2017-Wu-Metformin alters the gut microbiome of.pdf

Articles850VOLUME 23|NUMBER 7|JULY 2017 nAture medicineMetformin is the most prescribed pharmacotherapy for the treatment of individuals with type 2 diabetes(T2D)because of its relative safety,low cost,and beneficial effects on blood glucose and cardiovascu-lar mortality1,2.However,its mechanism of action remains unclear.Although metformin is generally considered to mediate its antihyper-glycemic effects by suppressing hepatic glucose output through the activation of AMP-activated protein kinase(AMPK)-dependent35 and AMPK-independent pathways68 in the liver,accumulating evi-dence indicates that it might also act through pathways in the gut9,10.For example,its glucose-lowering effect is more pronounced when given orally than when administered intravenously11.In addition,a study comparing metformin formulations with reduced and normal plasma exposure provided evidence to indicate that the lower bowel is a major site of action for metformin12.Furthermore,recent stud-ies in both rodents1315 and humans1618 suggest that gut microbial changes might contribute to the antidiabetic effect of metformin.So far,however,it is not known how metformin affects the gut micro-biota of individuals with treatment-naive T2D,nor how metformin interacts with gut bacteria.Here we performed a randomized,placebo-controlled,double-blind study in individuals with newly diagnosed T2D on a calorie-restricted diet,and we combined metagenomics and targeted metabolomics to investigate the effect of metformin on the composition and function of the gut microbiota.We also transferred human fecal samples to germ-free mice to study the effects of metformin-altered microbiota on host glucose metabolism,and we used an in vitro gut simulator to investigate metforminmicrobiota interactions directly.RESULTSMetformin alters the gut microbiota compositionTo investigate how metformin affects the composition of the gut microbiota,we randomized treatment-naive individuals with recently diagnosed T2D to receive either placebo(n=18)or 1,700 mg/d of metformin(n=22)for 4 months in a double-blind study.Clinical characteristics of these individuals before and after treatment are presented in Table 1.Both groups were recommended to consume a calorie-restricted diet for the 4-month study period(Supplementary Table 1);calorie intake was reduced by a median of 342 kcal/d,and 1Department of Molecular and Clinical Medicine,Wallenberg Laboratory,Institute of Medicine,University of Gothenburg,Gothenburg,Sweden.2Department of Diabetes,Endocrinology and Nutrition,Institut dInvestigaci Biomdica de Girona,Hospital Josep Trueta,Girona,Spain.3Departament de Medicina,Facultat de Medicina,University of Girona,Girona,Spain.4Centro de Investigacin Biomdica en Red de Fisiopatologa de la Obesidad y Nutricin(CIBEROBN),Instituto de Salud Carlos III,Madrid,Spain.5IRSD,Universit de Toulouse,INSERM,INRA,ENVT,UPS,Toulouse,France.6Barcelona Supercomputing Center(BSC),Joint BSCCRGIRB Research Program in Computational Biology,Barcelona,Spain.7Instituci Catalana de Recerca i Estudis Avanats(ICREA),Barcelona,Spain.8Institut National de la Sant et de la Recherche Mdicale(INSERM),Toulouse,France.9Universit Paul Sabatier(UPS),Unit Mixte de Recherche 1048,Institut de Maladies Mtaboliques et Cardiovasculaires,Toulouse,France.10Sahlgrenska University Hospital,Gothenburg,Sweden.11Novo Nordisk Foundation Center for Basic Metabolic Research,Section for Metabolic Receptology and Enteroendocrinology,Faculty of Health Sciences,University of Copenhagen,Copenhagen,Denmark.12These authors contributed equally to this work.Correspondence should be addressed to J.M.F.-R.(jmfrealidibgi.org)or F.B.(fredrik.backhedwlab.gu.se).Received 27 June 2016;accepted 19 April 2017;published online 22 May 2017;doi:10.1038/nm.4345Metformin alters the gut microbiome of individuals with treatment-naive type 2 diabetes,contributing to the therapeutic effects of the drugHao Wu1,12,Eduardo Esteve24,12,Valentina Tremaroli1,Muhammad Tanweer Khan1,Robert Caesar1,Louise Manners-Holm1,Marcus Sthlman1,Lisa M Olsson1,Matteo Serino5,Merc Planas-Flix6,Gemma Xifra24,Josep M Mercader6,David Torrents6,7,Rmy Burcelin8,9,Wifredo Ricart24,Rosie Perkins1,Jos Manuel Fernndez-Real24&Fredrik Bckhed1,10,11Metformin is widely used in the treatment of type 2 diabetes(T2D),but its mechanism of action is poorly defined.Recent evidence implicates the gut microbiota as a site of metformin action.In a double-blind study,we randomized individuals with treatment-naive T2D to placebo or metformin for 4 months and showed that metformin had strong effects on the gut microbiome.These results were verified in a subset of the placebo group that switched to metformin 6 months after the start of the trial.Transfer of fecal samples(obtained before and 4 months after treatment)from metformin-treated donors to germ-free mice showed that glucose tolerance was improved in mice that received metformin-altered microbiota.By directly investigating metforminmicrobiota interactions in a gut simulator,we showed that metformin affected pathways with common biological functions in species from two different phyla,and many of the metformin-regulated genes in these species encoded metalloproteins or metal transporters.Our findings provide support for the notion that altered gut microbiota mediates some of metformins antidiabetic effects.2017 Nature America,Inc.,part of Springer Nature.All rights reserved.ArticlesnAture medicine VOLUME 23|NUMBER 7|JULY 2017 851no significant differences were seen between the groups(P=0.90).A subset of the placebo group switched to receive metformin(850 or 1,700 mg/d;n=13)6 months after the start of the study;to validate our findings from the randomized study,we analyzed samples from this group after a further 6 months.As expected given the reduced calorie intake,body-mass index(BMI)decreased significantly in both the placebo and metformin groups over the initial 4-month study period(Fig.1a).However,sig-nificant decreases in%hemoglobin A1c(HbA1c)and fasting blood glucose were observed only in the group randomized to metformin treatment(Fig.1b,c).BMI did not decrease further in the switched subgroup after 6 months on metformin(Fig.1a),but%HbA1c and fasting blood glucose were significantly reduced by metformin in this subgroup(Fig.1b,c).To characterize the effects of metformin on the gut microbiome,we performed whole-genome shotgun sequencing of 131 fecal samples.On average,we obtained 38 million paired-end reads for each sample(ranging from 15 million to 116 million;Supplementary Table 2).The taxonomy and gene profiles were estimated by mapping the high-quality reads to nonredundant genome and gene catalogs imple-mented in the metagenomic data-utilization and analysis(MEDUSA)pipeline19,respectively.Only one bacterial strain was altered over the 4-month study period in the placebo group(Fig.1d),despite the reduction in BMI.By contrast,metformin treatment for 2 and 4 months resulted in significant alterations in the relative abundance of 81 and 86 bacterial strains,respectively,most of which belonged to-proteobacteria(for example,Escherichia coli)and Firmicutes(Fig.1d and Supplementary Table 3;false-discovery rate(FDR)0.1;Supplementary Fig.1).To investigate how different gut bacteria interact with each other,we performed a coabundance network analysis.We showed that 2 months of metformin treatment promoted an increased number of positive connections among microbial genera,especially those within Proteobacteria and Firmicutes(Fig.1f).We also iden-tified a few interphylum connections,such as between Shewanella(Proteobacteria)and Blautia(Firmicutes),a short-chain fatty acid(SCFA)-producing genus23.To test the effect of metformin on microbial growth,we mapped whole-genome shotgun reads to the genomes of common strains in the human gut to determine the ratio between DNA copy number near the replica-tion origin and DNA copy number near the terminus(termed the peak-to-trough ratio,PTR)of bacterial genomes24.After correction for FDR,we found that the PTR of only one bacterial species(Bifidobacterium adolescentis)was significantly increased by metformin(Fig.2a).In agreement,the PTR of B.adolescentis was also increased in the switched subgroup after 6 months on metformin(Fig.2a).Furthermore,in our cohort,we observed a negative correlation between the PTR of B.ado-lescentis and%HbA1c(Spearman coefficient rho=0.28,P 0.01).Consistent with this observation,in vitro analysis showed that met-formin directly promoted the growth of B.adolescentis in pure cultures (Fig.2b).We also showed that metformin directly promoted the growth of A.muciniphila,but not of E.coli,in pure cultures(Fig.2c,d).Metformin-altered microbiota improves glucose toleranceTo investigate whether metformin-altered microbiota could con-tribute to the glucose-lowering effect of metformin,we transferred fecal samples from three metformin-treated participants(before and 4 months after metformin,here termed M0 and M4 microbiota)to germ-free mice.All three of the metformin recipients responded similarly to metformin in terms of reduced%HbA1c,as compared to baseline,after 2 and 4 months on metformin.The mice were fed a high-fat diet for 1 week before and during colonization for 18 d.We did not observe any differences in body weight,body fat,or fasting insulin between mice that received M4 and M0 microbiota(Fig.3a,b and Supplementary Fig.2ac).However,we found improvements Table 1 Clinical characteristics for the 40 individuals with T2D enrolled in this studyPlacebo group(n=18)Metformin group(n=22)P0P2P4M0M2M4Age(years)54.9 1.952.6 2.0Sex(male/female)9/98/14Weight(kg)85.4 5.682.2 5.6+81.5 5.4#96.5 4.192.9 4.091.4 3.9Waist circumference(cm)106.1 4.4101.7 4.3*101.9 3.6111.5 2.7108.3 2.9*108.7 2.9HOMA8.0 1.58.9 1.68.1 1.88.3 1.26.2 0.96.0 0.8*Total cholesterol(mg/dl)205.8 8.8197.8 7.7190.7 6.9*206.0 7.4196.6 7.2198.8 7.5HDL-C(mg/dl)46.9 3.445.0 3.0 46.8 3.148.4 2.755.2 6.151.1 3.0*LDL-C(mg/dl)126.8 6.6124.8 5.7118.1 6.2*129.4 6.4117.4 6.2*121.5 6.8Triglycerides(mg/dl)151.9 18.7155.9 15.2 129.3 12.5129.0 17.8139.6 11.6135.9 12.7ALT(U/liter)33.2 7.224.3 2.9 22.3 2.135.5 3.528.0 1.8+32.8 3.2GGT(U/liter)38.4 5.4 28.2 2.1+26.3 2.0+44.0 6.031.3 3.2+34.1 3.9*CRP(mg/dl)0.4 0.1 0.6 0.10.5 0.10.4 0.10.4 0.10.4 0.1Statin treatment(n)3 4Antihypertensive treatment(n)2 3ALT,alanine transaminase;CRP,C-reactive protein;GGT,-glutamyl transferase;HDL-C,high-density lipoprotein cholesterol;HOMA,homeostatic model assessment;LDL-C,low-density lipoprotein cholesterol.*P 0.05;+P 0.01;#P 0.001 versus P0 or M0.Wilcoxon signedrank test;data are shown as means s.e.m.2017 Nature America,Inc.,part of Springer Nature.All rights reserved.Articles852VOLUME 23|NUMBER 7|JULY 2017 nAture medicinefUreaplasmaEggerthellaRothiaActinobacillusBordetellaAggregatibacterKlebsiellaTuricibacterEscherichiaCollinsellaLactococcusListeriaLeuconostocStreptococcusYersiniaLactobacillusUnclassified BacteroidetesPseudoflavonifractorFilifactorActinomycesMethanobrevibacterGemellaPaludibacterAcinetobacterHerbaspirillumPropionibacteriumEremococcusPeptostreptococcusParascardoviaTreponemaMacrococcusEnterobacterHoldemaniaAnaerococcusVeillonellaOenococcusFinegoldiaBifidobacteriumPseudomonasSyntrophobotulusCoriobacteriumUnclassified PeptostreptococcaceaeOscillibacterMobiluncusVibrioCupriavidusHaemophilusPasteurellaSalmonellaCoprococcusEnterococcusWeissellaHeliobacteriumPediococcusGranulicatellaCitrobacterSubdoligranulumCellulosilyticumOxalobacterCryptobacteriumSlackiaParvimonasSutterellaLachnoanaerobaculumMegasphaeraRahnellaEdwardsiellaPeptoniphilusThermosinusDickeyaMicrococcusRiemerellaDermacoccusSelenomonasAtopobiumErwiniaPectobacteriumCampylobacterAerococcusProteusShewanellaButyrivibrioPseudoramibacterHaloplasmaPhascolarctobacteriumDoreaCandidatus AzobacteroidesPaenibacillusXenorhabdusSerratiaBlautiaLawsoniaAlkaliphilusBacillusCronobacterPantoeaunclassified ErysipelotrichaceaeBurkholderiaSphingobacteriumRheinheimeraEthanoligenensRhodobacterStreptomycesCentipedaScardoviaAzospiraBrachyspiraunclassified RuminococcaceaeOribacteriumBorreliaParabacteroidesRalstoniaShuttleworthiaDesulfosporosinusProvidenciaSodalisTetragenococcusAchromobacterThermoanaerobacteriumBacteroidesNeisseriaPrevotellaSolobacteriumParaprevotellaStaphylococcusPositive correlation(M0)Positive correlation(M2)Negative correlationArcanobacteriumeActinobacteriaBacteroidetesFirmicutesProteobacteria202Column Z scoreRothia mucilaginosa DY18Bacillus cereus SJ1Enterobacter cloacae subsp.ATCC 13047Enterococcus casseliflavus EC10Lactobacillus ultunensis DSM 16047Lactobacillus delbrueckii subsp.lactis DSM 20072Bacteroides clarus YIT 12056Staphylococcus aureus subsp.aureus MRSA177Lactobacillus fermentum CECT 5716Peptoniphilus sp.oral taxon 375 str.F0436Ruminococcus sp.5_1_39BFAACronobacter turicensis z3032Enterobacter lignolyticus SCF1Citrobacter koseri ATCC BAA895Yersinia enterocolitica subsp.enterocolitica 8081Klebsiella pneumoniae 342Enterobacter asburiae LF7aEnterobacter cloacae subsp.cloacae NCTC 9394Enterobacter cloacae EcWSU1Acinetobacter baumannii TCDCAB0715Pseudomonas aeruginosa NCGM2.S1Citrobacter rodentium ICC168Citrobacter youngae ATCC 29220 Dermacoccus sp.Ellin185Escherichia albertii TW07627Escherichia coli TW10509Citrobacter sp.30_2Rheinheimera sp.A13LStreptococcus parasanguinis ATCC 15912Enterobacter cancerogenus ATCC 35316Salmonella bongori NCTC 12419Lactobacillus brevis subsp.gravesensis ATCC 27305Enterobacter aerogenes KCTC 2190Salmonella enterica subsp.SD3246Bacillus coahuilensis m44Klebsiella oxytoca KCTC 1686Salmonella enterica subsp.RSK2980Intestinibacter bartlettii DSM 16795Clostridium beijerinckii NCIMB 8052Clostridium sp.7_2_43FAAClostridium perfringens CPE str.F4969Clostridium botulinum E1 str.BoNT E BelugaClostridium butyricum E4 str.BoNT E BL5262Pseudoflavonifractor capillosus ATCC 29799Clostridium perfringens E str.JGS1987Peptostreptococcus stomatis DSM 17678Subdoligranulum variabile DSM 15176Lactobacillus jensenii 1153Lactobacillus johnsonii NCC 533Lactobacillus gasseri JVV03Oenococcus oeni AWRIB429Weissella cibaria KACC 11862Faecalibacterium cf.prausnitzii KLE1255Bifidobacterium adolescentis L232Ruminococcus flavefaciens FD1M4M2M0P4P2P0+#*#+#+#*#+#+*#+#+#*#*+*+#*#*#+#*#+*+#+#*#+*+*+*20253035404550adbcBMI(kg/m2)P0P2P4 P/M6 M0M2M4*5678HbA1c(%)P0P2P4 P/M6 M0M2M4*4681012Fasting blood glucose(mM)P0P2P4 P/M6 M0M2M4*log2fold-change(M2 versus M0)log2fold-change(M4 versus M0)r=0.92(P 2.2 1016)r=0.59(P=2.1 105)r=0.66(P=4.9 1010)420242024321012342024log2fold-change(P/M6 versus P4)Figure 1 Metformin treatment promotes rapid changes in the composition of the gut microbiota.(ac)Boxplots(with median)showing BMI,%HbA1c,and fasting blood glucose before treatment(P0 and M0)and after 2 and 4 months in individuals with T2D randomized to placebo(P2 and P4;n=18)or metformin(M2 and M4;n=22),and 6 months after metformin in a subgroup that switched from placebo to metformin after the randomized study period(P/M6;n=13)