiency. These data are consistent withcomparable plasma metabolic parameters between HFD-GRand HFD mice. These metabolic parameters include lipids,blood glucose, insulin, HOMA-IR, and liver enzymes suchas alanine transaminase (ALT) and aspartate transaminase(AST) (Supplementary Table 3).

Glucoraphanin Blocks HFD-Induced Reduction ofUcp1 Expression in WAT of Wild-Type Mice but Not inNrf22/2 MiceThe increased energy expenditure and body temperature ofHFD-GR mice suggest an increase in adaptive thermo-genesis. However, glucoraphanin supplementation hadlittle effect on the size and number of lipid droplets in theintrascapular brown adipose tissue (BAT) of HFD-fed wild-type mice (Supplementary Fig. 2A). In addition, the mRNAexpression of Ucps, PGC-1a, and deiodinase 2 in BAT andof Ucps in skeletal muscle was not altered by glucoraphaninsupplementation in NC- and HFD-fed wild-type mice (Sup-plementary Fig. 2B and C). In BAT, HFD increased Ucp1protein expression, but glucoraphanin did not alter theexpression in wild-type or Nrf22/2 mice (Fig. 4A).Brown-like adipocytes expressing Ucp1, also known as beigecells, exist in various WAT depots and can contribute tothermogenesis (26). Compared with NC, HFD significantlydecreased Ucp1 protein levels in epididymal and inguinalWAT of both wild-type and Nrf22/2 mice (Fig. 4A). Glucor-aphanin supplementation restored HFD-induced reductionin Ucp1 protein levels in epididymal and inguinal WAT ofwild-type mice but not in Nrf22/2 mice. To examinewhether the effects of glucoraphanin were fat cell autono-mous and Nrf2 mediated, we tested the effects of sulfora-phane, an active metabolite of glucoraphanin, on theexpression of brown fat–selective genes in primary beigeadipocytes obtained from inguinal WAT of wild-type andNrf22/2 mice. In wild-type beige adipocytes, treatmentwith sulforaphane induced the Nrf2 target gene NAD(P)H:quinone oxidoreductase 1 (Nqo1) (Fig. 4B) and antioxidant

genes (Supplementary Fig. 3A). Concurrently, sulforaphanesignificantly increased the mRNA expression of brown fat–selective genes, including Ucp1, Prdm16, Cidea, and Elovl3(Fig. 4B). In contrast, in Nrf2-deficient beige adipocytes,sulforaphane failed to activate Nrf2, as judged by unalteredmRNA expression of the target genes, and to promote theexpression of brown fat–selective genes (Supplementary Fig.3B and Fig. 4C). Of note, Nrf2-deficient beige adipocytesexhibited fewer differentiation levels associated with atten-uated lipid accumulation (Supplementary Fig. 3C) and lowermRNA expression of fatty acid binding protein 4 (Supple-mentary Fig. 3C) and brown fat–selective genes comparedwith wild-type beige adipocytes (Fig. 4C).

Glucoraphanin Reduces Hepatic Steatosisand Oxidative Stress in HFD-Fed MiceThe HFD caused hepatic steatosis and inflammation,eventually leading to steatohepatitis. As shown in Fig. 5A,the increase in liver weight caused by the 14-week HFD wasalleviated by glucoraphanin supplementation. Glucoraphaninalso attenuated HFD-induced hepatic steatosis (Fig. 5B). Ad-ditionally, compared with the HFD group, the lower levels ofplasma ALT, plasma AST, liver triglycerides, and liver FFAs inthe HFD-GR mice indicated that glucoraphanin alleviatedHFD-induced liver damage (Fig. 5C and D). The reductionin hepatic steatosis was accompanied by the decreased ex-pression of the following lipogenic genes: sterol regulatoryelement binding transcription factor 1c (Srebf1), fatty acidsynthase (Fasn), and peroxisome proliferator–activated re-ceptor g (Pparg) (Fig. 5E). Additionally, hepatic levels ofmalondialdehyde, a marker of lipid peroxidation, were in-creased by the HFD. Glucoraphanin attenuated lipid perox-idation (Fig. 5F) and decreased gene expression of theNADPH oxidase subunits gp91phox, p22phox, p47phox, andp67phox (Fig. 5E). The HFD led to a compensatory increasein the expression of genes involved in fatty acid b-oxidation(Ppara and Cpt1a) and antioxidative stress (Cat, Gpx1, andSod1) in the liver. However, glucoraphanin did not furtherincrease the expression of these genes in the liver of HFD-fed mice (Fig. 5E).

Table 1—Metabolic parameters in fed and fasted mice

NC NC-GR HFD HFD-GR

Plasma triglyceride (mg/dL) 139.2 6 9.4 140.7 6 8.1 144.6 6 5.4 123.6 6 9.2

Plasma total cholesterol (mg/dL) 161.2 6 7.7 161.4 6 3.5 212.6 6 2.1** 210.2 6 1.8**

Plasma FFA (mmol/L) 0.85 6 0.07 0.96 6 0.07 0.93 6 0.05 0.81 6 0.05

Blood glucose (mg/dL)Fed 126 6 6 124 6 3 152 6 4** 145 6 4*Fasted 77 6 3 76 6 3 131 6 9** 94 6 5##

Plasma insulin (ng/mL)Fed 1.4 6 0.2 1.0 6 0.2 5.3 6 0.5** 3.6 6 0.4**#Fasted 0.2 6 0.0 0.1 6 0.0 2.2 6 0.2** 1.3 6 0.2**##

HOMA-IR 1.0 6 0.1 0.6 6 0.2 17.7 6 1.6** 8.0 6 1.3**##

Data are mean 6 SEM (n = 9/group). Shown are blood glucose and plasma insulin levels of mice fed (ad libitum) or fasted for 16 h.Triglyceride, total cholesterol, and FFA levels were measured in fasting plasma. *P, 0.05, **P, 0.01 vs. NC; #P, 0.05, ##P, 0.01 vs.HFD.

1226 Antiobesity Effect of a Natural Nrf2 Inducer Diabetes Volume 66, May 2017

Glucoraphanin Suppresses HFD-InducedProinflammatory Activation of Macrophages in Liverand Adipose Tissue

In response to the HFD, liver-resident macrophages(Kupffer cells) increase the production of proinflamma-tory cytokines that promote insulin resistance andNAFLD in mice (27). In particular, chemokine (C-C motif)

ligand 2 (Ccl2) promotes the recruitment of chemokine(C-C motif) receptor 2 (Ccr2)–positive monocytic lineagesof myeloid cells into the liver (28). These recruited cellsproduce a large amount of proinflammatory mediatorsand activate a lipogenic program (28). Here, we found aprominent induction of tumor necrosis factor-a (Tnf-a),Ccl2, and Ccr2 in the liver of HFD-fed mice, which was

Figure 2—Glucoraphanin improves insulin sensitivity and glucose tolerance in HFD-fed mice. A: ITT (0.9 units/kg body weight for NC andNC-GR mice; 1.2 units/kg body weight for HFD and HFD-GR mice) after 11 weeks of feeding (n = 9/group). B: GTT (2 g/kg body weight)after 9 weeks (n = 9/group). Bar graphs represent area under the curve (AUC) calculations. C: Mice on the indicated diet for 14 weeks wereinjected intraperitoneally with saline (n = 2/group) or insulin (n = 6/group) (10 units/kg body weight), and sacrificed 10 min after injection.Total liver, quadriceps muscle, and epididymal fat depot (eWAT) lysates were immunoblotted for pS473 Akt and Akt, quantitated, andpresented as mean 6 SEM. *P < 0.05, **P < 0.01 vs. NC; #P < 0.05, ##P < 0.01 vs. HFD. A.U., arbitrary unit.

diabetes.diabetesjournals.org Nagata and Associates 1227

markedly reduced in glucoraphanin-treated mice (Fig. 6A).Glucoraphanin significantly suppressed HFD-induced in-flammatory pathways, such as c-Jun N-terminal kinase(JNK) and extracellular signal–regulated kinase (Erk)(Fig. 6B). Glucoraphanin tended to decrease levels ofp-NF-kB p65 (Ser536) in HFD-fed mice, although this de-crease was not statistically significant (Fig. 6B). Of note,in the liver of Nrf22/2 mice, glucoraphanin failed to

suppress HFD-induced inflammatory signal pathways(Supplementary Fig. 4). In addition, glucoraphanin signif-icantly decreased the HFD-induced hepatic expression ofmacrophage markers, including F4/80, Cd11b, and Cd68(Fig. 6C). Tissue macrophages are phenotypically hetero-geneous and have been characterized according to theiractivation/polarization state as M1-like proinflammatorymacrophages or M2-like anti-inflammatory macrophages

Figure 3—The antiobesity and insulin-sensitizing effects of glucoraphanin are abolished in Nrf22/2 mice. A: Body weight of Nrf22/2 micefed NC, NC-GR, HFD, or HFD-GR for 14 weeks. B–E: VO2 (B), VCO2 (C), energy expenditure (D), and RER (E) of Nrf22/2 mice fed theindicated diet for 3 weeks. F: Rectal temperature in mice Nrf22/2 fed the indicated diet for 6 weeks. G and H: ITT (0.9 units/kg body weightfor NC and NC-GR mice; 1.2 units/kg body weight for HFD and HFD-GR mice) (G) and GTT (2 g/kg body weight) (H) were performed after9 and 11 weeks of feeding, respectively. Bar graphs represent area under the curve (AUC) calculations. Data are mean 6 SEM (NC: n = 8;NC-GR: n = 8; HFD: n = 10; HFD-GR: n = 8). *P < 0.05, **P < 0.01 vs. NC.

1228 Antiobesity Effect of a Natural Nrf2 Inducer Diabetes Volume 66, May 2017

(29). Consistent with the decreased expression of macro-phage markers, glucoraphanin prevented macrophage(F4/80+CD11b+ cell) accumulation in the liver of HFD-fed

mice (Fig. 6D). Additionally, glucoraphanin decreased thenumber of M1-like liver macrophages expressing surfacemarkers (F4/80+CD11b+CD11c+CD2062) (Fig. 6E). In

Figure 4—Glucoraphanin blocks HFD-induced reduction of Ucp1 protein levels in white adipose depots of wild-type mice but not in Nrf22/2

mice. A: Immunoblot analyses of Ucp1 and tubulin expression by using lysates of BAT (0.5 mg protein), epididymal fat depots (eWAT) (5 mgprotein), and inguinal fat depots (ingWAT) (2 mg protein) from wild-type or Nrf22/2 mice on the indicated diet for 14 weeks. Bar graphsrepresent normalized data of Ucp1/tubulin from four independent blots, and data are mean 6 SEM (n = 8/group). *P < 0.05, **P < 0.01 vs.NC-fed wild-type or NC-fed Nrf22/2 mice; ##P < 0.01 vs. HFD-fed wild-type mice. B: Quantitative real-time PCR determination of mRNAlevels of Nqo1 and of genes involved in fat browning in the absence or presence of sulforaphane (SFN) (0.2, 1, 2, or 5 mmol/L) in primary beigeadipocytes isolated from ingWAT of wild-type mice (normalized against 36B4). Bar graph data are mean 6 SEM (n = 6/group). C: The sameexperiment as in B was repeated with primary beige adipocytes isolated from ingWAT of Nrf22/2 mice. *P < 0.05, **P < 0.01 vs. DMSO-treated wild-type adipocytes. The difference was determined by one-way ANOVA. Post hoc analysis was performed with Dunnett test. A.U.,arbitrary unit; Elovl3, fatty acid elongase 3; Cidea, cell death–inducing DFFA-like effector A; Prdm16, PR domain containing 16; WT, wild type.

diabetes.diabetesjournals.org Nagata and Associates 1229

contrast, glucoraphanin increased the number of M2-likeliver macrophages (F4/80+CD11b+CD11c2CD206+), re-sulting in a predominantly M2-like macrophage popula-tion (Fig. 6E). Moreover, glucoraphanin decreased the

mRNA expression of Tnf-a and NADPH oxidase in theepididymal WAT of HFD-fed mice (Supplementary Fig.5A). Although the HFD-induced expression of macrophagemarkers and macrophage accumulation in epididymal WAT

Figure 5—Glucoraphanin alleviates HFD-induced hepatic steatosis and oxidative stress. A: Livers of mice fed the indicated diet for14 weeks. Scale bar = 100 mm. Bar graphs represent the liver weights (n = 9/group). B: Hematoxylin-eosin staining (original magnifica-tion 3200; scale bar = 100 mm). C: Plasma ALT and AST levels. D: Liver triglyceride (TG) and FFA content. E: Quantitative real-time PCRdetermination of hepatic mRNA expression of genes involved in fatty acid synthesis, NADPH oxidase complex, antioxidant stress re-sponse, and fatty acid b-oxidation (normalized against 36B4). F: Hepatic levels of malondialdehyde. Data are mean 6 SEM. *P < 0.05,**P < 0.01 vs. NC; #P< 0.05, ##P< 0.01 vs. HFD. Cat, catalase; Cpt1a, carnitine palmitoyltransferase 1a; Gpx1, glutathione peroxidase 1;Ppara, peroxisome proliferator–activated receptor a; Sod1, superoxide dismutase 1.

1230 Antiobesity Effect of a Natural Nrf2 Inducer Diabetes Volume 66, May 2017

Figure 6—Glucoraphanin improves liver inflammation in HFD-fed mice. A: Relative mRNA expression of Tnf-a, Ccl2, and Ccr2 in the liver of micefed the indicated diet for 14 weeks (n = 8/group). B: Immunoblot analysis of p-JNK (Thr183/Tyr185), JNK, p-Erk (Thr202/Tyr204), Erk, p-NF-kB p65(Ser536), and NF-kB p65 with liver lysates. Each lane represents a liver lysate from a different animal (n = 8/group). Bar graphs represent normalizeddata of p-JNK/JNK, p-Erk/Erk, and p-NF-kB p65/NF-kB p65 from two independent experiments. Data are mean 6 SEM. C: Relative mRNAexpression of macrophage markers. D: FACS analysis of macrophages in the liver of mice fed HFD or HFD-GR (n = 8/group). Macrophages aredefined as propidium iodide-CD45+NK1.12CD32CD192TER1192CD11b+F4/80+ cells. Bar graph shows the number of liver macro-phages. E: M1- and M2-like macrophages are defined as CD11c+CD2062 and CD11c2CD206+, respectively. Bar graphs show thepercentage of M1- and M2-like macrophages and the M1/M2 ratio. Data are mean 6 SEM. *P < 0.05, **P < 0.01 vs. NC; #P < 0.05,##P < 0.01 vs. HFD. A.U., arbitrary unit.

diabetes.diabetesjournals.org Nagata and Associates 1231

were not altered by glucoraphanin (Supplementary Fig. 5Band C), the number of M1-like macrophages was signifi-cantly decreased in the epididymal WAT of HFD-GR mice(Supplementary Fig. 5D).

Glucoraphanin Decreases Circulating LPS and theRelative Abundance of Proteobacteria in the GutMicrobiomes of HFD-Fed MiceGut microbiota–derived LPS induces chronic inflamma-tion that eventually leads to insulin resistance in obesity,termed metabolic endotoxemia (1,2). On the basis of ourobservation that glucoraphanin alleviates inflammationin the liver and epididymal WAT of HFD-fed mice, wesubsequently investigated the effects of glucoraphaninon metabolic endotoxemia and gut microbiota. In accor-dance with previous studies (1,2), the HFD induced atwofold increase in circulatory LPS levels, which was re-duced by glucoraphanin supplementation (Fig. 7A). Fur-thermore, plasma and hepatic levels of the LPS markerLBP were significantly elevated by the HFD and reducedby glucoraphanin supplementation (Fig. 7B). A principalcomponent analysis distinguished cecal microbial com-munities based on diet and treatment, revealing thatthe metagenomes of HFD-fed mice formed a cluster dis-tinct from that formed by NC-fed mice (Fig. 7C). How-ever, samples from HFD-GR mice formed a cluster thatwas indistinguishable from that of NC or NC-GR mice(Fig. 7C). Of note, consistent with previous reports(30,31), further analysis at the phylum level demon-strated that the proportion of gram-negative Proteobac-teria was significantly elevated in the gut microbiomes ofHFD-fed mice, which was suppressed by glucoraphaninsupplementation (Fig. 7D and E). The increase in therelative abundance of Proteobacteria in HFD-fed miceis mostly explained by an increase in the relative abun-dance of bacteria from the family Desulfovibrionaceae(Fig. 7F), key producers of endotoxins in animal modelsof obesity (30). In fact, the relative abundance of Desul-fovibrionaceae was positively correlated with plasmaLPS levels (Fig. 7G). Furthermore, plasma LPS levelswere significantly and positively correlated with the he-patic mRNA levels of Tnf-a, gp91phox, and F4/80 (Fig.7H). Similarly, the liver expression of other markergenes was significantly and positively correlatedwith plasma LPS levels and with one another (Supple-mentary Table 4).

DISCUSSION

In the current study, we demonstrated that glucoraphanin,a stable precursor of the Nrf2 inducer sulforaphane,mitigated HFD-induced weight gain, insulin resistance,hepatic steatosis, oxidative stress, and chronic inflamma-tion in mice. The weight-reducing and insulin-sensitizingeffects of glucoraphanin were abolished in Nrf22/2 mice.Additionally, glucoraphanin lowered plasma LPS levels inHFD-fed mice and decreased the relative abundance ofDesulfovibrionaceae. At the molecular level, glucoraphaninincreased Ucp1 protein expression in WAT depots while

suppressing the hepatic mRNA expression of genes in-volved in lipogenesis, NADPH oxidase, and inflamma-tory cytokines. The data suggest that in diet-inducedobese mice, glucoraphanin restores energy expenditureand limits gut-derived metabolic endotoxemia, therebypreventing hepatic steatosis, insulin resistance, andchronic inflammation.

Consistent with previous reports demonstrating theantiobesity effects of synthetic Nrf2 inducers (9–11),we show that the oral administration of glucoraphaninmitigates HFD-induced weight gain (Fig. 1A). Thedose of glucoraphanin used in the current study(;12 mmol/mouse/day) is similar to that used in otherexperiments that investigated its antitumor effects inmice (14,32,33). Here, we show that the effect of glucor-aphanin on whole-body energy expenditure and the pro-tein expression of Ucp1 in WAT were abolished inNrf22/2 mice (Figs. 3D and 4A). A study that used adi-pocyte-specific PRDM16-deficient mice indicated thatadaptive thermogenesis in beige fat also contributes tosystemic energy expenditure (26). The mutant mice inthe aforementioned study, which exhibited markedly re-duced Ucp1 mRNA expression in inguinal WAT and min-imal effects on BAT, developed obesity and insulinresistance in response to an HFD. Thus, we believethat the increased energy expenditure in HFD-GR miceat least in part stems from an increase in beige fat, eventhough the expression of Ucps in BAT and skeletal mus-cle is not altered. Further analysis in Ucp1 knockoutmice will elucidate the relative contribution of beigefat to the Nrf2-mediated metabolic effects elicited byglucoraphanin.

Several studies have indicated that Nrf22/2 mice arepartially protected from HFD-induced obesity and are as-sociated with milder insulin resistance compared withwild-type counterparts (9,34,35). Recently, Schneideret al. (35) demonstrated mitigation of HFD-induced obe-sity in Nrf22/2 mice, which was 25% less body weightthan that of wild-type mice after 6 weeks of feeding.They also found that HFD-fed Nrf22/2 mice exhibited a20–30% increase in energy expenditure associated with anapproximately threefold upregulation of Ucp1 protein ex-pression in abdominal WAT. In the current study,Nrf22/2 mice gained less weight after 6 weeks of HFDfeeding than the HFD-fed wild-type mice (35.9 6 0.9 vs.39.0 6 0.8 g, Nrf22/2 wild type, respectively; P , 0.05)(Figs. 1A and 3A). The lower body mass of Nrf22/2 miceraises the possibility that the antiobesity effect ofglucoraphanin was completely phenocopied by Nrf2gene deficiency. However, several observations suggestthat Nrf22/2 mice only partially phenocopy the effectof glucoraphanin on weight gain reduction. First, in thecurrent study, the weight difference between Nrf22/2 andwild-type mice was only 8%, which is much less than thatin Schneider et al. Second, in Nrf22/2 mice, HFD inducedsignificant weight gain (Fig. 3A), glucose intolerance (Fig.3H), and insulin resistance compared with NC, as judged

1232 Antiobesity Effect of a Natural Nrf2 Inducer Diabetes Volume 66, May 2017

by increased HOMA-IR (Supplementary Table 3). Third,metabolic rate and energy expenditure of Nrf22/2 micewere comparable with those in wild-type mice (Figs. 1C–Eand 3B–D). Finally, Ucp1 protein levels in both epididymal

WAT and inguinal WAT of HFD-fed Nrf22/2 mice werelower than in HFD-GR wild-type mice (Fig. 4A). Takentogether, these findings suggest that Nrf2 gene deficiencyis not sufficient to block HFD-induced obesity by increasing

Figure 7—Glucoraphanin decreases circulating LPS and gram-negative Proteobacteria in the gut microbiomes of HFD-fed mice. A andB: Plasma levels of LPS (A) and LBP (B, left panel) in mice fed the indicated diet for 14 weeks (n = 6–7/group). Hepatic mRNA levels ofLbp (B, right panel) were determined by quantitative real-time PCR. C: A principal component analysis was performed by using thephylum-level taxonomic profiles of mouse cecal microbiota. Each plot represents the taxonomic profile data of individual mice. Thecloser the spatial distance among individual data points, the more similar they are with respect to both axes (PC1 and PC2). D: Relativeabundance distribution of OTU sequences (96% level). Percentage of total OTU sequences taxonomically assigned to bacterial phylafrom cecal contents of mice fed the noted diets for 14 weeks. E and F: Box plots of relative abundance of OTU sequences assigned tothe Proteobacteria phyla and Desulfovibrionaceae family. Data are mean 6 SEM (n = 6–7/group). *P < 0.05, **P < 0.01 vs. NC; #P < 0.05vs. HFD. G: Plasma LPS levels plotted against the relative abundance of the Desulfovibrionaceae family in the cecal contents. H: Correlationsbetween plasma LPS and hepatic mRNA expression levels of TNF-a, gp91phox, and F4/80. Spearman r correlation and corresponding P valueare shown. EU, endotoxin unit.

diabetes.diabetesjournals.org Nagata and Associates 1233

energy expenditure and Ucp1 expression in WAT depotsand to mask the effect of glucoraphanin. However, wecannot fully exclude the possibility that the effects ofglucoraphanin are mediated by Nrf2-independent mecha-nisms. Possible reasons for the discordance in metabolicphenotypes of Nrf22/2 between Schneider et al. and thecurrent study may be due to differences in knockoutmouse lines and experimental conditions (e.g., age ofmice at beginning of HFD feeding, composition of HFD,temperature in the metabolic chamber).

The current in vitro study of primary beige adipocytesrevealed that sulforaphane promotes the expression ofbrown fat–selective genes (Fig. 4B). Of note, the concen-tration of sulforaphane used in cell culture (0.2–5 mmol/L) is comparable with that detected in micefed NC-GR and HFD-GR (Supplementary Fig. 1B). More-over, we determined that Nrf2 acts as a positive regula-tor of beige adipocyte differentiation (Fig. 4C andSupplementary Fig. 3C). The fewer differentiation levelsin Nrf2-deficient beige adipocytes agree with previousreports demonstrating that Nrf2 induces white adipo-cyte differentiation through increasing the gene ex-pression of Pparg (34) and Cebpb (36), commontranscription factors that regulate the differentiation ofbrown, beige, and white adipocytes. Furthermore, Nrf2has been reported to bind NF-E2–binding sites in the 59flanking region of human and rodent Ucp1 genes (37).However, we cannot exclude the possibilities that glucor-aphanin affects sympathetic nervous activity or that hor-monal factors regulate fat browning (38). In addition,mitochondrial reactive oxidative species facilitate Ucp1-dependent respiration in BAT and whole-body energy ex-penditure by promoting the sulfenylation of a specificcysteine residue (Cys253) in Ucp1 (39). The molecularmechanism by which Nrf2 regulates the expression andthermogenic activity of Ucp1 in beige adipocytes requiresfurther investigation.

Glucoraphanin supplementation improved the sys-temic glucose tolerance and insulin sensitivity of HFD-fed mice. Although the molecular mechanism by whichsynthetic Nrf2 inducers enhance glucose uptake is un-clear, AMPK activation may mediate this enhancement inmouse skeletal muscle and adipose tissue (9–11). Thephosphorylation levels of AMPK (Thr172) and acetyl-CoAcarboxylase (Ser79) in peripheral insulin target tissueswere comparable between NC-GR and HFD-GR mice andvehicle-treated controls (Supplementary Fig. 6). Thesedata suggest that AMPK activation is not necessary forglucoraphanin to exert its insulin-sensitizing effect onHFD-fed mice. Additional studies that use the hyperinsu-linemic-euglycemic clamp technique are needed to deter-mine which tissues contribute to the insulin-sensitizingeffects of glucoraphanin.

The beneficial effects of glucoraphanin on hepaticlipid metabolism were not accompanied by AMPKactivation or the increased expression of fatty acidb-oxidation genes (Fig. 5E). Instead, glucoraphanin

mitigated HFD-induced oxidative stress and inflammationin the liver. In obesity, hepatic inflammation mediated bymacrophage/monocyte-derived proinflammatory cyto-kines promotes lipogenesis through the inhibition of in-sulin signaling and SREBP activation (40,41). In fact, thedepletion of Kupffer cells by clodronate liposomes ame-liorates hepatic steatosis and insulin sensitivity in HFD-fed mice (27). Furthermore, Ccl2- or Ccr2-deficient miceare protected from diet-induced hepatic steatosis, eventhough they still become obese (42,43). Moreover, thespecific ablation of M1-like macrophages restores insulinsensitivity in diet-induced obese mice (44), whereas thedeletion of Ppard, which promotes M2 activation, predis-poses lean mice to developing insulin resistance (45).Therefore, decreased hepatic macrophage accumulationand M2-dominant polarization of hepatic and adiposemacrophages account, at least in part, for the protectionfrom hepatic steatosis and insulin resistance in HFD-GRmice.

One of the most important findings of this study isthat glucoraphanin decreases the relative abundance ofgram-negative Proteobacteria, particularly family Desul-fovibrionaceae, while reducing circulatory LPS levels(Fig. 7). Studies have demonstrated a significant increasein Desulfovibrionaceae, potential endotoxin producers,in the gut microbiomes of both HFD-induced obesemice and obese human subjects compared with lean in-dividuals (30,31,46). We cannot exclude the possibilitythat other microbiota-derived products, such as bileacids and short-chain fatty acids, also mediate the met-abolic action of glucoraphanin. Whether the interactionbetween sulforaphane and gut microbiota is affected di-rectly or indirectly by altered host physiology remains tobe determined. However, several studies have suggestedthat sulforaphane can alter the gut microbiota directlybecause isothiocyanates (including sulforaphane) havebeen shown to exhibit antibacterial activity against Pro-teobacteria (47,48). This activity may proceed throughredox disruption and enzyme denaturation reactions in-volving the isothiocyanate reactivity group, -N = C = S,the thiol group (-SH) of glutathione, and proteobacterialproteins (47,48). Additionally, sulforaphane exhibits an-tibacterial activity against Helicobacter pylori, a memberof the phylum Proteobacteria (49). We are unaware ofprevious reports demonstrating that isothiocyanatesinhibit the proliferation of Desulfovibrionaceae. Themechanistic underpinnings of this antibacterial activityrequire elucidation.

In conclusion, the results of this study indicate thatglucoraphanin may be effective in preventing obesity andrelated metabolic disorders such as NAFLD and type2 diabetes. Another clinical study demonstrated thatsupplementation with a dietary dose of glucoraphanin(69 mmol/day) for 2 months significantly decreasedplasma liver enzymes, ALT, and AST, although bodymass did not change (18). Long-term treatment with ahigher dose of glucoraphanin (800 mmol/day), which

1234 Antiobesity Effect of a Natural Nrf2 Inducer Diabetes Volume 66, May 2017

can be safely administered without harmful adverse ef-fects (17), may be required to achieve an antiobesity effectin humans.

Acknowledgments. The authors thank M. Nakayama and K. Hara(Kanazawa University) for technical assistance. The authors also thank Editage(www.editage.jp) for English-language editing.Funding. This work was supported by Japan Society for the Promotion ofScience KAKENHI grant numbers 15K00813 (to N.N.), 15K12698 (to T.O.), and16H03035 (to T.O.).Duality of Interest. No potential conflicts of interest relevant to this articlewere reported.Author Contributions. N.N. collected data and wrote the manuscript.L.X., S.Ko., N.F., F.Z., Y.N., and M.N. collected data. Y.U., Y.A., and R.U. collecteddata and edited the manuscript. C.T., H.S., and S.Ka. contributed to thediscussion and reviewed the manuscript. T.O. contributed to the discussion andreviewed and edited the manuscript. T.O. is the guarantor of this work and, assuch, had full access to all the data in the study and takes responsibility for theintegrity of the data and the accuracy of the data analysis.Prior Presentation. Parts of this article were presented at the 76thScientific Sessions of the American Diabetes Association, New Orleans, LA,10–14 June 2016.

References1. Cani PD, Amar J, Iglesias MA, et al. Metabolic endotoxemia initiates obesityand insulin resistance. Diabetes 2007;56:1761–17722. Cani PD, Bibiloni R, Knauf C, et al. Changes in gut microbiota controlmetabolic endotoxemia-induced inflammation in high-fat diet-induced obesityand diabetes in mice. Diabetes 2008;57:1470–14813. Dietrich MO, Horvath TL. Limitations in anti-obesity drug development: thecritical role of hunger-promoting neurons. Nat Rev Drug Discov 2012;11:675–6914. Sanyal AJ, Chalasani N, Kowdley KV, et al.; NASH CRN. Pioglitazone, vitaminE, or placebo for nonalcoholic steatohepatitis. N Engl J Med 2010;362:1675–16855. Motohashi H, Yamamoto M. Nrf2-Keap1 defines a physiologically importantstress response mechanism. Trends Mol Med 2004;10:549–5576. Itoh K, Chiba T, Takahashi S, et al. An Nrf2/small Maf heterodimer mediatesthe induction of phase II detoxifying enzyme genes through antioxidant responseelements. Biochem Biophys Res Commun 1997;236:313–3227. Uruno A, Furusawa Y, Yagishita Y, et al. The Keap1-Nrf2 system preventsonset of diabetes mellitus. Mol Cell Biol 2013;33:2996–30108. Yates MS, Tran QT, Dolan PM, et al. Genetic versus chemoprotective activationof Nrf2 signaling: overlapping yet distinct gene expression profiles between Keap1knockout and triterpenoid-treated mice. Carcinogenesis 2009;30:1024–10319. Shin S, Wakabayashi J, Yates MS, et al. Role of Nrf2 in prevention of high-fat diet-induced obesity by synthetic triterpenoid CDDO-imidazolide. Eur JPharmacol 2009;620:138–14410. Saha PK, Reddy VT, Konopleva M, Andreeff M, Chan L. The triterpenoid2-cyano-3,12-dioxooleana-1,9-dien-28-oic-acid methyl ester has potent anti-diabetic effects in diet-induced diabetic mice and Lepr(db/db) mice. J Biol Chem2010;285:40581–4059211. Yu Z, Shao W, Chiang Y, et al. Oltipraz upregulates the nuclear factor(erythroid-derived 2)-like 2 [corrected](NRF2) antioxidant system and preventsinsulin resistance and obesity induced by a high-fat diet in C57BL/6J mice.Diabetologia 2011;54:922–93412. de Zeeuw D, Akizawa T, Audhya P, et al.; BEACON Trial Investigators.Bardoxolone methyl in type 2 diabetes and stage 4 chronic kidney disease.N Engl J Med 2013;369:2492–2503

13. Kelley MJ, Glaser EM, Herndon JE 2nd, et al. Safety and efficacy of weeklyoral oltipraz in chronic smokers. Cancer Epidemiol Biomarkers Prev 2005;14:892–899

14. Fahey JW, Zhang Y, Talalay P. Broccoli sprouts: an exceptionally rich sourceof inducers of enzymes that protect against chemical carcinogens. Proc Natl AcadSci U S A 1997;94:10367–1037215. Zhang Y, Talalay P, Cho CG, Posner GH. A major inducer of anticarcinogenicprotective enzymes from broccoli: isolation and elucidation of structure. Proc NatlAcad Sci U S A 1992;89:2399–240316. Shapiro TA, Fahey JW, Wade KL, Stephenson KK, Talalay P. Human metab-olism and excretion of cancer chemoprotective glucosinolates and isothiocyanatesof cruciferous vegetables. Cancer Epidemiol Biomarkers Prev 1998;7:1091–110017. Kensler TW, Ng D, Carmella SG, et al. Modulation of the metabolism ofairborne pollutants by glucoraphanin-rich and sulforaphane-rich broccoli sproutbeverages in Qidong, China. Carcinogenesis 2012;33:101–10718. Kikuchi M, Ushida Y, Shiozawa H, et al. Sulforaphane-rich broccoli sproutextract improves hepatic abnormalities in male subjects. World J Gastroenterol2015;21:12457–1246719. Wade KL, Garrard IJ, Fahey JW. Improved hydrophilic interaction chroma-tography method for the identification and quantification of glucosinolates.J Chromatogr A 2007;1154:469–47220. Nagata N, Matsuo K, Bettaieb A, et al. Hepatic Src homology phosphatase2 regulates energy balance in mice. Endocrinology 2012;153:3158–316921. Aune UL, Ruiz L, Kajimura S. Isolation and differentiation of stromal vascularcells to beige/brite cells. J Vis Exp 2013;(73):e5019122. Kitade H, Sawamoto K, Nagashimada M, et al. CCR5 plays a critical role inobesity-induced adipose tissue inflammation and insulin resistance by regulatingboth macrophage recruitment and M1/M2 status. Diabetes 2012;61:1680–169023. Kim SW, Suda W, Kim S, et al. Robustness of gut microbiota of healthyadults in response to probiotic intervention revealed by high-throughput py-rosequencing. DNA Res 2013;20:241–25324. Greaney AJ, Maier NK, Leppla SH, Moayeri M. Sulforaphane inhibits multipleinflammasomes through an Nrf2-independent mechanism. J Leukoc Biol 2016;99:189–19925. Myzak MC, Karplus PA, Chung FL, Dashwood RH. A novel mechanism ofchemoprotection by sulforaphane: inhibition of histone deacetylase. Cancer Res2004;64:5767–577426. Cohen P, Levy JD, Zhang Y, et al. Ablation of PRDM16 and beige adiposecauses metabolic dysfunction and a subcutaneous to visceral fat switch. Cell2014;156:304–31627. Huang W, Metlakunta A, Dedousis N, et al. Depletion of liver Kupffer cellsprevents the development of diet-induced hepatic steatosis and insulin re-sistance. Diabetes 2010;59:347–35728. Obstfeld AE, Sugaru E, Thearle M, et al. C-C chemokine receptor 2 (CCR2)regulates the hepatic recruitment of myeloid cells that promote obesity-inducedhepatic steatosis. Diabetes 2010;59:916–92529. Mantovani A, Sica A, Sozzani S, Allavena P, Vecchi A, Locati M. The che-mokine system in diverse forms of macrophage activation and polarization.Trends Immunol 2004;25:677–68630. Zhang C, Zhang M, Wang S, et al. Interactions between gut microbiota, hostgenetics and diet relevant to development of metabolic syndromes in mice. ISMEJ 2010;4:232–24131. Hildebrandt MA, Hoffmann C, Sherrill-Mix SA, et al. High-fat diet determinesthe composition of the murine gut microbiome independently of obesity. Gas-troenterology 2009;137:1716–24.e1, 232. Conaway CC, Wang CX, Pittman B, et al. Phenethyl isothiocyanate andsulforaphane and their N-acetylcysteine conjugates inhibit malignant progressionof lung adenomas induced by tobacco carcinogens in A/J mice. Cancer Res2005;65:8548–855733. Shen G, Khor TO, Hu R, et al. Chemoprevention of familial adenomatouspolyposis by natural dietary compounds sulforaphane and dibenzoylmethane aloneand in combination in ApcMin/+ mouse. Cancer Res 2007;67:9937–994434. Pi J, Leung L, Xue P, et al. Deficiency in the nuclear factor E2-related factor-2 transcription factor results in impaired adipogenesis and protects against diet-induced obesity. J Biol Chem 2010;285:9292–9300

diabetes.diabetesjournals.org Nagata and Associates 1235

35. Schneider K, Valdez J, Nguyen J, et al. Increased energy expenditure, Ucp1expression, and resistance to diet-induced obesity in mice lacking nuclear factor-erythroid-2-related transcription factor-2 (Nrf2). J Biol Chem 2016;291:7754–776636. Hou Y, Xue P, Bai Y, et al. Nuclear factor erythroid-derived factor 2-relatedfactor 2 regulates transcription of CCAAT/enhancer-binding protein b duringadipogenesis. Free Radic Biol Med 2012;52:462–47237. Rim JS, Kozak LP. Regulatory motifs for CREB-binding protein and Nfe2l2transcription factors in the upstream enhancer of the mitochondrial uncouplingprotein 1 gene. J Biol Chem 2002;277:34589–3460038. Kajimura S, Spiegelman BM, Seale P. Brown and beige fat: physiologicalroles beyond heat generation. Cell Metab 2015;22:546–55939. Chouchani ET, Kazak L, Jedrychowski MP, et al. Mitochondrial ROS regulatethermogenic energy expenditure and sulfenylation of UCP1. Nature 2016;532:112–11640. Ma KL, Ruan XZ, Powis SH, Chen Y, Moorhead JF, Varghese Z. In-flammatory stress exacerbates lipid accumulation in hepatic cells and fatty liversof apolipoprotein E knockout mice. Hepatology 2008;48:770–78141. Lawler JF Jr, Yin M, Diehl AM, Roberts E, Chatterjee S. Tumor necrosisfactor-alpha stimulates the maturation of sterol regulatory element bindingprotein-1 in human hepatocytes through the action of neutral sphingomyelinase.J Biol Chem 1998;273:5053–505942. Kanda H, Tateya S, Tamori Y, et al. MCP-1 contributes to macrophageinfiltration into adipose tissue, insulin resistance, and hepatic steatosis in obesity.J Clin Invest 2006;116:1494–1505

43. Weisberg SP, Hunter D, Huber R, et al. CCR2 modulates inflammatory andmetabolic effects of high-fat feeding. J Clin Invest 2006;116:115–12444. Patsouris D, Li PP, Thapar D, Chapman J, Olefsky JM, Neels JG. Ablation ofCD11c-positive cells normalizes insulin sensitivity in obese insulin resistantanimals. Cell Metab 2008;8:301–30945. Odegaard JI, Ricardo-Gonzalez RR, Red Eagle A, et al. Alternative M2 ac-tivation of Kupffer cells by PPARdelta ameliorates obesity-induced insulin re-sistance. Cell Metab 2008;7:496–50746. Xiao S, Fei N, Pang X, et al. A gut microbiota-targeted dietary interventionfor amelioration of chronic inflammation underlying metabolic syndrome. FEMSMicrobiol Ecol 2014;87:357–36747. Aires A, Mota VR, Saavedra MJ, Rosa EA, Bennett RN. The antimicrobialeffects of glucosinolates and their respective enzymatic hydrolysis products onbacteria isolated from the human intestinal tract. J Appl Microbiol 2009;106:2086–209548. Sofrata A, Santangelo EM, Azeem M, Borg-Karlson A-K, Gustafsson A,Pütsep K. Benzyl isothiocyanate, a major component from the roots of Salvadorapersica is highly active against gram-negative bacteria. PLoS One 2011;6:e2304549. Fahey JW, Haristoy X, Dolan PM, et al. Sulforaphane inhibits extracellular,intracellular, and antibiotic-resistant strains of Helicobacter pylori and preventsbenzo[a]pyrene-induced stomach tumors. Proc Natl Acad Sci U S A 2002;99:7610–7615

1236 Antiobesity Effect of a Natural Nrf2 Inducer Diabetes Volume 66, May 2017